JOURNAL OF THE NATIONAL CANCER INSTITUTE CER Estrogens as Endogenous Carcinogens 2000 I [TUTE in the Breast and Prostate Number 27 Contents Dedication 1 Preface 2 Ercole Cavaliexi, Eleanor G. Rogan Agenda Biosketches 7 Introductory Remarks 13 David G. Longfellow Symposium Overview ‘ | 15 Richard J. Santen Chapter 1: Developmental, Cellular, and Molecular Basis of Human Breast Cancer 17 Jose Russo, Yun-Fu Hu, Xiaoqi Yang, Irma H. Russo Chapter 2: The Role of Steroid Hormones in Prostate Carcinogenesis 39 Maarten C. Bosland Chapter 3: Endogenous Estrogens as Carcinogens Through Metabolic Activation 67 James D. Yager Chapter 4: Estrogens as Endogenous Genotoxic Agents—DNA Adducts and Mutations 75 Ercole Cavalieri, Krystyna Frenkel, Joachim G. Liehr, Eleanor Rogan, Deodutta Roy Chapter 5: Tissue-Specific Synthesis and Oxidative Metabolism of Estrogens 95 Colin R. Jefcoate, Joachim G. Liehr, Richard J. Santen, Thomas R. Sutter, James D. Yager, Wei Yue, Steven J. Santner, Rajeshwar Tekmal, Laurence Demers, Robert Pauley, Frederick Naftolin, Gil Mor, Lev Berstein Chapter 6: Estrogen Metabolism by Conjugation 113 Rebecca Raftogianis, Cyrus Creveling, Richard Weinshilboum, Judith Weisz Chapter 7: Molecular Epidemiology of Genetic Polymorphisms in Estrogen Metabolizing 125 Enzymes in Human Breast Cancer Patricia A. Thompson, Christine Ambrosone Chapter 8: Estrogen Receptor-Mediated Processes in Normal and Cancer Cells 135 Robert B. Dickson, George M. Stancel Chapter 9: Factors Critical to the Design and Execution of Epidemiologic Studies and 147 Description of an Innovative Technology to Follow the Progression From Normal to Cancer Tissue Montserrat Garcia—Closas, Susan E. Hankinson, Shuk-mei Ho, Donald C. Malins, Nayak L. Polissar, Stefan N. Schaefer, Yingzhong Su, Mark A. Vinson Chapter 10: Hope for Prevention—Perspective of the Cancer Advocate 157 Elizabeth A. Hart I Visit the Journal’s World Wide Web site at http://jnci.oupjournals.org/ I WM M "r” titan/rev Subscribe to the leading J journal in primary oncology research! The [oumal 0f the National Cancer Institute publishes the best papers in oncology from around the world. Internationally acclaimed as the source for the most up-to—date information from the rapidly changing fields of cancer research and treatment, the Journal of the National Cancer Institute is available by subscription from Oxford University Press. Each issue is packed with news about important cancer-related science, policy, politics, and people. And you’ll also find editorials and reviews by experts in the field, book reviews, and commentary on timely topics. Online Version www.jnci.oupjournals.org OURNAL And there’s more! Immediately on subscribing, you will become a member of the NCI Information Associates Program, which includes these outstanding benefits: 0 One copy of each new Ioumal of the National Cancer Institute Monograph* 0 24-hour access to PDQ®, the NCI’s comprehensive cancer information database 0 Listserv Subscription Service 0 ProtoCallTM Newsletter ONAL CER I TITUTE 0 Member Help Line Services NI'ORMATION ASSOCIATES |’ R O C. R A M For more information about the Information Associates Program, call 1-800-624-7890 (+301-496-7600 outside the USA). *Free to members on request prior to publication. I wish to subscribe to the [oumal of the National Cancer Institute. Name (please print) Address ZIP/Post code Country CI I enclose the correct amount (check payable to Oxford University Press). [3 Please bill me. Please charge my CI MasterCard B Visa D American Express D Diners DJCB card. Card number Illllllllllllllll Signature Exp. date 2000 SlleCl'iptiOIl Rates Volume 92 (24 issues) 1' Institutional subscribers will Individuals now receive all monographs I' automatically as they are published 8: 88: ianada 3 8m? yea 215533 55 with no need to order separately. urope ne year. If, however, you would prefer not Elsewhere C] One year: US$215 to receive this automatic service, please contact your nearest . . Journal Subscriptions Department Institutions at Oxford University Press. US & Canada B One year: US$235T P€m$§$tscribgfi Will have “1116 UK 8r Europe Cl One year: £210? oppo 1 0 or er monograp s . free of charge prior to publication. Elsewhere Cl One year. US$265T Single issues are available on request. Airexpedited delivery fee is included for subscriptions outside North America. Individual subscriptions must be paid for by personal check or credit card and sent to private addresses only. Canadian residents add 7% GST. To begin your subscription, call 1-800-852-7323 (USA Sr Canada). 01‘ call 919-677-0977. Fax: 919-677-1714. Outside the Americas, call +44 (0)1865 267907 or fax +44 (0)1865 267485. In the Americas, please return to: Oxford University Press, Journals Marketing (jnl), 2001 Evans Road, Cary, NC 27513, USA Elsewhere, please return to: Oxford University Press, Journals Marketing (jnl), Great Clarendon St., Oxford, OX2 6DP, UK OXfC {D l'NIVl RNIIY 1‘l<10’2 11012.0 *Student’s t tests were done for all possible comparisons. Lobular (Lob) areas showed significant differences between Lob 1 versus Lob 3 and Lob 2 versus Lob 3 (P<.005). TThe number of ductules per lobule was different (P<.Ol) in all the compari- sons. iThe area of each ductule per cross—section was significantly different in ductules of Lob 1 versus Lob 2 and Lob 3 (P<.01). §The number of cells per cross-section was significantly different in ductules of Lob 1 versus Lob 2 and Lob 3 (P<.01). Modified from (63). factors, responding selectively to given hormonal stimuli with either cell proliferation or differentiation. The type of response elicited is, in turn, modulated by specific topographic character- istics of the mammary parenchyma (70—78). In either case, the response of the mammary gland to these complex hormonal and metabolic interactions results in developmental changes that per- manently modify both the architecture and the biologic charac— teristics of the gland (72,73). Among all of the complex hor- monal influences, estrogens are considered to play a major role in promoting the proliferation of both the normal and the neo- plastic breast epithelium (71,72). Estradiol acts locally on the mammary gland, stimulating DNA synthesis and promoting bud formation. Although the influence of estrogens on the prolifera- tive activity of mammary epithelial cells has been traditionally considered to be mediated by at least three different mecha- nisms, a receptor-mediated (79—85), an autocrine/paracrine loop (85), and/or a negative feedback (86), it is generally accepted that the biologic activities of estrogens are mediated by the nuclear estrogen receptor (ER) that, on activation by cognate ligands, forms a homodimer with another ER—ligand complex and activates transcription of specific genes containing the es- trogen response elements (87). According to this classic model, the biologic responses to estrogens are mediated by a well— characterized ER. 19 100 [:1Lob1 Lob2 Lob3 60— % Lobular structures 20— O as Parous Parous Nulliparous Pre-Men Post-Men Pre & Post-Men Fig. 4. Percentage of lobules (Lob) type 1 (Lob 1), type 2 (Lob 2), and type 3 (Lob 3) in the breasts of parous premenopausal (Pre-Men.), parous postmeno— pausal (Post—Men), and of nulliparous premenopausal and postmenopausal (Pre & Post-Men) women. The recent cloning of a new type of ER, the ERB from the rat (88), mouse ( 89), and human (90) tissues, has required research- ers to rename the traditional ER as EROL. The presence of EROL in target tissue or cells is essential to their responsiveness to estrogen action. In fact, the expression levels of ERa in a par- ticular tissue have been used as an index of the degree of estro- gen responsiveness (91 ). A vast majority of human breast car- cinomas are initially positive for EROL, and their growth can be stimulated by estrogens and inhibited by antiestrogens (92—94). ERB and EROt share high sequence homology, especially in the regions or domains responsible for specific binding to DNA and the ligands {88—90). ERB can be activated by estrogen stimula- tion and blocked with antiestrogens (88,90,95). On activation, ERB can form homodimers as well as heterodimers with ERa (95-99). The existence of two ER subtypes and their ability to form DNA-binding heterodimers suggests three potential path— ways of estrogen signaling: via the EROL or ERB subtype in tissues exclusively expressing each subtype and via the forma- tion of heterodimers in tissues expressing both EROL and ERB (97). In addition, estrogens and antiestrogens can induce differ- ential activation of EROL and ERB to control transcription of genes that are under the control of an API element (100). The importance of the integrity of ERa in the mammary gland has been clearly elucidated by using the OL-ER knock—out (KO) mice (10]). The mammary glands of these animals are poorly developed. Nonetheless, the OL-ERKO mammary gland appears to possess the intrinsic tissue components necessary for pubertal development and pregnancy-induced maturation, but it fails to develop because of the loss of multiple stimuli that are downstream of ERoc action (10]). Unlike the dramatic underde— velopment observed in the mammary gland of the a-ERKO mouse, no such phenotype is observed in adult B-ERKO females (101). Although these studies are important for our understand- ing of the role of both ERs, extrapolation to the human breast 20 needs to be carefully done because these genetic-engineered mice reflect only a portion of the functional complexity of the human situation. Progesterone is another major, although controversial, player in mammary gland biology. This ovarian steroidal hormone also acts, in conjunction with estrogen, through its specific receptor PgR in the normal epithelium for regulating breast development. The role of these hormones on the proliferative activity of the breast, which is indispensable for its normal growth and devel- opment, has been for a long time, and still is, the subject of heated controversies. Although estrogen is known to stimulate cell proliferation, the breast epithelium of sexually mature and normally cycling women does not exhibit maximal proliferation during the follicular phase of the menstrual cycle (77, 78,102— 106), when estrogens reach peak levels of 200—300 pg/mL and progesterone is less than 1 ng/mL (107). Instead, the breast epithelium exhibits its maximal proliferative activity during the luteal phase, when progesterone levels reach 10—20 ng/mL and estrogen levels are twofold to threefold lower than those ob- served during the follicular phase (107). These observations are puzzling when analyzed to the light of in vitro and experimental data because estrogen stimulates the proliferation of cultured breast cells and breast tissues implanted in athymic nude mice. Progesterone, however, has no effect or even inhibits cell growth in the same models (105,106). In addition to its response to circulating hormones, the pro- liferative activity of the mammary epithelium in both rodents and humans varies with the degree of differentiation of the mam- mary parenchyma (68, 71—74,]08,109). In humans, the highest level of cell proliferation is observed in the undifferentiated Lob 1 present in the breast of young nulliparous females (71—74). The progressive differentiation of Lob 1 into Lob 2 and Lob 3, occurring under the hormonal influences of the menstrual cycle, and the full differentiation into lobules type 4 (Lob4) as the result of pregnancy leads to a concomitant reduction in the pro- liferative activity of the mammary epithelium (68, 71— 74,108,109). The relationship of lobular differentiation, cell proliferation, and hormone responsiveness of the mammary epithelium is just beginning to be unraveled. Of interest, the content of EROL and PgR in the lobular structures of the breast is directly proportional to the rate of cell proliferation. These three parameters are maxi- mal in the undifferentiated Lob l, decreasing progressively in Lob 2, Lob 3, and Lob 4 (Fig. 5, Table 2). The determination of the rate of cell proliferation, expressed as the percentage of cells that stain positively with Ki67 antibody, has revealed that pro- liferating cells are predominantly found in the epithelium lining ducts and lobules and less frequently in the myoepithelium and in the intralobular and interlobular stroma. Ki67-positive cells are most frequently found in Lob 1 (Fig. 5, Table 2). The per— centage of positive cells is reduced by threefold in Lob 2 and by more than 10-fold in Lob 3 (Figs. 5 and 6, Table 2) (73,110). ERa- and PgR-positive cells are found exclusively in the epi- thelium; the myoepithelium and the stroma are totally devoid of steroid receptor—containing cells. The highest number of cells positive for both receptors is found in Lob 1, decreasing pro- gressively in Lob 2 and Lob 3 (Fig. 5, Table 2) (110). To clarify the relationship between steroid receptor—positive cells and proliferating cells, we used a double—staining proce- dure, combining in the same tissue section anti—Ki67 and EROL, Ki67 and PgR, or EROL and PgR antibodies. Each antibody was Journal of the National Cancer Institute Monographs No. 27, 2000 1o 9 r 1: ER - PgR 8 r - Ki67 - ER+K167 7 — PgR+Ki67 a 6 - S 5 5 — § 1:. 4 — 3 _ 2 _ 1 _ 0 Lob.1 Lob.2 Lob.3 Fig. 5. Percentage of cells positive for estrogen receptor (ER), for progesterone receptor (PgR), for proliferating cells (K167), and for both ER and K167 (ER+Ki67), or PgR and K167 (PgR+Ki67) (ordinate). Cells were quantitated in lobule 1 (Lob l), Lob 2, and Lob 3 of the breast (abscissa). Reproduced with permission by Kluwer Academic Publishers (110). identified by its color reaction, brown with 3,3’-diaminobenzi- dine-HCl (DAB) or red with the alkaline phosphatase-vector red (110). This procedure allowed us to quantitatively determine the spatial relationship between those cells that are proliferating and those that react with either EROL or PgR antibodies. It was found that a higher percentage of cells reacted simultaneously with both EROL and PgR, appearing purple red in color (Fig. 6), whereas the number of cells positive for both EROL and Ki67 or PgR and Ki67 was very low (Table 2). The highest percentage of ERa-, PgR-, and Ki67-positive cells was observed in Lob 1 (Fig. 6, Table 2). The percentages of Ki67-, ERa-, and PgR- positive cells was reduced to 1.6%, 3.8%, and 0.7% in Lob 2, respectively. Their percentages became negligible in Lob 3 (Table 2). Of interest was the observation that even though there were similarities in the relative percentages of Ki67-, ERa- and PgR- positive cells and in the progressive reduction in the percentage of positive cells as the lobular differentiation progressed, those cells positive for Ki67 were not the same that reacted positively for EROL or PgR (Fig. 6) (110). Very few cells, less than 0.5% in Lob 1 and even fewer in Lob 2 and Lob 3, were positive for both K167 and ERa (Ki67+ER) or K167 and PgR (Ki67+PgR) (Table 2). Despite their low percentage, still double-labeled (Ki67+ER) cells were more numerous in Lob 1, decreasing gradually in Lob 2 and Lob 3. The percentage of cells exhibiting double labeling with Ki67 and PgR, however, were more numerous in Lob 2 than in Lob 1 but decreased to the same levels observed for ERor in Lob 3 (Table 2). The simultaneous immunocytochemical detection of prolif- erating cells and of those containing EROL and PgR in normal breast tissue led us to conclude that their number varies with the degree of lobular development of the organ and that steroid receptor content is linearly related to the rate of cell prolifera- tion. The use of a double-labeling immunocytochemical tech- nique has allowed us to demonstrate that the expression of the receptors occurs in cells other than the proliferating cells, con- firming results reported by others (104). The findings that pro- liferating cells are different from those that are EROL positive and PgR positive support data that indicate that estrogen controls cell proliferation by an indirect mechanism. This phenomenon has been demonstrated with the use of a supernatant of estrogen- treated Ech-positive cells that stimulates the growth of ERor— negative cell lines in culture. The same phenomenon has been shown in vivo in nude mice bearing ER-negative breast tumor xenografts (111,112). ERa-positive cells treated with antiestro- gens secrete tumor growth factor-B that inhibits the proliferation of ERa—negative cells (113). The fact that the highest prolifera- tive activity and the highest percentage of EROL- and PgR- positive cells are present in Lob 1 provides a mechanistic ex- planation for the higher susceptibility of these structures to be transformed by chemical carcinogens in vitro (4 7,1 14), support- ing as well the observations that Lob 1 is the site of origin of ductal carcinomas (115). ARCHITECTURAL PATTERN OF THE NORMAL BREAST AT MENOPAUSE Menopause supervenes as the consequence of the atresia of more than 99% of the 400 000 follicles that are present in the ovaries of a female fetus of a gestational age of 5 months (69). Gonadotropin-releasing hormone secretion is also implicated in this phenomenon, indicating that a hypothalamic process is in- volved in the development of menopause. The most character- istic sign of menopause is amenorrhea, which is the result of the almost complete cessation of ovarian estrogen and progesterone production. The years leading to the final menstrual period, until menopause sets in, generally at around the age of 51 years, constitute the perimenopause. During this period, many women Table 2. Distribution of Ki67, ER-cx, and PgR—positive cells in the lobular structures of the human breast Lobule type No. cells K167 ER PgR K167 + ER" Ki67 + PgR° Lob 1 19 3398 4.72 : 1.00%c 7.46 x 288'? 5.70 i 1.36k 0.48 1 0.28 0.09 x 0.01 Lob 2 8490b 1.58 t 0.45' 3.83 i 2.444 0.73 1 0.571 0.31 i 0.21 0.28 t 0.27 Lob 3 17 7500 0.40 1 0.1815’ 0.76 : 0.04J 0.09 i 0.04““ 0.01 i 0.01 0.01 i 0.01 aTotal number of cells counted in lobule l (Lob l) in breast tissue samples of 12 donors; btotal number of cells counted in Lob 2 in breast tissue samples of 5 donors; ctotal number of cells counted in Lob 3 in breast tissue samples of three donors; dproliferative activity determined by the percentage of cells Ki67 positive, expressed as the mean : standard deviation. Differences were significant in aLob 1 versus flob 2 (I = 1.98; P<.05). fLob 2 versus gLob 3 (t = 2.27; P<.04), and eLob 1 versus gLob 3 (t = 2.56; P<.01). Estrogen receptor (ER)-positive cells were significantly different in hLob 1 versus iLob 2 and jLob 3 (t = 2.04; P<.05). Progesterone receptor (PgR)»positive cells were significantly different in kLob 1 versus lLob 2 (I = 2.27; P<.05) and in kLob 1 versus rnLob 3 (t = 2.60; P<.03). “Percentage of cells positive for both Ki67 and ER, expressed as the mean:SD. °Percentage of cells positive for both K167 and PgR, expressed as the mean i standard deviation. Reprinted with permission by Kluwer Academic Publishers (110). Journal of the National Cancer Institute Monographs No. 27, 2000 21 Fig. 6. Ductal epithelium of the human breast. (A) Single—layered epithelium of a lobule 1 (Lob 1) ductule contains Ki67 positive cells (brown nuclei) and estrogen receptor (ER)-positive cells (red purple nuclei; x40); (B) the single— layered epithelium lining the ductule contains brown Ki67-positive cells, and red purple PgR-positive cells. The specificity of the reaction was verified by invert— ovulate irregularly, either because the rise in estrogen during the follicular phase is insufficient to trigger a LH surge or because the remaining follicles are resistant to the ovulatory stimulus (69). The increase in human longevity occurring in our society has caused a considerable increment in the number of women that will live one third or more of their lives after menopause, a period characterized by profound ovarian hormone deprivation. After menopause the breast undergoes regression in both nul- liparous and parous women. This regression is manifested as an increase in the number of Lob 1 and a concomitant decline in the number of Lob 2 and Lob 3. At the end of the fifth decade of life, the breast of both nulliparous and parous women is composed predominantly of Lob 1 (Fig. 4) (68). These observations have led us to conclude that the understanding of breast development requires a horizontal study in which all different phases of growth are taken into consideration. For example, the analysis of breast structures at a single given point, i.e., age 50 years, would lead one to conclude that the breasts of both nulliparous and parous women are identical. However, the phenomena occurring in prior years might have imprinted permanent changes in the breast, which affect its susceptibility to carcinogenesis but are no longer morphologically observable. Thus, from a quantitative point of View, the regressive phenomenon occurring in the breast at menopause differs between nulliparous and parous women. In the breast of nulliparous women, the most predominant structure is the Lob 1, which comprises 65%—80% of the total lobule type components and their relative percentage is independent of age. Second in frequency is the Lob 2, and the least frequent structure is the Lob 3, which represent 10%—35% and 0%—5% of the total 22 ing the order of the stains, i.e., (C) and (D) ER-positive cells, brown, Ki67- positive cells, purple red; (E) brown nuclei of PgR-positive cells and a Ki67- positive cell in mitosis appears stained purple red (DAB-Hematoxylin; x40). Reproduced with permission by Kluwer Academic Publishers (110). lobular population, respectively. In the breast of premenopausal parous women, however, the predominant lobular structure is the Lob 3, which comprises 70%—90% of the total lobular com— ponent. Only after menopause the number of Lob 3 declines and the relative proportion of the three lobular types approaches that observed in nulliparous women. Full lobular differentiation only occurs in the parous women, especially in those completing full-term pregnancy at a young age, but lobular differentiation in nulliparous women seldom reaches the Lob 3 and never the Lob 4 stages ( 68). These differences in the pattern of breast devel— opment between nulliparous and parous women greatly explain the protective effect induced by pregnancy from breast cancer development. They also highlight the need to determine whether the undifferentiated Lob 1 of nulliparous women differ from those of the parous postmenopausal woman in their ability to metabolize estrogens or in the ability of the cells to repair geno— toxic damage (116,] I 7). ARCHITECTURAL PATTERN OF THE BREAST WITH PROLIFERATIVE DISEASE Our studies of the pattern of breast development in tissues devoid of mammary pathology, such as those obtained from reduction mammoplasties, led us to establish certain criteria of normality specific for a given age and parity status of the donors. In these tissues, we identified parenchymal structures exhibiting variations in the degree of differentiation, rate of cell prolifera- tion, and content of ERa and PgR (118). To answer the question of whether breast lesions of either benign, premalignant, or ma— Journal of the National Cancer Institute Monographs N0. 27, 2000 lignant nature develop as a reflection of the stage of develop- ment of the breast, we compared parenchyma] structures present in 33 reduction mammoplasty specimens with those found in 45 breast biopsies performed because of mammographic abnor- malities or clinically suspicious breast masses (Table 3). Be- cause the initiation of the neoplastic process is inversely related to the degree of differentiation of the breast, which in turn is a function of reproductive history, the patient populations were subdivided according to parity status (Table 3). In this study, we confirmed previous observations that in the reduction mammo- plasty specimens (RM) the breast of nulliparous women of all ages was composed predominantly of Lob 1, whereas the breast of parous premenopausal women contained a higher concentra- tion of Lob 3 (Table 3) (118). Those breast tissues obtained from biopsies had an architectural pattern different from that obtained from RM for women of comparable parity status because parous women that had a breast biopsy contained a higher percentage of Lob l and a lower percentage of Lob 3 than the parous popu- lation of the RM group (Table 3) (118). The patient population that had breast biopsies was also sub~ divided into subgroups, based on the histopathologic diagnosis of their lesions. One group of 21 patients had no pathology present (normal breast or control group), one group of 15 pa- tients had ductal hyperplasia (DH group), four patients had blunt duct adenosis (BDA group), and five patients had sclerosing adenosis (SAD group) (Table 4). Tissue sections from all these groups were analyzed for lobular architecture, type of pathologic lesions, and proliferative activity of the breast (Figs. 7 and 8) (118). The breast tissues of the groups classified as normal breast (control) and DH had a significantly higher percentage of Lob 1 than Lob 2 and Lob 3 (P<.0008 and P<.0001, respec— tively) (Fig. 7). The breast tissues containing BDA were also characterized by having a higher percentage of Lob 1, whereas the SAD group had a higher percentage of Lob 2 (Table 4, Fig. 7) (118). In all of the groups, the percentage of Lob 3 was significantly lower than that of Lob 1, although the relative percentage of Lob 3 was significantly higher in breast biopsies with BDA and SAD than in the normal breast biopsies or in those diagnosed with DH (P<.05) (Fig. 7) (118). The number of proliferating epithelial cells, determined by immunostaining of the Ki67 nuclear antigen, was higher in Lob 1 than in Lob 2 and Lob 3 (P<.001), with a similar pattern in the normal breast, DH, and SAD groups, although the differences were not statistically significant between Lob 1 and Lob 2 in the SAD group (Fig. 8). In the BDA group, however, the rate of cell proliferation was higher in Lob 2 (P<.01) than in Lob l and Lob 3 (Fig. 8) (118). Although the proliferative activity was on an average higher in Lob 1 than in Lob 3, the differences were not statistically significant (Fig. 8) (118). These data allowed us to conclude that breast tissues obtained from biopsies performed because of mammographic or clinical abnormalities, even in the absence of cancer, have architectural and cell kinetic patterns different from the normal breast tissues obtained from reduction mammoplasties. More important is the observation that even in those cases in which no pathology or only benign lesions were diagnosed, the pattern of breast development in biopsies was more similar to that of the cancer—bearing breast than it is to the population not requiring a biopsy. Our findings that in DH-containing biopsies Lob l are the most frequent structures present and have the highest rate of cell proliferation support our postulate that DH originates from Lob l (115). Lob 2 and Lob 3, which are the sites of origin of more differentiated lesions, such as BDA and SAD, are more prominently represented and are more proliferative in those biopsies containing these types of pathologic lesions. It is of importance to clarify that parity does not seem to influence the pattern of development in DH-containing breast tissues. A similar observation has been made in cancer-bearing breasts or in breasts of nulliparous women in terms of lobular composition. BREAST DEVELOPMENT, HORMONEs, AND THE PATHOGENEsrs OF BREAST CANCER From our studies associating normal breast development and the pathogenesis of both experimental and spontaneous mam— mary carcinogenesis emerged an important concept: that the Lob l, the most undifferentiated structure found in the breast of young nulliparous women, is equivalent to the terminal ductal lobular unit, a structure originally identified by Wellings et al. (119) as the site of origin of ductal carcinomas (Fig. 9) (63,70). This observation was supported by comparative studies of nor— mal and cancer-bearing breasts obtained at autopsy. We ob- served that the nontumoral parenchyma of those breasts that had developed a malignancy contain a significantly higher number of hyperplastic terminal ducts, atypical Lob l, and ductal carci- nomas in situ originating from Lob 1 than those breasts free of malignancies. These findings indicate that the Lob l is affected by both preneoplastic and neoplastic processes (115). More dif- ferentiated lobular structures have been found to be affected by neoplastic lesions as well, although they originate tumors whose histologic type and malignancy are in an inverse relationship with the degree of differentiation of the parent structure (52,115,120,121). The finding that the most undifferentiated structures originate the most aggressive neoplasms is clinically important because these structures are more numerous in the breasts of nulliparous women who are, in turn, at a higher risk of developing breast cancer (68). Table 3. Lobular architecture of the breast—comparison of percentages of structures found in reduction mammoplasties and breast biopsies with proliferative breast disease Group (parity) No. of cases Age, y Lob l (%) Lob 2 (%) Lob 3 (%) RM (all) 33 29.4 x 8.2 22.45 i 23.73 37.25 i 28.61° 38.41 i 34.22e RM (nulliparous) 9 22.9 t 6.7 45.87 i 27.40 47.17 i 2201 6.94 1 7.01 RM (parous) 24 31.9 i 2.3 16.92 : 8.26g 35.45 i 3.14i 47.86 i 33.4k PBD (all) 45 46.6 i 1.5 65.66 :- 34.15b 24.64 i 20.64d 9.68 i 6.31 " PBD (nulliparous) 10 42.5 i 10.3 70.99 x 33.3 25.26 i 24.74 3.75 i 1.6 PBD (parous) 35 48.9 i 11.8 65.25 i 37.3h 21.10 i 8.07j 13.62 i 3.101 The differences between a and b is P<.0000005; the difference between C and d is P<.O4; between C and f is P<.00009; between 3 and h P TTA) of codon 254 in exon 7 of this gene was detected (Table 6) (148). There was a coexistence of a mutation of TP53 and instability of microsat— ellite DNA in the intragenic TP53 in these cells, manifested by additional bands with slower mobility. We used polymerase chain reaction (PCR) amplification of microsatellite DNA length polymorphism to detect allelic loss as well as microsatellite instability (MSI). These microsatellites are highly polymorphic, flanked by unique sequences that can serve as primers for PCR amplification. They have been proven to be useful markers for investigating multiple areas of MSI and loss of heterozygosity (LOH) and should be applicable to allelotyping as well as re— gional mapping of deletions in specific chromosomal regions. We have studied 466 markers that represent approximately 4.6% of the 10000 microsatellite markers identified. Microsatellite PCR analysis of MCF-lOF cells did not reveal LOH with any of the markers analyzed in this study when compared with their parental MCF-IOM cells. However, MCF-lOF cells did show MSI in chromosome 11 in the locus DIIS392 and in chromo— some 17 in the loci represented by markers D17SS49, TP53, D17S786, and D17SS20 (149). Neoplastic transformation of MCF-lOF cells with chemical carcinogens (e.g., BPl and BPlE cells) is associated with ge- netic instability on chromosomes 11 and 13, in addition to that observed on chromosome 17, which has been detected in asso- ciation with immortalization (150—153). MSI was found on chromosome 11 by using marker D1 IS912 and expressed as an allelic expansion in the BPl and BPlE cells, representing an additional location affected in the early stage of transformation of HBECs (Fig. 14) (151). On chromosome 13, MSI was found in both BPl and BPIE cells by using markers D13S260 and D135289 at 13ql2—13 (flanking the BRCA2 locus) (Fig. 14) (150). In addition, we have also observed other genomic alter— ations on chromosomes 9 and 16 (154). Activation of Telomerase in the Immortalized MCF-10F Cells There is evidence that the repetitive TTAGGG sequences located at the ends of human chromosomes (i.e., telomeres) may act as a molecular mitotic clock (155). It is generally believed that each successive genomic replication is accompanied by gradual shortening of 50—200 base pairs (bp) because of incom- plete replication of the 3’ ends, and cellular senescence occurs when telomeres reach a critically short length that replication of the genome cannot be maintained (156). The stabilization of the telomeric sequences at the ends of chromosomes, which is re- quired for the continuous proliferation of immortal cells, in- volves the activation of the enzyme telomerase, which adds TTAGGG repeats to the 3’ ends of chromosomes (157,158). The genetic nature of cellular senescence implicates activation of telomerase as a key element of cell immortalization (144,158). Elevated levels of telomerase activities have been detected in a number of immortal cell lines and human tumor tissues (159,160). We have observed telomerase activity in immortal MCF-10F but not in the mortal MCF-lOM cells (161), suggest- ing that telomerase activation may play a role in the spontaneous immortalization of MCF-lOF cells. Increase of H-Ferritin in MCF-IOF Cells In efforts to identify genes underlying the process of immor- talization, we have performed subtractive hybridization and dif- ferential display analysis between immortal MCF—lOF and its parental mortal MCF-IOM cells. With the use of a 10F(+)/ 10M(_) subtractive complementary DNA (cDNA) library, we isolated more than 15 clones, one of which contains sequences identical to H-ferritin (162). We observed marked increases in messenger RNA (mRNA) levels of ferritin H in immortal MCF- 10F cell lines (particularly in late passages) (Fig. 15) and in tissues exhibiting an increase in growth rate, such as ductal hyperplasia, carcinoma in situ, and invasive carcinoma (Fig. 16) (162). An increase in transcript signal was also confirmed by in situ hybridization of breast tissues containing lesions represen- tative of progressive stages of neoplastic evolution (Fig. 17). The levels of expression were undetectable in normal tissues; they increased progressively from moderately elevated in DH, a stage of cell progression from normal to neoplasia that may be a histopathologic parallel of cell immortalization, greater expres- sion in carcinoma in situ, with the highest transcript levels being detected in infiltrating ductal carcinoma (Fig. 17). Ferritin is a large protein found in most cell types of verte- brates, as well as of invertebrates, plants, and bacteria (163). The main function of ferritin is iron storage. This function, in turn, can be subdivided into iron storage for other cells (specialized- cell ferritin), iron storage for intracellular needs (normal house- keeping ferritin), and iron storage for intracellular protection Control DMBA Fig. 13. Representative colonies formed in agar-methocel by chemically transformed primary cultures of human breast epithelial cells obtained from women with familial history of breast cancer. The cells were treated with the chemical carcinogens benz[a]pyrene (BP) or 7,12—dimethylbenz[a]anthracene (DMBA) prior to plating (phase contrast photograph, x508). Reproduced with permission by the Society for In Vitra Biology (147). Journal of the National Cancer Institute Monographs No. 27, 2000 29 Table 6. Comparison of the sequence of p53 exon 7 in MCF—lOM and MCF—lOF cells Codon N 0. Cells 253 254 255 256 257 258 259 260 261 262 263 MCF-10M* Antisense TGG TAG TAG TGT GAC CTT CTG AGG TCC agt cct Sense ACC ATG ATG ACA CTG GAA GAC TCC AGG tea gga Aminoacid Thr Ileu Ileu Thr Leu Glu Asp Ser Ser MCF- 10F]L Antisense TGG TTA GTA GTG TGA CCT T CT GAG GTC CAG TCC Sense ACC AAT CAT CAC ACT GGA AGA CTC CAG GTC AGG Aminoacid Thr Asn His His Thr Gly Arg Leu Gln Val Arg *MCF-lOM cells exhibit a wild-type sequence. TMCF-IOF cells show a frame-shift mutation by insertion of a base at codon 254. Reprinted with permission from International Journal of Oncology (148). 123456 123456 123456 Fig. 14. Microsatellite instability detected in chromosome 11 and 13 with markers D1 189129 (left panel), D13S26O (central panel), and D138289 (right panel) in MCF—lOM (p22), lane 1; MCF-lOF (P130), lane 2; BP1 (p27, p52), lane 3; BPlE (p25), lane 4; BPlE (p30), lane 5, and BPlE (P60), lane 6. S#130 240 229 MCF-10F BP1 E BP1 ET D3 D31 MCF7 HBL-100 BT—20 T-47D +1.1kb _ «cm Fig. 15. Northern analysis of primary cultures of breast epithelial cells S#130, 240, and 229; immortal MCF-lOF cells; benz[a]pyrene-transfonned cells BPlE and BPlET; dimethylbenz[a]anthracene—transformed S3 and D31 cells; and the neoplastic cell lines MCF—7, HBL-lOO, BT-20, and T—47D. An increased signal for H-ferritin messenger RNA is observed in MCF-IOF cells; a much higher intensity in BPlE, BPlET, D3, and D3—1; and the highest signal intensity is observed in the malignant cell lines MCF—7, HBL-lOO, BT-20, and T—47D. Reproduced with permission by Wiley—Liss, Inc., a subsidiary of John Wiley & Sons, Inc. (162). from iron overload (stress housekeeping ferritin) (163). Iron is required for DNA synthesis necessary for cell growth and mu]— tiplication (164—166). It is also required for electron transport and for oxygen metabolism, generating harmful activated oxy- gen species capable of damaging DNA, lipids, and proteins (167). The iron—catalyzed conversion of H202 is a major route to the synthesis of highly reactive OH radicals that inflict damage on the nucleotide bases of DNA, inducing mutations and in- creasing the risk of cancer (42,168). 30 B1 M1 32 M2 N1 N2 H1 H2 4-1.1kb _+a°tm Fig. 16. Northern analysis of total RNA from breast cancer tissue (M1 and M2), normal tissue from the same patient (B1 and B2), normal control breast tissue (N1 and N2), and tissue with ductal hyperplasia (H1 and H2). A high signal intensity is seen in the malignant tissue, very low levels are detected in normal tissue samples, and an increase in signal above normal levels is seen in tissue showing ductal hyperplasia. Reproduced with permission by Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc. (162). Normally, the concentration of iron capable of catalyzing these reactions is tightly controlled and regulated, and a critical homeostatic balance is achieved by the synthesis of ferritin (163,169). Many compounds, such as flavins and xanthine oxi- dase, are capable of reductively releasing iron from ferritin (170,171). Once released from ferritin, iron in the ferrous (Fe++) state is capable of participating in free-radical reactions (Fenton reactions) leading to oxidative damage (Fig. 18). Thus, a dis- ruption in normal iron homeostasis may lead to an increase in the level of reactive iron and to a corresponding increase in oxygen free-radical generation and DNA damage (1 72). Oxidative stress has also been implicated in metastasis, because it results in loss of cell adhesion, a prerequisite for cell detachment and subse- quent host tissue invasion (173,174). Progression of invasive breast cancer to the metastatic state is Journal of the National Cancer Institute Monographs No. 27, 2000 32:53 éiypeegniamha Fig. 17. In situ hybridization (upper row) and Northern blot analysis of Ferritin H chain messenger RNA (lower row) in normal breast lobules (Normal Breast), Ductal Hyperplasia, Ductal Carcinoma in situ, and Invasive Carcinoma. Tissue sections were counterstained with hematoxylin and photographed in dark field (final magnification x100). 'OH 6—— Fe++ @— Ferritin H . . Change Immune Ox1dative Surface Antigen DNA Damage Alter Cell—Cell Interactions Escape Immune Mutation DNA Surveillance Synthesis Neoplastic ; Clonal . ‘ . Transformation Expansion ‘9 InvaSion _'—> Metasta31s Fig. 18. The possible role of increased ferritin H chain gene in the neoplastic transformation of human breast epithelial cells. Adapted from (162). linked to OH radical-induced DNA damage (175—177). Thus, ferritin—dependent oxidative damage to DNA may be one of the mechanisms contributing to immortalization of HBECs. An in- crease in ferritin H chain levels may provide iron necessary for the clonal selection and uncontrolled growth of cells. It has been shown that iron and its binding proteins participate in a variety of reactions required for cell proliferation (56,178) and are criti— cal for the activity of the enzyme ribonucleotide reductase, a rate—limiting enzyme in DNA synthesis ( 165,179). Ferritin has been shown to have an immunosuppressive effect on host im- mune response in cancer patients (180,181). Placental isoferritin, an acidic form of ferritin. and its p43 super heavy chain have Journal of the National Cancer Institute Monographs No. 27. 2000 been reported to be synthesized by breast cancer cells but are absent in normal breast epithelium (182). Breast cancer- associated p43 induces alterations in the expression of cell- surface molecules in neoplastic cells, which in turn could have an effect on the modulation of the cells’ adhesive interactions (183). Cytokines, such as tumor necrosis factor, interleukin-10L, and the NF—KB family of transcription factors, specifically in- duce synthesis of ferritin H by selectively increasing ferritin H transcription (Fig. 18) (183—185). Our observations that the immortalization of HBECs was associated with an increased ferritin H chain gene transcription led us to postulate that this increase might have contributed to 31 the immortalization of HBEC; probably through one or more of the following mechanisms: a) providing a source of iron re— quired by rapidly dividing cells for clonal expansion, b) provid- ing iron capable of participating in free—radical reactions leading to oxidative DNA damage and mutation, or c) affecting immune surface antigens and thus providing immortal cells a growth advantage by allowing them to escape immune surveillance (Fig. 18). However, the possibility that ferritin H chain gene induction may be a consequence of the immortalized condition of the cells, rather than its cause, cannot be ruled out. In either instance, it may prove to be a valuable marker of cell immortalization or an early indicator of malignant transformation. Reversion of Immortalized and Transformed Phenotypes To investigate the functional role of genomic changes on chromosomes 11 and 17 that were detected in the immortalized and transformed cells, single normal human fibroblast A9— derived chromosome 11 or 17, tagged with a neomycin-resistant gene, was transferred into 6 x 106 transformed BPlE cells. Sur— viving cells or clones of microcell hybrids from chromosome 11 or 17 were designated BPlE-l lneo or BPlE-17neo cells. A total of 16 colonies was isolated from BPlE-l lneo and BPlE—l7neo cells each. The transfer efficiency in BPlE cells was approxi- mately 2.6 x 10“6 cells (149). During a selection period of up to 6 months, BPlE-17neo cells, and to a lesser degree BPlE-l lneo cells, exhibited altered cellular morphology and growth pattern, such as contact inhibition and cellular senescence (149). In ad- dition to the acquired ability to survive in the G-418 selection medium that indicates the active function of the neomycin- resistance gene tagged on the donor chromosome 11 or 17, the physical presence of these chromosomes was further con- firmed by dual-color fluorescence in situ hybridization (FISH) analysis. As shown in Fig. 19, A and B, the metaphases of the BPlE-l lneo#145 cells were confirmed to contain an extra chro- mosome 11 that had a stronger signal with the painting probe (red). Anchorage-independent growth in agar—methocel gel was reduced from 17% in control BPlE cells to 7% in the BPlE-l lneo#145 cells, whereas BPlE-l7neo D100 cells failed is a Fig. 19. Dual—color fluorescence in situ hybridization (FISH) for the detection of chromosome 11 in metaphases of BPlE—lneo #145 cells. (A) A representative field of a metaphase spread showing a pair of host chromosome ll (red stained) and a donor (partial) chromosome 11 (red stained with green signal). This clone contained one extra donor chromosome 1 1, in addition to the pair of host chro— mosome 11. (B) Representative field of a metaphase spread showing a pair of chromosome 11 (red stained) in BPlE cells. 32 to form any colony (100% reduction), reflecting a more potent suppression of the BPlE cells by chromosome 17 than that by chromosome 11. These data indicated that the introduced chro- mosome 11 caused a partial growth inhibition, whereas chromo- some 17 produced a nearly complete growth suppression of the BPlE cells (149). Another phenotypic reversion induced by the chromosome transfer was the recovery of the ability of cells to form ductule-like structures in collagen gel, a property exhibited by BPlE-17neo D100 cells, similar to that of MCF-lOF cells, whereas BPlE cells grew in loosely arranged clusters or as isolated cells. These observations confirmed the reversion of the transformed phenotype by chromosome 17 transfer. Microsatellite analysis showed that the preexisting instability in the parental BPlE cells at loci D17SS49 (l7pl3.3), TP53 (l7p13.1), Dl7S786 (17p13.1), and D178520 (17p12.0) was reverted in BPlE-l7neo D100 cells, which acquired an allelic pattern similar to that of the mortal MCF-lOM cells. In con— trast, the instability of these markers was not restored in the BPlE~11neo #145 cells. These data indicated a specific effect associated with transfer of chromosome 17. Surprisingly, none of the corresponding donor alleles observed in the A9—l7neo cells could be detected in the BPlE-17neo D100 cells, suggesting that other untested regions of the donor chromosome 17 might be the responsible ones for the phenotypic reversion and the restored microsatellite stability. In summary, our data provide supportive evidence for the hypothesis that MSI within or near genes can confer instability to these genes and alter their expression or functions. However, further investigation is required for determining what is their functional role in the initiation and progression of neoplasia. TP53 has been considered as the guardian of the genome by allowing cells to undergo DNA repair prior to entering a new cell cycle (186—188). The observation that introduction of an unaffected chromosome 17 can correct instability on the corre- sponding chromosome, including that of marker TP53, in addi— tion to the reversion of transformed phenotypes in the trans- formed BPlE cells, suggests that other important genes on chromosome 17 may control this process. FUTURE PERSPECTIVES In the paradigm described above, it is clear that if estrogens play a role in the early stages of cell immortalization and trans— formation, this experimental system will allow us to demonstrate such phenomena. The demonstration of the ability of the mam- mary epithelial cells to metabolize estradiol and/or to accumu- late “genotoxic” metabolites could profoundly influence our un- derstanding of the neoplastic transformation of the mammary epithelium (189). Metabolic biotransformation of estradiol oc- curs in human mammary explant cultures composed of a mixture of epithelial and stromal cells (190,191). Treatment of normal mouse mammary epithelial cells with the mutagenic polycyclic hydrocarbon DMBA results in production of l6a-hydroxy- estrone. This predominant metabolite of estrogen increases un- scheduled DNA synthesis, cellular proliferation, and anchorage- independent growth, all phenomena indicative of preneoplastic transformation (192). Because normal HBECs are susceptible to be transformed by environmental carcinogens that require metabolic activation (116,117) and many of the enzymes (e.g., CYP 1A1) that catalyze the oxidation of drugs, alkaloids, and environmental pollutants also catalyze the hydroxylation of es- trogens (136,193,! 94), we hypothesize that HBECs, regardless Journal of the National Cancer Institute Monographs No. 27, 2000 of their ER status, are capable of metabolic activation of estro- gen and, thus, susceptible to estrogen-induced carcinogenesis. It is possible that the rates of metabolic activation of estrogen might vary among HBECs with different carcinogenic suscep- tibility. We postulate that the susceptibility of cells to be trans- formed by estrogens would depend on their rate of proliferation, genetic predisposition, and mortal status of the cells, rather than their ER contents, similar to what has been observed with chemi- cal carcinogens (47,II4,I46,147). If these assumptions are true, the efficiency and extent of estrogen-induced neoplastic trans- formation will be high in immortalized MCF-lOF cells, moder— ate in those HBECs derived from breasts of women with family history of breast cancer, and low in cells derived from the breast of parous women and of those women with no family history of breast cancer. The independence of ER contents in estrogen- induced carcinogenesis would support the postulate that meta— bolic activation of estrogen is involved in the neoplastic trans— formation of susceptible HBECs. Alternatively, estrogen or its metabolites may not initiate neoplastic transformation, but they may act by promoting neoplastic progression in chemically transformed HBECs by increasing the genomic instability of the cells. In essence, estrogen, on metabolic activation, might serve as an initiator and/or as a promoter of carcinogenesis in the human breast. However, the metabolism of estrogen in normal HBECS and the carcinogenic potential of estrogen and its me- tabolites in the human breast are virtually unknown. REFERENCES (1) Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1999. CA Cancer J Clin 1999;49:8—31. (2) Mettlin C. Global breast cancer mortality statistics. CA Cancer J Clin 1999;49:138-44. (3) Pike MC, Spicer DV, Dahmoush L, Press MF. Estrogens, progestogens, normal breast cell proliferation, and breast cancer risk. Epidemiol Rev 1993;15:17—35. (4) Bernstein L, Ross RK. Endogenous hormones and breast cancer risk. Epidemiol Rev 1993;15:48—65. (5) Hu, YF, Russo IH, Russo J. Prevention of human breast cancer. J Wom— en’s Cancer. In press 2000. (6) Henderson BE, Ross R, Bernstein L. Estrogens as a cause of human cancer: the Richard and Hinda Rosenthal Foundation award lecture. Can- cer Res 1988;48:246—53. (7) Greenwald P, Barlow JJ, Nasca PC, Burnett WS. Vaginal cancer after maternal treatment with synthetic estrogens. N Engl J Med 1971;285: 390—2. (8) Nissen ED, Kent DR. Liver tumours and oral contraceptives. Obstet Gy- necol 1975;46:460-7. (9) Herbst AL. Clear cell adenocarcinoma and current status of DES-exposed females. Cancer 1981;48(2 suppl):484—8. (10) Shaw RW. Adverse long-term effects of oral contraceptives: a review. Br J Obstet Gynaecol 1987;94:889794. ( I I ) Chilvers C, Mant D, Pike MC. Cervical adenocarcinoma and oral contra- ceptives. Br Med J 1987;295:1446—7. (12) Beral V, Hannaford P, Kay C. Oral contraceptive use and malignancies of the genital tract. Results from the Royal College of General Practitioners’ Oral Contraceptive Study. Lancet 1988;221331—5. ([3) Yan ZJ, Roy D. Mutations in DNA polymerase [3 mRNA of stilbene estrogen-induced kidney tumors in Syrian hamster. Biochem Mol Biol Int 1995;37:175—823. (14) Zhu BT, Liehr JG. Quercetin increases the severity of estradiol-induced tumorigenesis in hamster kidney. Toxicol Appl Pharmacol 1994;125: 149—58. (15) Tsutsui T, Barrett JC. Neoplastic transformation of cultured mammalian cells by estrogens and estrogenlike chemicals. Environ Health Perspect 1997;105(suppl 3):6l9—24. (16) Adlercreutz H, Gorbach SL, Goldin BR, Woods MN, Dwyer JT, Ha- Journal of the National Cancer Institute Monographs No. 27, 2000 malainen E. Estrogen metabolism and excretion in Oriental and Caucasian women. J Natl Cancer Inst 1994;86:1076—82. (17) Nandi S, Guzman RC, Yang J. Hormones and mammary carcinogenesis in mice rats and humans: a unifying hypothesis. Proc Natl Acad Sci U S A 1995;92:3650’7. ( 18) Liehr JG, Ulubelen AA, Strobe] l-IW. Cytochrome P—450-mediated redox cycling of estrogens. J Biol Chem 1986;261:16865—70. (19) Roy D, Liehr JG. Temporary decrease in renal quinone and reductase activity induced by chronic administration of estradiol to male Syrian hamsters. Increased superoxide formation by redox cycling of estrogen. J Biol Chem 1988;263:3646—51. (20) Ball P, Knuppen R. Catecholestrogens (2» and 4-hydroxyoestrogens): chemistry, biosynthesis, metabolism, occurrence and physiological sig- nificance. Acta Endocr Suppl (Copenh) 1980;232:1—127. (2]) Zhu BT, Bui QD, Weisz J, Liehr JG. Conversion of estrone t0 2- and 4-hydroxyestrone by hamster kidney and liver microsomes: implications for the mechanism of estrogen—induced carcinogenesis. Endocrinology 1994;135:1772—9. (22) Ashburn SP, Han X, Liehr JG. Microsomal hydroxylation of 2- and 4—fluoroestradiol to catechol metabolites and their conversion to methyl ethers: catechol estrogens as possible mediators of hormonal carcinogen— esis. Mol Pharmacol 1993;43:534—41. (23) Bmeggemeier RW, Katlic NE, Palmer CW Jr, Stevens JM. Catechol estrogen formation in MCF—7 cell culture and effects of bromoestrogen inhibitors. Mol Cell Endocrinol 1989;64:161—7. (24) N iwa T, Bradlow HL, Fishman J, Swaneck GE. Determination of estradiol 2— and l6—alpha-hydroxylase activities in MCF-7 human breast cancer cells in culture using radiometric analysis. J Steroid Biochem 1989;33: 311—4. (25) Spink DC, Hayes CL, Young NR, Christou M, Sutter TR, Jefcoat CR, et al. The effects of 2,3,7,8—tetrachlorodibenzo-p—dioxin on estrogen metabo- lism in MCF-7 breast cancer cells: evidence for induction of a novel l7B—estradiol 4-hydroxylase. J Steroid Biochem Mol Biol 1994;51:251—8. (26) Liehr JG, Ricci MJ. 4-Hydroxylation of estrogens as a marker of human mammary tumors. Proc Natl Acad Sci U S A 1996;93:3294—6. (27) Knuppen R, Ball P, Emons G. Importance of A-ring substitution of es— trogens for the physiology and pharmacology of reproduction. J Steroid Biochem 1986;24:193—8. (28) Creveling CR, Inoue K. Catechol—0—methyltransferase: factors relating to the carcinogenic potential of catecholestrogens. Polycyclic Aromat Compd 1994;6:253—9. (29) Dwivedy I, Devanesan P, Cremonesi P, Rogan E, Cavalieri E. Synthesis and characterization of estrogen 2,3— and 3,4—quinones. Comparison of DNA adducts formed by the quinones versus horseradish peroxidase- activated catechol estrogens. Chem Res Toxicol 1992;5:828—33. (30) Stalford AC, Maggs JL, Gilchrist TL, Park BK. Catecholestrogens as mediators of carcinogenesis: correlation of aromatic hydroxylation of es- tradiol and its fluorinated analogs with tumor induction in Syrian ham- sters. Mol Pharmacol 1994;45:1259—67. (31) Weisz J, Bui QD, Roy D, Liehr JG. Elevated 4—hydroxylation of estradiol by hamster kidney microsomes: a potential pathway of metabolic activa- tion of estrogens. Endocrinology 1992;131:655—61. (32) Bunyagidj C, McLachlan JA. Catechol estrogen formation in mouse uterus. J Steroid Biochem 1988;31:795—801. (33) Paria BC, Chakraborty C, Dey SK. Catechol estrogen formation in the mouse uterus and its role in implantation. Mol Cell Endocrinol 1990;69: 25—32. (34) Bui Q, Weisz J. Identification of microsomal, organic hydroperoxide- dependent catechol estrogen formation: comparison with NADPH depen- dent mechanism. Pharmacology l988;36:356~64. (35) Zhu BT, Liehr JG. Inhibition of the catechol-0—methyltransferase— catalyzed 0-methylation of 2— and 4-hydroxyestradiol by catecholamines: implications for the mechanism of estrogen—induced carcinogenesis. Arch Biochem Biophys 1993;304:248—56. (36) Zhu BT, Liehr JG. Inhibition of catechol 0-methyltransferase—catalyzed 0—methylation of 2— and 4—hydroxyestradiol by quercetin. Possible role in estradiol—induced tumorigenesis. J Biol Chem 1996;271:1357—63. (37) Bradlow HL, Fishman J, Telang NT, Osborne MP. Re: Estrogen metabo— lism and excretion in Oriental and Caucasian women [letter]. J Natl Can— cer Inst 1994;86:1643—5. 33 (38) Schneider J, Kinne D, Fracchia A, Pierce V, Anderson KE, Bradlow HL, et a1. Abnormal oxidative metabolism of estradiol in women with breast cancer. Proc Natl Acad Sci U S A 1982;79:3047—51. (39) Bradlow HL, Hershcopf J, Martucci C, Fishman J. 16a—hydroxylation of estradiol: a possible risk marker for breast cancer. Ann N Y Acad Sci 1986;464:138—51. (40) Osborne MP, Bradlow HL, Wong GY, Telang NT. Upregulation of es- tradiol C—160L—hydroxylation in human breast tissue: a potential biomarker of human breast cancer risk. J Natl Cancer Inst 1993;85:1917—20. (41) Malins DC, Haimanot R. Major alterations in the nucleotide structure of DNA in cancer of the female breast. Cancer Res 1991;51:5430—2. (42) Malins DC, Holmes EH, Polissar NL, Gunselman SJ. The etiology of breast cancer. Characteristic alteration in hydroxyl radical—induced DNA base lesions during oncogenesis with potential for evaluating incidence risk. Cancer 1993;71:3036—43. (43) Sipe HJ Jr, Jordan SJ, Hanna PM, Mason RP. The metabolism of 17(3— estradiol by lactoperoxidase: a possible source of oxidative stress in breast cancer. Carcinogenesis 1994;15:2637—43. (44) Black DM. The genetics of breast cancer. EurJ Cancer 1994;30Asl957—61. (45) Van de Vijver MJ. Molecular genetic changes in breast cancer. In: Wein- berg RA, editor. Oncogenes and the molecular origins of cancer. Cold Spring Harbor (NY): Cold Spring Harbor Press; 1993. p. 25—6. (46) Murray AW, Edwards EM, Hii CS. Tumor promotion: biology and mo— lecular mechanisms. In: Cooper CS, Grove PL, editors. Chemical carci— nogenesis and mutagenesis. Berlin (Germany): Springer Verlag; 1990. p. 135—49. (47) Russo J, Calaf G, Russo IH. A critical approach to the malignant trans- formation of human breast epithelial cells with chemical carcinogens. Crit Rev Oncog 1993;42403—17. (48) Russo J, Barnabas N, Zhang PL, Adesina K. Molecular basis of breast cell transformation. Radiat Oncol Investig 1996;3z424—9. (49) Russo J, Yang X, Hu YF, Bove BA, Huang Y, Silva ID, et a1. Biological and molecular basis of human breast cancer. Front Biosci 1998;3z D944—60. (50) Tait L, Soule H, Russo J. Ultrastructural and immunocytochemical char- acterization of an immortalized human breast epithelial cell line, MCF—lO. Cancer Res 1990;50:6087—94. (51) Russo J, Russo IH. A comprehensive approach to mammary carcinogen- esis. Proc Int Assoc Breast Cancer Res 1989,15. (52) Russo J, Russo IH. Role of differentiation on transformation on human breast epithelial cells. In: Medina D, Kidwell W, Heppner G, Anderson E, editors. Cellular and molecular biology of mammary cancer. New York (NY): Plenum Publishing; 1987. p. 399—417. (53) Russo J, Mills MJ, Moussalli MJ, Russo IH. Influence of human breast development on the growth properties of primary cultures. In Vitro Cell Develop Biol 1989;25:643—9. (54) Ham RG, McKeehan WL. Development of improved media and culture conditions for clonal growth of normal diploid cells. In Vitro 1978;14: 11—22. (55) Hammond SL, Ham RG, Stampfer MR. Serum—free growth of human mammary epithelial cells: rapid clonal growth in defined medium and extended serial passage with pituitary extract. Proc Natl Acad Sci U S A 1984;81:5435—9. (56) McGrath CM, Soule HD. Calcium regulation of normal human mammary epithelial cell growth in culture. In Vitro 1984;20:652—62. (57) Mills MJ, Mousalli MJ, Russo J. DNA-labeling index and in Vitro growth of mammary epithelium are modulated by the stage of gland development. Proceedings of the International Association of Breast Cancer Research; 1987', Miami. (58) Mousalli MJ, Mills MJ, Russo J. Influence of mammary gland develop— ment on its epithelium’s growth potential in tissue culture. Proceedings of the International Association of Breast Cancer Research; 1987; Miami. (59) Soule HD, McGrath CM. A simplified method for passage and long-term growth of human mammary epithelial cells. In Vitro Cell Dev Biol 1986; 2226—12. (60) Briand P, Peterson OW, Van Deurs B. A new diploid nontumorigenic human breast epithelial cell line isolated and propagated in chemically defined medium. In Vitro Cell Dev Biol 1987;23:181—8. (6]) Gaffney EV. A cell line (HBL-lOO) established from human breast milk. Cell Tissue Res 1982;227:563—8. 34 (62) Steuer AF, Rhim JS, Hentosh PM, Ting RC. Survival of human cells in the aggregate form: potential index of in vitro cell transformation. J Natl Cancer Inst 1977;58:917—21. (63) Russo J, Russo IH. Development of human mammary gland. In: Neville MC, Daniel CW, editors. The mammary gland development regulation and function. New York (NY): Plenum Publishing; 1987. p. 67—93. (64) Tanner JM. The development of the reproductive system. In: Tanner JM, editor. Growth at adolescence with a general consideration of the effect of hereditary and environmental factors upon growth and maturation from birth to maturity. Oxford (UK): Blackwell Scientific; 1962. p. 28—39. (65) Vorherr H. Development of the female breast. In: Vorherr H, editor. The breast: morphology, physiology and lactation. New York (NY): Academic Press; 1974. p. 1—18. (66) Dabelow A. die Milchdruse. In: Bargmann W, editor. Handbuch der mik- roskopishen anatomic des menschen. Vol 3, Part 3 haut and sinnes organs. Berlin (Germany): Springer—Verlag; 1957. p. 277—485. (67) Russo J, Tay LK, Russo IH. Differentiation of the mammary gland and susceptibility to carcinogenesis. Breast Cancer Res Treat 1982;225—73. (68) Russo J, Rivera R, Russo IH. Influence of age and parity on the devel- opment of the human breast. Breast Cancer Res Treat 1992;23:211—8. (69) Edwards RG, Howles CM, Macnamee C. Clinical endocrinology of re- production. In: Baulieu EE, Kelly PA, editors. Hormones: from molecules to disease. New York (NY): Chapman and Hall; 1990. p. 457—76. (70) Russo IH, Russo J. Mammary gland neoplasia in long-term rodent studies. Environ Health Perspect 1996;104:938—67. (71) Russo J, Russo IH. Role of hormones in human breast development: the menopausal breast. In: Progress in the management of menopause. Lon- don (UK): Parthenon Publishing: 1997. p. l—10. ( 72) Russo IH, Russo J. Role of hormones in cancer initiation and progression. J Mammary Gland Biol Neoplasia 1998;3249—61. (73) Russo J, Russo IH. Role of differentiation in the pathogenesis and pre- vention of breast cancer. Endocr Related Cancer 1997;4z7—21. (74) Calaf G, Alvarado ME, Bonney GE, Amfoh KK, Russo J. Influence of lobular development on breast epithelial cell proliferation and steroid hormone receptor content. Int J Oncol 1995;721285—8. (75) Lippman ME, Dickson RB, Gelmann EP, Rosen N, Knabbe C, Bates S, et a1. Growth regulation of human breast carcinoma occurs through regulated growth factor secretion. J Cell Biochem 1987:3511—16. (76) Meyer J S. Cell proliferation in normal human breast ducts fibroadenomas and other ductal hyperplasias measured by nuclear labeling with tritiated thymidine. Effects of menstrual phase, age, and oral contraceptive hor— mones. Hum Pathol 1977;8267—81. (77) Masters JR, Drife JO, Scarisbrick JJ. Cyclic variations of DNA synthesis in human breast epithelium. J Natl Cancer Inst 1977;58:1263—5. (78) Ferguson DJ, Anderson TJ. Morphological evaluation of cell turnover in relation to the menstrual cycle in the “resting” human breast. Br J Cancer 1981;44:177—81. (79) Anderson TJ, Ferguson DJ, Raab GM. Cell turnover in the “resting” human breast: influence of parity, contraceptive pill, age and laterality. Br J Cancer 1982;46:376—82. (80) Russo J, Russo IH. Estrogens and cell proliferation in the human breast. J Cardiovasc Pharrnacol 1996;28 Suppl 5:19—23. (81) Kumar V, Green S, Stack G, Berry M, Jin JR, Chambon P. Functional domains of the human estrogen receptor. Cell 1987;51:941—51. (82) King RJ. Effects of steroid hormones and related compounds on gene transcription. Clin Endocrinol (Oxf) 1992;36:1—14. (83) Huseby RA, Maloney TM, McGrath CM. Evidence for a direct growth— stimulating effect of estradiol on human MCF-7 cells in Vitro. Cancer Res 1984;44:2654—9. (84) Huff KK, Knabbe C, Lindsey R, Kaufman D, Bronzert D, Lippman ME, et al. Multihormonal regulation of insulin—like growth factor-I-related pro— tein in MCF—7 human breast cancer cells. M01 Endocrinol 1988;2z200—8. (85) Dickson RB, Huff KK, Spencer EM, Lippman ME. Induction of epider- mal growth factor—related polypeptides by 17B—estradiol in MCF—7 human breast cancer cells. Endocrinology 1986;118:138—42. (86) Soto AM, Sonnenschein C. Cell proliferation of estrogen-sensitive cells: the case for negative control. Endocr Rev 1987;8z44—52. (87) Tsai MJ, O’Malley BW. Molecular mechanisms of steroid/thyroid recep- tor superfamily members. Annu Rev Biochem 1994;63:451—236. (88) Kuiper GG, Enmark E, Pelto-Huikko M, Nilsson S, Gustaffson JA. Clon- Joumal of the National Cancer Institute Monographs No. 27, 2000 ing of a novel receptor expressed in rat prostate and ovary. Proc Natl Acad Sci U S A 1996;93:5925—30. (89) Tremblay GB, Tremblay A, Copeland NG, Gilbert DJ, Jenkins NA, La- brie F, et al. Cloning, chromosomal localization, and functional analysis of the murine estrogen receptor B. M01 Endocrinol 1997;11:353—65. (90) Mosselman S, Polman J, Dijkema R. ER B: identification and character- ization of a novel human estrogen receptor. FEBS Lett 1996;392:49—53. (91) Katzenellenbogen BS. Dynamics of steroid hormone receptor action. Annu Rev Physiol 1980;42:17—35. (92) Topper J, Freedman C. Multiple hormone interactions in the developmen- tal biology of the mammary gland. Physiol Rev 1980;60:1049—60. (93) Dickson RB, Lippman ME. Cellular and molecular biology: In: Lippman ME, Dickson RB, editors. Breast cancer. Boston (MA): Kluwer Academic Publishers; 1988. p. 119—65. (94) Jordan VC. Tamoxifen: the herald of a new era of preventive therapeutics. J Natl Cancer Inst 1997;89:747—9. (95) Kuiper GG, Carlsson B, Grandien K, Enmark E, Haggblad J, Nilsson S, et al. Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors 0t and B. Endocrinology 1997:1382 863—70. (96) Cowley SM, Hoare S, Mosselman S, Parker MG. Estrogen receptors 0L and B form heterodimers on DNA. J Biol Chem 1997;272:19858—62. (97) Kuiper GG, Gustafsson JA. The novel estrogen receptor-B subtype: po- tential role in the cell— and promotor-specific actions of estrogens and anti-estrogens. FEBS Lett 1997;410:87—90. (98) Pace P, Taylor J, Suntharalingam S, Coombes RC, Ali S. Human estrogen receptor B binds DNA in a manner similar to and dimerizes with estrogen receptor 0L. J Biol Chem 1997;272:25832—8. (99) Ogawa S, Inoue S, Watanabe T, Hiroi H, Orimo A, Hosoi T, et al. The complete primary structure of human estrogen receptor B (hERB) and its heterodimerization with ER or in vivo and in vitro. Biochem Biophys Res Commun 1998;243:122—6. (100) Paech K, Webb P, Kuiper GG, Nilsson S, Gustafsson J, Kushner PJ, et al. Differential ligand activation of estrogen receptors ERor and ERB at APl sites. Science 1997;277:1508—10. (101) Couse JF, Korach KS. Estrogen receptor null mice: what have we learned and where will they lead us? Endocr Rev 1999;20:358—417. (102) Going JJ, Anderson TJ, Battersby S, MacIntyre CC. Proliferative and secretory activity in human breast during natural and artificial menstrual cycles. Am J Pathol 1988;130:193—204. (103) Potten CS, Watson RJ, Williams GT, Tickle S, Roberts SA, Harris M, et al. The effect of age and menstrual cycle upon proliferative activity of the normal human breast. Br J Cancer 1988;58:163—70. (104) Clarke RB, Howell A, Potten CS, Anderson E. Dissociation between steroid receptor expression and cell proliferation in the human breast. Cancer Res 1997;57:4987—91. (105) Laidlaw IJ, Clarke RB, Howell A, Owen AW, Potten CS, Anderson E. The proliferation of normal human breast tissue implanted in athymic nude mice is stimulated by estrogen but not progesterone. Endocrinology 1995;136:164—71. (106) Clarke RB, Howell A, Anderson E. Estrogen sensitivity of normal human breast tissue in vivo and implanted into athymic nude mice: analysis of the relationship between estrogen—induced proliferation and progesterone re- ceptor expression. Breast Cancer Res Treat 1997;45:121—33. (107) Goodman HM, editor. Basic medical endocrinology. 2“d ed. New York (NY): Raven Press; 1994. p. 288—90. (108) Russo J, Russo IH. Influence of differentiation and cell kinetics on the susceptibility of the rat mammary gland to carcinogenesis. Cancer Res 1980:40(8 Pt l):2677‘87. (109) Russo J, Russo IH. Biology of disease. Biological and molecular bases of mammary carcinogenesis. Lab Invest 1987;57:112—37. (110) Russo J, Ao X, Grill C, Russo IH. Pattern of distribution of cells positive for estrogen receptor 0L and progesterone receptor in relation to prolifer- ating cells in the mammary gland. Breast Cancer Res Treat 1999:53: 217—27. (111) Clarke R, Dickson RB, Lippman ME. Hormonal aspects of breast cancer. Growth factors, drugs and stromal interactions. Crit Rev Oncol Hematol 1992;12:1—23. (112) Knabbe C, Lippman ME, Wakefield LM, Flanders KC, Kasid A, Derynck R, et a1. Evidence that transforming growth factor—B is a hormonally Journal of the National Cancer Institute Monographs No. 27, 2000 regulated negative growth factor in human breast cancer cells. Cell 1987; 48:417—28. (113) Dickson RB, Lippman ME. Growth factors in breast cancer. Endocr Rev 1995;16:559—89. (114) Russo J, Reina D, Frederick J, Russo IH. Expression of phenotypical changes by human breast epithelial cells treated with carcinogens in vitro. Cancer Res 1988;48:2837~57. (115) Russo J, Gusterson BA, Rogers A, Russo IH, Wellings SR, van Zwieten MJ. Comparative study of human and rat mammary tumorigenesis. Lab Invest 1990;62:244—78. (116) Tay LK, Russo J. Formation and removal of 7,]2-dimethylbenz[a]- anthracene—nucleic acid adducts in rat mammary epithelial cells with different susceptibility to carcinogenesis. Carcinogenesis 1981;2z1327—33. (117) Tay LK, Russo J. In vitro metabolism of 7,l2-dimethylbenz[a]anthracene in mammary epithelial cells from rats with different susceptibilities to carcinogenesis. Carcinogenesis 1983;4z733—8. (118) Russo J, Russo IH. Differentiation and breast cancer development. In: Heppner G, Bittar ED, editors. Advances in oncobiology. Greenwich (CT): JAI Press; 1998. p. 1—10. (119) Wellings SR, Jensen HM, Marcum RG. An atlas of subgross pathology of the human breast with special reference to possible precancerous lesions. J Natl Cancer Inst 1975;55:231—73. (120) Russo J, Russo IH. Development pattern of human breast and suscepti- bility to carcinogenesis. Eur J Cancer Prev 1993;2(suppl 3):85—100. (121) Russo J, Russo IH. Toward a physiological approach to breast cancer prevention. Cancer Epidemiol Biomarker Prev 1994;3z353—64. (122) Russo J, Romero AL, Russo IH. Architectural pattern of the normal and cancerous breast under the influence of parity. Cancer Epidemiol Biomar- ker Prev l994;3:219—24. ( 123 ) Khan SA, Rogers MA, Khurana KK, Meguid MM, Numann PJ. Estrogen receptor expression in benign breast epithelium and breast cancer risk. J Natl Cancer Inst 1998;90:37—41. (124) Molino A, Micciolo R, Turazza M, Bonetti F, Piubello Q, Corgnati A, et al. Prognostic significance of estrogen receptors in 405 primary breast cancers: a comparison of immunohistochemical and biochemical methods. Breast Cancer Res Treat 1997;45:241—9. (125) Habel LA, Stanford JL. Hormone receptors and breast cancer. Epidemiol Rev 1993;15:209—19. (126) Harlan LC, Coates RJ, Block G, Greenberg RS, Ershow A, Forrnan M, et a1. Estrogen receptor status and dietary intakes in breast cancer patients. Epidemiology 1993;4:25—3l. ( 127) Moolgavkar SH, Day NE, Stevens RG. Two-stage model for carcinogen- esis: epidemiology of breast cancer in females. J Natl Cancer Inst 1980; 65:559-69. (128) Vladusic EA, Homby AE, Guerra-Vladusic FK, Lupu R. Expression of ERB in breast cancer cells: does it play an important role? Proc Am Assoc Cancer Res 1997;38:297a. (129) Dotzlaw H, Leygue E, Watson PH, Murphy LC. Expression of estrogen receptor—B in human breast tumors. J Clin Endocrinol Metab 1997;82: 2371—4. (130) Vladusic EA, Homby AE, Guerra-Vladusic FK, Lupu R. Expression of estrogen receptor B messenger RNA variant in breast cancer. Cancer Res 1998;58:210—4. (131) Hu YF, Lau KM, Ho SM, Russo J. Increased expression of estrogen receptor B in chemically transformed human breast epithelial cells. Int J Oncol 1998;12:1225—8. (132) Lau KM, Leav 1, Ho SM. Rat estrogen receptor-or and —B, and progester- one receptor mRNA expression in various prostatic lobes and microdis- sected normal and dysplastic epithelial tissues of the Nobel rats. Endo— crinology 1998;139:424—7. (I33) Watanabe T, Inoue S, Ogawa S, Ishii Y, Hiroi H, Ikeda K, et al. Agonistic effect of tamoxifen is dependent on cell type, ERE-promotor context, and estrogen receptor subtype: functional difference between estrogen recep- tors u and B. Biochem Biophys Res Commun 1997;236:140—5. (134) Aronica SM, Kraus WL, Katzenellenbogen BS. Estrogen action via the CAMP signaling pathway: stimulation of adenylate cyclase and cAMP— regulated gene transcription. Proc Natl Acad Sci U S A 1994;91:8517—21. (135) Pappas TC, Gametchu B, Watson CS. Membrane estrogen receptors iden- tified by multiple antibody labeling and impeded-ligand binding. FASEB J 1995;9z404—10. 35 (136) Spink DC, Eugster HP, Lincoln DW 2nd, Schuetz JD, Schuetz EG, Johnson JA, et al. 17B-estradiol hydroxylation catalyzed by human cyto- chrome P450 IAl: a comparison of the activities induced by 2,3,7,8- tetrachlorodibenzo-p-dioxin in MCF-7 cells with those from heterologous expression of the cDNA. Arch Biochem Biophys 1992;293:342—8. (I37) Cristofalo VJ, Wallace JM, Rosner BA. Glucocorticoid enhancement of proliferative activity in WI—38 cells. In: Sata SD, Ross R, editors. Hor— mones and cell culture book B. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory; 1979. p. 875—87. (138) Hennings H, Michael D, Cheng C, Steinert P, Holbrook K, Yuspa SH. Calcium regulation of growth and differentiation of mouse epidermal cells in culture. Cell 1980;19:245—54. (I39) Hennings H, Holbrook K. Extracellular calcium regulates growth and terminal differentiation of cultured mouse epidermal cells. In: Boynton AL, McKeehan WL, Whitfield JF, editors. Symposium on ions, cell pro- liferation, and cancer. New York: Academic Press; 1982. p. 497—514. (140) Hennings H, Kruszewski FH, Yuspa SH, Tucker RW. Intracellular cal- cium alterations in response to increased external calcium in normal and neoplastic keratinocytes. Carcinogenesis 1989;10:777—80. (141) Stoner GD, Babcock MS. Influence of growth factors on proliferation of normal and chemically transformed rat esophageal epithelial cells. Proc Am Assoc Cancer Res 1982;23:42a. (142) Huschtscha Ll, Holiday R. Limited and unlimited growth of SV40- transformed cells from human diploid MRC-5 fibroblasts. J Cell Sci 1983; 63:77—99. ([43) Linder S, Marshall H. Immortalization of primary cells by DNA tumor Viruses. Exp Cell Res 1990;191:14. (I44) Shay JW, Wright WE, Werbin H. Defining the molecular mechanisms of human cell immortalization. Biochim Biophys Acta 1991;1072zl—7. (145) Soule HD, Maloney TM, Wolman SR, Peterson WD Jr, Brenz R, Mc- Grath CM, et a1. Isolation and characterization of a spontaneously im- mortalized human breast epithelial cell line, MCF-lO. Cancer Res 1990; 50:6075—86. (146) Calaf G, Russo J. Transformation of human breast epithelial cells by chemical carcinogens. Carcinogenesis 1993;14:483—92. (I47) Hu YF, Russo IH, Zalipsky U, Lynch HT, Russo J. Environmental chemi- cal carcinogens induce transformation of breast epithelial cells from women with familial history of breast cancer. In Vitro Cell Dev Biol Anim 1997;33:495—8. (I48) Barnabas N, Moraes R, Calaf G, Estrada S, Russo J. Role of p53 in MCF—lOF cell immortalization and chemically—induced neoplastic trans- formation. Int J Oncol 1995;7z1289—96. (149) Yang X, Tahin Q, Hu YF, Russo IH, Balsara BR, Mihaila D, et a1. Functional roles of chromosomes 11 and 17 in the transformation of human breast epithelial cells in Vitro. Int J Oncol 1999;15:629—38. (I50) Huang Y, Bove B, Wu Y, Russo IH, Yang X, Zekri A, et a1. Microsatellite instability during the immortalization and transformation of human breast epithelial cells in vitro. M01 Carcinog 1999;24:118—27. ([5]) Huang Y, Bove B, Russo IH, Yang X, Russo J. Microsatellite instability is induced by benzo[a]pyrene and 7,12-dimethyl benz[a]anthracene in human breast epithelial cells in vitro. Proc Am Assoc Cancer Res 1998; 39:537. ([52) Yang X, Huang Y, Russo IH, Barrett JC, Russo J. Chromosomes 11 and 17 revert transformed phenotypes induced by benzo[a]pyrene (BP) in human breast epithelial cells. Proc Am Assoc Cancer Res 1998;39: 1263a. (153) Yang XQ, Russo IH, Huang YJ, Russo J. Microsatellite instability of D17SSl3 on chromosome 17 is associated with progression of breast Cancer. Int J Oncol 1997;11:41—6. (154) Wu Y, Barnabas N, Russo IH, Yang X, Russo J. Microsatellite instability and loss of heterozygosity in chromosomes 9 and 16 in human breast epithelial cells transformed by chemical carcinogens. Carcinogenesis 1997;18:1069—74. (I55) Harley CB. Telomere loss: mitotic clock or genetic time bomb? Mutat Res 1991;256:271—82. ( I 56) Hopfer U, J acobberger JW, Gruenert DC, Eckert RL, J at PS, Whitsett JA. Immortalization of epithelial cells. Am J Physiol 1996;270(1 Pt 1):C1-C11. (I57) Blackburn EH. Structure and function of telomeres. Nature 1991;350: 569—73. ( 158) Shay JW, Wright WE. The reactivation of telomerase activity in cancer progression. Trends Genet 1996;12:129—31. 36 (159) Shay JW, Bacchetti S. A survey of telomerase activity in human cancer. Eur J Cancer 1997;33:787—91. (160) Avilion AA, Piatyszek MA, Gupta J, Shay JW, Bacchetti S, Greider CW. Human telomerase RNA and telomerase activity in immortal cell lines and tumor tissues. Cancer Res 1996;56:645—50. ([61) Higgy NA, Russo J, Mgbonyebi P, Salicioni AM, Russo IH. Human chorionic gonadotropin inhibits telomerase activity in human breast epi- thelial cells in Vitro. Proc Am Assoc Cancer Res 1998;39:3678a. (I62) Higgy NA, Salicioni AM, Russo IH, Zhang PL, Russo J. Differential expression of human ferritin H chain gene in immortal human breast epithelial MCF-IOF cells. Mol Carcinog 1997;20:332—9. (I63) Theil EC. Ferritin: structure, gene regulation, and cellular function in animals, plants, and microorganisms. Annu Rev Biochem 1987;56: 289—315. (164) Anison P. Current concepts in iron metabolism. Clin Lab Haematol 1982; 1 12241—57. (I65) Reichard P, Ehrenberg A. Ribonucleotide reductase—a radical enzyme. Science 1983;221:514—9. (166) Ponka P, Schulman HM, Cox TM. Iron metabolism in relation to heme synthesis. In: Dailey HA, editor. Biosynthesis of heme and chlorophyllus. New York (NY): McGraw—Hill; 1990; p. 393—405. (167) Reif DW. Ferritin as a source of iron for oxidative damage. Free Radic Biol Med 1992;12:417—27. (168) Guyton KZ, Kensler TW. Oxidative mechanisms in carcinogenesis. Br Med Bull 1993;49:523—44. (169) Klausner RD, Rouault TA, Harford J B. Regulating the fate of mRNA: the control of cellular iron metabolism. Cell 1993;72:19—28. (170) Dognin J, Crichton RR. Mobilisation of iron from ferritin fractions of defined iron content by biological reductants. FEBS Lett 1975;54:234—6. (171) Biemond P, Swaak AJ, Beindorf CM, Koster JF. Superoxide-dependent and -independent mechanisms of iron mobilization from ferritin by xan- thine oxidase. Implications for oxygen—free-radical-induced tissue de— struction during ischaemia and inflammation. Biochem J 1986;239: 169—73. (1 72) Zhao MJ, Jung L. Kinetics of the competitive degradation of deoxyribose and other molecules by hydroxyl radicals produced by the Fenton reaction in the presence of ascorbic acid. Free Radic Res 1995;23:229—43. (I73) Fidler IJ, Nicolson GL. Concepts and mechanisms of breast cancer me— tastasis. In: Bland KI, Copeland EM, editors. The breast: comprehensive management of benign and malignant diseases. Philadelphia (PA): Saun- ders; 1991. p. 395—408. ( I 74) Kohn EC, Liotta LA. Molecular insights into cancer invasion: strategies for prevention and intervention. Cancer Res 1995;55:1856—62. (I75) Malins DC, Polissar NL, Gunselman SJ. Progression of human breast cancers to the metastatic state is linked to hydroxyl radical-induced DNA damage. Proc Natl Acad Sci U S A 1996;93:2557—63. ( J 76) Han X, Liehr JG. Microsome-mediated 8-hydroxylation of guanine bases of DNA by steroid estrogens: correlation of DNA damage by free radicals with metabolic activation to quinones. Carcinogenesis 1995;16:2571—4. (177) Frenkel K, Wei L, Wei H. 7,12-dimethylbenz[a]anthracene induces oxi- dative DNA modification in viva. Free Radic Biol Med 1995;19:373—230. (178) Weinberg ED. Roles of iron in neoplasia. Promotion, prevention, and therapy. Biol Trace Elem Res 1992;34:123—40. (179) Fan H, Villegas C, Wright JA. A link between ferritin gene expression and ribonucleotide reductase R2 protein, as demonstrated by retroviral vector mediated stable expression of R2 cDNA. FEBS Lett 1996;382:145—8. (180) Moroz C, Lahat N, Biniaminov M, Ramot B. Ferritin on the surface of lymphocytes in Hodgkin’s disease patients. A possible blocking substance removed by levamisole. Clin Exp Immunol 1977;29:30—5. (18]) Keown P, Descamps—Latscha B. In Vitro suppression of cell—mediated immunity by ferroproteins and ferric salts. Cell Immunol 1983;80:257—66. (I82) Rosen HR, Moroz C, Reiner A, Reinerova M, Stierer M, Svec J, et al. Placental isoferritin associated p43 antigen correlates with features of high differentiation in breast cancer. Breast Cancer Res Treat 1992;24:17—26. (183) Rosen HR, Flex D, Stierer M, Moroz C. Monoclonal antibody CM—H-9 detects circulating placental isoferritin in the serum of patients with vis- ceral metastases of breast cancer. Cancer Lett 1991;59:145—51. (I84) Kwak EL, Larochelle DA, Blaumont C, Torti SV, Torti FM. Role of NF-KB in the regulation of ferritin H by tumor necrosis factor-a. J Biol Chem 1995;270:15285—93. Journal of the National Cancer Institute Monographs No. 27, 2000 (185) Rogers JT. Ferritin translation by interleukin—6: the role of sequences upstream of the start codons of the heavy and light subunit genes. Blood 1996;87:2525—37. (186) Kastan MB, Onyekwere O, Sidransky D, Vogelstein B, Craig RW. Par- ticipation of p53 protein in the cellular response to DNA damage. Cancer Res 1991;51:6304—11. (187) el—Deiry WS, Harper JW, O’Connor PM, Velculescu VE, Canman CE, Jackman J, et a1. WAF 1/CIP1 is induced in p53-mediated G1 arrest and apoptosis. Cancer Res 1994;54:1169—74. (188) Agarwal ML, Taylor WR, Chemov MV, Chernova OB, Stark GR. The p53 network. J Biol Chem 1998;273:1—4. (189) Russo J, Hu YF, A0 X, Grill C, Russo IH. Critical appraisal of estrogens as carcinogenic agents. In: Birkhauser MH, Rosenbaum H, editors. IV European Congress on Menopause. Paris: Editions ESKA; 1998. p. 279—87. (190) Telang NT, Axelrod DM, Bradlow HL, Osborne MP. Metabolic biotrans- formation of estradiol in human mammary explant cultures. Ann N Y Acad Sci 1990;586:70—8. Journal of the National Cancer Institute Monographs No. 27, 2000 (191) Fishman J, Osborne MP, Telang NT. The role of estrogen in mammary carcinogenesis. Ann N Y Acad Sci 1995;768:91-100. (192) Telang NT, Suto A, Wong GY, Osborne MP, Bradlow HL. Induction by estrogen metabolite 16cx-hydroxyestrone of genotoxic damage and aber- rant proliferation in mouse mammary epithelial cells. J Natl Cancer Inst 1992;84:634—8. (193) Ryan DE, Thomas PE, Reik LM, Levin W. Purification, characterization and regulation of five rat hepatic microsomal cytochrome P—450 enzymes. Xenobiotica 1982;12:727—44. ( I 94 ) Wilson NM, Christou M, Turner CR, Wrighton SA, Jefcoate CR. Binding and metabolism of benzo[a]pyrene and 7,12—dimethylbenz[a]anthracene by seven purified forms of cytochrome P—450. Carcinogenesis 1923452 1475—83. NOTE Supported by Public Health Services grants ROlCA6728 (National Cancer Institute) and ESO7280 (National Institute of Environmental Health Sciences). National Institutes of Health, Department of Health and Human Services. 37 Chapter 2: The Role of Steroid Hormones in Prostate Carcinogenesis Maarten C. Bosland Carcinoma of the prostate is the most frequently diagnosed malignancy and the second leading cause of death as a result of cancer in men in the United States and in many other Western countries. Notwithstanding the importance of this malignancy, little is understood about its causes. The epide- miology of prostate cancer strongly suggests that environ- mental factors, particularly diet and nutrition, are major determinants of risk for this disease, and evidence is mount- ing that there are important genetic risk factors for prostate cancer. Human prostate carcinomas are often androgen sen- sitive and react to hormonal therapy by temporary remis- sion, followed by relapse to an androgen-insensitive state. These well-established features of prostate cancer strongly suggest that steroid hormones, particularly androgens, play a major role in human prostatic carcinogenesis, but the pre- cise mechanisms by which androgens affect this process are unknown. In addition, the possible involvement of estrogenic hormones is not entirely clear. The purpose of this overview is to summarize the literature about steroid hormonal fac- tors, androgens and estrogens, and prostate carcinogenesis. From these literature observations, a multifactorial general hypothesis of prostate carcinogenesis emerges with andro- gens as strong tumor promoters acting via androgen recep- tor-mediated mechanisms to enhance the carcinogenic activ- ity of strong endogenous genotoxic carcinogens, such as reactive estrogen metabolites and estrogen- and prostatitis- generated reactive oxygen species and possible weak envi- ronmental carcinogens of unknown nature. In this hypoth- esis, all of these processes are modulated by a variety of environmental factors such as diet and by genetic determi- nants such as hereditary susceptibility and polymorphic genes that encode for steroid hormone receptors and en- zymes involved in the metabolism and action of steroid hor- mones. [J Natl Cancer Inst Monogr 2000;27:39—66] Carcinoma of the prostate is the most frequently diagnosed malignancy and the second leading cause of death as a result of cancer in men in the United States and in many Western coun: tries (not counting nonmelanoma skin cancer) (I). Notwithstand— ing the importance of this malignancy, little is understood about its causes. Steroid hormones, particularly androgens, are sus- pected to play a major role in human prostate carcinogenesis, but the precise mechanisms by which androgens affect this process and the possible involvement of estrogenic hormones are not clear. A causal relation between androgens and prostate cancer development is generally considered to be biologically very plausible because the vast majority of human prostate cancers are androgen sensitive and respond to hormonal therapy by tem- porary remission, later followed by relapse to a hormone- refractory state. The purpose of this overview is to summarize the literature about steroid hormonal factors and prostate carci— nogenesis. Although the objective of this overview is not to be Journal of the National Cancer Institute Monographs No. 27, 2000 comprehensive, an attempt is made to be complete, especially where crucial aspects of this hormonal involvement are con- cerned. In contrast, the prostate is a rare site of tumor development in carcinogenesis bioassays in rodents (2,3) and in aging male labo— ratory rodents, with the exception of ventral prostatic neoplasms in some rat strains (4—9). Prostate cancer is also rare in male farm and companion animals, With the notable exception of the dog, which is the only species besides man that develops this malignancy. As will be discussed in this overview, steroid hor- mones can induce and can substantially enhance prostate carci- noma development in rodents, and this phenomenon has been exploited to further our knowledge about the involvement of hormonal factors and mechanisms in prostate cancer etiology. In this overview, the epidemiologic evidence for a role of steroid hormonal factors in prostate carcinogenesis is summa- rized first, followed by review of experimental data, a discussion of the possible mechanisms whereby steroid hormones, andro- gens as well as estrogens, may be involved in prostate cancer causation, and overall conclusions and suggestions for future research. As will be demonstrated, there is no lack of hypotheses about the role of steroid hormones in prostate cancer etiology, but the available data are often contradictory and incomplete, and an in-depth overall mechanistic understanding of how ste— roid hormonal factors are involved in prostate carcinogenesis is very limited. EPIDEMIOLOGIC EVIDENCE FOR INVOLVEMENT 0F STEROID HORMONES The epidemiology of prostatic cancer has been reviewed in depth elsewhere ( 10—15). Prostate cancer risk factors that are associated with hormonal factors are summarized in subsequent sections, and epidemiologic and other studies related to the me- tabolism and action of steroid hormones are reviewed in detail. Besides hormonal factors, there are only a few established risk factors for prostate cancer. These risk factors are briefly sum— marized below to put the relative importance of steroid hor- monal factors in perspective. Prostate cancer incidence and mortality rates have increased in the United States over the few decades preceding the frequent use of prostate-specific antigen (PSA) for early detection (10). Even though incidence rates have increased substantially since the mid—19805 because of the use of PSA “screening” for early detection (1), incidence has declined over the period from 1990 through 1996 by an average of 2% per year and mortality by 1.6% per year (16). Because of the increasing use of PSA for Correspondence to: Maarten C. Bosland, D.V.Sc., Ph.D., Departments of Environmental Medicine and Urology, New York University School of Medi- cine, 550 First Ave., New York, NY, 10016 (e—mail: maarten.bosland@ med.nyu.edu). See “Note” following “References.” © Oxford University Press 39 early detection and treatment in this period, it is too early to separately determine changes in prostate cancer rates and the impact of PSA screening on rates. In 1999, prostate cancer was the most frequent malignancy in US. males with 179 300 new cases expected, and it was the second most frequent cause of death as a result of cancer with 37 000 deaths expected (17). Many studies have demonstrated that prostate cancer is more frequent in men with a family history of prostate cancer, as summarized elsewhere (10,11,15,18—20). This familial aggrega- tion appears to be similar in African-American and in European- American men (21,22). However, inherited risk for prostate can- cer can only explain a small proportion of prostate cancer cases, less than 10% (20). Besides a variety of genetic alterations iden- tified with varying frequency in human prostate carcinomas, as summarized by Dong et al. (23), a few susceptibility loci linked to inherited prostate cancer risk have been identified on chro- mosome 1 (24—28) and on the X chromosome (29,30). Breast and prostate cancers cluster in some families, and there is some evidence that BRCA1 and BRCA2 mutations are involved in this clustering (18,31). However, none of these loci have thus far been associated with hormonal factors. Evidence is limited that a history of venereal disease (11,32— 34) and a history of prostatitis (34,35) are risk factors. An as- sociation between prostate cancer risk and the prior occurrence of benign prostatic hypertrophy (BPH) is biologically unlikely, even though steroid hormones are also implicated in BPH. Pros— tate cancer and BPH originate from different parts of the prostate (all BPH is found in the transition zone, and more than 80% of all cancers develop in the peripheral zone), and their epidemi- ology is dissimilar (11). Although, in some studies (34,36,37), a relationship between smoking and risk for prostate cancer has been found, no such relationship has been observed in the vast majority of studies (37—40). In addition, smoking appears to have no effect on cir— culating levels of testosterone and other hormones that may be involved in prostate carcinogenesis (41,42). Most studies (11,43) addressing alcohol consumption as a potential risk factor for prostate cancer did not find evidence for an association. One notable exception is a study by Hayes et al. (44) that found a positive association in a US. case—control study, which was limited to heavy use of alcohol. Possible reasons for the asso- ciation observed in this study are discussed by Lumey et al. (43). One of these reasons may be that prostate cancer risk is elevated in alcoholics with liver disease (11,43). This risk elevation is possibly related to the impaired clearance of estrogens described in men with liver cirrhosis (11,45,46). Increased risk has been observed for a variety of occupations in studies of occupational factors and prostate cancer (11,14,47,48), including armed services personnel (1]) and workers in the nuclear industry (11,15,48,49). Although prostate cancer risk in survivors of the atomic bomb in Japan appears not to be elevated (50), there is a rather strong international correlation between prostate cancer incidence and indoor radon levels reported (51). Thus, prostate cancer risk may be associ- ated with exposure to ionizing radiation, but the evidence is equivocal. Associations between exposures and prostate cancer risk observed in the rubber industry are limited to one or a few plants (11). The evidence for a positive association between farming and prostate cancer risk is weak to inconclusive (11,14,47,48,52). There is only very weak, if any, evidence for an association of cadmium exposure and prostate cancer risk 40 (11,53,54). Hormonal factors are most likely not involved in any of these (possible) associations between risk and occupational factors. Risk Factors Associated With Possible Hormonal Mechanisms The results of a variety of epidemiologic studies have led to suggestions for several risk factors that may be related to a hormonal mechanism. These risks include dietary factors, va- sectomy, sexual factors, the level of physical activity, and obe— srty. Diet and Nutrition The associations between dietary factors and prostate cancer risk have been extensively summarized elsewhere (10,11,15,55— 57). Considerable consistency across studies indicates that a high intake of fat, particularly total fat and saturated fat, is a risk factor for prostate cancer, but the strength of the associations is modest at best (57,58) and may be greater for African- Americans than for European-Americans (59). Results from Ha- waiian case—control studies suggest that as much as 25% of prostate cancer in the United States may be attributable to a high saturated-fat intake (60). However, Whittemore et al. (22) esti- mated that dietary fat intake may account for only 10%—15% of the difference in prostate cancer occurrence between European- Americans and African—Americans or Asians. The mechanism that could underlie an enhancing effect of fat on prostate carci- nogenesis is not understood, but several hypotheses, including hormonal mediation, have been discussed elsewhere (11,15,57,61). In addition, a high intake of protein and energy and a low intake of dietary fiber and complex carbohydrates have been found to be associated with the increased risk for prostate cancer in some studies (10,11,15,55). Associations with prostate cancer risk reported for individual nutrients or foods are not very strong. However, migration from low-risk areas, such as Japan, to high—risk countries, such as the United States, increases risk considerably (10,11). These changes in risk are thought to be due to differences in environ- ment, including lifestyle and particularly dietary habits (10,11). It is, therefore, conceivable that the combined effects of dietary factors on prostate carcinogenesis are more important than the separate effects of any individual dietary factor (62). This idea is supported by the lack of any effect of dietary fat per se on the induction of prostate cancer in animal models, whereas epide- miologic studies (11,15,62) rather consistently show a positive association between prostate cancer risk and intake of dietary fat. Older studies (63—67) of the effects on hormonal status of dietary changes and of the consumption of vegetarian or health food diets, which have been summarized previously (11), did not separately address the effects of dietary fat. However, they clearly indicate that diet can influence circulating hormone lev- els by changing androgen production rates and/or the metabo- lism and clearance of androgens and estrogens. In a study re- ported by Dorgan et al. (68), controlled changes in fat and fiber were applied to healthy men. The combination of a high-fat, low—fiber diet increased both total testosterone (by 13%) and testosterone bound to sex hormone—binding globulin (SHBG; by 15%) in the plasma as well as urinary testosterone excretion (13%), compared with a low-fat, high-fiber diet. However, uri- nary excretion of estrone, estradiol, and the 2-hydroxy metabo- Joumal of the National Cancer Institute Monographs No. 27, 2000 lites of these estrogens was lower. All of these studies indicate that diet can affect steroid hormone status, but no studies have addressed the separate effects of single dietary factors, such as fat intake. Complete consistency is lacking among epidemiologic stud- ies of prostate cancer risk and intake of dietary vitamin A and B—carotene (10,11,15,69). It is possible that retinoids and/or car- otenes enhance rather than inhibit prostatic carcinogenesis under certain circumstances or in certain populations (69), although animal and in vitro studies suggest a protective effect of reti- noids (11). In two more recent experiments on prostate cancer chemoprevention in a rat model, 9-cis-retinoic acid, a major retinol metabolite in mammalian species, strongly inhibited the induction of prostate cancer (70), but N—(4-hydroxy- phenyl)retinamide (4-HPR), a synthetic retinoid, did not have any effect ( 71 ). 9-cis-Retinoic acid is unique in that it is a pan— agonist for retinoic acid receptors, binding both retinoic acid receptor (RAR) and retinoid X receptor (RXR) receptors. In vitro, however, both 9-cis—retinoic acid and 4-HPR inhibit the growth and induce apoptosis of the androgen-sensitive human prostate cancer LNCaP cell line, and so does all-trans-retinoic acid, which only binds to RAR (72—74). There are indications that 4-HPR acts via a nonreceptor mechanism (74). The specific mechanism is not known by which retinoids and/or carotenoids may inhibit or enhance prostate carcinogenesis, but inhibition seems biologically more plausible than enhancement, as dis— cussed previously (11). The retinoic acid and androgen receptors both belong to the steroid receptor superfamily (75). This cir- cumstance raises the intriguing possibility that retinoids may be able to bind to and activate mutated forms of the androgen receptor or that the retinoic acid RAR and/or RXR may activate transcription of androgen—regulated genes. Studies on the regu- lation by sex steroids and retinoic acid of glutathione S— transferase in hamster smooth muscle tumor cells (76) and on androgen—receptor gene expression in human breast cancer cells (77) suggest that such mechanisms may exist. Vasectomy Vasectomy has been identified as a possible risk factor for prostate cancer in seven case—control studies (34, 78—83 ) and in three cohort studies (84—86). The range of risk ratios in the case—control studies was 1.4 to 5.3. No elevation of risk for prostate cancer following vasectomy was found in six other case—control studies (87-92) and in two retrospective cohort studies (93—95). Although a meta-analysis (96) of 14 studies indicated that there is no causal relation between vasectomy and prostate cancer, further studies, particularly cohort studies, will be required to definitively establish whether or not vasectomy is a true risk factor for prostate cancer (58,97—99). Three mechanisms by which vasectomy could enhance risk have been proposed: elevation of circulating androgens, immu- nologic mechanisms involving antisperm antibodies, and reduc- tion of seminal fluid production (34,78, 79,85,90,98,100). Most studies (101—105) that investigated the effect of vasectomy on pituitary—gonadal function did not find any effect, but some studies (90,100,106—110) found slight, but statistically signifi- cant, changes in circulating levels of certain hormones. Four groups (34,100,108,111) reported slightly elevated circulating testosterone levels, but only in two of these groups (100,108) was the increase statistically significant. M0 et al. (100) also found elevated levels of 50t-dihydrotestosterone (DHT), the ac- Journal of the National Cancer Institute Monographs No. 27, 2000 tive metabolite of testosterone in the prostate, in vasectomized men. John et al. (90) reported a decrease in SHBG, and Honda et al. (34) observed an increase in the ratio of testosterone to SHBG. These results suggest an elevation of circulating free testosterone following vasectomy, which may be a critical factor associated with risk for prostate cancer. A possible specific mechanism whereby vasectomy could influence the hypotha— lamic—pituitary—gonadal axis is not known. Sexual Factors Attempts have been made in several case~control studies (11,32—34,112—114) to investigate the possibility that sexual factors play a role in prostate cancer etiology. The results of these studies suggest that prostate cancer risk may be associated with the level of sexual activity, but no conclusive evidence exists for such a relation (11). Tsitouras et al. (115) reported a significant positive association between the level of sexual ac- tivity (intercourse and masturbation) and circulating total testos- terone levels in men between the ages of 60 and 79 years as well as an absence of a decrease in testosterone levels with aging in sexually active men. These findings suggest that a hormonal mechanism may underlie a possible association between prostate cancer risk and sexual activity suggested by the aforementioned case—control studies. Physical Activity and Anthropometric Correlates of Risk Contradictory indications are found that the level of physical activity may be a possible risk factor for prostate cancer, but the evidence for such an association is inconclusive at present (15,62,116). Sports exercise may decrease, as well as increase, circulating androgen concentrations or have no effect, depending on such factors as the time of blood sampling in relation to the exercise, the level of exercise, and the training protocol followed (117,118). Therefore, it is possible that the type and extent of physical activity influence circulating androgen concentrations and, thereby, perhaps prostate cancer risk. At present evidence is contradictory that obesity or an increased body mass index is a risk factor for prostate cancer (15,62,119). Severson et al. (120) observed a significant increase in prostate cancer risk with in- creasing upper-arm circumference and upper-arm muscle area but not fat area. A positive association between prostate cancer risk and muscle mass, but not fat mass, may suggest exposure to endogenous or exogenous androgenic hormones or other ana- bolic factors (120,121). Indeed, evidence is available that body mass index is inversely associated with plasma testosterone and SHBG levels and positively associated with estradiol levels (119,122,123), as discussed elsewhere (11,42). Epidemiologic Studies of Endogenous Hormones and Hormone Metabolism As indicated earlier, a causal relation between androgens and prostate cancer development is generally considered biologically plausible because this malignancy develops in an androgen- dependent epithelium and is usually androgen sensitive. In ad- dition, a few case reports (124—129) are available of prostate cancer in men who used androgenic steroids as anabolic agents or for medical purposes, suggestive of a causal relationship. Studies (11,15,130) comparing the endocrine status of human prostate cancer patients with that of control subjects are prob- ably not very informative about the endocrine status prior to the 41 onset of the disease and are, therefore, not meaningful to explore this relationship; in addition, the presence of the malignancy may by itself alter hormonal status. Indeed, the results of such studies do not provide a consistent pattern as summarized by Andersson et al. (130), which is confirmed in other case—control studies (119,131—133). These types of studies will, therefore, not be discussed here. Nested case—control studies of steroid hormonal factors in ongoing cohort studies, as well as studies comparing healthy males in populations that are at high risk for prostate cancer with populations at lower risk, are likely to be more meaningful. These studies are summarized in the following sections. The two major hypotheses for these studies were that increased risk for prostate cancer would be associated with either an increased testicular production of testosterone or an increased conversion of testosterone to DHT because of an increased Sa-reductase activity (134—136). Studies have focused on the notion that func- tional genetic polymorphisms in the Sa-reductase gene or in genes involved in testosterone biosynthesis (the CYP17 gene) or DHT catabolism (the 3B-hydroxysteroid dehydrogenase gene) could be responsible for increased testosterone production or increased DHT levels (136). In addition, polymorphisms have been discovered in the androgen receptor gene that can have functional significance for androgen receptor activity (137—139). Such polymorphisms have been postulated to be critical deter- minants of prostate cancer risk at the population or individual level by affecting intraprostatic DHT concentrations and andro— gen receptor transactivation (18,136). Steroid Hormonal Factors in Populations That Difler in Risk For Prostate Cancer Circulating of steroid hormone levels. A summary of the results of studies that compared circulating levels of steroid hormones in very high—risk African-Americans with those in high-risk European—Americans, lower-risk Asian-Americans, and very low—risk Asians living in Asia or African black men is provided in Table 1. The details of each study are summarized in the following paragraphs. Ahluwalia et al. (140) studied 170 African-Americans and 55 black-Nigerian men who were matched control subjects in a case~control study of prostate cancer and were older than 50 years. Plasma levels of testosterone and estrone were signifi- cantly higher in the African-American men than in the Nigerian men, whereas levels of DHT and estradiol were not different. Similar differences were found for the prostate cancer case pa- tients. Hill et al. (63—65) compared the hormonal status of small groups (n = 11—20) of 40— to 55—year-old African-American, European-American, and black (rural) South African men con- suming their customary diets (the effects of diet changes were also studied in these men; see earlier section on “Diet and Nu- trition”). In a separate study (141), African-American, Euro— pean-American, and black South African boys (ages 12—15 years) and young African-American and black South African men (ages 18—21 years) were compared. In the older men, plasma levels of the testosterone processor dehydroepiandros- terone (DHEA) were significantly lower in the two groups of black men than in the white men, whereas estrone levels were higher. Plasma levels of the testosterone processor androstene- dione and estradiol were significantly higher in the African men than in the two American groups, whereas no differences were 42 noted among these groups in testosterone levels. In the study with the 12- to 15-year-old boys and young men, similar find— ings were obtained for testosterone and DHEA. However, an- drostenedione levels were significantly lower (not higher) in the African than in the American subjects, and estradiol was lower in young black boys (12—14 years old) than in white boys but higher in older black boys (12—14 years old) and young black men than in white boys and men. These data suggest a complex interaction between ethnic background and environmental dif- ferences that change over the years of sexual maturation. In these studies by Hill and colleagues among South African black men, the 18- to 21-year-old men were different from those in 40- to 55-year-old men for androstenedione and DHEA. This diver- gence suggests that it is probably important for the interpretation of hormonal profiles to separately consider younger and older men. Ross et al. (134) compared 50 healthy young African- American men (at very high risk for prostate cancer) and 50 young European-American males (at half the risk of the black men). Total circulating testosterone was 19% higher, and free testosterone was 21% higher in the group of black subjects than in the group of white subjects. Serum estrone concentrations were also significantly higher (16%) in the black than in the white group. No significant differences were seen between the groups in circulating estradiol and SHBG levels. The authors estimated that the 19%—21% difference in circulating testoster- one is sufficiently large to explain the twofold difference in prostate cancer risk between white and black men in the United States. This study suggests an association between prostate can- cer risk and high concentrations of circulating androgens and, possibly, estrogens. Henderson et al. (142) compared circulating hormone levels in 20 pregnant African-American with 20 European-American women in their first trimester. Serum testosterone levels were 47% higher in black women than in white women, and estradiol levels were 37% higher. No significant differences were ob— served in circulating SHBG and human chorionic gonadotro- phin, or in relevant pregnancy characteristics, such as the sex ratio of the offspring. These findings suggest that African— American males are exposed to higher androgen concentrations than European—American males even before birth. The U.S. black and white men from the study by Ross et al. (134) were compared with 54 Japanese men of the same age (mean age, 19—23 years) in a follow-up study (135). The serum testosterone levels of Japanese men were not lower than those of the U.S. whites and blacks but were intermediate between these two groups, whereas their SHBG levels were significantly lower. This finding may suggest a higher percentage of free testosterone in the Japanese (at very low risk) than in the U.S. men (at high risk), but free testosterone was not measured. Com— pared with the Japanese men, the two U.S. groups had signifi- cantly higher circulating levels of the conjugated androgen me- tabolites androsterone glucuronide (4l%—50% higher) and 30L,17B-androstanediol glucuronide (25%—31% higher). This finding suggests that, in comparison with the high—risk U.S. groups, the low—risk Japanese population has a lower testoster- one metabolism, most likely a lower activity of the enzyme SOL-reductase that converts testosterone to DHT and the testos— terone precursor androstenedione to androsterone. However, the markedly higher levels of androsterone glucuronide in U.S. men could also be indicative of a higher testosterone production in comparison with Japanese men. Journal of the National Cancer Institute Monographs No. 27, 2000 Table 1. Summary of 11 studies of circulating steroid hormone levels (given as percentage of a referent group) in men from different ethnic groups* Asian-American Asian Mean agei African- European (South) African Hormone StudyT (range, y) American (American) Japanese Chinese Japanese Chinese black men Testosterone 1 250 y 100 Lower§ 2 43—49 (40—55) 108 100 108 3 12—15 Samell 100 Same 18—21 100 Same 4 20 (18—22) 118.6§ 100 5 20, 19, 23‘][ 111.3 100 104.7 6 25 (18—47) 100 Same 7 63, 70 (50—79) 100 85.7§ 8 38 (30—50) 103.5 100 101.6 9 70 (35—89) 105.2 100 110.5§ 108.8§ 10 20—39 100 Higher Lower# Free testosterone 4 20 (18—22) 121.2§ 100 6 25 (18—47) 100 ~100 9 70 (35—89) 104.0 100 106.7 112.0§ 10 20—39 100 Same Same % Free testosterone 4 20 (18—22) 103.2 100 Bioavailable testosterone 6 25 (18—47) 100 Same 9 70 (35—89) 103.3 100 105.1 111.4§ SHBG 4 20 (18—22) 104.8 100 5 20, 19, 23‘][ 109.2 100 71.4§ 7 63, 70 (50—79) 100 115.7§ 9 70 (35—89) 106.0 100 107.8§ 96.6 10 20—39 100 Higher Same# Dihydrotestosterone (DHT) 1 250 y 100 Same 6 25 (18—47) 100 112.6 7 63, 70 (50—79) 100 89.0 9 70 (35—89) 107.0§ 100 107.2§ 99.1 DHT/testosterone ratio 7 63, 70 (50—79) 100 107.0** 9 70 (35—89) 102.0 100 96.0 89.9§ Androsterone—glucuronide 5 20, 19, 2391 93.8 100 66.6§ 6 25 (18—47) 100 59.8§ Androstanediol—glucuronide 5 20, 19,23‘][ 95.4 100 76.4§ 6 25 (18—47) 100 56.8§ DHEA 2 43—49 (40—55) 76.9§ 100 67.6§ 3 12—15 Same to lower§ 100 Much lower§ 18—21 100 Same DHEA—sulfate 9 25 (18—47) 100 68.4§ 11 250% 83.8—99.4 100 Androstenedione 2 43—49 (40—55) 96.8 100 128.6§ 3 12—15 Same to higher§ 100 Much lower§ 18—21 100 L0wer§ 6 25 (18—47) 100 75.6§ Estrone 1 250 y 100 Same 2 43—49 (40—55) ll7.4§ 100 150.8§ 4 20 (18—22) 116.1§ 100 Estradiol-17B 1 250 y 100 Same 2 43—49 (40—55) 103.1 100 128.3§ 3 12—14 Lower§ 100 Same to lower§ 15 Higher§ 100 Higher§ 18~2l 100 Higher§ 4 20 (18—22) 109.8 100 7 63, 70 (50—79) 100 87.0§ *Values are presented as the percentage of the value (set at 100%) in a referent group [European(-American) group, except in study 1, in which African-Americans are used as the reference group]. A two—sided P value of less than .05 was considered significant. Unless designated, none of the other differences are statistically significant. "Studies: 1) Ahluwalia et al. (140); 2) Hill and colleagues (63—65); 3) Hill et al. (141); 4) Ross et al. (134); 5) Ross et al. (135); 6) Lookingbill et al. (143); de Jong et al. (144); 8) Ellis and Nyborg (I45); 9) Wu et al. (146); 10) Santner et al. (148); ll) Corder et al. (147). :tMean (occasionally median) age (in years) is given with the range (if available) in parentheses. For studies 5 and 7, the mean age is given for each ethnic group consecutively. If mean age is not available, only the age range is given. §Statistically significantly different from the European(—American) group (referent). llIf exact numerical information is not available, significant higher or lower values are indicated as such and absence of significant differences are indicated by “same.” ‘][Exact age range is not available. #Significantly different from Chinese—Americans. **Significant only after age adjustment. Journal of the National Cancer Institute Monographs No. 27, 2000 43 Lookingbill et al. (143) reported a similar observation, com- paring 53 normal healthy U.S. Caucasians and 57 Chinese males in Hong Kong between the ages of 24 and 26 years. The Cau- casian men had 67% higher serum levels of androsterone gluc- uronide and 76% higher levels of 3u,17B-androstanediol gluc— uronide than the Chinese men did. Circulating levels of testosterone, free testosterone, or DHT were not significantly different, but Caucasian men had 46% higher serum levels of the androgen precursor DHEA sulfate and 32% higher levels of androstenedione. These data are also suggestive of a higher 50L- reductase activity in high-risk Caucasians than in low-risk Chi- nese men, and they suggest an increased production of androgen precursors in the Caucasians. In contrast to the observations of Ross et al. (134,135) and those of Lookingbill et al. (143), De Jong et al. (144) found 71% higher circulating total testosterone levels in 123 Caucasian— Dutch men (high risk) than in 91 Japanese men (low risk). The men in these studies were considerably older (50—79 years) than those studied by the previous two other groups. DHT levels were not different, but the ratio of DHT to testosterone was 10% lower in Dutch men than in Japanese men, possibly reflecting lower Sa-reductase activity; no data were presented on androgen metabolites. Serum levels of estradiol were 15% higher (signifi- cant) in the Dutch men than in the Japanese men. SHBG levels were not different, but the ratio of testosterone to SHBG con- centrations was 34% higher in Dutch men than in Japanese men, which suggests higher amounts of free testosterone in Dutch men, but this parameter was not measured separately. Ellis and Nyborg (I45) studied 4462 U.S. Army Vietnam veterans, ages 31—50 years, and compared serum testosterone levels in 3654 non-Hispanic white men (mean 6.37 ng/mL) with those in 525 African-Americans (6.58), 200 Hispanics (6.33), 34 Asian/Pacific Islanders (6.89), and 49 Native Americans (6.31). The serum testosterone levels in the African-American men were significantly higher than those in the non-Hispanic white men, but the differences among the other groups were not sig- nificant. The serum testosterone difference between black and white men was larger in men between 31 and 35 years of age (6.6%) than for men ages 35—40 years (3.7%) or ages 40—50 years (0.5%). No other hormones were measured in this study. Wu et al. (146) conducted a population-based study, compar- ing circulating hormone levels in 1127 healthy men: 325 Afri- can-American men, 411 European-American men, 126 Chinese— Americans, and 275 Japanese-Americans with a median age of 69.6 years (range, 35—89 years), 8.2% of whom were 60 years or younger. Serum levels of total testosterone were slightly, but significantly, higher (9%—11%) in Asian-Americans than in Eu- ropean-American men, whereas they were intermediate and not significantly different from the two other groups in African- Americans. The same pattern was found for serum levels of bioavailable testosterone (not bound to SHBG) and the percent- age of free testosterone (not bound to either SHBG or albumin), but only the 11%—12% difference between Chinese-American and European—American men was significant. SHBG levels were not different among the four groups. In comparison with Euro- pean-American men, DHT levels were 7% higher (significant) in high-risk African-Americans and low-risk J apanese-Americans, but similar in Chinese-Americans. The ratio of DHT to testos- terone was 10% lower (significant) in Chinese-Americans than in European-Americans, but not significantly different in Afri- can-Americans and J apanese-Americans who had slightly higher 44 and lower ratios, respectively, than European-Americans. These data do not appear to provide clear support for the notions of a relation between increased SOL-reductase activity or testosterone production and prostate cancer risk, but this study did not in- clude more direct indicators, such as androsterone glucuronide and 30L,l7B—androstanediol glucuronide. DHEA sulfate was measured by Corder et al. (147) in stored serum samples of 90 African—American and 91 European- American men with prostate cancer and equal numbers of matched controls who were identified in a nested case—control study in a cohort of men in the Kaiser Permanente Medical Care Program in Northern California. Regardless of age, no signifi— cant differences were found between the two groups in DHEA sulfate levels, which were lower in men 57 years and older than in younger men. Santner et al. (148) conducted the only study to date in which androgen production and metabolism by Sa-reductase were de- termined in a direct fashion in populations with different risk for prostate cancer. A radioisotope method involving intravenous administration of tritiated testosterone was used to measure the conversion of testosterone to DHT in healthy European- Americans (ages 22—27 years), Chinese-Americans (ages 20—37 years), and Chinese men living in Beijing, China (ages 24—39 years). No differences in conversion of testosterone to DHT were found among these three groups. Circulating testosterone and SHBG levels were lower in the Beijing Chinese than in the two U.S. groups, and the differences with the U.S. Chinese subjects were significant, whereas no differences were found in free testosterone. There was a nonsignificant trend toward lower calculated metabolic conversion rates of testosterone comparing European-Americans with the Chinese groups and U.S. Chinese with Beijing Chinese. Calculated testosterone production rates were lower in Beijing Chinese than in the two U.S. groups, the difference with American Chinese being significant. The ratios of urinary SB— to SOL—reduced steroids, which are an indicator of overall Sa—reductase activity, were also not different in 20 Eu- ropean-American male students compared with 20 Chinese stu- dents living in Hong Kong. Urinary excretion of androsterone, etiocholanolone, and total ketosteroids was lower in the Chinese than in the U.S. students, which was significant when the data were combined with those of 20 female students from each of the two populations. Taken together, these data indicate that Set-reductase activity is not different in Asian and Caucasian men and is not affected by the environment in which Asian men live. However, these results suggest that the living environment influences testosterone production in Asian men. Polymorphisms in genes involved in steroid hormone me- tabolism and action. Studies have addressed the hypothesis (137—139) that functional polymorphisms in the 5a-reductase gene, in genes involved in testosterone biosynthesis or DHT catabolism, and in the androgen receptor gene could be associ— ated with the differences in prostate cancer risk among various populations. These studies are summarized in the following paragraphs. The SRD5A2 gene, which encodes for human type II 50L- reductase enzyme, is expressed in the prostate and is located on chromosome 2p23 (149,150) and contains polymorphic TA di- nucleotide repeats in its transcribed 3’ untranslated region (151). Reichardt et al. (152) demonstrated that TA(0) [87 base pairs (bp)] is the most common allele and was homozygous in 81% of non—Hispanic, white Americans (11 = 68), 78% of Asian- Journal of the National Cancer Institute Monographs No. 27, 2000 Americans (n = 37), and 67% of African-Americans (n = 94). The next most common allele TA(9) (103-105 bp) is heterozy— gous with the TA(O) allele and occurred in 19% of the non- Hispanic, white American men, 22% of the Asian-Americans, and 15% of the African—Americans. The TA(18) allele (212—131 bp) was only found in African-Americans (18%) as heterozy- gous with the TA(O) allele in all except one who was homozy- gous. Thus, the longer alleles are unique to African-Americans and may be related to their extremely high risk for prostate cancer. However, the functional significance of these polymor- phisms is not yet known. Makridakis et al. ([53) identified another polymorphism in the SRD5A2 gene, the presence of a valine to leucine mutation at codon 89. If this mutation occurs in a homozygous state, it confers 28% lower SOL-reductase activity as measured in Asian men with this genotype compared with heterozygous men and men without the mutation. These researchers observed that the frequency of the 89 valine—valine genotype was 59% in African- American men (n = 95), 57% in non—Hispanic white Americans (11 = 49), 48% in Latino Americans (11 = 40), and 29% in Asian-Americans (n = 102). The 89 valine—leucine genotype occurred in 37%—39% of African-Americans, non-Hispanic white Americans, and Latino Americans, and in 49% of Asian- Americans. The frequency of the 89 leucine—leucine genotype (lower SOL-reductase activity) was 3%—4% in African-American and non-Hispanic white Americans, 15% in Latino Americans, and 22% in Asian-Americans. A recent report from another, larger study by Lunn et al. (154) is essentially consistent with these findings. In this study, the frequency of the 89 valine— valine genotype was 65% in African-American men (n = 118), 41% in European-Americans (n = 176), and 15% in Asians (Taiwanese) (n = 108). The 89 valine—leucine genotype oc— curred in 32% of African-Americans, 50% of European- Americans, and in 57% of Asians. The frequency of the 89 leucine—leucine genotype (lower SOL-reductase activity) was 2.5% in African—Americans, 8.5% in European-Americans, and 28% in Asian~Americans. The higher frequency of the 89 leucine—leucine genotype in Latino American men and particu- larly Asians may be related with the lower risk for prostate cancer found in these two ethnic groups, and the low frequency of 86 leucine alleles in African-Americans may be related to their extremely high risk. However, there appears not to be a relation between plasma concentrations of 3a-androstanediol- glucuronide as an indicator of SOL-reductase activity and the three different SRD5A2 gene codon 89 genotypes (155). Makridakis et al. (156) also identified another polymorphism in the SRD5A2 gene, a mis-sense alanine to threonine mutation at codon 89. An in vitro construction of the mutant enzyme displayed a substantial increase in activity (Vmax). The fre- quency of the mutation was very low, 1.0% and 2.3%, in healthy, high—risk African-Americans and lower—risk Hispanic men, respectively. Although no data were presented on other ethnic/racial groups, it seems unlikely that this mutation is re- sponsible for the large ethnic/racial variations in prostate cancer risk. The CYP17 gene, which encodes for the cytochrome P450C170L enzyme that has both 170t-hydroxylase and 17,20- lyase activity in the adrenal and testicular biosynthesis of an- drogens, is located on chromosome 10q24.3 (15 7). This gene is polymorphic with two common alleles, the wild-type CYP17A1 allele and the CYP17A2 allele containing a single base pair Journal of the National Cancer Institute Monographs No. 27, 2000 mutation in the untranscribed 5’ region of exon 1 (157). This mutation creates an additional Spl site in the promoter region, suggestive of increased expression potential (157). The func- tional significance of this polymorphism in men is not known, but premenopausal and postmenopausal women with the A2 allele have been reported to have higher circulating estradiol and progesterone levels than women homozygous for the A1 allele (158,159). Circulating levels of DHEA and androstenedione, but not testosterone, were increased in postmenopausal women (159). Lunn et al. (154) recently reported that the frequencies of the A1/A1 and Al/A2 genotype were between 40% and 44%, and the frequency of the A2/A2 genotype was 16%—17% in both African-American men (n = 115) and European-Americans (n = l 15), accounting for an A2 allele frequency of 0.36—0.38. In Asians (Taiwanese; n = 110), however, the Al/Al genotype occurred in 24%, the Al/A2 genotype in 49%, and the A2/A2 genotype in 27%, with an A2 allele frequency of 0.52. The frequency differences between the Asians and the two American groups were statistically significant and are perhaps related to the low risk for prostate cancer in Asian men. Verreault et al. (160) reported complex dinucleotide poly- morphisms in the 3rd intron of the human HSD3B2 gene, located on chromosome 1p13, which encodes type II 3B-hydr0xysteroid dehydrogenase, which is expressed in the adrenals and testes, and catabolizes DHT (I61). Devgan et al. (162) reported that the frequency of HSD3B2 alleles differs between African- American, European-American, and Asian men. One minor al- lele is unique for African-American men (6% allele frequency), whereas the most common allele is more frequent in European- Americans (52%) than among African-American or Asian men (34%—37%). The second most common allele is more frequent in African-Americans (25%) than in either Asians (15%) or European-American men (11%). As with the TA dinucleotide polymorphisms in the SRD5A2 gene, the functional significance of these HSD3B2 gene polymorphisms is not known. The human androgen receptor gene, which is located on the X chromosome, also contains polymorphisms that are found as 8—31 CAG and 8—17 GGC (or GGN) microsatellite repeats in exon 1 encoding for the N-terminal domain of the protein where transactivation activity resides (139). The CAG repeat length has been demonstrated to determine transactivation activity of the androgen receptor, with 40 or more repeats being associated with human androgen insensitivity syndromes, such as spinal and bulbar muscular atrophy, and reduction of repeat length leading to increased transactivation activity in vitro (137—139). The functional significance of the GGC repeat length is not clear. Irvine et al. (163) reported that 75% of African-Americans (n = 44) had CAG repeat lengths of less than 22, whereas 62% of European-Americans (n = 39) and 49% of Asian-Americans (n = 39) had such shorter alleles. Very short alleles (<17 CAG repeats) occurred almost exclusively in African-Americans. The most common GGC allele (16 repeats) was found in 70% of Asian-Americans, 57% of European-Americans, and only 20% of African—Americans. The frequency of short GGC repeats (<16) was 61% in African-Americans, 27% in Asian-Americans, and 11% in European-Americans. GGC repeats longer than 16 were rare in the Asian-American men (3%) but more frequent in African-Americans (20%) and European-Americans (32%). Sar— tor et al. (164) essentially confirmed the findings on CAG re- peats in a sample of African-Americans (n = 65) and European- American men (n = 130). Mean and median number of CAG 45 repeats was 19 in African—Americans and 21 in European- Americans, and 57% of the African—American men had less than 20 repeats, whereas only 28% of European—American men had such short repeats. Ekman et al. (165), however, did not find significant differences in the distribution of CAG repeats com- paring Swedish and Japanese men with BPH but without cancer (n = 38 and 33, respectively), but Swedish men with prostate cancer (11 = 118) had somewhat shorter CAG repeats (mean, 15.9; median, 15) than Japanese prostate cancer patients (n = 34; mean, 17.5; median, 17). In conclusion, in two studies short CAG repeat alleles in the androgen receptor gene, which are probably associated with greater androgen receptor transactiva— tion activity, were most frequent in the highest-risk popula— tion (African-Americans) and least frequent in the lowest-risk group (Asian-Americans), whereas the frequency was interme— diate in intermediate-risk European-Americans. The high fre~ quency of short GGC repeats found in African-Americans may also be related with their extremely high risk for prostate cancer, but the functional significance of this polymorphism is not yet known. . Summary and conclusions. When examining Table 1, few clear or convincing patterns emerge about associations between circulating hormone concentrations and prostate cancer risk at the population level. Two studies examined levels of androste- rone glucuronide and 3a,17B—androstenediol glucuronide, which are considered (166—168) indicators of Sci-reductase activity, particularly 3a,l7B-androstenediol glucuronide, which is a di- rect metabolite of DHT. In both studies, the levels of these Sci-reduced androgen metabolites were lower in low—risk Asian populations than in high—risk European-Americans (135,143). These findings suggest lower Soc-reductase activity in the Asians and consequently reduced formation of DHT and androgenic stimulation of the prostate. This notion is supported by the re— ported higher frequency in Asians than in European-Americans or African-Americans of a polymorphism in the 50L’reductase (SRD5A2) gene that appears to be associated with lower 504— reductase activity (a valine to leucine mutation at codon 89) (136,153). However, no differences were found between Asians and European-Americans in a study in which overall conversion of testosterone into DHT was directly measured (148). Further- more, androsterone glucuronide and 3a,17,or-androstenediol glucuronide levels were not higher in very high—risk African- Americans than in intermediate-risk European-Americans, and circulating levels of DHT and the ratio of DHT to testosterone were not different in ethnic populations (Asian, black, and white) that differ in prostate cancer risk (143,144,146). Thus, the relation between Sat—reductase activity and prostate cancer risk at the population level remains unclear at present. Circulating levels of testosterone and/or free testosterone were slightly higher in African—Americans than in European- Americans in five of six studies that examined this question, but this finding is statistically significant in only one study (134). Furthermore, lower as well as higher testosterone concentrations have been found in lower-risk Asian or African men compared with European-Americans or African-Americans, although tes- tosterone levels were lower in Asians living in Asia than in American populations regardless of ethnicity in two of three studies. Thus, these studies in ethnic/racial groups provide at present no substantive evidence in support of the hypothesis of a causal positive relation between elevated SOL—reductase activity and prostate cancer risk at the population level and only very 46 limited evidence for elevated (free) testosterone levels being associated with prostate cancer risk. The only other patterns appearing in the data in Table 1 are that levels of estrogens are slightly higher (in five of five studies) and those of DHEA (sulfate) lower (in three of three studies) in black Africans and African-Americans than in men of European descent—hardly any data are available on Asians in this regard (63—65,134,140,141,144). The biologic significance of these ob— servations is unclear, but they may be related to the high sus- ceptibility of black men to prostate cancer when they live in the American environment. However, the above summarized endo- crine differences between very high-risk African-Americans and high-risk European—Americans were not consistent in younger and older men, and they were not similar to the differences observed between the high-risk U.S. populations and the low- risk African black men (63-65,140,141). These inconsistencies raise the possibility that the factors and endocrine mechanisms that determine the difference in risk between African black men and African-Americans are dissimilar from those that determine the risk difference between African-Americans and European— Americans (11). Finally, CAG repeat length polymorphism in the androgen receptor gene was found to be associated with prostate cancer risk in two studies. Short CAG repeat alleles are probably as- sociated with greater androgen receptor transactivation activity. Such short CAG repeat alleles were most frequent in African- Americans (very high—risk), least frequent in Asian-Americans (low risk), and intermediate in European-Americans (intermedi- ate risk). Another androgen receptor polymorphism in GGC re- peats may also be related with risk for prostate cancer, but the functional significance of this polymorphism is unknown. Association of Steroid Hormonal Factors With Prostate Cancer Risk in Population-Based Case—Control Studies Circulating of steroid hormone levels in nested case— control studies. A summary of the results of population-based, nested case—control studies that examined the association be- tween circulating levels of steroid hormones and risk for prostate cancer is provided in Table 2. The details of each study are summarized in the following paragraphs. One study by Carter et al. (169) concerned only 16 case subjects and contained consid- erable bias because of storage effects on hormone measure— ments, which were recognized but not controlled for. This study is, therefore, not further discussed here. Nomura et al. (170) compared 98 prostate cancer case pa— tients with matched control subjects from a cohort of 6860 Ha— waiian-Japanese men, which were followed for an average of 14 years. No significant differences were found between case pa- tients and control subjects or associations with risk for serum levels of testosterone, DHT, estrone, estradiol, and SHBG, mea- sured once at the start of the cohort study (free testosterone levels were not determined). An elevation in risk was only ob- served for an increasing ratio of testosterone to DHT, which was borderline significant. This latter observation perhaps suggests an inverse relation between (peripheral) Sci-reductase activity and prostate cancer risk. Barrett-Connor et al. (171) followed a Californian cohort of 1008 white, upper-middle class men between the ages of 40 and 79 years for a period of 14 years, during which time 57 cases of prostate cancer occurred (26 deaths and 31 incident cases). No significant relation was found between the risk for prostate can— Journal of the National Cancer Institute Monographs No. 27, 2000 cer and baseline serum concentrations of testosterone, estrone, and SHBG. However, RR increased linearly with an increasing serum level of androstenedione, a testosterone precursor. RR also increased linearly with an increasing serum level of estra- diol, but this finding was not statistically significant. Hsing and Comstock (I72) and Comstock et al. (173) re— ported results of a population-based, nested case—control study in a cohort of 25 620 men (98% European-American) in Mary- land. Blood samples were obtained in 1974, and 98 cases of prostate cancer were identified in the first 13 years of follow-up (81 cases in 12 years of follow-up for DHEA and DHEA sul- fate). Men 70 years and older as well as men younger than 70 years were studied separately (except for DHEA and DHEA sulfate). No significant differences were found between case patients and control subjects or associations with risk for base- line serum testosterone, DHT, DHEA, DHEA-sulfate, estrone, or estradiol. The ratio of testosterone to DHT was higher in case patients than in control subjects of all ages, and, for men younger than 70 years but not for older men, risk for prostate cancer increased with an increasing testosterone/DHT ratio; both find- ings were borderline significant (0.0520 6.3, mean 14 (5—23) 10, mean 4.1, median <8—24 Age at entry, date/time BMI, age, smoking, BMI, smoking, alcohol, —— — Smoking, BMI, other of blood sampling alcohol, exercise, all diabetes hormones other hormones Mean Median Median Mean Mean Mean 100.2 102.8 104.7 97.1 98.2 101.6 OR = 1.03 OR = 2.60§H OR = 1.00 OR = 0.83 RR = 0.8 RR = 1.23‘][# 103.6 100.0 102.6 OR = 1.09 OR: 1.14 RR: 1.1 91.7 96.9 100.9 OR = 0.46§1| RR = 0.8 RR = 1.12 103.4 94.4 98.0 95.7 OR = 0.82 OR = 0.71 OR = 0.83 RR =0.7 0.32 for >62 y||) 99.0 102.7 99.2 Increased** OR = no data OR = 2.35§1| OR = 1.31 RR: 1.7 97.4 101.2 (OR = 1.37) (OR = 1.13) 109.2 104.6 106.4 102.1 102.3 OR = 0.85 OR-1.60‘][ OR = 1.16 OR = 1.10 RR = 1.2 106.3 RR = l 2 104.7 99.4 92.0 OR = 1.24 RR = 1.0 RR = 0.92 1000 RR = 0.8 97.7 100.0 OR = 0.56§H RR = 1.1 >"Values presented are hormone values (means or medians) of cases as percentage of the values in controls (set at 100%) and risk estimates [as either odds ratios (OR) or relative risks (RR)] of highest tertile or quartile relative to the lowest tertile, quartile, 0r quintile (set at 1.00). The indicated values are for cases presented as the percentages of the values in controls (set at 100%) or risk estimates indicating the relation with prostate cancer risk. Results of statistical analysis are indicated only when significant at a (twovsided) P value of less than .05, considering tests for differences between the lowest (referent) and highest tertile or quartile as well as for trend. TThis study ca1culated prostate cancer rates for case patients and compared them with population data, which were also used to present median hormone levels for case patients and control subjects. RRs were calculated for an increase in hormone concentration equal to 1 standard deviation. iThe risk estimates by hormone level were adjusted for the concentrations of all other hormones considered; this adjustment was not done in any of the other studies. §Statistically significant difference between lowest (referent) and highest tertile or quartile different from controls. HStatistically significant trend. l][Borderline significant trend (.0516 1.00 Referent 56/75 $16 1.60* 1.07—2.41 201/175 Platz et al. (187) 582 794 Not 23 1.00 Referent 244/369 23 1.20 0.97—1.49 338/425 *Odds ratio (OR) or relative risk (RR) is significantly different from referent value at the P<.05 level. CI = confidence interval. Table 5. Summary of three case—control studies of the interaction of CAG and GGC/GGN repeat length polymorphisms in the androgen receptor circulating levels in relation to risk for prostate cancer No. of No. of CAG/GGC—N repeat N (case subjects/ Study (reference No.) case subjects control subjects comparison OR/RR 95% CI control subjects) Irvine et al. (163) 57 37 22/16 1.00 Referent 34/28 <22/not 16 2.10 (P = .08) Not presented 23/9 Stanford et al. (184) 257 250 ZZZ/>16 1.00 Referent 22/32 >22/S16 1.15 0.56—2.35 32/41 <22/>16 1.54 0.83——2.86 97/93 <22/216 2.05* 1.09—3.84 98/77 Trend: P = .008 Platz et al. (187) 582 794 >23/not 23 1.00 Referent 66/119 >23/23 1.17 0.77—1.77 90/133 21—23/not 23 1.39 0.93—2.06 75/116 21—23/23 1.22 0.82—1.83 152/185 <21/not 23 1.49* 1.02—2.15 103/134 <21/23 1.62* 1.07—2.44 96/107 *Odds ratio (OR) or relative risk (RR) is significantly different from referent value at the P<.05 level. CI = confidence interval. cancer risk and elevated levels of testosterone and androstene- dione or decreased levels of SHBG and estradiol were found each in only a single study [(175) for testosterone, SHBG, and estradiol; (171) for androstenedione], and they were not ob- served in eight (testosterone), four (SHBG and estradiol), or three (androstenedione) other studies. It is possible that relevant associations may have been missed in most studies, because the data for each individual hormone were not adjusted for concen— trations of other hormones studied, even though there are many intercorrelations between circulating levels of these hormones. Only in the study by Gann et al. (175) were these types of adjustments applied, after which risk was significantly associ— ated with increased circulating testosterone levels and testoster— one/DHT ratio, as well as decreased concentrations of SHBG and estradiol, and, in men older than 61 years, DHT. A meta- analysis study by Eaton et al. (188) used most but not all studies included in this overview, as well as a study that was discounted here (169) and some unpublished data. They found no signifi- cant differences for the ratios of mean hormone levels between case patients and control subjects, with the exception of slightly elevated levels of 3a,l7B-androstenediol glucuronide. This analysis is essentially in agreement with the analysis of this overview, with the only consistent finding slightly elevated ra- tios between case patients and control subjects of 3a,17B- androstenediol glucuronide in five of five studies (Table 2). However, Eaton et al. (188) did not take into account the risk estimates produced by these studies, which seriously limits its conclusions. The results of three nested case—control studies on the rela— Joumal of the National Cancer Institute Monographs No. 27, 2000 tion between prostate cancer risk and two different polymor- phisms in the human type II 5a—reductase enzyme gene (SRD5A2) do not support the notion of an association between risk and increased Set-reductase activity (154,155,180). How— ever, an infrequent polymorphism associated with increased 50(- reductase activity was more common in case patients than in control subjects in one case—control study ( 156), which indicates that associations between polymorphisms in the SRD5A2 gene and prostate cancer risk cannot be discounted. Data on a relation between prostate cancer risk and a polymorphism in the CYP17 gene, which encodes for the 170t-hydroxylase and 17,20—1yase activity involved in androgen biosynthesis, are contradictory (154,181). Furthermore, the functional significance of this poly- morphism in males is not yet known. In four of five similar nested case—control studies of polymorphisms in trinucleotide repeats in the promoter region of the androgen receptor gene, an association was found between risk and short CAG repeat al— leles—short CAG repeat lengths are associated with greater an— drogen receptor transactivation activity. However, this associa- tion was weak and significant in only one study. An association between risk and polymorphisms in androgen receptor GGC or GGN repeat lengths is not clear because results of three studies were inconsistent, and the functional significance of these poly- morphisms is not known. There is possibly an interaction be- tween CAG and GGC/GGN repeat length in relation to prostate cancer risk, but results of the three studies examining this pos- sible interaction were inconsistent. Short CAG repeats were also correlated with advanced disease and/or early onset of prostate cancer (182—186,189,190). 51 Epidemiologic Evidence for Involvement of Steroid Hormones: Summary and Conclusions Taken together, the results of the above summarized studies do not provide unequivocal or strong evidence for any particular association between prostate cancer risk and circulating levels of hormones or polymorphisms in genes that encode for proteins involved in steroid hormone action or metabolism. Only a few associations with prostate cancer risk have been observed con— sistently (in at least three studies), and they are weak at best: 1) slightly, but mostly not significantly, higher circulating testos- terone and estrogen levels and lower DHEA (sulfate) levels in high-risk African-American men as compared with lower-risk European-American men, and 2) a CAG repeat length polymor- phism in the androgen receptor gene with short repeat lengths associated with increased risk and increased receptor transacti— vation activity. The evidence for involvement of activity of the enzyme Sa—reductase, which is critical in androgen action in the prostate, is inconsistent and contradictory. Difi‘iculties in Interpretation Several important points should be considered in interpreting these observations: First, there are numerous potential problems with most studies that measure circulating hormone levels, such as the usually large intra- and interassay variability in the im- munoassays used (122,191—194). Typically, only single blood samples are available, and within-subject variations over time and possible differences in circadian rhythms cannot be taken into account. Another problem is that there are many interrela- tionships between various hormones (144,174), which only an occasional study has taken into account during data analysis (175). Second, young Japanese men and Chinese men from Hong Kong are probably at least partially westernized in their lifestyle (194), and they can, therefore, not simply be compared with older Asian men. Young men that are hormonally studied today may have a prostate cancer risk that is different from the currently recorded risk in older men of the same population, as suggested by the rising prostate cancer rates in Japan (194). Third, the factors that cause the differences in prostate cancer risk between black, white, and Asian men in the United States may be different from those that determine differences in pros- tate cancer risk between Asian or African populations and popu— lations in the United States or West European countries, as in— dicated earlier. Another crucial issue is that circulating hormone levels and polymorphisms in critical genes provide very little information about concentrations at the molecular targets of these hormones in the prostate gland or about steroid hormone metabolic pro- cesses within the prostate. For example, less than 10% of circu- lating DHT is produced by the prostate, and a substantial pro- portion of serum 3a,l7B-androstenediol glucuronide is derived from nonprostate sources; these two steroids are, therefore, not very good indicators of prostatic SOL-reductase activity (166— 168). Also, aromatase activity has been identified in the human prostate and the LNCaP prostate cancer cell line (195—199), although there are reports of contradictory findings (200,201) that may be related to methodologic differences. In addition, estrogen levels in the human prostate exceed those found in the circulation (202). Thus, local formation of estrogens in the pros- tate may occur and may contribute to disregulation of growth. Finally, although there is some information about the functional 52 significance of some polymorphisms in genes encoding for pro- teins involved in steroid hormone action or metabolism, their influence on the prostatic activity of steroid hormone metabo- lizing enzymes or the activity of steroid hormone receptors in the prostate is not known. In View of the highly complex and often tissue-specific mechanisms of regulation of gene expression, it is likely that these polymorphisms have only limited and probably cell type—specific influences on these regulatory processes. Hypotheses From the studies summarized above, four possible hypoth- eses emerge about steroid hormone factors associated with prostate cancer risk: 1) slightly elevated (bioavailable) testoster- one serum levels, as indicated by studies comparing healthy low- and high-risk men (134,135,140,144,148); 2) increased peripheral and possibly prostatic activity of SOL—reductase (135,143,153,175); 3) slightly increased serum levels of estro- gens, as indicated by studies comparing healthy low- and high— risk men (63—65,134,140,144); and/or 4) increased androgen receptor transactivation activity related to short CAG repeats in the promoter region of the androgen receptor gene (163,182— 184). However, there are contradictory data for each of these hypotheses, as indicated earlier and documented in Tables 1—5, and the observed associations were at best weak. Circulating Androgens and Estrogens Two of these four hypotheses implicate higher bioavailable circulating androgen levels in high-risk men compared with low- risk populations (191), which may be related to increased an- drogen production or exposure (136). For example, although body mass index or obesity does not appear to be a risk factor, there are some indications that muscle mass is positively asso- ciated with risk, perhaps reflecting exposure to endogenous an- drogens or anabolic steroids. However, studies of polymor- phisms in the CYP17 gene (which encodes for the cytochrome P450C170L-hydroxylase and 17,20-lyase activity involved in an- drogen biosynthesis) do not support the notion of a relation between risk and androgen production rates. As will be detailed later, the results of several animal model studies strongly support this contention, but more research is needed to confirm and further define this association in humans and to establish its underlying biologic mechanisms (increased androgen produc— tion or SOL-reductase activity and decreased DHT catabolism) (191). Furthermore, elevated androgen levels do not universally occur in all high-risk groups. Meikle and colleagues (203,204) studied brothers (ages 47—75 years) and sons (ages 22—43 years) of prostate cancer patients who have a threefold to fourfold excess risk compared with unrelated control subjects of the same age ranges. They reported that serum levels of testosterone and DHT were significantly lower, rather than higher as one might expect, in these blood relatives of prostate cancer patients. Be- cause circulating testosterone levels may thus be lower in men with a family history of prostate cancer than in other men, hor- monal involvement in familial aggregation of prostate cancer risk seems paradoxical and the involvement of androgens in hereditary prostate cancer may be different from that in sporadic prostate cancer. Zumoff et al. (205) observed that circulating levels of testosterone, but not DHT, were markedly lower in prostate cancer patients younger than 65 years than in those patients 65 years and older. However, control subjects had tes- Journal of the National Cancer Institute Monographs No. 27, 2000 tosterone levels that were similar to those Of prostate cancer patients of 65 years and older. In several of the studies, sum- marized in Tables 1 and 2, findings were markedly different when comparing younger (18 to 25—40 years) and older healthy men (>40 years) or comparing younger (<62—70 years) and older men (>62—70 years) with prostate cancer and their age—matched controls. Circulating testosterone levels are also known to para- doxically decrease with aging, whereas prostate cancer risk in- creases (144—146,206). At the same time, SHBG levels increase with age and estrogen levels remain constant or increase (144,146,206). Thus, bioavailable estrogens and particularly tes- tosterone decrease with increasing age and increasing risk for prostate cancer. This situation may explain the lower prostatic concentrations of DHT with aging reported by Krieg et al. (202) but is in contrast to increasing prostatic estrogen levels they observed with aging. These observations suggest that the role of androgens and estrogens in prostate carcinogenesis may differ in younger men (early onset prostate cancer) and in older men (late-onset cancer) and may be different in men that are at high risk because of familial predisposition and those at high risk associated with their ethnic background or living environment. It is also possible that risk-increasing effects of elevated circulat- ing levels of androgens and estrogens may be effectuated early in life (134,135,141,143) or even before birth (142,207), rather than in the one or two decades preceding the diagnosis of pros- tate cancer. Another risk factor may be increased androgen receptor ac— tivity related to genetic polymorphisms in the androgen receptor gene. Although body mass index or obesity does not appear to be a risk factor, there are some indications that muscle mass is positively correlated with risk, perhaps reflecting exposure to endogenous androgens or anabolic steroids. Heavy alcohol use accompanied with liver disease may increase risk and be related with decreased clearance of estrogens and elevated circulating estrogen levels. Estrogen levels were also elevated in healthy black men living in the United States compared with European- American men, and this is perhaps associated with the very high risk for prostate cancer of black men living in the United States. However, no association between risk and circulating estrogen levels was found in nested case—control studies (in predomi- nantly European-American cohorts). Thus, the epidemiologic evidence for involvement of androgenic and estrogenic steroid hormones in human prostate carcinogenesis remains inconclu- sive (191). Interactions of Environmental, Hormonal, and Racial/Ethnic Factors The strongest single risk factor for prostate cancer appears to be a western lifestyle, particularly western dietary habits, includ- ing a high-fat intake. It is conceivable that dietary risk factors, such as fat, exert their enhancing effects mediated by a hormonal mechanism that involves androgens (11,15,208). For example, heavy alcohol use accompanied with liver disease may increase risk and be related with decreased clearance of estrogens and elevated circulating estrogen levels. However, it is unlikely that lifestyle is the sole factor that explains the differences in prostate cancer risk between Asian and American populations (10,11,136). Genetic factors are probably also important deter- minants of racial/ethnic disparity of sporadic prostate cancer, such as 5a-reductase activity or increased androgen receptor Journal of the National Cancer Institute Monographs No. 27, 2000 activity related to polymorphisms in the SOL-reductase or andro— gen receptor genes. The single most important combination of risk factors is to be of sub-Saharan African descent and to reside in the United States—African-Americans, as a group, have twice the risk of European-Americans. The reasons for the black—white disparity in prostate cancer rates in the United States are not understood. Environmental exposures (in the broadest sense of the term) are probably responsible for a large fraction of this disparity (10,11,15,209). A relation with similar racial disparities in ex- posure to potential carcinogens and high-risk dietary habits has been proposed (11,22,59). However, genetic factors, such as polymorphisms in the SOL-reductase or androgen receptor genes may be related with increased SOL-reductase activity or androgen receptor activity. However, environmentally influenced hor- monal mechanisms may be involved as well, possibly acting in utero (207). For example, young, African—American men and pregnant, African—American women have been reported to have higher circulating levels of androgens and estrogens than Euro- pean-Americans (142). Estrogen levels were also elevated in healthy black men living in the United States compared with European—American men, and this finding is perhaps associated with the very high risk for prostate cancer of black men living in the United States. However, no association between risk and circulating estrogen levels was found in nested case—control studies (in predominantly European-American cohorts). Conclusions The epidemiologic evidence for involvement of androgenic and estrogenic steroid hormones in human prostate carcinogen- esis remains inconclusive (191). The most promising hormonal risk factor candidates are elevated circulating testosterone and estrogen levels and polymorphisms in the androgen receptor gene associated with increased receptor transactivation activity. In addition, hormonal effects Of dietary factors, such as fat, may play a critical role in prostate carcinogenesis in humans, as well as, perhaps, still unexplored/unknown polymorphisms in genes encoding for proteins involved in steroid hormone metabolism and hormone action. PROSTATE CANCER AND STEROID HORMONES IN LABORATORY ANIMALS Spontaneously occurring prostate tumors are rare in most species (5—7,210), with exception of the dog and, particularly, humans. It is not understood why prostate cancer is so common in men, whereas it is very rare in almost all other species. There are compelling reasons to implicate hormones, particularly an— drogenic and estrogenic steroids, in human prostate carcinogen- esis, as indicated earlier. The same steroid hormones are also very powerful factors in the induction of prostate cancer in ro- dent species in which spontaneous prostate neoplasms are rare (15,56,211,212). Pertinent studies concerning the role of andro- gens and estrogens in experimental prostate carcinogenesis are summarized in the following sections. It is important to first point out that the various lobes of the rat prostate differ in their propensity to develop prostate carci- nomas, either spontaneously or induced by carcinogens or hor- mones (15,210,211). The rodent prostate, unlike the human or canine prostate, consists of distinct paired lobes: the ventral, dorsal, lateral, and anterior lobes; the dorsal and lateral lobes are 53 frequently referred to as the dorsolateral prostate, and the ante— rior lobe is more commonly termed the coagulating gland. In the human and canine prostate, these lobes have merged into one gland, in which different zones have been defined (213). A homologue of the rodent ventral lobe is not present in the human gland (214). Hormonal Induction Testosterone Long-term administration of testosterone induces a low to moderate (5%—56%) incidence of prostate cancer in several rat strains (210,215—220) but not in all strains (221). The induced tumors were adenocarcinomas in all studies but one, in which some squamous cell carcinomas were also observed (218). These carcinomas appeared to develop from the dorsolateral prostate and/or coagulating gland but not from the ventral pros- tate lobe (210,215—219). The prostate carcinoma incidence in most of these studies was low (5%—20%) (215,218—220). Only the studies reported by Pollard et a1. (216,222—225) with the use of the Lobund—Wistar strain sometimes had higher carcinoma incidences, but the incidences varied considerably (0%—60%). In the only other studies with the Lobund—Wistar strain, a low incidence of prostate cancer was found but a high incidence of seminal vesicle tumors was found (219,226). The actual dose of testosterone considerably fluctuated over time in many of these studies from five to 10 times control values down to control values (221,226), but, even when the level of circulating testos- terone was kept steadily elevated by twofold to threefold, pros- tate carcinomas were induced (220). These data indicate that testosterone acts as a complete carcinogen for the rat prostate. Estrogens and Testosterone Noble (215) was the first to demonstrate that testosterone is carcinogenic for the rat prostate. He also established that se- quential treatment with testosterone and estrogens was even more effective than testosterone per se in the Noble (or NBL) rat strain that he developed. Long—term treatment of NBL rats with a combination of testosterone and estradiol leads to a 100% incidence of adenocarcinomas, which develop from the periure- thral ducts of the dorsolateral and anterior prostate (227—229). The development of these tumors is preceded by the appearance of epithelial dysplasia in these ducts and in the acini of the dorsolateral prostate in 100% of treated animals {228—230). Car- cinomas developing from the acinar dysplasia in the periphery of the prostate gland have not been observed, but the absence of malignant progression of these lesions, which are morphologi- cally similar to human prostatic intraepithelial neoplasia, has not been established with certainty ( 228,229). When diethylstilbes- trol (DES) was combined with testosterone, the treatment re- sulted in widespread dysplasia in the ventral prostate, but less or no dysplasia in the dorsolateral prostate (230). Long—term treat- ment of NBL rats with DES and testosterone induced a low carcinoma incidence in the dorsolateral prostate and some early— stage carcinomas (carcinoma in situ) in the ventral lobe (229). When the combined testosterone and estradiol treatment was given to Sprague—Dawley rats, dysplasia developed in the same high frequency as in NBL rats, but the incidence of carcinomas was considerably lower (228,229). Thus, a very high incidence of prostate cancer results from the addition of estrogen to tes- 54 tosterone treatment, which by itself produces prostate cancer in 35%—40% of treated NBL rats. Perinatal Estrogen Exposure Carcinogenic effects of perinatal exposure to DES on the accessory sex glands in male experimental animals have been described in mice, rats, and hamsters (15,231—233). McLachlan and colleagues (231,234) found that 25% of the male offspring of CD-1 mice that had been treated with DES on days 9—16 of gestation had nodular enlargements of the coagulating gland, ampullary glands, and colliculus seminalis at an age of 9—10 months. In one animal, a lesion was found in the area of the coagulating gland and colliculus seminalis that resembled early neoplasia (234). Of eight prenatally DES-exposed male mice that survived for 20—26 months, one had an adenocarcinoma of the coagulating gland, three had hyperplasia of the coagulating gland, two had hyperplasia of the ventral prostate, one had a carcinoma of the seminal vesicle, and two had squamous meta- plasia of the seminal vesicle (231,232). No such lesions occurred in control animals. Prenatal DES exposure of mice also induces testicular tumors (particularly of the rete testis) and non- neoplastic lesions in the testes and epididymis (235). Treatment of Han : NMRI mice with DES or estradiol on the first 3 days of life resulted in a 90%—100% incidence of epithe- lial dysplasia of the periurethral glands and of the periurethral proximal parts of the dorsolateral prostate, coagulating glands, and seminal vesicles after 12—18 months (236,237). Subsequent treatment with DHT and estradiol from 9—12 months of age increased the severity of the dysplasia when the prostates were examined at 12 months, suggesting permanent estrogen hyper— sensitivity of these tissues. Arai et al. (232) treated Wistar rats with DES for the first 30 days of life. One group was neonatally castrated and the second group remained intact. Two of 11 cas- trated, DES-exposed rats developed squamous cell carcinomas in the area of the dorsolateral prostate, coagulating gland, and ejaculatory ducts, and all these animals had papillary hyperplasia and squamous metaplasia of the coagulating gland and ejacula— tory duct. Squamous metaplasia was also found in some of eight noncastrated DES-exposed rats, but no hyperplasia or neoplasia developed. Vorherr et al. (238) obtained similar results in rats exposed prenatally and/0r neonatally to DES. The results of these studies demonstrate that prenatal and neonatal estrogen exposure of rodents can be carcinogenic for the prostate. Data also suggest that these treatments may imprint permanent alterations in the hormonal sensitivity of the prostate that may play a role in the carcinogenic effect of perinatal es- trogen exposure. Induction by Chemical Carcinogens and Hormones Exposure to Chemical Carcinogens Combined With Hormonal Stimulation of Cell Proliferation Very few reports are available of induction of prostate tumors by chemical carcinogens administered systemically or via the oral or inhalation routes. Only two organic chemical carcinogens induce prostate adenocarcinomas on systemic administration, without any additional concomitant or subsequent treatment, N- nitrosobis(oxopropyl)amine (BOP) and 3,2’-dimethy1-4- aminobiphenyl (DMAB) (239,240). Direct application of chemi— cal carcinogens to prostate tissue in experimental animals produces sarcomas or squamous cell carcinomas (7,241). Journal of the National Cancer Institute Monographs No. 27, 2000 Short-term hormonal stimulation of cell proliferation in the prostate at the time of carcinogen administration has been demonstrated to increase the sensitivity of the target cells for tumor induction. Dorsolateral prostate adenocarcinomas have been produced at 5%—25% incidence when prostatic cell prolif- eration was stimulated in combination with treatment with in- direct-acting carcinogens (such as DMAB and 7,12-dimethyl- benz[a]anthracene) and direct-acting chemical carcinogens (such as N—methyl-N-nitrosourea [MNU]); these carcinogens do not induce these tumors when administered alone (220,242— 245). However, in some studies, only a very small or no enhanc- ing effect was found of stimulation of prostatic cell prolifera— tion on prostate carcinoma induction by carcinogens (218, 221,246,247). Nevertheless, stimulation of cell proliferation ap- pears to be co-carcinogenic for prostate cancer induction by many chemical carcinogens. Testosterone as Tumor Promoter of Prostate Carcinogenesis Long-term administration of testosterone to rats markedly enhances prostatic carcinogenesis following initial treatment with chemical carcinogens that target the prostate because of tissue-specific metabolism (DMAB and BOP) and/or concurrent hormonal stimulation of prostatic cell proliferation (70,71,211, 217—221,223,225,248). This enhancement may not occur if cer- tain requirements are not adequately met (210,211,218). For example, after a single injection of BOP or MNU given to F344 rats without concurrent stimulation of prostatic cell proliferation, long—term testosterone treatment did not enhance prostatic car— cinogenesis (221). High incidences (66%—83%) of adenocarci- nomas of the dorsolateral and/or anterior prostate, but not the ventral prostate, were induced by chronic treatment with testos- terone, following a single administration of MNU or BOP given during stimulation of prostatic cell proliferation in Wistar rats, or during and after 10 repeated biweekly injections of DMAB in F344 rats (70, 71,2IO,211,218,220,221,248). This effect is some- what strain dependent, because when the same treatments were given to Lobund—Wistar rats, rather variable prostate carcinoma incidences of between 50% and 97% were reported by Pollard and colleagues (217,223,225) and only a 21%—24% incidence was found by Hoover et al. (219) and Tamano et al. (226). The enhancing effect of testosterone on prostate carcinogen- esis is remarkably confined to the dorsolateral and anterior pros- tate, and no tumors occur in the ventral prostate. In fact, long- term testosterone treatment produces a shift of the site of DMAB- and BOP-induced carcinoma occurrence from exclu- sively the ventral lobe to predominantly the dorsolateral and anterior lobes (218,221,248). The dose—response relationship between testosterone dose and prostate carcinoma yield is very steep. A slight (less than 15-fold) elevation of circulating tes- tosterone levels is sufficient for a near-maximal enhancement of the tumor response, and a twofold to threefold elevation is suf- ficient for a maximal response. These concentrations are within the normal range of circulating testosterone levels in the rat (220). Thus, testosterone is a powerful tumor promoter for the rat prostate. Efi‘ects of Testosterone on Prostate Cancer Induction by Cadmium and Ionizing Radiation Cadmium can be carcinogenic for the rat ventral prostate as demonstrated by Waalkes et a1. (249,250). The selective sensi- Journal of the National Cancer Institute Monographs No. 27, 2000 tivity of the ventral prostate lobe for the carcinogenic action of cadmium is most likely due to its lack of cadmium-binding proteins (251). A single injection of cadmium chloride produced in situ (noninvasive) carcinomas in the ventral lobe provided that cadmium—induced testicular toxicity was avoided, either by keeping the cadmium dose below 5 mg/kg, by intramuscular rather than subcutaneous administration of the cadmium, or by antagonizing the testicular toxicity of cadmium by simultaneous administration of sufficient amounts of zinc. These observations indicate that cadmium induces proliferative lesions in the rat ventral prostate only when testicular function, conceivably tes- tosterone production, is intact. In addition, these data suggest that androgens also act as tumor promoters in this system, but this hypothesis has not been tested. Other mechanisms may also be involved, because, for example, testosterone considerably in- creases cadmium disposition and retention in the rat ventral prostate (252). Local x-ray exposure of the pelvis has been shown to induce prostate carcinomas in ICR/JCL mice (253) and in Sprague— Dawley rats (254). Prostate carcinomas (33% incidence) devel- oped only in rats that were castrated and received androgen replacement prior to irradiation, but intact and only castrated rats did not develop prostate cancer following irradiation. These ob- servations suggest that testosterone treatment was required for tumor development, perhaps as tumor promoter ( 254 ). Prostate Cancer and Steroid Hormones in Laboratory Animals: Summary and Conclusions Stimulation of prostatic epithelial cell proliferation by andro- gens during exposure to chemical carcinogens increases the sus- ceptibility of the rat prostate to cancer induction in a co— carcinogenic fashion. Testosterone appears to be a weak complete carcinogen, but it is a very strong tumor promoter for the rat prostate at near-physiologic plasma concentrations (220). The very powerful tumor-promoting activity of androgens per— haps explains their weak complete carcinogenic activity on the rat prostate. A slight elevation of circulating testosterone can lead to a marked increase in prostate cancer in rat models. This observation is highly relevant in view of the aforementioned possible weak association between human prostate cancer risk and slightly elevated circulating androgen levels found in some epidemiologic studies (191). Thus, the experimental data pro- vide strong support for the concept that minimal increases in circulating androgens may have substantial enhancing effects on prostate cancer risk. The addition of estradiol to chronic treat- ment with testosterone strongly enhances the carcinogenic ac- tivity of the androgen for the rat dorsolateral prostate. The sen- sitivity for the carcinogenic effects of this hormone combination appears to be confined to the periurethral, proximal ducts of the dorsolateral and anterior prostate. The estradiol plus testosterone treatment also induces acinar lesions that are similar to human prostatic intraepithelial neoplasia. These observations strongly suggest a critical role for estrogens in prostate carcinogenesis. Perinatal estrogen exposure is also carcinogenic for the male rodent accessory sex glands. The periurethral, proximal ducts of the dorsolateral and anterior prostate and seminal vesicle and the intraprostatic urethral epithelium appear to be the most sensitive rodent male genital tract tissue to the carcinogenic effects of perinatal estrogen exposure. Of interest in this regard is the report by Driscoll and Taylor (255) of hypertrophy and squa- mous metaplasia of the prostatic utricle and prostatic ducts in 55 55%—71% of 31 infants that had been exposed to DES in utero and had died perinatally from unrelated causes. Such squamous metaplastic changes have also been reported to occur in human fetal prostatic tissue transplanted into nude mice that were sub- sequently treated with DES (256). These human observations suggest that the DES findings in rodents may have human rel- evance. MECHANISMS OF HORMONAL PROSTATE CARCINOGENESIS As stipulated before, there are compelling reasons to assume that androgens play a critical role in prostate carcinogenesis, and there is experimental evidence to suggest that estrogens are in- volved as well (56). Because of the hormonal nature of these steroids, receptor mediation has been proposed as the major mechanism by which androgens and estrogens act in the causa— tion of prostate cancer (257). For androgens, mechanisms other than those mediated by androgen receptors seem unlikely, ex- cept for the generation of estrogens via aromatization. For es- trogens, however, nonreceptor—mediated genotoxic effects are conceivable, in addition to receptor-mediated processes (56). These various potential mechanisms are discussed in the follow- ing sections. Stromal—Epithelial Interactions First, it is important to emphasize that considerable evidence indicates interactions between epithelial and stromal cells in the normal prostate. Such interactions are undoubtedly critical and may be essential in prostate carcinogenesis as well, because prostatic mesenchyme is known to be a mediator of androgen action in the developing and adult rodent prostate and possibly the human prostate (258,259). No studies, however, have di- rectly addressed the role of stromal—epithelial interaction in hu- man or rodent prostate carcinogenesis. Krieg et al. (202) studied steroid hormone concentrations in stromal and epithelial com— partments of normal human prostates from subjects varying from 20 to 80 years of age. DHT concentrations in the epithe— lium decreased considerably with aging, but they remained stable in stromal cells, whereas testosterone concentrations ap- peared unaffected by age in either compartment. These data suggest that the activity of SOL-reductase in the epithelium de- creases with aging but remains intact in the stroma. However, concentrations of estradiol and estrone in the stroma, but not the epithelium, increased markedly with aging. These observations suggest that the prostatic stroma is an important site for both androgen and estrogen action and metabolism, such as aro- matase activity, which seems to increase with aging because estrogens accumulate with aging and androgen levels remain stable. This is unlike the concentrations of estrogens and andro- gens in the circulation or in epithelial cells, where both decrease. Thus, it is conceivable that with aging and increasing risk for prostate cancer the prostatic stroma continues to be an important androgen signal mediator to the epithelium and is an increas- ingly important local producer of estrogens. Role of Androgens in Prostate Carcinogenesis The results of the earlier summarized rodent experiments clearly indicate carcinogenic and strong tumor-promoting prop- erties of androgens, and the results of a limited number of epi— demiologic studies provide some support for the notion that androgens may have such effects in humans. However, the 56 mechanisms of the carcinogenic and tumor-promoting effects of androgens on the rodent prostate are not known with certainty. The very steep relationship between testosterone dose and pros- tate carcinoma response in rat models is suggestive of involve- ment of an androgen receptor-mediated mechanism (220). Other mechanisms may nevertheless be involved as well. For example, Ripple et al. (260) observed increases in indicators of oxidative stress in androgen-sensitive LNCaP human prostate cancer cells exposed to DHT, although it is possible that these effects were androgen receptor mediated. Stimulation of Cell Proliferation and Carcinogenic and Tumor-Promoting Efiects ofAndrogens The postulated role of androgens in human prostate carcino- genesis has been ascribed to their androgen receptor—mediated stimulating effects on prostatic cell proliferation (136,257). No direct evidence, however, is available that elevation of circulat- ing testosterone leads to increased cell proliferation in the hu- man prostate. It has been well established that androgen admin- istration to castrated rodents causes elevation of prostatic cell proliferation similar to that observed in cell cycle synchroniza- tion experiments with cells in vitro. However, the increase in prostatic cell proliferation caused by testosterone or DHT ad- ministration in castrated rodents is only transient, and after a few days cell turnover returns to its normal very low levels (261). Thus, continued androgen treatment of rodents does not result in permanently elevated cell proliferation rates in the male acces- sory sex glands, but rather appears to support differentiation. Furthermore, DHT may even suppress prostatic cell proliferation in intact rats (228). Thus, a mere continuous stimulation of cell proliferation is unlikely to be the major mechanism of the en- hancing effects of testosterone on prostatic cancer induction in rodents and possibly humans. Conceivably other, nonhormonal factors affect prostatic cell proliferation. For example, over the lifetime of a man, the pros- tate undergoes repeated inflammatory insults (prostatitis) with reactive cell proliferation and generation of reactive oxygen spe- cies as possible consequences (262,263), and sexual activity conceivably also affects prostatic cell turnover. Support for a cell proliferation hypothesis is provided by rodent experiments that indicate that increased prostatic cell proliferation at the time of exposure to carcinogens can enhance the sensitivity of the tissue to the carcinogenic effects of these agents (220, 242—245). Stimulation of cell proliferation during carcinogen exposure in- creases the likelihood that promutagenic DNA damage, such as carcinogen—DNA adducts, will be fixed as permanent mutations. In humans, increased cell proliferation may thus enhance the carcinogenic effects of low-level exposure to environmental and endogenous carcinogens. The rate of cell proliferation at the time of carcinogen expo— sure may be only one of several androgen-related factors that determine sensitivity of the prostate to cancer induction by carcinogens through androgen-receptor mediated mechanisms. For example, Sukumar et al. (264) have hypothesized that pros- tate cells that harbor critical genetic alterations, such as activat— ing point mutations in oncogenes or inactivating alterations in tumor suppressor genes, may be selectively sensitive to in- duction of the cell proliferation, rather than cellular differentia- tion, by androgens. However, this hypothesis has not been criti- cally tested. These cells could thus have a selective growth advantage over normal cells, which do not respond to chronic Journal of the National Cancer Institute Monographs No. 27, 2000 testosterone treatment with sustained proliferation (221). It is also possible that androgens, in addition to other factors, influ- ence the effectiveness of indirect-acting carcinogens that are metabolically activated and otherwise metabolized in the pros- tate itself. Role of Androgen Metabolism and Androgen Receptor Sensitivity Pertinent to any hypothesis implicating androgens in prostate carcinogenesis are considerations related to androgen receptor function and androgen metabolism, from steroid biosynthesis, to conversion of testosterone to DHT and to DHT catabolism. Ross et al. (136) have developed the idea that genetically determined differences in the activities of steroid biosynthetic enzymes, 50(- reductase, and enzymes that metabolize DHT, as well as in androgen receptor activity are major determinants of risk both at the population and at individual levels [see also (18)]. Func— tional polymorphisms in the genes that encode for these en- zymes and the androgen receptor have been hypothesized to underlie this notion (18,136). The evidence for these polymorphisms being important in human prostate carcinogenesis has been summarized in detail and evaluated together with the results of endocrinologic studies in earlier sections. The overall conclusions were that to date there is inconsistent and conflicting evidence that functional polymorphisms in the SOL-reductase gene and differences in 50L- reductase activity are important determinants of prostate cancer risk. However, there is stronger evidence to suggest that risk is associated with a functional polymorphism in the androgen re- ceptor gene, short lengths of CAG repeats in the transactivation domain of the protein that are linked with increased transacti- vation activity in vitro (137—139,163,164,182—185). However, this association is weak at best. Several other polymorphisms identified in genes encoding for the androgen receptor and other androgen metabolizing enzymes studied have been unevenly distributed among populations that differ in prostate cancer risk (HZ—156,160,162) or to be associated with risk in case—control studies (153—155,163,170,182). These studies concerned poly- morphisms with unknown functional significance in genes en- coding the type II 3B-hydroxysteroid dehydrogenase that catabolizes DHT, the cytochrome P450C170t enzyme that has l7a-hydroxylase and 17,20-lyase activity involved in androgen biosynthesis, and CCG or GGN repeats in the promoter region of the androgen receptor gene. The study of these types of poly- morphisms is a rapidly evolving field of investigation and will no doubt lead to significant and relevant new findings in the near future (136). The observation of slightly, but mostly not significantly, higher circulating testosterone levels in high-risk African— American men compared with lower-risk European men sug- gests that their rates of androgen biosynthesis may be higher. Although lower as well as higher testosterone concentrations have been found in lower-risk Asian or African men compared with European- or African—Americans, testosterone levels were lower in Asians living in Asia than in American populations regardless of their ethnicity in the only two studies that included Asian populations. In addition, directly measured testosterone production rates were lower in Chinese in China than in both Chinese-Americans and European-Americans (148). These observations are consistent with the hypothesis that environmen— tal factors, such as diet, determine prostate cancer risk at the Journal of the National Cancer Institute Monographs No. 27, 2000 population level by influencing androgen production such that they are lower in low-risk than in high—risk circumstances (208). Assessing the role of androgen in prostate carcinogenesis is complicated by the fact that the prostatic stroma is an important site for androgen action and metabolism in the prostate in addi- tion to the epithelium. For example, epithelial DHT concentra- tions decline dramatically with aging, but they remain stable in the stroma even though the source for intraprostatic DHT, cir— culating testosterone, also diminishes with aging. This observa- tion suggests that stromal Sci—reductase activity remains stable, whereas epithelial activity of this enzyme declines with aging. These observations illustrate the difficulties in interpreting the results of studies of circulating androgenic (or other) hormone levels of genomic polymorphisms in relevant genes, because they do not necessarily provide relevant information about what is going on at the level of the prostatic epithelial cell and its important immediate environment, the prostatic stroma. Role of Estrogen in Prostate Carcinogenesis The results of the earlier summarized epidemiologic studies provide limited evidence for an association between prostate cancer risk and circulating levels of estrogens, which appear to be higher in men of African descent younger than 50 years of age than in European-American men. This observation suggests that estrogens may be involved in prostate carcinogen— esis, because men of African descent living in an American environment have the highest risk for prostate cancer of any population. Most of the direct evidence in support for a role of estrogens in prostate carcinogenesis comes from studies with treatment of NBL rats with testosterone and estradiol (229,265). It appears that the estrogen-related mechanisms underlying this effect are a mixture of estrogen receptor-mediated and nonreceptor pro- cesses, which are discussed in the following paragraphs. In ad— dition, there is evidence to suggest that the mechanisms involved in hormonal induction of rat prostate cancer, originating from periurethral prostatic ducts, are different from those involved in the induction by testosterone and estradiol of dysplastic lesions developing in the dorsolateral prostate acini. As alluded to earlier, there is evidence for the presence of the CYP19 enzyme aromatase in the human prostate, which could provide a local source of estrogens from conversion of testos- terone (195—199), but there are contradictory reports (200,201). The local production of estrogens in the human prostate is pos- sibly a stromal process, and stromal aromatase activity may increase with aging (202). Data on the presence of aromatase in the rodent prostate are also somewhat contradictory, because aromatase activity has been reported in the rat ventral prostate and a transplantable rat prostate carcinoma (266), but it was not detectable in mouse prostate (267). These discrepancies, which may be due to interspecies or methodologic differences, point to the need for further research. Estrogen Receptor—Mediated Mechanisms Estrogen receptors are found in the prostate, and Lau et al. (268) demonstrated that both the estrogen receptor-0L and -[3 are present in the rat prostate. Thus, direct receptor-mediated effects of estrogens on the prostate are plausible. However, rodent studies that used antiestrogen treatments (tamoxifen and ICI- 182,780) have yielded contradictory results about the involve— 57 ment of estrogen receptor mechanisms in prostate carcinogene- sis. The prostate tumor-promoting effects of testosterone may involve estrogen generated by aromatization. However, simul- taneous administration of testosterone and tamoxifen failed to alter the prostate carcinogenesis-enhancing effect of the andro- gen in an experiment in rats injected with DMAB prior to the hormone treatment (269). However, ICI-l82,780 blocked the induction of epithelial dysplasia in the prostatic periphery in NBL rats treated with testosterone and estradiol for 16 weeks (270); the effects of this antiestrogen on induction of periure- thral prostate carcinomas are not known. Leav et al. (228) and Ofner et al. (230) showed that dorso- lateral prostatic tissue with epithelial dysplasia from NBL rats treated with testosterone and estradiol for 16 weeks accumulates estradiol and 5a—androstane-3B,17B—diol, a weak estrogenic agonist; this accumulation of estrogenic species does not occur in the ventral lobe, which also does not develop dysplasia. In rats treated with testosterone and DES for 16 weeks, dysplasia de- veloped more distinctly in the ventral than in the dorsolateral prostate, as indicated earlier. This development coincided with a preferential accumulation of estradiol and SOL-androstane- 38,17B-diol in the ventral prostate (230). These observations suggest that increased levels of estradiol and the weakly estro- genic androgen metabolite in prostatic target tissue may be caus- ally related with the development of hormone—induced dysplasia and perhaps carcinomas in the NBL rat model (230). In tissue with epithelial dysplasia from the dorsolateral pros- tatic periphery that was derived from rats treated with testoster- one and estradiol for 16 weeks, elevated levels of nuclear, but not cytosolic, type II (intermediate-affinity) estrogen-binding sites, but not type I (high-affinity) binding sites, have been found (228,271). The type II estrogen receptor is a cell proliferation marker believed to be a key factor in normal and aberrant growth regulation in female estrogen target tissues. These data indicate that protracted stimulation of cell proliferation may be involved in the formation of hormone-induced rat prostate dysplasia (228,271). Indeed, mitotic indices in testosterone plus estrogen— treated NBL rat dorsolateral prostate were increased over control values; this increased mitotic activity was largely confined to the dysplastic lesions (228,271). One well-established effect of estrogen treatment in rodents is stimulation of prolactin secretion. This finding raises the possi— bility that some or even all estrogen effects on the rodent pros- tate may be mediated through elevation of prolactin secretion, and there is some experimental support for this notion. Trans- plantation of a prolactin—producing pituitary tumor into rats treated with DMAB enhanced the formation of atypical hyper— plasia, a preneoplastic lesion, but not carcinomas, in the ventral prostate, and treatment with bromocriptine, a prolactin secre- tion-suppressing agent, counteracted this effect (272). Bromo- criptine also lowered the formation of ventral prostatic atypical hyperplasia and carcinomas in rats treated with only DMAB. Development of epithelial dysplasia in the dorsolateral prostatic periphery of NBL rats treated with testosterone and estradiol for 16 weeks was also blocked by bromocriptine, but effects on periurethral carcinoma development were not studied (273). Thus, there is evidence to suggest that prolactin may modulate the induction of preneoplastic lesions in the rat prostate, but the relevance of these findings for prostate cancer development are not clear. In conclusion, several lines of evidence are available to sug- 58 gest that estrogen receptor-mediated mechanisms contribute to the induction of prostate cancer by hormonal treatments, but conclusive data in this regard are largely lacking. Nonreceptor Mechanisms Estrogens have been shown to be capable of producing DNA damage in target tissues susceptible to estrogen-induced carci- nogenesis, independent of their interaction with the estrogen receptor, as discussed in detail elsewhere in this monograph. In the kidney of male hamsters treated with DES, Liehr and col- leagues (275) have found a direct DES—DNA adduct and indi- rect estrogen-generated DNA adducts perhaps of endogenous origin and of undetermined structure detectable by 32P- postlabeling (276). Both observations are thought to be related to the formation of catechol estrogens that undergo redox cycling during which reactive intermediates and reactive oxygen species are generated and lipid peroxidation can be initiated (274). Simi- lar observations have been made in the prostate of NBL rats treated for 16 weeks with testosterone plus estradiol. This treat— ment enhanced the formation of a chromatographically unique endogenous adduct selectively in the periurethral region of the rat dorsolateral prostate, which is the site of the carcinogenic effect of this treatment [(277) Bosland et al., unpublished data; see also Chapter 4]. Ho and Roy (278) reported increased single- strand DNA breaks and accumulation of fluorescent lipid per- oxidation products in the dorsolateral prostate of NBL rats after this treatment, but they did not separately analyze the periure- thral and peripheral areas of the prostate. In addition, substan- tially elevated levels of 8-hydroxydeoxyguanosine and, to a lesser extent, lipid hydroperoxides have been found at the peri- urethral tissue but not in the peripheral area of these glands (Bosland et al., unpublished data; see also Chapter 4). Lower, but still elevated, levels of the endogenous DNA adduct detect- able by 32P-postlabeling, 8-hydroxydeoxyguanosine, and lipid hydroperoxides were also found in the periurethral prostate of rats treated with only testosterone, perhaps due to formation of estrogens by aromatization (Bosland et al., unpublished data). The enhancement of endogenous DNA adduct formation, oxi- dative DNA damage, and lipid peroxidation selectively at the site of tumor formation and preceding it strongly suggests that these effects are causally involved in the carcinogenic effect of the hormone treatment. It is likely, but as yet unproven, that the exogenously administered estrogens or formation of estrogen Via aromatization of testosterone and a genotoxic mechanism are critical to the carcinogenic effect of this hormone combination for the prostate, rather than other mechanisms, including recep- tor mediation. This hypothesis implies that catechol estrogen formation oc- curs at the relevant site within the prostate, as indicated earlier and discussed in detail elsewhere in this monograph. Lane et al. (212) demonstrated that microsomes isolated from testosterone plus estradiol-treated NBL rat dorsolateral prostate do not ap- pear to be able to generate the catechol estrogens 2-hydroxy and 4-hydroxy-estradiol and —estrone. However, because periurethral prostate tissue was not incorporated in the analysis and the rel- evance of such microsomal assays for the in viva situation is unclear, these data do not refute the possibility of catechol es- trogen formation in the periurethral prostate. Furthermore, it is conceivable that the mechanisms of induction of dysplasia in the dorsolateral glandular prostate (estrogen receptor-mediated, pos- sibly not involving estrogen-generated genotoxic processes) are Journal of the National Cancer Institute Monographs No. 27, 2000 different from those involved in generation of the periurethral prostatic carcinomas (estrogen-generated genotoxicity, but pos- sibly no estrogen receptor mediation). This idea leads to the hypothesis that l) testosterone acts as a tumor promoter and estrogens act as genotoxic “tumor initiators” in the testosterone plus estradiol-treated NBL rat model of (periurethral) prostate carcinogenesis, and 2) the androgen also acts as enhancer of induction of dysplasia (periphery) in this model, which requires conjunct action of estrogen via estrogen receptors. However, these hypotheses remain to be critically tested. The human relevance of these findings in the testosterone plus estradiol-treated NBL rat model remains unclear at present. However, oxidative DNA damage and lipid peroxidation reflec- tive of reactive oxygen damage have been observed in the hu- man prostate (2 79), and signs of increased oxidative stress have been found in patients with prostate cancer as compared with control subjects (280). Whether these observations are estrogen- exposure related or associated with other risk factors, such as a high fat diet (263), is not known, but they suggest that endog- enous oxidative stress may be important in human prostate car- cinogenesis and are consistent with involvement of estrogen- generated oxidative DNA damage. Perinatal Estrogen Exposure: Imprinting As summarized earlier, perinatal estrogen exposure of mice resulted in epithelial dysplasia of the periurethral proximal parts of the dorsolateral and anterior prostate and of the seminal vesicles (236,237) as well as carcinomas in these areas (234). In addition, mice that were neonatally estrogenized hyperre- sponded to secondary estrogen treatment (estradiol) with the development of considerable squamous metaplasia in these same tissues, but control subjects responded with little or no squamous change (236). These very same tissue areas possess estrogen receptors, which indicate their estrogen sensitivity (236). How- ever, the activity of estradiol hydroxysteroid oxidoreductase, a marker of estrogen sensitivity, and incorporation of tritiated thy— midine in epithelial compartments of these tissues were not changed in response to secondary treatments with estradiol in neonatally estrogenized mice (236,237). In response to second- ary androgen treatment (DHT), tritiated thymidine incorporation was markedly increased selectively in stromal cells of the ante— rior and ventral prostate, indicating a lasting effect of neonatal estrogen exposure on the androgen responsiveness of the stromal component of the mouse prostate (236). These observations sug- gest that perinatal estrogen exposure of mice imprints lasting alterations in estrogen and androgen responsiveness of the male accessory sex glands. The exact mechanism of these complex imprinting effects is not clear. Perinatal estrogen treatment may act indirectly on the male accessory sex glands by imprinting permanent alterations in the secretion of pituitary hormones and testicular androgen, or directly by, e.g., imprinting altered expression of androgen, es— trogen, and prolactin receptors or changes in steroid metabolism in the accessory sex gland, which all may result in modified development of these glands (281—284). For example, in neona— tally estrogenized mice, luteinizing hormone and follicle— stimulating hormone plasma levels were found to be elevated (282), whereas circulating testosterone levels were decreased (281,284) or unaltered (282). However, no abnormalities in cir- culating estrogen and androgen levels were found in boys that had been exposed to DES in utero (285). Prostatic DHT forma- Journal of the National Cancer Institute Monographs No. 27, 2000 tion by Sa-reductase was found to be impaired in adult mice neonatally treated with DES (267). Nuclear androgen receptor levels in these mice were decreased in the dorsal and ventral prostate but not affected in the lateral lobe, and the number of androgen receptor-positive stromal cells was increased in all three lobes (284). The significance of these findings for the carcinogenic effects of perinatal estrogen exposure to mice is not clear. Although the exact mechanisms of the carcinogenic ef- fects of perinatal estrogen exposure for the prostate remain un- clear, there appear to be lasting direct and indirect effects of this treatment on the mouse prostate. The human relevance of the findings in mice remains unclear at present, but in utero estrogen exposure is likely to occur in humans (I42). OVERALL CONCLUSIONS With the exception of “exposure” to a western lifestyle, in- cluding a high-fat diet, an African-American “environment,” and, perhaps, venereal disease and unknown factors related to farming and employment in armed services and nuclear industry, there are no known exogenous exposures that are associated with prostate cancer risk, and none of these circumstances con— stitute exposures to specific chemicals or factors. Familial ag— gregation of prostate cancer risk is consistently observed and confers a considerable increase in risk but explains less than 10% of all cases. Putative susceptibility loci have been identi- fied, but there are no indications that these loci are related to hormonal factors. This lack of known specific risk factors is remarkable in view of the high frequency of this malignancy in western countries. It may indicate that there are many exogenous risk factors for prostate cancer that are too ubiquitous and over- lapping to be detectable by epidemiologists. However, it is pos- sible that there are strong endogenous determinants of prostate cancer risk that are “overpowering” most exogenous risk factors in epidemiologic analyses. Androgenic hormones and androgen receptor mechanisms are prime candidates to be such important endogenous factors, but the epidemiologic evidence in favor of this view is weak. El— evation of bioavailable and bioactive androgens in the circula- tion and in the target tissue as an important risk factor is bio- logically very plausible. The results of several animal model studies strongly support this contention. Some experiments in— dicate that substantial enhancement of prostate carcinogenesis can be produced by very small elevations of circulating testos- terone, which, if also valid for humans, may explain why the epidemiologic associations between circulating androgen levels and prostate cancer are weak at best. Evidence is also available indicating that increased transactivation activity of the androgen receptor may be associated with increased prostate cancer risk, both at the population and individual levels. However, more research is needed to confirm and further define these associa- tions in humans and to further unravel the biologic mechanisms underlying the increased risk that may be associated with el- evated circulating androgen levels and increased androgen re- ceptor sensitivity. African-American men have a twofold higher risk than Eu- ropean-American men do. The unknown environmental and pos- sibly genetic factors that determine the high prostate cancer risk in African—American men may act through modifying their hor- monal status. Indeed, circulating levels of androgens and, in men younger than 50 years, estrogens appear to be higher in men of African descent than in European-American men. Such endo- 59 crine mechanisms perhaps act as early as in utero, because cir- culating levels of androgens and estrogens have been shown to be slightly higher in young men and in pregnant African- American women than in European—American women. Hormonal stimulation of prostatic epithelial cell proliferation enhances the susceptibility of the rat prostate to chemical car- cinogens. Testosterone at near—physiologic plasma concentra- tions is a weak complete carcinogen and a strong tumor pro- moter for the rat prostate. The very strong tumor-promoting activity of androgens possibly explains their weak complete car— cinogenic activity. The mechanism of the tumor-inducing and -promoting activities of androgens for the rat prostate is un— known. It is unlikely that chronic stimulation of prostatic cell proliferation rates by androgens is involved. However, it is pos— sible that prostatic epithelial cells that carry critical genetic al— terations have a selective growth advantage over normal cells and do not respond to androgens by differentiation, as normal cells would, but by proliferation. Chronic exposure to testosterone plus estradiol is strongly carcinogenic for the dorsolateral prostate of some rat strains, whereas testosterone alone is only weakly carcinogenic. The mechanism of this carcinogenic effect in the rat prostate is in- completely understood, but it appears to involve estrogen- generated oxidative stress and genotoxicity and also requires androgen- and estrogen receptor-mediated processes, such as changes in sex steroid metabolism and receptor status. There is evidence for the presence of the enzyme aromatase in the human and rat prostate, providing a local source of estrogens, which in humans seem to increase in activity with aging. Perinatal estro— gen exposure is carcinogenic for the rodent male accessory sex glands. Hyperplastic and squamous metaplastic changes have been reported in human genital tract tissue following prenatal DES exposure, indicating that prenatal exposure to DES may also target the human prostate. The mechanisms of these prena— tal estrogen effects are not clear, but they may involve perma- nently imprinted changes in hormone production and tissue hor— mone sensitivity. From these observations, the following multifactorial general hypothesis of prostate carcinogenesis emerges: Androgens act as strong tumor promoters via androgen receptor-mediated mecha- nisms to enhance the carcinogenic activity of strong endogenous genotoxic carcinogens, including reactive estrogen metabolites and estrogen- and prostatitis-generated reactive oxygen species, and possibly unknown weak environmental carcinogens. All these processes are modulated by a variety of environmental factors, such as diet, and by genetic determinants, such as he- reditary susceptibility genes and polymorphic genes that encode receptors and enzymes involved in the metabolism and action of steroid hormones. FUTURE RESEARCH NEEDS This overview clearly indicates that, although steroid hor— monal factors are strongly implicated in prostate carcinogenesis, we know very little about their involvement. Considerable re- search is needed to further our understanding of this relationship. Some promising areas for future research are summarized be- low. One aspect to be mentioned up front is that the African- American population offers unparalleled but vastly underex- ploited opportunities for such research, which may also lead to new insights in the possible prevention of prostate cancer in this underrepresented but disproportionally affected group. 60 1) To resolve the uncertainties about the importance of circu- lating hormone levels, additional, large, nested case—control studies are needed by using cohorts of men belonging to diverse racial/ethnic and other groups that differ substantially in risk of prostate cancer with serial measurements of circu- lating steroid and other hormones (both over time to assess consistency and trends and within 24-hour periods to assess circadian rhythm variations). 2) To address the functional significance of polymorphisms in genes encoding for enzymes involved in steroid hormone biosynthesis and metabolism, studies are needed of correla— tions between circulating hormone levels and these polymor— phisms in relation to prostate cancer risk. 3) Even more important than the studies mentioned above, there is an urgent need to develop strategies to examine the rela- tionships of circulating hormone levels and genetic polymor- phisms in genes encoding for relevant enzymes with in- traprostatic hormone levels, activities of steroid hormone metabolizing enzymes, and androgen receptor mechanisms. 4) To determine the importance of estrogen-generated gene damage, studies in humans and animal models are needed of DNA damage and mutations in the prostate associated with exposure to estrogens (and possibly other steroid hormones) in relation to risk of prostate cancer. 5) To assess the involvement of gene—environment interactions in prostate carcinogenesis, studies are needed in humans and animal models of the effects of diet and other environmental factors on circulating and prostatic hormone levels, intrapros- tatic activities of steroid hormone metabolizing enzymes, and prostatic androgen receptor function. 6) To expand understanding of the importance of genetic poly- morphisms in prostate carcinogenesis, intensification is needed of searches for new relevant polymorphisms in genes encoding for enzymes involved in steroid hormone biosyn- thesis and metabolism, as well as factors involved in andro- gen receptor function, including determination of their func— tion. 7) Finally, to facilitate many of the research needs listed above, there is the need for establishing banks of adequate DNA, serum, and prostatic tissue samples in large, well-docu— mented, relevant cohorts of aging men. REFERENCES ( I ) American Cancer Society. Cancer Facts and Figures. Atlanta (GA): American Cancer Society; 1999. (2) Huff J, Cirvello J, Haseman J, Bucher J. Chemicals associated with site— specific neoplasia in 1394 long—term carcinogenesis experiments in labo— ratory rodents. Environ Health Perspect 1991;93:247—70. (3) Gold LS, Slone TH, Manley NB, Bernstein L. Target organs in chronic bioassays of 533 chemical carcinogens. Environ Health Perspect 1991; 93:233—46. (4) Kirkman H, Kempson RL. Tumours of the testis and accessory male sex gland. In: Turusov VS, editor. Pathology of Tumours in Laboratory Ani- mals, Vol. III, Tumours of the Hamster. Lyon (France): IARC;1982: 175—90. (5) Bosland MC. Adenocarcinoma, prostate, rat. In: Jones TC, Mohr U, Hunt RD, editors. Genital System. Berlin (Germany): Springer—Verlag 1987: 252—60. (6) Bosland MC. Adenoma, prostate, rat. In: Jones TC, Mohr U, Hunt RD, editors. Genital System. Berlin (Germany): Springer-Verlag. 1987:261—6. (7) Bosland MC. Adenocarcinoma, seminal vesicle/coagulating gland, rat. In: Jones TC, Mohr U, Hunt RD, editors. Genital System. Berlin (Germany): Springer Verlag, 1987:27275. (8) Mitsumori K, Elwell MR. Proliferative lesions in the male reproductive Journal of the National Cancer Institute Monographs No. 27, 2000 system of F344 rats and B6C3F1 mice: incidence and classification. En- viron Health Perspect 1988:7721 1—21. (9) Ito N, Shirai T. Tumors of the accessory male sex organs. In: Turusov V, Mohr U, editors. Pathology of Tumors in Laboratory Animals. Vol 1, Tumors of the Rat. Lyon (France): IARC; 1990:421—31. (IO) Nomura AM, Kolonel LN. Prostate cancer: a current perspective. Am J Epidemiol 1991;13:200—27. (11) Bosland MC. The etiopathogenesis of prostatic cancer with special ref— erence to environmental factors. Adv Cancer Res 1988: 5021—106. (12) Meikle AW, Smith JA Jr. Epidemiology of prostate cancer. Urol Clin North Am 1990;17: 709—18. (13) Muir CS, Nectoux J, Staszewski J. The epidemiology of prostatic cancer. Geographical distribution and time-trends. Acta Oncol 1991;30:133—40. (14) van der Gulden, JWJ, Kolk JJ, Verbeek AL. Prostate cancer and work environment. J Occup Med 1992;34:402—9. (I5) Bosland MC. The male reproductive system. In: Waalkes MP, Ward JM, editors. Carcinogenicity. New York (NY): Raven Press; 1994:339—402. (I6) Wingo PA, Ries LA, Giovino GA, Miller DS, Rosenberg HM, Shopland DR, et a1. Annual report to the nation on the status of cancer, 1973—1996, with a special section on lung cancer and tobacco smoking. J Natl Cancer Inst 1999;91:675—90. (17) Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1999. CA Cancer] Clin 1999;49:8—31. (18) Shibata A, Whittemore AS. Genetic predisposition to prostate cancer: possible explanations for ethnic differences in risk. Prostate 1997;32: 65—72. (19) Steinberg GD, Carter BS, Beaty TH, Childs B, Walsh PC. Family history and the risk of prostate cancer. Prostate 1990;17:337—47. (20) Carter BS, Beaty TH, Steinberg GD, Childs B, Walsh PC. Mendelian inheritance of familial prostate cancer. Proc Natl Acad Sci U S A 1992; 89:3367—71. (21) Hayes RB, Liff J, Pottern LM, Greenberg RS, Schoenberg JB, Schwartz AG, et al. Prostate cancer risk in US. blacks and whites with a family history of cancer. Int J Cancer 1995;60:361—4. (22) Whittemore AS, Kolonel LN, Wu AH, John EM, Gallagher RP, Howe GR, et a1. Prostate cancer in relation to diet, physical activity, and body size in blacks, whites, and Asians in the United States and Canada. J Natl Cancer Inst 1995;87:652—61. (23) Dong JT, Isaacs WB, Isaacs JT. Molecular advances in prostate cancer. Curr Opin Oncol 1997;92101—7. (24) Smith JR, Freije D, Carpten JD, Gronberg H, Xu J, Isaacs SD, et al. Major susceptibility locus for prostate cancer on chromosome 1 suggested by a genome-wide search. Science 1996;274:1371—4. (25) Gronberg H, Xu J, Smith JR, Carpten JD, Isaacs SD, Freije D, et a1. Early age at diagnosis in families providing evidence of linkage to the hereditary prostate cancer locus (HPCl) on chromosome 1. Cancer Res 1997;57: 4707—9. (26) Gronberg H, Isaacs SD, Smith JR, Carpten JD, Bova GS, Freije D, et a1. Characteristics of prostate cancer in families potentially linked to the hereditary prostate cancer 1 (HPCl) locus. JAMA 1997;278:1251—5. (27) Eeles RA, Durocher F, Edwards S, Teare D, Badzioch M, Hamoudi R, et al. Linkage analysis of chromosome lq markers in 136 prostate cancer families. The Cancer Research Campaign/British Prostate Group UK. Familial Prostate Cancer Study Collaborators. Am J Hum Genet 1998;62: 653—8. (28) Gibbs M, Chakrabarti L, Stanford JL, Goode EL, Kolb S, Schuster EF, et al. Analysis of chromosome lq42.2—43 in 152 families with high risk of prostate cancer. Am J Hum Genet 1999;64:1087—95. (29) Monroe KR, Yu MC, Kolonel LN, Coetzee GA, Wilkins LR, Ross RK, et a1. Evidence of an X—linked or recessive genetic component to prostate cancer risk. Nat Med l995;l:827—9. (30) Xu J, Meyers DA, Freije D, Isaacs S, Wiley K, Nusskem D, et al. Evi— dence for a prostate cancer susceptibility locus on the X chromosome. Nat Genet 1998;20:175—9. (3]) Lehrer S, Fodor F, Stock RG, Stone NN, Eng C, Song HK, et al. Absence of 185delAG mutation of the BRCA1 gene and 6174de1T mutation of the BRCA2 gene in Ashkenazi Jewish men with prostate cancer. Br J Cancer 1998;78:771—3. (32) Mandel JS, Schuman LM. Sexual factors and prostatic cancer: results from a case—control study. J Gerontol 1987;42:259—64. Journal of the National Cancer Institute Monographs No. 27, 2000 (33) Ross RK, Shimizu H, Paganini—Hill A, Honda G, Henderson BE. Case— control studies of prostate cancer in blacks and whites in southern Cali- fornia. J Natl Cancer Inst 1987;78:869—74. (34) Honda GD, Bernstein L, Ross RK, Greenland S, Gerkins V, Henderson BE. Vasectomy, cigarette smoking, and age at first sexual intercourse as risk factors for prostate cancer in middle-aged men. Br J Cancer 1988; 57:326—31. (35) Mills PK, Beeson WL, Phillips RL, Fraser GE. Cohort study of diet, lifestyle, and prostate cancer in Adventist men. Cancer 1989;64:598—604. (36) Hsing AW, McLaughlin JK, Schuman LM, Bjelke E, Gridley G, Wa— cholder S, et al. Diet, tobacco use, and fatal prostate cancer: results from the Lutheran Brotherhood Cohort Study. Cancer Res 1990;50:6836—40. (37) Hsing AW, McLaughlin J K, Hrubec Z, Blot WJ, Fraumeni JF Ir. Tobacco use and prostate cancer: 26-year follow-up of US veterans. Am J Epide— miol 1991;133:437—41. (38) Hayes RB, Pottern LM, Swanson GM, Liff IM, Schoenberg J, Greenberg RS, et al. Tobacco use and prostate cancer in blacks and whites in the United States. Cancer Causes Control 1994;52221—6. (39) Lumey LH. Prostate cancer and smoking: a review of case—control and cohort studies. Prostate 1996;29:249—60. (40) Lumey LH, Pittman B, Zhang EA, Wynder EL. Cigarette smoking and prostate cancer: no relation with six measures of lifetime smoking habits in a large case—control study among US. whites. Prostate 1997;33: 195—200. (41) Handelsman DJ, Conway AJ, Boylan LM, Turtle JR. Testicular function in potential sperm donors: normal ranges and the effects of smoking and varicocele. Int J Androl 1984;7:369—82. (42) Meikle AW, Bishop DT, Stringham JD, Ford MH, West DW. Relation— ship between body mass index, cigarette smoking, and plasma sex steroids in normal male twins. Genet Epidemiol 1989;6z399—412. (43) Lumey LH, Pittman B, Wynder EL. Alcohol use and prostate cancer in US. whites: no association in a confirmatory study. Prostate 1998;36:250—5. (44) Hayes RB, Brown LM, Schoenberg JB, Greenberg RS, Silverman DT, Schwartz AG, et a1. Alcohol use and prostate cancer risk in US blacks and whites. Am J Epidemiol 1996;143:692—7. (45) Chopra 1], Tulchinsky D, Greenway FL. Estrogen—androgen imbalance in hepatic cirrhosis. Studies in 13 male patients. Ann Intern Med 1973;79: 198—203. (46) Gordon GG, Olivo J, Rafil F, Southren AL. Conversion of androgens t0 estrogens in cirrhosis of the liver. J Clin Endocrinol Metab 1975;40: 1018—26. (47) Krstev S, Baris D, Stewart PA, Hayes RB, Blair A, Dosemici M. Risk for prostate cancer by occupation and industry: a 24-state death certificate study. Am J Ind Med 1998;34: 413—20. (48) Krstev S, Baris D, Stewart P, Dosemici M, Swanson GM, Greenberg RS, et a1. Occupational risk factors and prostate cancer in US. blacks and whites. Am J Ind Med 1998;34:421—30. (49) Checkoway H, Mathew RM, Shy CM, Watson I R, Tankersley WG, Wolfe SH, et a1. Radiation, work experience, and cause specific mortality among workers at an energy research laboratory. Br J Indust Med 1985;42: 525—33. (50) Shimizu Y, Kato H, Schull WI. Risk of cancer among atomic bomb survivors. J Radiat Res 1991:2254—63. (51) Eatough JP, Henshaw DL. Radon and prostate cancer. Lancet 1990;335: 1292. (52) Blair A, Zahm SH, Pearce NE, Heineman EF, Fraumeni JF Jr. Clues to cancer etiology from studies of farmers. Scand J Work Environ Health 1992;18:209-15. (53) IARC. Monographs on the Evaluation of Carcinogenic Risks to Humans, Supplement 7, Overall Evaluations of Carcinogenicity: An Updating of IARC Monographs. Vol 1—42. Lyon (France): IARC; 1987. (54) Waalkes MP, Oberdorster G. Cadmium carcinogenesis. In: Foulkes ED, editor. Biological effects of heavy metals, Vol II. Boca Raton (FL): CRC Press: 1990:129—58. (55) Kolonel LN. Nutrition and prostate cancer. Cancer Causes Control 1996; 7:83—94. (56) Bosland MC. Hormonal factors in carcinogenesis of the prostate and testis in humans and animal models. In: Huff J, Boyd I, Barrett JC, editors. Cellular and Molecular Mechanisms of Hormonal Carcinogenesis: Envi- ronmental Influences, New York (NY): Wiley-Liss; 1996:309—52. 61 (57) Kolonel LN, Nomura AM, Cooney RV. Dietary fat and prostate cancer: current status. J Natl Cancer Inst 1999;91:414—28. (58) Hayes RB. Are dietary fat and vasectomy risk factors for prostate cancer? J Natl Cancer Inst 1995;87:629—31. (59) Hayes RB, Ziegler RG, Gridley G, Swanson C, Greenberg RS, Swanson GM, et al. Dietary factors and risks for prostate cancer among blacks and whites in the United States. Cancer Epidemiol Biomarkers Prev l999;8: 25—34. (60) Hankin JH, Zhao LP, Wilkens LR, Kolonel LN. Attributable risk of breast, prostate, and lung cancer in Hawaii due to saturated fat. Cancer Causes Control 1992:3217—23. (61) National Research Council. Diet and Health: Implications for Reducing Chronic Disease Risk. Washington (DC): National Academy Press, Na— tional Academy of Sciences; 1989:314—6. (62) Bosland MC, Oakley-Girvan I, Whittemore AS. Dietary fat, calories, and prostate cancer risk. J Natl Cancer Inst 1999;91:489—91. (63) Hill P, Wynder EL, Garbaczewski L, Games H, Walker AR. Diet and urinary steroids in black and white North American and black South African men. Cancer Res 1979;39:5101—5. (64) Hill P, Wynder EL, Garbaczewski L, Garnes H, Walker AR, Helman P. Plasma hormones and lipids in men at different risk for coronary heart disease. Am J Clin Nutr 1980;33:1010—8. (65) Hill P, Wynder EL, Games H, Walker AR. Environmental factors, hor- mone status, and prostatic cancer. Prev Med 1980;9z657—66. (66) Hill P, Wynder EL, Garbaczewski L, Walker AR. Effect of diet on plasma and urinary hormones in South African black men with prostate cancer. Cancer Res 1982;42:3864—9. (67) Hamalianen E, Adlercreutz H, Puska P, Pietinen P. Diet and serum sex hormones in health men. J Steroid Biochem 1984;20:459—64. (68) Dorgan JF, Judd JT, Longcope C, Brown C, Schatzkin A, Clevidence BA, et al. Effects of dietary fat and fiber on plasma and urine androgens and estrogens in men: a controlled feeding study. Am J Clin Nutr 1996;64:850—5. (69) Bosland MC. Possible enhancement of prostate carcinogenesis by some chemopreventive agents. J Cell Biochem l992;(suppl 16H):l35—7. (70) McCormick DL, Rao KV, Steele VE, Lubet RA, Kelloff GJ, Bosland MC. Chemoprevention of rat prostate carcinogenesis by 9-cis-retinoic acid. Cancer Res 1999;59:521—24. (71) McCormick DL, Rao KV, Dooley L, Steele VE, Lubet RA, Kelloff GJ, et al. Influence of N-methyl-N-nitrosourea, testosterone, and N—(4- hydroxyphenyl)-all-trans—retinamide on prostate cancer induction in Wistar—Unilever rats. Cancer Res 1998;58:3282—8. (72) Esquenet M, Swinnen JV, Heyns W, Verhoeven G. Control of LNCaP proliferation and differentiation: actions and interactions of androgens, 1alpha,25—dihydr0xycholecalciferol, all-trans retinoic acid, 9-cis retinoic acid, and phenylacetate. Prostate 1996;28:182—94. (73) Kuiper JW, Walden P, Bosland MC. Effects of 9-cis-retinoic acid on cell proliferation, differentiation, and apoptosis in LNCaP prostate cancer cells. Proc Am Assoc Cancer Res 1999;40:304. (74) Shen JC, Wang Tl", Chang S, Hursting SD. Mechanistic studies of the effects of the retinoids N—(4-hydroxyphenyl)retinamide on prostate cancer cell growth and apoptosis. Mol Carcinog 1999;24:160—8. (75) Fuller PJ. The steroid receptor superfamily: mechanisms of diversity. FASEB J 1991;523092—9. (76) Schwartz DA, Norris JS. Glucocorticoid, androgen, and retinoic acid regulation of glutathione S—transferase gene expression in hamster smooth muscle tumor cells. Biochem Biophys Res Comm 1992;184:1108—13. (77) Hall RE, Tilley WD, McPhaul MJ, Sutherland RL. Regulation of andro- gen receptor gene expression by steroids and retinoic acid in human breast—cancer cells. Int J Cancer 1992;52:778—84. (78) Mettlin C, Natarajan N, Huben R. Vasectomy and prostate cancer risk. Am J Epidemiol 1990;132:1056—61. (79) Rosenberg L, Palmer JR, Zauber AG, Warshauer ME, Stolley PD, Shapiro S. Vasectomy and the risk of prostate cancer. Am J Epidemiol 1990;132: 1051—5. (80) Perlman JA, Spirtas R, Kelaghan J. Vasectomy and the risk of prostate cancer. Am J Epidemiol 1991;134:107—8. (81) Spitz MR, Fueger JJ, Babaian RJ, Newell GR. Vasectomy and the risk of prostate cancer. Am J Epidemiol 1991;134:108—9. (82) Peterson DE, Remington PL, Anderson HA. Re: “Vasectomy and the risk of prostate cancer”. Am J Epidemiol 1992;135:324—5. 62 (83) Hayes RB, Pottern LM, Greenberg R, Schoenberg J, Swanson GM, Liff J, et al. Vasectomy and prostate cancer in US blacks and whites. Am J Epidemiol 1993;137:263—9. (84) Giovannuci E, Tosteson TD, Speizer FE, Vessey MP, Colditz GA. A long-term study of mortality in men who have undergone vasectomy. N Engl J Med 1992;326:1392—8. (85) Giovannuci E, Ascherio A, Rimm EB, Colditz GA, Stampfer JD, Willett WC. A prospective cohort study of vasectomy and prostate cancer in US men. JAMA 1993;269:873—7. (86) Giovannuci E, Tosteson TD, Speiger FE, Ascherio A, Vessey MP, Colditz GA. A retrospective cohort study of vasectomy and prostate cancer in US men. JAMA 1993;269:878—82. (87) Ross RK, Paganini-Hill A, Henderson BE. The etiology of prostate can- cer: what does the epidemiology suggest? Prostate 1983;4:333—44. (88) Newell GR, Fueger JJ, Spitz MR, Babaian R]. A case—control study of prostate cancer. Am J Epidemiol 1989;130:395—8. (89) Rosenberg L, Palmer JR, Zauber AG, Warshauer ME, Strom BL, Harlap S, et al. The relation of vasectomy to the risk of cancer. Am J Epidemiol 1994;140:431—8. (90) John EM, Whittemore AS, Wu AH, Kolonel LN, Hislop TG, Howe GR, et al. Vasectomy and prostate cancer: results from a multiethnic case— control study. J Natl Cancer Inst 1995;87:662—9. (9]) Zhu K, Stanford JL, Daling JR, McKnight B, Stergachis A, Brawer MK, et al. Vasectomy and prostate cancer: a case—control study in a health maintenance organization. Am J Epidemiol 1996;144:717—22. (92) Platz EA, Yeole BB, Cho E, Jussawalla DJ, Giovannucci E, Ascherio A. Vasectomy and prostate cancer: a case—control study in India. Int J Epi- demiol 1997;26:933—8. (93) Sidney S. Vasectomy and the risk of prostatic cancer and benign prostatic hypertrophy. J Urol 1987;138:795—7. (94) Sidney S, Quesenberry CP, Sadler MC, Guess HA, Lydick EG, Cattolica EV. Vasectomy and the risk of prostate cancer in a cohort of multiphasic health-checkup examinees: second report. Cancer Causes Control 1991; 2:113—6. (95) Nienhuis H, Goldacre M, Seagroatt V, Gill L, Vessey M. Incidence of disease after vasectomy: a record linkage retrospective cohort study. BMJ 1992;304:743—6. (96) Bemal-Delgado E, Latour-Perez J, Pradas-Amal F, Gomez—Lopez L1. The association between vasectomy and prostate cancer: a systematic review of the literature. Fertil Steril 1998;70:191—200. (97) Guess HA. Invited commentary: vasectomy and prostate cancer. Am J Epidemiol 1990;132:1062—5. (98) Howards SS, Peterson HB. Vasectomy and prostate cancer. Chance, bias, or a causal relationship? JAMA 1993;269:913—4. (99) Peterson HB, Howards SS. Vasectomy and prostate cancer: the evidence to date. Fertil Steril 1998;70:201—3. (100) Mo ZN, Huang X, Zhang SC, Yang JR. Early and late long-term effects of vasectomy on serum testosterone, dihydrotestosterone, luteinizing hor- mone and follicle—stimulating hormone levels. J Urol 1995;154:2065—9. (101) Wieland RG, Hallberg MC, Zorn EM, Klein DE, Luria SS. Pituitary— gonadal function before and after vasectomy. Fertil Steril 1972;23: 779—8l. (102) Rosemberg E, Marks SC Jr, Howard PH Jr, James LP. Serum levels of follicle stimulating and luteinizing hormones before and after vasectomy in men. J Urol 1974;111:626—9. (103) Vanna MM, Varma RR, Johanson AJ, Kowarski A, Migeon CJ. Long— term effects of vasectomy on pituitary—gonadal function in man. J Clin Endocrinol Metab 1975;40:868—7 1. (104) Johnsonbaugh RE, O’Connell K, Engel SB, Edison M, Sode J. Plasma testosterone, luteinizing hormone, and follicle-stimulating hormone after vasectomy. Fertil Steril 1975;26:329—30. (105) de la Torre B, Hedman M, Jensen F, Pederson PH, Diczfalusy E. Lack of effect of vasectomy on peripheral gonadotrophin and steroid levels. Int J Androl 1983;6:125—34. (106) Kobrinsky NL, Winter JS, Reyes Fl, Faiman C. Endocrine effects of vasectomy in man. Fertil Steril 1976;27:152—6. (107) Purvis K, Saksena SK, Cekan Z, Diczfalusy CK, Ginger J. Endocrine effects of vasectomy. Clin Endocrinol (Oxf) 1976;5:263—72. (108) smith KD, Tcholakian RK, Chowdhury M. Endocrine studies in vasec— tomized men. In: Lepow IH, Crozier R, editors. Vasectomy, Immunologic Journal of the National Cancer Institute Monographs N0. 27, 2000 and Pathophysiologic Effects in Animals and Man. New York (NY): Academic Press; 1979: 183—97. (109) Whitby RM, Gordon RD, Blair BR. The endocrine effects of vasectomy: a prospective five-year study. Fertil Steril 1979;31:518—20. (110) Alexander NJ, Free MJ, Paulsen CA, Bushbom R, Fulgham DL. A com- parison of blood chemistry, reproductive hormones, and the development of antisperm antibodies after vasectomy in men. J Androl 1980:1240—50. (111) Skegg DC, Mathews JD, Guillebaud J, Vessey MP, Biswas S, Ferguson KM, et al. Hormonal assessment before and after vasectomy. Br Med J 1976;12621—2. (112) Fincham SM, Hill GB, Hanson J, Wijayasinghe C. Epidemiology of pros- tatic cancer: a case—control study. Prostate 1990;17:189—206. (113) Yoshida O, Oishi K, Ohno Y, Schroeder FH. A comparative study on prostatic cancer in the Netherlands and Japan. In: Sasaki R, Aoki K, editors. Proceedings Monbusho International Symposium. Comparative Study of Etiology & Prevention of Cancer. Nagoya (Japan): Nagoya Press; 1990:73—84. (114) Hayes RB, de Jong FH, Raatgever J, Bogdanovicz J, Schroeder FH, van der Maas P, et al. Physical characteristics and factors related to sexual development and behaviour and the risk for prostate cancer. Eur J Cancer Prev l992;l:239—45. (115) Tsitouras PD, Martin CE, Harman SM. Relationship of serum testosterone to sexual activity in healthy elderly men. J Gerontol 1982;37:288—93. (116) Giovannuci E, Leitzmann M, Speigelman D, Rimm EB, Colditz GA, Stampfer MJ, et al. A prospective study of physical activity and prostate cancer in male health professionals. Cancer Res 1998;58:5117—22. (117) Adlercreutz H, Harkonen M, Kuoppasalmi K, Naveri H, Huhtaniemi I, Tikkanen H. Effect of training on plasma anabolic and catabolic steroid hormones and their response during physical exercise. Int J Sports Med 1986;7(suppl l):27—8. (118) Keizer H, Janssen GME, Menheere P, Kranenburg G. Changes in basal plasma testosterone, cortisol, and dehydroepiandrosterone sulfate in pre- viously untrained males and females preparing for a marathon. Int J Sports Med 1989;10(suppl 3):Sl39—45. (119) Demark—Wahnefried W, Lesko SM, Conaway MR, Robertson CN, Clark RV, Lobaugh B, et al. Serum androgens: associations with prostate cancer risk and hair patterning. J Androl 1997;18:495—500. (120) Severson RK, Grove J S, N omura AM, Stemmermann GN. Body mass and prostatic cancer: a prospective study. BMJ 1988;297:713—5. (121) Landry GL, Primos WA Jr. Anabolic steroid abuse. Adv Pediatr 1990; 37:185—205. (122) Kato I, Nomura A, Stemmermann GN, Chyou PH. Determinants of sex hormone levels in men as useful indices in hormone-related disorders. J Clin Epidemiol 1992;45:1417—21. (123) Tymchuk CM, Tessler SB, Aronson WJ, Barnard RJ. Effects of diet and exercise on insulin, sex hormone-binding globulin, and prostate—specific antigen. Nutr Cancer 1998;31:127—31. ( 124) Sandeman TF. The possible dangers of androgens used for male climac- teric [letter]. Med J Aust 1975;12634—5. (125) Guinan PD, Sadoughi W, Alsheik H, Ablin RJ, Alrenga D, Bush IM. Impotence therapy and cancer of the prostate. Am J Surg 1976;131: 599—600. (126) Roberts JT, Essenhigh DM. Adenocarcinoma of prostate in 40—year—old body—builder. Lancet 1986;2z742. (127) Reif J, Pearce N, Fraser J. Cancer risks in New Zealand farmers. Int J Epidemiol 1989;18:768—74. (128) Jackson JA, Waxman J, Spiekerman AM. Prostatic complications of tes- tosterone replacement therapy. Arch Intern Med 1989;149:2365—6. ( 129) Ebling DW, Ruffer J, Whittington R, Vanarsdalen K, Broderick GA, Malkowicz SB, et al. Development of prostate cancer after pituitary dys- function: a report of 8 patients. Urology 1997;49:564—8. (130) Andersson SO, Adami HO, Bergstrom R, Wide L. Serum pituitary and sex steroid hormone levels in the etiology of prostatic cancer—a population- based case—control study. Br J Cancer 1993;68:97—102. (131) Carlstrom K, Stege R. Testicular and adrenocortical function in men with prostatic cancer and in healthy age-matched control. Br J Urol 1997,79: 427—31. (132) Signorello LB, Tzonou A, Mantzoros CS, Lipworth L, Lagiou P, Hsieh CC, et a1. Serum steroids in relation to prostate cancer risk in a case— control study (Greece). Cancer Causes Control 1997;8z632—6. Journal of the National Cancer Institute Monographs N0. 27, 2000 (133) Wolk A, Andersson SO, Bergstrom R. Prospective study of sex hormone levels and risk of prostate cancer. J Natl Cancer Inst 1997;89:820. (134) Ross R, Bernstein L, Judd H, Hanisch R, Pike M, Henderson B. Serum testosterone levels in healthy young black and white men. J Natl Cancer Inst 1986;76:45—8. (135) Ross RK, Bernstein L, Lobo RA, Shimizu H, Stanczyk FZ, Pike MC, et al. 5—alpha-reductase activity and risk of prostate cancer among Japanese and US white and black males. Lancet 1992;339:887—9. (136) Ross RK, Pike MC, Coetzee GA, Reichardt JK, Yu MC, Feigelson H, et a]. Androgen metabolism and prostate cancer: establishing a model of genetic susceptibility. Cancer Res 1998;58:4497—504. (137) Mhatre AN, Trifiro MA, Kaufman M, Kazemi-Esfarjani, Figlewicz D, Rouleau G, et al. Reduced transcriptional regulatory competence of the androgen receptor in X—linked spinal and bulbar muscular atrophy. Nat Genet 1993;5z184—8. (138) Chamberlain NL, Driver E, Miesfeld RL. The length and location of CAG trinucleotide repeats in the androgen receptor N-terrninal domain affect transactivation function. Nucleic Acids Res 1994;22:3181—6. (139) Jenster G, van der Korput HA, Trapman J, Brinkmann A0. Identification of two transcription activation units in the N-terminal domain of the human androgen receptor. J Biol Chem 1995;270:7341—6. (140) Ahluwalia E, Jackson MA, Jones GW, Williams AO, Rao MS, Raj guru S. Blood hormone profiles in prostate cancer patients in high-risk and low- risk populations. Cancer 1981; 48:2267-73. (141) Hill P, Garbaczewski L, Walker AR. Age, environmental factors and prostatic cancer. Med Hypotheses 1984;14:29—39. (142) Henderson BE, Bernstein L, Ross RK, Depue RH, Judd HL. The early in utero oestrogen and testosterone environment of blacks and whites: po— tential effects on male offspring. Br J Cancer 1988;57:216—8. (143) Lookingbill PD, Demers LM, Wang C, Leung A, Rittmaster RS, Santen RJ. Clinical and biochemical parameters of androgen action in normal healthy Caucasian versus Chinese subjects. J Clin Endocrinol Metab 1991;72:1242—8. (144) de J ong FH, Oishi K, Hayes RB, Bogdanowicz JF, Raatgever JW, van der Maas P], et al. Peripheral hormone levels in controls and patients with prostate cancer or benign prostatic hyperplasia: results from the Dutch- Japanese case—control study. Cancer Res 1991;51:3445—50. (145) Ellis L, Nyborg H. Racial/ethnic variations in male testosterone levels: a probable contributor to group differences in health. Steroids 1992;57: 72—5. (146) Wu AH, Whittemore AS, Kolonel LN, John EM, Gallagher RP, West DW, et al. Serum androgens and sex hormone-binding globulins in rela— tion to lifestyle factors in older African-American, white, and Asian men in the United States and Canada. Cancer Epidemiol Biomarkers Prev 1995;42735—41. (147) Corder EH, Friedman GD, Vogelman JH, Orentreich N. Seasonal varia- tion in vitamin D, vitamin D-binding protein, and dehydroepiandros— terone: risk of prostate cancer in black and white men. Cancer Epidemiol Biomarkers Prev 1995;42655—9. (148) Santner SJ, Albertson B, Zhang GY, Zhang GH, Santulli M, Wang C, et al. Comparative rates of androgen production and metabolism in Cauca— sian and Chinese subjects. J Clin Endocrinol Metab 1998;83:2104—9. (149) Labrie F, Sugimoto Y, Luu—The V, Simard J, Lachance Y, Bachvarov D, et al. Structure of human type II SOL-reductase gene. Endocrinology 1992; 13121571—3. (150) Thigpen AE, Davis DL, Milatovich A, Medonca BM, Imperato-McGinley J, Griffin JE, et al. Molecular genetics of steroid Soc—reductase 2 defi— ciency. J Clin Invest 1992;90:799—809. (151) Davis DL, Russell DW. Unusual length polymorphism in human steroid Set—reductase type 2 gene (SRD5A2). Hum Mol Genet 1993;2z820. (152) Reichardt JK, Makridakis N, Henderson BE, Yu MC, Pike MC, Ross RK. Genetic variability of the human SRD5A2 gene: implications for prostate cancer risk. Cancer Res 1995;55:3973—5. (153) Makridakis N, Ross RK, Pike MC, Chang L, Stanczyk FZ, Kolonel LN, et al. A prevalent missense substitution that modulates activity of prostatic steroid Soc—reductase. Cancer Res 1997;57:1020—2. (154) Lunn RM, Bell DA, Mohler JL, Taylor JA. Prostate cancer risk and polymorphism in 17 hydroxylase (CYPI7) and steroid reductase (SRDSAZ). Carcinogenesis 1999;20:1727—31. (155) Febbo PG, Kantoff PW, Platz EA, Casey D, Batter S, Giovannucci E, et 63 al. The V89L polymorphism in the Sa-reductase type 2 gene and risk of prostate cancer. Cancer Res 1999;59:5878—81. (156) Makridakis NM, Ross RK, Pike MC, Crocitto LE, Kolonel LN, Pearce CL, et al. Associations of mis—sense substitution in the SRD5A2 gene with prostate cancer in African—American and Hispanic men in Los Angeles, USA. Lancet 1999;354:975—8. (157) Carey AH, Waterworth D, Patel K, White D, Little J, Novelli P, et al. Polycystic ovaries and premature male pattern baldness are associated with one allele of the steroid metabolism gene CYP17. Hum Mol Genet 1994;3z1873—6. (158) Feigelson HS, Shames LS, Pike MC, Coetzee GA, Stanczyk FZ, Hender— son BE. Cytochrome P4500170t gene (CYP17) polymorphism is associ- ated with serum estrogen and progesterone concentrations. Cancer Res 1998;58:585—7. (159) Haiman CA, Hankinson SE, Speigelman D, Colditz GA, Willett WC, Speizer FE, et al. The relationship between a polymorphism in CYP17 with plasma hormone levels and breast cancer. Cancer Res 1999;59: 1015—20. (160) Verreault H, Dufort I, Simard J, Labrie F, Luu-The V. Dinucleotide repeat polymorphisms in the HSD3B2 gene. Hum Mol Genet 1994;3:384. (161) Labrie F, Simard J, Luu—The V, Pelletier G, Belanger A, Lachance Y, et al. Structure and tissue-specific expression of 38—hydroxysteroid de- hydrogenase/5—ene—4—ene isomerase genes in human and rat classical and peripheral steroidogenic tissues. J Steroid Biochem Mol Biol 1992;41: 421—35. (162) Devgan SA, Henderson BE, Yu MC, Shi CY, Pike MC, Ross RK, et al. Genetic variation of 3B—hydroxysteroid dehydrogenase type II in three racial/ethnic groups: implications for prostate cancer risk. Prostate 1997; 33:9—12. (163) Irvine RA, Yu MC, Ross RK, Coetzee GA. The CAG and GGC micro- satellites of the androgen receptor gene are in linkage disequilibrium in men with prostate cancer. Cancer Res 1995;55:1937—40. (164) Sartor O, Zheng Q, Eastham JA. Androgen receptor gene CAG repeat length varies in a race—specific fashion in men without prostate cancer. Urology 1999;53:378—80. (165) Ekman P, Gronberg H, Matsuyama H, Kivineva M, Bergerheim US, Li C. Links between genetic and environmental factors and prostate cancer risk. Prostate 1999;39:262—8. (166) Horton R. Dihydrotestosterone is a peripheral paracrine hormone. J An— drol 1982;13:23—7. (167) Moghissi E, Ablan F, Horton R. Origin of plasma androstanediol gluc- uronide in men. J Clin Endocrinol Metab 1984;59:417—21. (168) Horton R, Hawks D, Lobo R. 30L,17[3-androstanediol glucuronide in plasma. J Clin Invest 1992;69:1203—6. (169) Carter HB, Pearson JD, Metter EJ, Chan DW, Andres R, Fozard JL, et al. Longitudinal evaluation of serum androgen levels in men with and with— out prostate cancer. Prostate 1995;27:25—31. (I70) Nomura A, Heilbrun LK, Stemmermann GN, Judd HL. Prediagnostic serum hormones and the risk of prostate cancer. Cancer Res 1988;48:3515—7. (171) Barrett-Connor E, Garland C, McPhillips JB, Khaw KT, Wingard DL. A prospective, population—based study of androstenedione, estrogens, and prostatic cancer. Cancer Res 1990;50:169—731. ( I 72 ) Hsing AW, Comstock GW. Serological precursors of cancer: serum hor- mones and risk of subsequent prostate cancer. Cancer Epidemiol Bio- markers Prev 1993;2227—32. (173) Comstock GW, Gordon GB, Hsing AW. The relationship of serum de— hydroepiandrosterone and its sulfate to subsequent cancer of the prostate. Cancer Epidemiol Biomarkers Prev 1993;2:219—21. ( I 74 ) Nomura AM, Stemmermann GN, Chyou PH, Henderson BE, Stanczyk, FZ. Serum androgens and prostate cancer. Cancer Epidemiol Biomarkers Prev 1996;5z621—5. (175) Gann PH, Hennekens CH, Ma J, Longcope C, Stampfer MJ. Prospective study of sex hormones levels and risk of prostate cancer. J Natl Cancer Inst 1996;88:1118—26. (I 76) Guess HA, Friedman GD, Sadler MC, Stanczyk FZ, Vogelman JH, Im- perato-McGinley, et al. Sa—Reductase activity and prostate cancer: a case—control study using stored sera. Cancer Epidemiol Biomarkers Prev 1997;6:21—4. (I77) Vatten LJ, Usrin G, Ross RK, Stanczyk FZ, Lobo RA, Harvei S, et al. Androgens in serum and the risk of prostate cancer: a nested case—control 64 study from the Janus serum bank in Norway. Cancer Epidemiol Biomark- ers Prev l997:6:967—9. (178) Dorgan JF, Albanes D, Virtamo J, Heinonen OP, Chandler DW, Galma- rini M, et a1. Relationships of serum androgens and estrogens to prostate cancer risk: results from a prospective study in Finland. Cancer Epidemiol Biomarkers Prev 1998;7:1069—74. (179) Heikkila R, Aho K, Heliovaara M, Hakama M, Mamiemi J, Reunanen A, et al. Serum testosterone and sex hormone-binding globulin concentra— tions and the risk of prostate carcinoma: a longitudinal study. Cancer 1999;86:312—5. (I80) Kantoff PW, Febbo PG, Giovannucci E, Krithivas K, Dahl DM, Chang G, et al. A polymorphism of the Set-reductase gene and its association with prostate cancer: a case—control analysis. Cancer Epidemiol Biomarkers Prev 1997;6:189—92. (181) Wadelius M, Andersson SO, Johansson JE, Wadelius C, Rane A. Prostate cancer associated with CYP17 genotype. Pharmacogenetics 1999;9:635—9. (I82) Giovannuci E, Stampfer MJ, Krithavas K, Brown M, Dahl D, Brufsky A, et al. The CAG repeat within the androgen receptor gene and its relation- ship to prostate cancer. Proc Natl Acad Sci U S A 1997;94:3320—3. (183) Ingles SA, Ross RK, Yu MC, Irvine RA, La Pera G, Haile RW, et al. Association of prostate cancer risk with genetic polymorphisms in vitamin D receptor and androgen receptor. J Natl Cancer Inst 1997;89:166—70. (184) Stanford JL, Just JJ, Gibbs M, Wicklund KG, Neal CL, Blumenstein BA, et al. Polymorphic repeats in the androgen receptor gene: molecular mark- ers of prostate cancer risk. Cancer Res 1997;57:1194—8. (I85) Correa-Cerro L, Wohr G, Haussler J, Berthon P, Drelon E, Mangin P, et al. (CAG)nCAA and GGN repeats in the human androgen receptor gene are not associated with prostate cancer in a French—German population. Eur J Hum Genet 1999;7:357—62. (186) Hakimi JM, Schoenberg MP, Rondinelli RH, Piantadosi S, Barrack ER. Androgen receptor variants with short glutamine or glycine repeats may identify unique subpopulations of men with prostate cancer. Clin Cancer Res 1997;3z1599—1608. (187) Platz EA, Giovannucci E, Dahl DM, Krithivas K, Hennekens CH, Brown M, et al. The androgen receptor gene GGN microsatellite and prostate cancer risk. Cancer Epidemiol Biomarkers Prev 1998;7:379—84. (188) Eaton NE, Reeves GK, Appelby PN, Key TJ. Endogenous sex hormones and prostate cancer: a quantitative review of prospective studies. Br J Cancer 1999;80:930—4. (189) Bratt O, Borg A, Kristoffersson U, Lundgren R, Zhang QX, Olsson H. CAG repeat length in the androgen receptor gene is related to age at diagnosis of prostate cancer and response to endocrine therapy, but not to prostate cancer risk. Br J Cancer 1999;81:672—6. ([90) Hardy DO, Scher HI, Bogenreider T, Sabbatini P, Zhang ZF, Nanus DM, et al. Androgen receptor CAG repeat lengths in prostate cancer: correla— tion with age of onset. J Clin Endocrinol Metab 1996;81:4400—5. (19]) Hsing AW. Hormones and prostate cancer: where do we go from here? J Natl Cancer Inst 1996;88:1093—5. ( I 92 ) Gail MH, Fears TR, Hoover RN, Chandler DW, Donaldson JL, Hyer MB, et al. Reproducibility studies and interlaboratory concordance for assays of serum hormone levels: estrone, estradiol, estrone sulfate, and proges- terone. Cancer Epidemiol Biomarkers Prev 1996;5z835—44. ( I 93 ) McShane LM, Dorgan JF, Greenhut S, Damato JJ. Reliability and validity of sex serum hormone measurements. Cancer Epidemiol Biomarkers Prev 1996;52923—8. (I94) Wynder EL, Fujita Y, Harris RE, Hirayama T, Hiyama T. Comparative epidemiology of cancer between the United Sates and Japan. A second look. Cancer 1991;67:746—63. (195) Matzkin H, Soloway MS. Immunohistochemical evidence of the existence and localization of aromatase in human prostatic tissues. Prostate 1992; 21:309—14. ([96) Tsugaya M, Harada N, Tozawa K, Yamada Y, Hayashi Y, Tanaka S, et al. Aromatase mRNA levels in benign prostatic hyperplasia and prostate cancer. Int J Urol 1996;32292—6. (197) Castagnetta LA, Granata OM, Bellavia V, Amodio R, Scaccianoce E, Notarbartolo M, et al. Product of aromatase activity in intact LNCaP and MCF—7 human cancer cells. J Steroid Biochem Mol Biol 1997;61:287—92. (198) Hiramatsu M, Maehara I, Ozaki M, Harada N, Orikasa S, Sasano H. Aromatase in hyperplasia and carcinoma of the human prostate. Prostate 1997;31:118—24. Journal of the National Cancer Institute Monographs No. 27, 2000 (199) Negri-Cesi P, Poletti A, Colciago A, Magni P, Martini P, Motta M. Presence of Sa-reductase isozymes and aromatase in human prostate can- cer cells and in benign prostate hyperplastic tissue. Prostate 1998;34: 283—91. (200) Smith T, Chisholm GD, Habib FK. Failure of human benign prostatic hyperplasia to aromatise testosterone. J Steroid Biochem 1982;17:119— 20. (201) Brodie AM, Son C, King DA, Meyer KM, Inkster SE. Lack of evidence for aromatase in human prostatic tissues: effects of 4-hydroxyandrostene— dione and other inhibitors on androgen metabolism. Cancer Res 1989;49: 6551—5. (202) Krieg M, Nass R, Tunn S. Effect of aging on endogenous level of 50(- dihydrotestosterone, testosterone, estradiol, and estrone in epithelium and stroma of normal and hyperplastic human prostate. J Clin Endocrinol Metab 1993;77:375—231. (203) Meikle AW, Stanish WM, Taylor N, Edwards CQ, Bishop CT. Familial effects on plasma sex-steroid content in man: testosterone, estradiol and sex-hormone-binding globulin. Metabolism 1982;31:6—9. (204) Meikle AW, Smith JA, West DW. Familial factors affecting prostatic cancer risk and plasma sex-steroid levels. Prostate 1985;62121—8. (205) Zumoff B, Levin J, Strain GW, Rosenfeld RS, O’Connor J, Freed SZ, et a]. Abnormal levels of plasma hormones in men with prostate cancer: evidence toward a “two—disease” theory. Prostate 1982;3z579—88. (206) Hammond GL. Endogenous steroid levels in the human prostate from birth to old age: a comparison of normal and diseased tissues. J Endocrinol 1978;78:7—19. (207) Ekbom P. Growing evidence that several human cancers may originate in utero. Sem Cancer Biol 1998;8:237—44. (208) Ross RK, Henderson BE. Do diet and androgens alter prostate cancer risk via a common etiologic pathway? J Natl Cancer Inst 1994;86:252—4. (209) Powell IJ. Prostate cancer and African-American men. Oncology 1997; 11:599—605. (210) Bosland MC. Animal models for the study of prostate carcinogenesis. J Cell Biochem Suppl 1992;16Hz89—98. (211) Bosland MC. Chemical and hormonal induction of prostate cancer in animal models. Urol Oncol l996;2:103—10. (212 ) Lane KE, Ricci MJ, Ho SM. Effect of combined testosterone and estra— diol-17E» treatment on the metabolism of E2 in the prostate and liver of noble rats. Prostate 1997;30:256—62. (213) McNeal JE, Redwine EA, Freiha FS, Stamey TA. Zonal distribution of prostatic adenocarcinoma. Correlation with histologic pattern and direc- tion of spread. Am J Surg Pathol 1988;12:897—906. (214) Price D. Comparative aspects of development and structure in the pros— tate. Monogr Natl Cancer Inst 1963;12:1—27. (215) Noble RL. Prostate carcinoma of the Nb rat in relation to hormones. Int Rev Exp Pathol 1982;23:113—59. (216) Pollard M, Luckert PH, Schmidt MA. Induction of prostate adenocarci- nomas in Lobund Wistar rats by testosterone. Prostate 1982;31563—8. (217) Pollard M, Luckert PH. Production of autochthonous prostate cancer in Lobund—Wistar rats by treatments with N-nitroso—N—methylurea and tes- tosterone. J Natl Cancer Inst 1986;77:583—7. (218) Pour PM, Stepan K. Induction of prostatic carcinomas and lower urinary tract neoplasms by combined treatment of intact and castrated rats with testosterone propionate and N—nitrosobis(2—oxopropyl)amine. Cancer Res 1987;47:5699—706. (219) Hoover DM, Best KL, McKenney BK, Tamura RN, Neubauer BL. Ex— perimental induction of neoplasia in the accessory sex organs of male Lobund—Wistar rats. Cancer Res 1990;50:142—6. (220) Bosland MC, Dreef—Van Der Meulen, HC, Sukumar S, Ofner P, Leav 1, Han X, et a1. Multistage prostate carcinogenesis: the role of hormones. In: Harris CC, Hirohashi S, Ito N, Pitot HC, Sugimura T, Terada M, Yokota J, editors. Multistage carcinogenesis. Japan Sci Soc Press, Tokyo/CRC Press. Boca Raton (FL): CRC Press 1992;109—23. (221) Shirai T, Tamano S, Kato T, Iwasaki S, Takahashi S, Ito N. Induction of invasive carcinomas in the accessory sex organs other than the ventral prostate of rats given 3,2’—dimethyl—4-aminobiphenyl and testosterone propionate. Cancer Res 1991;51:1264-9. (222) Pollard M, Luckert PH. Promotional effects of testosterone and dietary fat on prostate carcinogenesis in genetically susceptible rats. Prostate 1985; 611—5. Journal of the National Cancer Institute Monographs No. 27, 2000 (223) Pollard M, Luckert PH. Autochthonous prostate adenocarcinomas in Lobund—Wistar rats: a model system. Prostate 1987;11:219—27. (224) Pollard M, Luckert PH, Snyder D. Prevention and treatment of experi- mental prostate cancer in Lobund—Wistar rats. I. Effects of estradiol, dihydrotestosterone, and castration. Prostate 1989;15:95—103. (225) Pollard M, Luckert PH, Snyder DL. The promotional effect of testoster- one on induction of prostate-cancer in MNU—sensitized L-W rats. Cancer Lett 1989;45:209—12. (226) Tamano S, Rehm S, Waalkes MP, Ward JM. High incidence and histo— genesis of seminal vesicle adenocarcinomas and lower incidence of pros— tate carcinomas in Lobund—Wister prostate cancer rat model using N- nitrosomethylurea and testosterone. Vet Pathol 1996;33:557—67. (227) Drago JR. The induction of NB rat prostatic carcinomas. Anticancer Res 1984;4z255—6. (228) Leav I, Merk FB, Kwan PW, Ho SM. Androgen—supported estrogen- enhanced epithelial proliferation in the prostates of intact Noble rats. Prostate 1989;15:23—40. (229) Bosland MC, Ford H, Horton L. Induction at high incidence of ductal prostate adenocarcinomas in NBL/Cr and Sprague—Dawley Hsd:SD rats treated with a combination of testosterone and estradiol-173 or diethyl— stilbestrol. Carcinogenesis 1995;16:131 l—7. (230) Ofner P, Bosland MC, Vena RL. Differential effects of diethylstilbestrol and estradiol—17B in combination with testosterone on rat prostate lobes. Toxicol Appl Pharmacol 1992;112:300—9. (231) McLachlan J A. Rodent models for perinatal exposure to diethylstilbestrol and their relation to human disease in the male. In: Herbst A, Bern HA, editors. Developmental Effects of Diethylstilbestrol (DES) in Pregnancy. Stuttgart (Germany):Thieme Verlag, 1981:148—57. (232) Arai Y, Mori T, Suzuki Y, Bern HA. Long-term effects of perinatal exposure to sex steroids and diethylstilbestrol on the reproductive system of male mammals. Int Rev Cytol 1983;84:235—68. (233) Santti R, Newbold RR, Makela S, Pylkkanen L, McLachlan JA. Devel- opmental estrogenization and prostatic neoplasia. Prostate 1994;24:67—78. (234) McLachlan JA, Newbold RR, Bullock B. Reproductive tract lesions in male mice exposed prenatally to diethylstilbestrol. Science 1975:1902 991—2. (235) Newbold RR, Bullock BC, McLachlan JA. Testicular tumors in mice exposed in utero to diethylstilbestrol. J Urol 1987;138:1446—50. (236) Pylkkanen L, Santti R, Newbold R, McLachlan JA. Regional differences in the prostate of the neonatally estrogenized mouse. Prostate 1991;18: 117—29. (237) Pylkkanen L, Santti R, Maentausta O, Vihko R. Distribution of estradiol- 17B hydroxysteroid oxidoreductase in the urogenital tract of control and neonatally estrogenized male mice: immunohistochemical, enzymehisto- chemical, and biochemical study. Prostate 1992;20:59—72. (238) Vorherr H, Messer RH, Vorherr UF, Jordan SW, Kornfeld M. Teratogen— esis and carcinogenesis in rat offspring after transplacental and transmam— mary exposure to diethylstilbestrol. Biochem Pharmacol 1979;28: 1865—77. (239) Pour PM. Prostatic cancer induced in MRC rats by N-nitrosobis(2- oxopropyl)—amine and N—nitrosobis(2-hydroxypropyl)amine. Carcinogen— esis l983;4:49—55. (240) Shirai T, Sakata T, Fukushima S, Ikawa E, Ito N. Rat prostate as one of the target organs for 3,2'-dimethyl-4-aminobiphenyl—induced carcinogen- esis: effects of dietary ethinyl estradiol and methyltestosterone. Jpn J Cancer Res 1985;76:803—8. (241) Rivenson A, Silverrnan J. The prostatic carcinoma in laboratory animals: a bibliographic survey from 1900 to l977. Invest Urol 1979;16:468—72. (242) Shirai T, Fukushima S, Ikawa E, Tagawa T, Ito N. Induction of prostate carcinoma in situ at high incidence in F344 rats by combination of 3,2’- dimethyl-4-aminobiphenyl and ethinyl estradiol. Cancer Res 1986;46: 6423—6. ( 243 ) Ito N, Shirai T, Tagawa Y, Nakamura A, Fukushima S. Variation in tumor yield in the prostate and other target organs of the rat in response to varied dosage and duration of administration of 3,2'—dimethyl—4-aminobiphenyl. Cancer Res 1988;48:4629—32. (244) Bosland MC, Prinsen MK. Induction of dorsolateral prostate adenocarci— nomas and other accessory sex gland lesions in male Wistar rats by a single administration of N—methyl-N-nitrosourea, 7,12—dimethyl— benz[a]anthracene, and 3,2’-dimethy]-4-aminobiphenyl after sequential 65 treatment with cyproterone acetate and testosterone propionate. Cancer Res 1990; 50:691—9. (245) Bosland MC, Prinsen MK, Kroes R. Adenocarcinomas of the prostate induced by N-nitroso—N—methylurea in rats pretreated with cyproterone acetate and testosterone. Cancer Lett 1983;18:69—78. (246) Shirai T, Ikawa E, Imaida K, Tagawa Y, Ito N. Proliferative response of rat accessory sex organs to dietary sex hormones after castration of initial dietary administration of estrogen. J Urol 1987;138:216—8. (247) Takai K, Kakizoe T, Tobisu K, Ohtani M, Kishi K, Sato S, et a1. Sequen— tial changes in the prostate of rats treated with chlormadinone acetate, testosterone and N-nitroso-N—methylurea. J Urol 1988;139:1363—6. (248) Shirai T, Imaida K, Iwasaki S, Mori T, Tada M, Ito N. Sequential obser- vation of rat prostate lesion development induced by 3,2'—dimethyl—4— aminobiphenyl and testosterone. Jpn Cancer Res 1993;84:20—5. (249) Waalkes MP, Rehm S, Riggs CW, Bare RM, Devor DE, Poirier LA, et a1. Cadmium carcinogenesis in male Wistar [Crl:(WI)BR] rats: dose—re— sponse analysis of tumor induction in the prostate and testes and at the injection site. Cancer Res 1988;48:4656—63. (250) Waalkes MP, Rehm S, Riggs CW, Bare RM, Devor DE, Poirier LA, et a1. Cadmium carcinogenesis in male Wistar [Crl:(WI)BR] rats: dose—re- sponse analysis of effects of zinc on tumor induction in the prostate, in the testes, and at the injection site. Cancer Res 1989;49:4282—8. (251) Waalkes MP, Perantoni A. Apparent deficiency of metallothionein in the Wistar rat prostate. Toxicol Appl Pharmacol 1989;101:83—94. (252) Waalkes MP, Rehm S, Perantoni A0, Coogan TP. Cadmium exposure in rats and tumours of the prostate. In: Nordberg GF, Herber, RFM, Alessio L, editors. Cadmium in the Human Environment: Toxicity and Carcino- genicity. Lyon (France): IARC; 19922391400. (253) Hirose F, Takizawa S, Watanabe H, Takeichi N. Development of adeno— carcinoma of the prostate in ICR mice locally irradiated X-rays. Jpn J Cancer Res 1976;67:407—11. (254) Takizawa S, Hirose F. Role of testosterone in the development of radia— tion—induced prostate carcinoma in rats. Gann 1978;69:723—6. (255) Driscoll SG, Taylor SH. Effects of prenatal maternal estrogen on the male urogenital system. Obstet Gynecol 1980;56:537—42. (256) Sugimura Y, Cunha GR, Yonemura CU, Kawamura J. Temporal and spatial factors in diethylstilbestrol—induced squamous metaplasia of the developing human prostate. Hum Pathol 1988;19:133—9. (257) Henderson BE, Ross RK, Pike MC, Casagrande JT. Endogenous hor— mones as a major factor in human cancer. Cancer Res 1982;42:3232—9. (258) Cunha GR, Foster B, Thomson A, Sugimura Y, Tanji N, Tsuji M, et a1. Growth factors as mediators of androgen action during the development of the male urogenital tract. World J Urol 1995;13:264—76. (259) Hayward SW, Rosen MA, Cunha GR. Stromal-epithelial interactions in the normal and neoplastic prostate. Br J Urol l997;79(suppl 2):18—26. (260) Ripple MO, Henry WF, Rago RP, Wilding G. Prooxidant—antioxidant shift induced by androgen treatment of human prostate carcinoma cells. J Natl Cancer Inst 1997;89:40—8. (261) Tuohimaa P. Control of cell proliferation in male accessory sex glands. In: Spring—Mills E, Hafez ES, editors. Male Accessory Sex Glands. Amster- dam (The Netherlands): Elsevier/North-Holland Biomedical Press; 1980: 131—53. (262) Orsilles MA, Depiante-Depaoli M. Oxidative stress-related parameters in prostate of rats with experimental autoimmune prostatitis. Prostate 1998; 34:270—4. (263) De Marzo AM, Coffey DS, Neslon WG. New concepts in tissue speci— ficity for prostate cancer and benign prostatic hyperplasia. Urology l999;53(supp1 3a):29—39. (264) Sukumar S, Armstrong B, Bruyntjes JP, Leav I, Bosland MC. Frequent activation of the Ki—ras oncogene at codon 12 in N—methyl-N—nitrosourea- induced rat prostate adenocarcinomas and neurogenic sarcomas. M01 Car— cinog 1991;4:362—8. (265) Ho SM, Lane K. Sex hormone-induction and dietary modulation of pros- tatic adenocarcinoma (PA) in animal models. Urol Oncol 1996;22110—5. (266) Marts SA, Padilla GM, Petrow V. Aromatase activity in microsomes from rat ventral prostate and Dunning R3327H rat prostatic adenocarcinoma. J Steroid Biochem 1987;26:25—9. 66 (267) Santti R, Newbold RR, McLachlan JA. Androgen metabolism in control and neonatally estrogenized male mice. Reprod Toxicol 1991;5z149—55. (268) Lau KM, Leav I, Ho SM. Rat estrogen receptor-(x and -B, and progester- one receptor mRNA expression in various prostatic lobes and microdis— sected normal and dysplastic epithelial tissues of the Noble rats. Endo— crinology 1998;139:424—7. (269) Miyata E, Kawabe M, Sano M, Takesada Y, Takahashi S, Shirai T. Effects of tamoxifen, an antiestrogen, on rat prostate carcinogenesis by 3,2’-dimethyl-4-amniobiphenyl and testosterone do not support an estro- gen role in testosterone promotion. Prostate 1997;31:9—13. (2 70) Thompson C], Ho SM, Lane K, Leav I. The role of prolactin in the genesis of dysplasia in the rat prostate. Proc Am Assoc Cancer Res 1999;40:382. (271) Ho SM, Yu M, Leav I. The conjoint actions of androgens and estrogens in the induction of proliferative lesions in the rat prostate. In: Li J], Nandi S, Li SA, editors. Hormonal Carcinogenesis. Proceedings of the First International Symposium. New York (NY): Springer—Verlag; 1992: 18—25. (272) Nakamura A, Shirai T, Ogawa K, Wada S, Fujimoto NA, Ito A, et al. Promoting action of prolactin released from a grafted transplantable pi— tuitary tumor (MtT/F84) on rat prostate carcinogenesis. Cancer Lett 1990; 53:151—7. (273) Lane KE, Leav I, Ziar J, Bridges RS, Rand WM, Ho SM. Suppression of testosterone and estradiol-17B-induced dysplasia in the dorsolateral pros— tate of Noble rats by bromocriptine. Carcinogenesis 1997;18:1505—10. (274) Yager JD, Liehr JG. Molecular mechanisms of estrogen carcinogenesis. Ann Rev Pharmacol Toxicol 1996;36:203—32. (275) Gladek A, Liehr JG. Mechanism of genotoxicity 0f diethylstilbestrol in vivo. J Biol Chem 1989;264:16847—52. (276) Liehr JG, Avitts TA, Randerath E, Randerath K. Estrogen-induced en- dogenous DNA adduction: Possible mechanism of hormonal cancer. Proc Natl Acad Sci U S A 1986;83:5301—5. (277) Han X, Liehr JG, Bosland MC. Induction of a DNA adduct detectable by 32P-postlabeling in the dorsolateral prostate of NBL/Cr rats treated with estradiol-17B and testosterone. Carcinogenesis 1995;16:951—4. (278) Ho SM, Roy D. Sex hormone—induced nuclear DNA damage and lipid peroxidation in the dorsolateral prostates of Noble rats. Cancer Lett 1994; 84:155—62. (279) Olinksi R, Zastawny TH, Foksinski M, Barecki A, Dizdaroglu M. DNA base modifications and antioxidant enzyme activities in human benign prostatic hyperplasia. Free Radicals Biol Med 1995;18:807—13. (280) Hietanen E, Bartsch H, Bereziat JC, Camus AM, McClinton S, Eremin O, et a1. Diet and oxidative stress in breast, colon and prostate cancer pa- tients: a case—control study. Eur J Clin Nutr 1994;48:575—86. (28]) Andre JM, Berger JM, De Turckheim M, Veyssiere G. Estimation of testosterone and androstenedione in the plasma and testes of cryptorchid offspring of mice treated with oestradiol during pregnancy. J Reprod Fertil 1975;44:235—47. (282) Dalterio S, Bartke A, Steger R, Mayfield D. Neonatal exposure to DES in BALB/c male mice: effects on pituitary-gonadal function. Pharmacol Bio- chem Behavior 1985;22:1019—24. (283) Edery M, Turner T, Dauder S, Young G, Bern HA. Influence of neonatal diethylstilbestrol treatment on prolactin receptor levels in the mouse male reproductive system. Proc Soc Exp Biol Med 1990;194:289—92. (284) Prins GS. Neonatal estrogen exposure induces lobe-specific alterations in adult rat prostate androgen receptor expression. Endocrinology 1992:1302 2401—12. (285) Ross RK, Garbeff P, Paganini-Hill A, Henderson BE. Effect of in utero exposure to diethylstilbestrol on age at onset of puberty and on postpu— bertal hormone levels in boys. Can Med Assoc J 1983;128:1197—8. NOTE Supported in part by Public Health Service grants CA58088, CA 75293, CA13343 (National Cancer Institute), and ES 00260 (National Institute of En— vironmental Health Sciences), National Institutes of Health, Department of Health and Human Services. Journal of the National Cancer Institute Monographs No. 27, 2000 Chapter 3: Endogenous Estrogens as Carcinogens Through Metabolic Activation James D. Yager A common thread linking the main risks for developing breast cancer in women is cumulative, excessive exposure to estrogen. The standard paradigm to account for this asso- ciation focuses on increased cell proliferation caused by es- trogen through estrogen receptor-mediated signal transduc- tion accompanied by increased probability for mutation to occur during DNA synthesis. This chapter provides an over- view of the mounting evidence, provided from cell culture and whole animal experimental studies, in support of a role for the oxidative metabolites of estrogen, in particular, the catechol estrogens, in the development of estrogen carcino- genesis. This provides a paradigm for how estrogens may contribute to the development of human breast cancer. The chapters that follow will fill in the details. Evidence shows that the catechols themselves are signaling molecules that work through the estrogen receptor. In addition, upon fur- ther oxidation, the catechols can give rise to reactive qui- nones capable of forming direct adducts with glutathione and purines in DNA and of redox cycling to generate reac- tive oxygen species that can cause oxidative damage. Estra- diol and estrone, as well as their 4-hydroxy catechols, are carcinogenic in the Syrian golden hamster kidney, and ethi- nyl estradiol is a strong promoter of hepatocarcinogenesis in the rat. Increased oxidative DNA damage has been de- tected in target tissues after estrogen treatment in both animal model systems. Furthermore, several recent molecu- lar epidemiologic studies have found that a polymorphism associated with a low-activity form of catechol-O-methyl- transferase, an enzyme involved in the inactivation of cat- echol estrogens, is associated with an increased risk for de- veloping breast cancer. The increased risk is observed in certain women, although the studies are not consistent on which subgroup of women (e.g., premenopausal or post- menopausal) is at increased risk, and one study detected no increased risk. Reasons for such discrepancies are discussed in light of factors, such as genetic polymorphisms and envi- ronmental/lifestyle susceptibility factors, which control the tissue-specific balance within cells among the estrogen me- tabolites. It is concluded that such factors will have to be identified through additional mechanistic studies and that, as they are identified, they can be incorporated into future molecular epidemiologic studies designed to determine their actual impact on cancer risk in human populations. [J Natl Cancer Inst Monogr 2000;27:67—73] For a substantial fraction of breast cancer cases in women, well-established risk factors, revealed by epidemiologic studies, include early age at menarche, late first full—term pregnancy, nulliparity, late menopause, family history of breast cancer, so- cioeconomic status, and perhaps estrogen replacement therapy (1—5). A common thread linking these factors is cumulative, excessive lifetime exposure to estrogen, suggesting that this ex- Journal of the National Cancer Institute Monographs No. 27, 2000 posure has an important role in the cause of breast cancer. Al- though a number of environmental chemicals are suspected of contributing to breast cancer, no single environmental chemical has been identified as a strong “smoking gun” for causing breast cancer (6). However, a consequence of excessive estrogen ex- posure may include unwanted cell division. The standard para- digm providing a general mechanistic explanation for the asso- ciation of cumulative, excessive estrogen exposure and breast cancer risk was aptly stated by Feigelson and Henderson (4) and is shown in Fig. 1. The notion is that the proliferative stimulus provided by l7B—estradiol (E2) leads to the appearance of spon- taneous mutations; thus, the key contribution of E2 is the stimu- lation of breast epithelial cell proliferation (Chapter 8). How- ever, an important aspect of estrogen toxicology is its tissue- specific, cellular oxidative metabolism by several specific cytochrome P450 isoforms and various peroxidases (7—10). Mounting evidence suggests that the oxidative metabolites may contribute to estrogen carcinogenesis (Chapters 4 and 5). Among the metabolites formed during the process of estrogen biotransformation and elimination (Fig. 2), some are estrogenic ( I I ) and some may be protective through their antioxidant prop- erties and/or growth and angiogenesis inhibitory activities (12— 14). On the other hand, the more reactive quinone metabolites are able to form direct adducts with DNA (15) and/or can cause oxidative damage to lipids (16) and DNA through redox cycling processes that produce reactive oxygen species (ROS) [( 7, 8); Chapter 4]. Increased production of ROS could also lead to disruption of cellular redox homeostasis and, as a consequence, could alter transcription factor function, causing inappropriate alterations in the regulation of gene expression (J 7). The possible contribution of these metabolites to estrogen carcinogenesis has received relatively little attention compared with that given to estrogen receptor-mediated processes. How- ever, accumulating evidence, much of which was presented in this symposium, supports an expansion of the standard mecha- nistic paradigm for the causal association of estrogen exposure and breast cancer (Fig. 1). Thus, while estrogen-induced cell proliferation undoubtedly has an important role in estrogen car- cinogenesis, complementary pathways involving indirect and/or direct genotoxicity originating from estrogen metabolites, in par- ticular, the 4-hydroxy catechol metabolite, are also likely to make important contributions. Furthermore, since other metabo- lites, such as 2-methoxyE2, may have protective effects, a bal- ance among these metabolites is likely required to maintain ho- meostasis. In this chapter, I will provide a brief overview of some evi— dence in support of a role for estrogen metabolites in estrogen Correspondence to: James D. Yager, Ph.D., Division of Toxicological Sci— ences, Department of Environmental Health Sciences, The Johns Hopkins Uni— versity School of Hygiene and Public Health, 615 North Wolfe St., Baltimore, MD 21205 (e-mail: jyager@jhsph.edu). See “Notes” following “References.” © Oxford University Press 67 Fig. 1. Standard and modified paradigms for estrogen carcinogenesis. Estrogen, and perhaps progesterone “...affect the rate of cell division and thus manifest their effect on the risk of breast cancer by causing proliferation of breast epithelial cells. Proliferating cells are susceptible to genetic errors during DNA replication which, if uncorrected, can ultimately lead to a malignant phenotype.” (Feigelson and Henderson, Carcinogenesis, 17:2279-84, 1996) While estrogen-induced cell proliferation undoubtedly has important role in the carcinogenic process, mounting evidence supports a complimentary pathway involving: Indirect and direct genotoxicity originating from estrogen metabolites, i.e. 4—OHE1 -Indirect: Oxidative DNA damage via Redox Cycling —v ROS ~Direct: Estrogen-quinone DNA adducts °Pr0tective effects: Perhaps through 2-methoxy catechol estrogen-mediated growth inhibition, apoptosis and anti—angiogenesis Standard Paradigm Modified Paradigm carcinogenesis. I will place particular emphasis on their potential for causing oxidative DNA damage in association with the car- cinogenic process [see also two reviews (7,8)]. ESTROGEN OXIDATIVE METABOLISM A more specific scheme for the oxidative metabolism of E2 is shown in Fig. 3. The chemical structures of the estrogen cat- echols, semiquinones, quinones, and DNA adducts are presented in Chapter 4. The oxidative metabolites that have been shown to exhibit estrogenic and genotoxic effects include 160t-hydroxy- estrone [which will not be considered further in this monograph, but see (18—20)] and the 2-hydroxy- and 4—hydroxyE2/estrone (E1) catechols (7,8). The formation of the catechols is catalyzed by specific cytochrome P450 isoforms, including CYP1A1/1A2, CYPlBl, and CYP3A4 (7,8,21—24). The tissue specificity of estrogen metabolism results from the tissue-specific basal ex- pression and inducibility of these enzymes and from differences among the P450 isoforms in their kinetic parameters for E2 (23). This will be discussed in detail for CYP1A1 and CYPlBl in Chapter 5 of this monograph, along with a discussion of the potential role for estrogen overproduction by aromatase in breast carcinogenesis. ESTROGEN CARCINOGENESIS IN THE SYRIAN GOLDEN HAMSTER KIDNEY Extensive evidence for a role for catechol estrogen (CE) me- tabolites in estrogen carcinogenesis has come from work done with the use of the male Syrian golden hamster kidney carcino- genesis model, principally in the laboratories of Liehr et a1. (25) and Li and Li (26). The data in Table 1 demonstrate that E2, E1, and their 4-hydroxy CE (4-OHE2 and 4-OHE1), but not their 2-OH catechols, are carcinogenic in this model. Additional sup— port for a role for the CEs was provided by the finding that quercetin, which is both a competitive and a noncompetitive inhibitor of the phase II enzyme catechol—0—methyltransferase (COMT), increased the number of large renal tumors and the incidence of abdominal metastases in E2-treated hamsters (27,28). Much additional data support a role for CE metabolites in estrogen carcinogenesis in this model. For example, hamster kidney microsomes have been shown to biotransform estrogens to their 2-OHE2 and 4-OHE2 metabolites (29,30), whereas in vivo treatment caused the appearance of DNA strand breaks (3]) and increased the levels of 8-hydroxy-deoxyguanosine (8— OHdG) (32). Han and Liehr (33) also reported that, using ham- i Fig. 2. Estrogens as carcinogens. ER = estrogen receptor; Ox = oxidative. ER—mediated Signal Transduction Non ER-mediated Signal Transduction 68 Journal of the National Cancer Institute Monographs No. 27, 2000 Estradiol (E2) Fig. 3. Oxidative metabolism of es— trogens. 2-OHCE and 4—OHCE = 2-hydroxy and 4-hydroxy catechol ‘3 , Cu(II)-> Cu(I) ' l ' MT = ‘ . . . . esmgens’ reSpemve y. CO COMT, sulfotransferases, \ Llpld Peroxrdation catechol—0—methyltransferase; GSH . UDP-glucuronosyltransferases ‘. = glutathione. ..--V 16a-OHE1 Estrone (E1) DNA Damage - Adducts - Oxidative via: ' Redox cycling via: . 0 P450 ox/red Glutathione S- Conjugates Glucuronides, Sulfates 2- & 4-Methoxyestrogens, Transferases A‘ GSH Conjugates Table 1. Syrian golden hamster renal carcinogenicity of catechol estrogens % animals with renal carcinomas Estrogen* Liehr et a1. (25) Li and Li (26) None 0 0 E2 80 100 4—OHE2 80 100 2—OHE2 0 0 El — 80 4—OHEl — 33 2-OHEl — 0 ”(E2 = 17B-estradiol; 4—OHE2 = 4—hydroxyE2; 2-OHE2 = 2-hydroxyE2; El = estrone; 4—OHE1 = 4-hydroxyE1; 2-OHEl = 2—hydroxyE,. ster liver microsomes 4-OHE2 but not 2-OHE2 caused increased 8-OHdG levels. Similarly, Liehr and co-workers (34—36) dem- onstrated formation of 4-OHE2 by microsomes from normal and tumor tissues of human breast, uterus, cervix, and ovary, whereas other investigators (37,38) have detected increased lev— els of 8—OHdG in human breast tumor tissue. These associations support the hypothesis that ROS generated through redox cy— cling processes involving CE metabolites may contribute to es- trogen carcinogenesis of the human breast and perhaps other tissues. ESTROGEN CARCINOGENESIS IN RAT LIVER A role for oxidative DNA damage originating from estrogen metabolism is also supported by results from studies on the mechanisms Of 2,3,7.8-tetrachlor0dibenzo—p-dioxin (TCDD) and ethinyl estradiol (EE) carcinogenesis in rat liver. TCDD is a potent carcinogen in certain laboratory animals (39). Tritscher et a]. (40) reported detecting higher levels Of 8—OHdG in nuclear DNA from livers of TCDD-treated intact rats than in nuclear DNA from livers of TCDD-treated ovariectomized rats. Intact female rats are more sensitive tO TCDD—induced hepatocarcino- genesis than are ovariectomized female rats or male rats (39,40). Since TCDD is a potent inducer of the P450s involved in the oxidative metabolism of E2, these results are consistent with a Journal Of the National Cancer Institute Monographs No. 27, 2000 contribution of endogenous estrogens and increased oxidative DNA damage arising from estrogen metabolites in TCDD car- cinogenesis in female rat liver, although other unknown mecha- nisms could be contributing to or could be responsible for this carcinogenic process in this experimental model. Prolonged exposure of women to BE in the form of oral contraceptives has been associated with increased risk for de— veloping hepatic tumors (41). A number of laboratories have been involved in studies on the mechanisms of EB hepatocar— cinogenesis. The effects of EB on rat liver are summarized in Fig. 4 (7). EE is a strong promoter of hepatocarcinogenesis initiated by diethylnitrosamine, enhancing the development Of altered hepatic foci, nodules, and carcinomas ( 7). In non— nitrosamine-initiated rats, EE alone is a weak carcinogen. As- sociated with this carcinogenic process, EE causes a transient increase in DNA synthesis, followed by growth inhibition, which provides a period of negative selective pressure during which resistant hepatocytes (initiated) begin clonal expansion (promotion) (42 ). In addition, increased oxidative DNA damage occurs during EE treatment, as shown by data (Table 2) com- piled from those presented in a report by Ogawa et al. (43). In that study, female Wistar rats were treated with EB at the doses shown, which were sufficient to cause the development of he- patocellular carcinomas after 12 months. These results Show an association between increased 8-OHdG and increased incidence of carcinomas. Furthermore, simultaneous treatment with each of three antioxidant vitamins inhibited tumor development and reduced the levels of oxidative DNA damage. SUPPORT FROM MOLECULAR EPIDEMIOLOGIC STUDIES FOR A ROLE OF ENDOGENOUS CE METABOLITES IN HUMAN BREAST CANCER DEVELOPMENT Evidence from model experimental systems mentioned above and described in the following chapters supports a role for CE metabolites in estrogen carcinogenesis. In humans, however, the growing body of available evidence is indirect and, thus, is only suggestive. As mentioned above, Liehr and Ricci (35) found 4—hydroxylase activity in normal and tumor breast tissues, Sutter 69 DNA Synthesis (Estrogen Receptor Mediated) Fig. 4. Effects of ethinyl estradiol on rat liver. Adapted from Fig. 3 in (7). Reprinted with per- mission from the Annual Review of Pharmacology and Toxicology, Vol. 36, / ©1996 by Annual Re— views www.AnnualRe- views.org. : Initiation I? ' Estrogen Exposure Oxidative DNA DaKiage & Adducts _ Ersiawntflypgpleaicflate. (Caused by Additive Growth at Low Dose) Cancer Nodules Foci Persistent DN Synthesis (Caused by Cyt toxicity at high dose) Mitosuppression > (Negative Selective Pressure) I A A E A p Estrogen Metabolism Adapted from Fig. 3 in Yager and Liehr (7) Table 2. Enhanced oxidative DNA damage in liver nuclear DNA from ethinyl estradiol (EE)—treated rats* 8-oxodG/106 dG, 1 mo, mean : Hepatocellular carcinoma incidence, 12 mo, % Treatment standard deviation (No/total No.) Control 3.5 i 0.7 0 (0/24) EB, 75 rig/day 7.1 i 0.91L 8.7 (2/23)T BB, 750 rig/day 8.4 i 0.7T 38.5 (lO/26)T EB, 75 rig/day + vitamin C, 6.0 + 1.7 0 (0/19) 1g/kg diet Vitamin E, 500 mg/kg diet 5.4 i 1.9 4.5 (1/22)’; B-Carotene, 5.5 i 1 8 4.8 (1/21):l: 250 mg/kg diet *Adapted from Ogawa et a1. (43). oxodG = oxodeoxyguanosine; dG = deoxyguanosine. TP<.05 versus control. iP<.05 versus EE alone. (see Chapter 5) has observed expression of a 4—hydroxylase, CYPlBl, in human breast tissue, and Malins (37); Chapter 9) has detected increased oxidative DNA damage in breast tumor tissue. In addition, the results from some molecular epidemiol— ogy studies also provide support for a contribution of CE me~ tabolites to the development of breast cancer. A number of stud- ies, guided by the hypothesis that increased breast cancer risk is associated with exposures to certain environmental chemicals, have examined the association between risk and genetic poly- morphisms in several genes encoding biotransformation en- zymes. These include genes encoding CYP1A1 (44—46), N- acetyltransferase 2 (47), and glutathione-S-transferase (GST) isoforms Ml (null), Tl (null), and P1 (low—activity allele) (45,46,48). The results have been mixed, depending on the sub- ject cohort. COMT is a gene involved in the phase II metabolism 70 of catechols, such as catecholamines and flavanoids. However, COMT also catalyzes the O-methylation of both 2-OH- and 4-OH-catechols formed from the oxidative metabolism of en— dogenous E2 and E1. O-Methylation of CEs inactivates their estrogenic potential and blocks their ability to undergo further oxidation to more reactive semiquinone and quinone metabolites that can directly adduct DNA and/or participate in redox cycling to produce superoxide, as described above. COMT is polymor- phic, and 25% of Caucasians are homozygous for an allele en- coding a low-activity form of the enzyme. The ability of CE metabolites to contribute to estrogen carcinogenesis, suggested by the experimental studies mentioned above and in Chapters 4 and 5 of this monograph, led to the hypothesis that women homozygous for the low-activity COMT allele would be at in- creased risk for breast cancer. At the time of this conference, two studies, one published (49) and another one that was in press but is now published (50), presented evidence that the gene encod- ing a low-activity form of COMT was associated with an in— creased risk for developing breast cancer in certain women. These data will be discussed in greater detail in Chapter 7. Briefly, in a prospective, nested, case—control study from a large western Maryland cohort, Lavigne et a1. (49) found that the risk for developing postmenopausal breast cancer associated with the low-activity COMT allele was increased in postmeno- pausal women Who were also either heavy (body mass index >24.27 kg/mz) or GSTM1 null or GSTPl low activity. In a hospital-based case—control study of subjects from western New York, Thompson et a1. (50) found an increased breast cancer risk associated with the low-activity COMT allele, but only in pre- menopausal women, and they found that the risk was increased further in those women who had increased body mass indices (>23 kg/mz). Two additional studies have been published. In one study (51), no increase in risk was detected; in contrast, in the Journal of the National Cancer Institute Monographs No. 27, 2000 other study (52 ), an increased breast cancer risk associated with the low-activity COMT allele was found in postmenopausal women. Why might there be such differences in the risk conferred by the same genetic polymorphism in different cohorts? One pos— sibility is that these studies simply detected random findings. At this point, this possibility cannot be ruled out, for the number of cases and controls in these studies was small. On the other hand, the studies were hypothesis driven, and an increased risk asso— ciated with the low-activity allele has now been detected in three of four studies, one in premenopausal and two in postmeno- pausal women. Another possibility to account for the differences among the various cohorts may relate to the strength of the effect. According to a classification by Rebbeck et al. (53), mutations (detected as genetic polymorphisms) in some genes, such as BRCAl, confer a high risk for an individual. However, because these allele frequencies are low, the overall attributable risk that they represent is small. On the other hand, mutations in phase I and II enzyme genes involved in xenobiotic (but also endobiotic, i.e., endogenous molecules) metabolism might con- fer a low relative cancer risk for an individual. But, because these mutations seem to be common among individuals, they represent a high attributable risk category of genes. In addition, the specificities of many of these enzymes are overlapping, their activities within the cells can be altered by environmental agents, their polymorphic allele frequencies within a population can differ depending on ethnicity (54), and they function in somewhat redundant pathways. Thus, the balance in the cell of the metabolite products of these enzymes could be differentially affected by interactions among various genetic and environmen- tal factors, as illustrated in Fig. 5. The intent of this scheme is to show that both genetic and environmental/lifestyle factors can act either as susceptibility or as protective influences with regard to risk of developing a particular disease. With regard to the oxidative metabolism and conjugation of estrogens, several en- zymes are involved, including specific cytochrome P450 iso- forms, sulfotransferases, COMT, and GST (for the products of oxidative damage). The levels of these enzymes in a given in- dividual or population may be influenced (induced or inhibited) by xenobiotic exposures. In addition, these enzymes are poly- morphic, and the distribution of polymorphisms may vary among different populations as a result of ethnic compositions. Since the oxidative metabolites of estrogen appear to contribute to and protect from disease, it is, therefore, not surprising to detect differences in single-gene genetic polymorphism/disease associations among different study populations. In summary, most of the data from the studies cited above and to be discussed in the following chapters that implicate metabolites of endogenous estrogens in estrogen carcinogenesis are from in vitro studies and in vivo studies in which rodent models were used. The results have provided important insight that has led to the development of a new paradigm (see Fig. l) for the contribution of CE metabolites to estrogen carcinogene- sis. It predicts that the level of and balance among the parent hormone and the catechol metabolites should be determined by the level of expression and activity of certain key enzymes in- cluding aromatase (Chapter 5), several cytochrome P4505, par- ticularly CYPlBl (Chapter 5), and various protective enzymes, such as COMT (Chapters 6 and 7). Expression of these enzymes is likely to be tissue specific as well as developmental stage specific and to be affected by both environmental and endog- enous factors. These parameters need to be thoroughly defined. Expression levels should be determined, along with the levels of the estrogen metabolites and selected biologic end points for their potential effects, e.g., gene expression, DNA damage, and mutagenesis. For several of these enzymes, genetic polymor— phisms that affect activity have been discovered; e.g., the COMT polymorphism decreases enzyme activity. The effects of these polymorphisms on estrogen metabolite levels and on the bio- logic end points also require thorough investigation. However, an important question pertains to the appropriate experimental model systems to use. One can envision using cultured human cells originating from various human tissues, e.g., breast, pros- tate, and ovary, along with cells genetically engineered to ex- press these enzymes or their polymorphic forms alone and in combination. One can also envision using knockout mice to determine the effects of the absence of particular genes, e.g., aromatase, CYPlBl, and COMT, on tissue estrogen metabolite levels and on the biologic end points including cancer. From the knockout mice, it should be possible to create transgenics ex- pressing the human genes. By use of bacterial artificial chromo- some (bac) vectors that accept up to 300-kilobase inserts, it is Genetic & EnvironmentallLife-Styl Susceptibility Factors Fig. 5. Genetic, environmental, and lifestyle factors affect risk. RISK Genetic & Environmental/Life-Style Protective Factors Journal of the National Cancer Institute Monographs No. 27, 2000 71 possible to include most of the regulatory regions of these genes and, thus, perhaps achieve their tissue-specific and developmen- tal stage-specific expression. Results from studies such as these should more directly test the paradigm for the role of the CE metabolites. However, data obtained from human tissues will ultimately be required to be certain that the conclusions drawn from the model systems apply. Thus, efforts should be devoted to obtaining normal and tumor human tissues, e.g., breast and prostate, for simultaneous analysis of estrogen metabolites and the relevant biologic end points, e.g., DNA damage, CYPlBl genotype and expression, and COMT genotype and expression. It will only be through such future in—depth mechanistic studies that we will be able to dissect the genetic and environmental “factors” affecting the pathways that provide adverse or protec- tive “states,” characterized by a correct or incorrect balance of oxidative metabolites impacting cancer risk. Identification of these factors should also allow more focused, mechanistically based molecular epidemiologic studies to be conducted in the future. REFERENCES (1) Harris JR, Lippman ME, Veronesi U, Willett W. Breast cancer (1). N Engl J Med 1992;327:319—28. (2) Madigan MP, Ziegler RG, Benichou J, Byrne C, Hoover RN. Proportion of breast cancer cases in the United States explained by well—established risk factors. J Natl Cancer Inst 1995;87:1681—5. (3) Henderson BE, Pike MC, Bernstein L, Ross RK. Breast Cancer. In: Schot— tenfeld D, Fraumeni JF Jr, editors. Cancer epidemiology and prevention. New York (NY): Oxford University Press; 1996. p. 1022—39. Feigelson HS, Henderson BE. Estrogens and breast cancer. Carcinogenesis 1996; 17 :2279—84. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,41 1 women without breast cancer [published erratum appears in Lancet 1997; 350:1484]. Lancet 1997;347:1047—59. Wolff MS, Collman GW, Barrett JC, Huff J. Breast cancer and environ— mental risk factors: epidemiological and experimental findings. Ann Rev Pharmacol Toxicol 1996;36:573-96. (7) Yager JD, Liehr JG. Molecular mechanisms of estrogen carcinogenesis. Ann Rev Pharmacol Toxicol 1996;36:203—32. (8) Zhu BT, Conney AH. Functional role of estrogen metabolism in target cells: review and perspectives. Carcinogenesis 1998:1921—27. (9) Levin M, Weisz J, Bui QD, Santen RJ. Peroxidatic catecholestrogen pro— duction by human breast cancer tissue in vitro. J Steroid Biochem 1987; 28:513—20. (10) Sipe HJ Jr, Jordan SJ, Hanna PM, Mason RP. The metabolism of 1713— estradiol by lactoperoxidase: a possible source of oxidative stress in breast cancer. Carcinogenesis 1994;15:2637—43. (11) Schutze N, Vollmer G, Tiemann I, Geiger M, Knuppen R. Catecholestro- gens are MCF-7 cell estrogen receptor agonists. J Steroid Biochem Mol Biol 1993;46:781—9. (12) Fotsis T, Zhang Y, Pepper MS, Adlercreutz H, Montesano R, Nawroth PP, et al. The endogenous oestrogen metabolite 2—methoxyestradiol inhibits angiogenesis and suppresses tumour growth. Nature 1994:3681 237—9. (13) Yue TL, Wang X, Louden CS, Gupta S, Pillarisetti K, Gu JL, et al. 2—Me— thoxyestradiol, an endogenous estrogen metabolite, induces apoptosis in endothelial cells and inhibits angiogenesis: possible role for stress-activated protein kinase signaling pathway and Fas expression. Mol Pharmacol 1997; 51:951—62. (14) Zhu BT, Conney AH. Is 2—methoxyestradiol an endogenous estrogen me— tabolite that inhibits mammary carcinogenesis? Cancer Res 1998;58: 2269—77. (15) Cavalieri EL, Stack DE, Devanesan PD, Todorovic R, Dwivedy I, Higgin- botham S, et al. Molecular origin of cancer: catechol estrogen-3,4-quinones (4 \_, (5 \— (6 \. 72 as endogenous tumor initiators. Proc Natl Acad Sci USA 1997;94: 10937—42. (16) Wang MY, Liehr JG. Induction by estrogens of lipid peroxidation and lipid peroxide-derived malonaldehyde—DNA adducts in male Syrian hamsters: role of lipid peroxidation in estrogen—induced kidney carcinogenesis. Car- cinogenesis 1995;16:1941—5. ( I 7) Li Y, Zhu H, Stansbury KH, Tmsh MA Role of reactive oxygen species in multistage carcinogenesis. In: Thomas CE, Kalyanaraman B, editors. Oxy- gen radicals and the disease process. Amsterdam (The Netherlands): Har- wood Academic Publishers; 1997. p. 237—77. (18) Swaneck GE, Fishman J. Covalent binding of the endogenous estrogen 160L-hydr0xyestrone to estradiol receptor in human breast cancer cells: characterization and intranuclear localization. Proc Natl Acad Sci U S A 1988;85:7831—5. (19) Davis DL, Telang NT, Osborne MP, Bradlow HL. Medical hypothesis: bifunctional genetic—hormonal pathways to breast cancer. Environ Health Perspect 1997;105(Supp1 3):571—6. (20) Telang NT, Suto A, Bradlow HL, Wong GY, Osborne MP. Genotoxic damage and aberrant proliferation in mouse mammary epithelial cells. Re- cent Prog Horm Res 1993;48:481—8. (2]) Brian WR, Sari MA, Iwasaki M, Shimada T, Kaminsky LS, Guengerich FP. Catalytic activities of human liver cytochrome P-450 IIIA4 expressed in Saccharomyces cerevisiae. Biochemistry 1990;29:11280—92. (22) Stresser DM, Kupfer D. Catalytic characteristics of CYP3A4: requirement for a phenolic function in ortho hydroxylation of estradiol and mono—0— demethylated methoxychlor. Biochemistry 1997;36:2203—10. (23) Hayes CL, Spink DC, Spink BC, Cao JQ, Walker NJ, Sutter TR. 178- estradiol hydroxylation catalyzed by human cytochrome P450 1B1. Proc Natl Acad Sci U S A 1996;93:9776—81. (24) Martucci CP, Fishman J. P450 enzymes of estrogen metabolism. Pharma- col Ther 1993;57:237—57. (25) Liehr JG, Fang WF, Sirbasku DA, Ari-Ulubelen A. Carcinogenicity of catechol estrogens in Syrian hamsters. J Steroid Biochem 1986;24:353—6. (26) Li JJ, Li SA. Estrogen carcinogenesis in Syrian hamster tissues: role of metabolism. Fed Proc 1987;46:1858—63. (27) Zhu BT, Liehr JG. Inhibition of catechol O-methyltransferase-catalyzed O-methylation of 2- and 4-hydroxyestradiol by quercetin. Possible role in estradiol-induced tumorigenesis. J Biol Chem 1996;271:1357—63. (28) Zhu BT, Liehr JG. Quercetin increases the severity of estradiol-induced tumorigenesis in hamster kidney. Toxicol Appl Pharmacol 1994;125:149—58. (29) Li SA, Klicka JK, Li J]. Estrogen 2- and 4-hydr0xylase activity, catechol estrogen formation, and implications for estrogen carcinogenesis in the hamster kidney. Cancer Res 1985;45:181—5. (30) Zhu BT, Bui QD, Weisz J, Liehr JG. Conversion of estrone to 2- and 4-hydroxyestrone by hamster kidney and liver microsomes: implications for the mechanism of estrogen—induced carcinogenesis. Endocrinology 1994;135:1772—9. (3]) Han X, Liehr JG. DNA single-strand breaks in kidneys of Syrian hamsters treated with steroidal estrogens: hormone—induced free radical damage pre- ceding renal malignancy. Carcinogenesis 1994;15:997—1000. (32) Han X, Liehr JG. 8-Hydroxy1ation of guanine bases in kidney and liver DNA of hamsters treated with estradiol: role of free radicals in estrogen— induced carcinogenesis. Cancer Res 1994;54:5515—7. (33) Han X, Liehr JG. Microsome-mediated 8—hydroxylation of guanine bases of DNA by steroid estrogens: correlation of DNA damage by free radicals with metabolic activation to quinones. Carcinogenesis 1995;16:2571—4. (34) Liehr JG, Ricci MJ, Jefcoate CR, Hannigan EV, Hokanson JA, Zhu BT. 4-Hydroxy1ation of estradiol by human uterine myometrium and myoma microsomes: implications for the mechanism of uterine tumorigenesis. Proc Natl Acad Sci U S A 1995;92:9220—4. (35) Liehr JG, Ricci MJ. 4-Hydroxylation of estrogens as marker of human mammary tumors. Proc Natl Acad Sci U S A 1996;93:3294—6. (36) Somasunderam A, Liehr JG. 4-Hydroxy1ation of estradiol by microsomes of human neoplastic and normal ovarian and cervical tissues [abstract]. Proc Am Assoc Cancer Res 1997;38:451—2. (37) Malins DC, Haimanot R. Major alterations in nucleotide structure of DNA in cancer of the female breast. Cancer Res 1991;51:5430—2. (38) Musarrat J, Arezina—Wilson J, Wani AA. Prognostic and aetiological rel— evance of 8-hydroxyguanosine in human breast carcinogenesis. Eur J Can— cer 1996;32Azl209—14. Journal of the National Cancer Institute Monographs No. 27, 2000 (39) Lucier GW, Tritscher A, Goldsworthy T, Foley J, Clark G, Goldstein J, et al. Ovarian hormones enhance 2,3,7,8-tetrachlorodibenzo—p—dioxin- mediated increases in cell proliferation and preneoplastic foci in a two- stage model for rat hepatocarcinogenesis. Cancer Res 1991;51:1391—7. (40) Tritscher AM, Seacat AM, Yager JD, Groopman JD, Miller BD, Bell D, et a1. Increased oxidative DNA damage in livers of 2,3,7,8-tetrachloro- dibenzo-p—dioxin treated intact but not ovariectomized rats. Cancer Lett 1996;98:219—25. (41) Palmer JR, Rosenberg L, Kaufman DW, Warshauer ME, Stolley P, Shapiro S. Oral contraceptive use and liver cancer. Am J Epidemiol 1989;130: 878—82. (42) Yager JD, Zurlo J, Sewall CH, Lucier GW, He H. Growth stimulation followed by growth inhibition in livers of female rats treated with ethinyl estradiol. Carcinogenesis 1994;15:2117—23. (43) Ogawa T, Higashi S, Kawarada Y, Mizumoto R. Role of reactive oxygen in synthetic estrogen induction of hepatocellular carcinomas in rats and preventive effect of vitamins. Carcinogenesis 1995;16:831—6. (44) Taioli E, Trachman J, Chen X, Toniolo P, Garte SJ. A CYP1A1 restriction fragment length polymorphism is associated with breast cancer in African— American women. Cancer Res 1995;55:3757—8. (45) Bailey LR, Roodi N, Verrier CS, Yee CJ, Dupont WD, Parl FF. Breast cancer and CYP1A1, GSTMl, and GSTTl polymorphisms: evidence of a lack of association in Caucasians and African Americans. Cancer Res 1998; 58:65—70. (46) Ambrosone CB, Freudenheim JL, Graham S, Marshall JR, Vena JE, Bra— sure JR, et al. Cytochrome P4501A1 and glutathione S-transferase (M1) genetic polymorphisms and postmenopausal breast cancer risk. Cancer Res 1995;55:3483—5. (47) Ambrosone CB, Freudenheim JL, Graham S, Marshall JR, Vena JE, Bra- sure JR, et a1. Cigarette smoking, N—acetyltransferase 2 genetic polymor- phisms, and breast cancer risk. JAMA 1996;276:1494—501. (48) Helzlsouer KJ, Selmin 0, Huang HY, Strickland PT, Hoffman S, Alberg AJ, et a1. Association between glutathione S-transferase M1, P1, and T1 Journal of the National Cancer Institute Monographs No. 27, 2000 genetic polymorphisms and development of breast cancer. J Natl Cancer Inst 1998;90:512—8. (49) Lavigne JA, Helzlsouer KJ, Huang HY, Strickland PT, Bell DA, Selmin O, et al. An association between the allele coding for a low activity variant of catechol-0-methyltransferase and the risk for breast cancer. Cancer Res 1997;57:5493—7. (50) Thompson PA, Shields PG, Freudenheim JL, Stone A, Vena JE, Marshall JR, et al. Genetic polymorphisms in catechol-0—methyltransferase, meno- pausal status, and breast cancer risk. Cancer Res 1998;58:2107—10. (51) Millikan RC, Pittman GS, Tse CK, Duell E, Newman B, Savita D, et al. Catechol-0-methyltransferase and breast cancer risk. Carcinogenesis 1998; 19:1943—7. (52) Huang CS, Chem HD, Chang K], Cheng CW, Hsu SM, Shen CY. Breast cancer risk associated with genotype polymorphism of the estrogen- metabolizing genes CYP17, CYP1A1, and COMT: a multigenic study on cancer susceptibility. Cancer Res 1999;59:4870—5. (53) Rebbeck TR, Walker AH, Phelan CM, Godwin AK, Buetow KH, Garber JE, et a1. Defining etiologic heterogeneity in breast cancer using genetic biomarkers. Prog Clin Biol Res 1997;396:53—61. (54) Garte S. The role of ethnicity in cancer susceptibility gene polymorphisms: the example of CYP1A1. Carcinogenesis 1998;19:1329—32. NOTES Supported by Public Health Service (PHS) grants CA36701 and CA77550 (National Cancer Institute) and PHS grant ESO3819 (National Institute of En- vironmental Health Sciences), National Institutes of Health, Department of Health and Human Services. Several predoctoral and postdoctoral trainees have received support from PHS training grant E50714] from the National Institute of Environmental Health Sciences. I thank all of the students and postdoctoral fellows who have conducted the research in my laboratory that is referred to in this chapter. I am also grateful to Dr. Jackie A. Lavigne for her critical comments on this chapter. 73 . e i. A, 3.; .2... .n {1. L. lite; f ;. Chapter 4: Estrogens as Endogenous Genotoxic Agents—DNA Adducts and Mutations Ercole Cavalieri, Krystyna Frenkel, Joachim G. Liehr, Eleanor Rogan, Deodutta Roy Estrogens induce tumors in laboratory animals and have been associated with breast and uterine cancers in humans. In relation to the role of estrogens in the induction of cancer, we examine formation of DNA adducts by reactive electro- philic estrogen metabolites, formation of reactive oxygen species by estrogens and the resulting indirect DNA damage by these oxidants, and, finally, genomic and gene mutations induced by estrogens. Quinone intermediates derived by oxi- dation of the catechol estrogens 4-hydroxyestradiol or 4-hy- droxyestrone may react with purine bases of DNA to form depurinating adducts that generate highly mutagenic apu- rinic sites. In contrast, quinones of 2-hydroxylated estrogens produce less harmful, stable DNA adducts. The catechol es- trogen metabolites may also generate potentially mutagenic oxygen radicals by metabolic redox cycling or other mecha- nisms. Several types of indirect DNA damage are caused by estrogen-induced oxidants, such as oxidized DNA bases, DNA strand breakage, and adduct formation by reactive aldehydes derived from lipid hydroperoxides. Estradiol and the synthetic estrogen diethylstilbestrol also induce numeri- cal and structural chromosomal aberrations and several types of gene mutations in cells in culture and in vivo. In conclusion, estrogens, including the natural hormones estra- diol and estrone, must be considered genotoxic carcinogens on the basis of the evidence outlined in this chapter. [J Natl Cancer Inst Monogr 2000;27:75—93] Estrogens, including the natural hormones estradiol (E2) and estrone (E1), induce tumors in various organs of several labora- tory animal species and strains [reviewed in (1,2)]. In humans, exogenous estrogen-containing medications or elevated concen- trations of circulating endogenous estrogens increase the risk of uterine and mammary cancers [reviewed in (1,2)]. Nevertheless, synthetic or steroidal estrogens or their metabolites failed to induce gene mutations in several classical bacterial and mam- malian gene mutation assays (3—7) and were, therefore, classi- fied as epigenetic carcinogens ( 8,9). Two possible mechanisms of tumor induction by estrogens were subsequently advanced. Estrogen—induced cell transformation and tumor development were proposed to be mediated by l) estrogen receptor—based proliferation of cells carrying spontaneous replication errors [(8,10); Chapter 8] and 2) disruption of spindle formation and subsequent numeric chromosomal changes (9). An increasing body of experimental evidence stands, how- ever, in contradiction to these two hypotheses of hormonal tu- morigenesis, including the following data: 1) In human mam- mary epithelial cells, estrogen receptors are expressed in cells different and distinct from proliferating cells carrying prolifera- tion markers (11,12). 2) Aneuploidy and other karyotypic changes were detected in Syrian hamster embryo cells predis- posed to immortalization and progression to tumorigenicity; Journal of the National Cancer Institute Monographs No. 27, 2000 nude mice inoculated with cells carrying such chromosomal a1- terations, however, did not produce tumors (13). Therefore, ad- ditional genetic changes (mutations) were postulated to be re- quired for tumor induction (13). 3) Compared with hamsters treated only with E2, tumor formation is decreased in animals exposed to E2 plus inhibitors of estrogen metabolism (14,15) or to hormonally potent estrogens with poor metabolic conversion to catechol metabolites (16,17). These data support a tumor- initiating role for catechol estrogens (CE). 4) A large body of evidence is accumulating that estrogens induce various types of DNA damage in vitro and in vivo. As outlined below, CE can, indeed, mediate this damage. 5) The classification of estrogens as epigenetic carcinogens is contradicted by preliminary evi- dence of estrogen-induced gene mutations (reviewed below). In this chapter, we discuss the induction of DNA damage and gene mutations. First, we focus on the direct adduction of es- trogen metabolites to DNA in vitro and in vivo, second on the generation of reactive oxygen species (ROS) by estrogen me- tabolites and various types of DNA damage induced indirectly by estrogens and, finally, on estrogen-induced gene mutations. ONCOGENIC MUTATIONS BY DEPURINATING CARCINOGEN—DNA ADDUCTS AS A MODEL OF ESTROGEN-INDUCED MUTATIONS The origin of cancer represents one of the most intriguing scientific mysteries. Cancer is a disease of mutated critical regu- latory genes and abnormal cell proliferation (18). Understanding the origin of these mutations opens the door to strategies for controlling and preventing cancer. One possible approach to investigating the origin of cancer has been to gain a fundamental understanding of the properties of molecules that induce this disease. During the last 25 years, polycyclic aromatic hydrocar- bons (PAH) have been investigated by Cavalieri and Rogan (19,20) as model carcinogenic compounds. The purpose of studying these molecules has been threefold: First, they repre- sent a good model for understanding the mechanism of tumor initiation by chemicals; second, they have some geometric re- semblance to endogenous estrogens; and third, both PAH and estrogens contain aromatic rings. Afliliations of authors: E. Cavalieri, Eppley Institute, University of Nebraska Medical Center, Omaha, NE; K. Frenkel, New York University School of Medi- cine, NY; J. G. Liehr, Stehlin Foundation for Cancer Research, Houston, TX; E. G. Rogan, Eppley Institute, University of Nebraska Medical Center; D. Roy, University of Alabama at Birmingham, School of Public Health. Correspondence to: Ercole Cavalieri, D.Sc., Eppley Institute for Research in Cancer and Allied Diseases, 986805 Nebraska Medical Center, Omaha, NE 68198—6805 (e-mail: ecavalie@unmc.edu). See “Notes” following “References.” © Oxford University Press 75 Comprehensive studies of PAH have led to an understanding of their mechanism of tumor initiation (19,20). PAH are acti- vated by two main pathways: one-electron oxidation to produce reactive intermediate radical cations and monooxygenation to afford bay-region diol epoxides (19,20). The reactive interme— diates formed by these two mechanisms, radical cations and diol epoxides, can bind to DNA to produce adducts that initiate the process of tumor formation, as illustrated in Fig. 1 for dibenzo- [a,l]pyrene (DB [a, HP). DNA adducts are obtained by reaction of the metabolically activated PAH with the nucleophilic groups of the two purine bases, adenine (Ade) and guanine (Gua). These adducts can be either stable or depurinating. The stable adducts are those that remain covalently bonded to DNA unless removed during repair, whereas the depurinating adducts are the ones that are spontaneously released from DNA by destabilization of the glycosidic bond (Fig. 2). Stable DNA adducts are formed when PAH react with the exocyclic amino group of Ade or Gua, whereas depurinating adducts are obtained when PAH cova- lently bond at the N-3 or N-7 position of Ade or the N-7 or, sometimes, the 08 position of Gua. Among the various approaches to the study of carcinogenesis by PAH, identification and quantitation of their DNA adducts Procarcinogen P450 1 P450 or Peroxidases Metabolic Activation One-electron Oxudatlon P450 Ultimate Carcinogen (Electrophile) OH . . Dihydrodiol Epoxide Rad'ca' 03"“ Covalent Binding to DNA Initiation of Cancer Fig. 1. Metabolic activation of DB[a,l]P by the diol epoxide and radical cation pathways. 76 have been the most fruitful in unraveling the mechanism of tumor initiation by these compounds. Through comprehensive studies of the DNA adducts of the potent carcinogenic PAH, benzo[a]pyrene (BP), 7,l2-dimethylbenz[a]anthracene (DMBA), and DB[a,l]P (Fig. 3), Cavalieri and Rogan (20) and Chakravarti et a1. (21) have discovered that there is an associa- tion between depurinating adducts and oncogenic mutations, suggesting that these adducts are the primary culprits in the tumor initiation process. This discovery was made by identifying and quantifying the DNA adducts formed in mouse skin by BP, DMBA, and DB[a,l]P and, at the same time, determining the mutations in the Harvey (H)-ras oncogene in mouse skin papil- lomas initiated by these three PAH, as shown in Fig. 4 (21). When mouse skin was treated with DMBA, 79% of the adducts were depurinating Ade adducts and 20% were depurinating Gua adducts (20,22). For DB[a,l]P, 81% were depurinating Ade ad- ducts and 18% were depurinating Gua adducts (20). In contrast, mouse skin treated with BP produced 46% depurinating Gua adducts and 25% depurinating Ade adducts (20,23). Examina— tion of the ras oncogene mutations in papillomas induced by DMBA or DB[a,l]P demonstrates that, in both cases, an A ——> T transversion (CAA —> CTA) consistently occurs (Table 1; Fig. 5) (21). These mutations associate with the predominant forma- tion of depurinating Ade adducts by these two PAH. About twice as many of the papillomas induced with BP contain G —> T mutations at codon 13 in ras (GGC —> GTC) compared with the number of tumors with a codon 61 CAA ——> CTA mutation (21,24). The ratio of mutations is consistent with the profile of depurinating Gua and Ade adducts formed by BP in the target tissue (Table 1). This pattern of ras mutations suggests that the oncogenic mutations in mouse skin papillomas induced by these PAH are generated by misreplication or misrepair of the apurinic sites derived from loss of the depurinating adducts (21). For example, an A —> T transversion can be attributed to loss of a depurinating Ade adduct and generation of an apurinic site. If the apurinic site is not correctly repaired in the next round of DNA replication, the most likely base to be inserted opposite the apurinic site is Ade (Fig. 6). When the coding strand of the DNA is then replicated, a thymine is inserted opposite the new Ade, resulting in the A —> T mutation observed in codon 61 of the ras oncogene in tumors initiated by PAH forming predominantly depurinating Ade adducts. When a Gua adduct is lost by depu- rination, leaving an apurinic site in the DNA, the preferential insertion of Ade in the opposite DNA strand leads to a G —> T transversion at the site of the adduct. It is also possible that the ras mutations are generated by misrepair, rather than misreplication, of the apurinic sites. Strong evidence for misrepair is provided by the observation of codon 61 CAA —> CTA transversions in mouse skin DNA 1 day after treatment with DB[a,l]P (Table 2) (25), when the cells are unlikely to have divided. The A —> T transversions are present in 0.1% of the cells by day 1, increase to about 5% by day 3, and then decrease to background levels by day 9. Subsequently, the A —> T mutation is detected in increasing levels as papillomas begin to develop. Because thousands of apurinic sites are spontaneously formed per cell each day, repair of apurinic sites induced by PAH might be expected. The level of apurinic sites arising from treatment with PAH is, however, 15—120 times higher than those formed spontaneously, suggesting that this large increase in apurinic sites could overwhelm the capacity of the cell to repair them Journal of the National Cancer Institute Monographs No. 27, 2000 o -------- —N cl) \(N / [ fN o ' d7 N / N—H ------- N/ \ N / H20 H20 NH N=< / N\ N=< O Guanine Hhkhl-r ----- 0 dB 6 HN—R O=P—O — — I (N HN—H--"" \ Electrophile 0—: 0 CN9 NHz N / \ \ | / \ H20 N ------ H—N H20 (W N N a )_N\ :03/ N_/ / . ' O dR Fig. 2. Formation of stable and depu- Ademne Thymine O R rinating DNA adducts and generation FIT—09 O=l|=—Oe 0f apurinic Sltes‘ o -> Depurinating adduct ('3 :9 Stable adduct | ’é F: NHg H20 OH H20 @‘3 NfiN (9 \ + < K + H ‘— @‘D \N J O 0 H20 N | e Apurinic Site of DNA I e o=F|'—o 0=Fl’—o 0K at l __o_ before replication occurs (20,21). Furthermore, the apparent nonrepair 0f apurinic sites induced by treatment with PAH may also be due to the presence of stable adducts that could interfere with error—free repair of apurinic sites. Thus, apurinic sites can generate the mutations that play the critical role in the initiation of cancer, and formation of depurinating adducts has become the common denominator for recognizing the potential of a chemi- cal to initiate cancer. The evidence that depurinating PAH—DNA adducts play a major role in tumor initiation has provided the impetus for dis- covering the estrogen metabolites that form depurinating DNA adducts and can be potential endogenous initiators of many hu- man cancers (26,27). Dibenzo[a,l]pyrene CH8 0 CH3 7,12-dimethylbenz[a]anthracene Benzo[a]pyrene Fig. 3. Structures of three potent carcinogenic polycyclic aromatic hydrocarbon. Journal of the National Cancer Institute Monographs No. 27, 2000 CATECHOL ESTROGEN-3,4-QUINONES AND APURINIC SITES IN CANCER INITIATION CEs are among the major metabolites of E1 and E2, as dis— cussed in Chapter 5. If these metabolites are oxidized to catechol estrogen quinones (CE-Q), they may react with DNA to form depurinating adducts. It is hypothesized that these adducts gen- erate apurinic sites leading to mutations, which may initiate breast, prostate, and possibly other human cancers. The estro- gens E1 and E2 are biochemically interconvertible by the enzyme 17 B-estradiol dehydrogenase. These two estrogens are metabo- lized via two major pathways: formation of CE (Fig. 7) and, to a lesser extent, 160L-hydroxylation (not shown). The catechols formed are the 2-hydroxylated and 4-hydroxylated estrogens (28,29). Generally, these two CEs can be inactivated by 0- methylation catalyzed by catechol-0-methyltransferases (COMT) (28). Other possible conjugations of CE, such as gluc- uronidation and sulfation (not shown), may also play a role in inactivation/protection (Chapter 6). If formation of the 4-hy- droxylated metabolites is excessive (see below) and/or produc— tion of these methyl, glucuronide, or sulfate conjugates is insuf— ficient and, thus, the cells are not totally protected from CE toxicity, competitive catalytic oxidation to semiquinones (CE- SQ) and CE-Q can occur. CE-SQ and CE-Q may conjugate with glutathione (GSH), catalyzed by S-transferases. If this inactivat- ing process is incomplete, CE-Q may react with DNA to form stable and depurinating adducts (26,27). To determine the possible genotoxic effects of CE-Q, they were reacted with the nucleosides 2’-deoxyguanosine (dG) and 2’-deoxyadenosine (dA) and the nucleobase Ade (26,30). An acetonitrile solution of E1 (E2)—3,4-Q was mixed with dG, dis— solved in acetic acid/water (l : 1) (Fig. 8). The adduct 4-OHE1(E2)-l(or,B)-N7Gua was formed during 5 hours at room 77 After 4h, kill the mouse and excise treated area promote twice weekly with phorbol ester _Pmnpflmuon|m Mouse skin of skin e - . pidermls Extract\‘ Afier 7days, Mouse skin papillomas 1. Harvest Purify DNA Stable adducts analyzed / by 32P—postlabeling Depurinating adducts (N3 Ade, N7Ade, N7Gua, C8Gua) analyzed by HPLC with fluorescence detector with solvents c-Harvey-ras mutations G —'T in codon 13 papillomas A —~T in codon 61 2. purify DNA 3. PCR amplify the exon 1&2 region of c-Harvey-ras 4. analyze mutations by base sequencing Fig. 4. Determination of DNA adducts and Harvey-ras mutations in mouse skin. Table 1. Correlation of depurinating adducts with H-ras mutations in mouse skin papillomas* H-ras mutations Major DNA No. of mutations/ PAH adducts (%) No. of mice Codon DMBA N7Ade (79) 4/4 CAA —> CTA 61 DB[a,l]P N7Ade (32) 4/5 CAA ~> CTA 61 N3Ade (49) BP C8Gua + N7Gua (46) 10/20 GGC —> GTC l3 N7Ade (25) 5/20 CAA -—> CTA 61 *PAH = polycyclic aromatic hydrocarbons; DMBA = 7,12-dimethyl- benz[a]anthracene; DB[a,l]P = dibenzo[a,l]pyrene; and BP = benzo[a]pyrene. temperature (26), and it is a mixture of two conformational isomers resulting from the restricted rotation of the Gua moiety around the N7(Gua)—C1(estrogen) bond. The reaction of the CE- 3,4-Q with dG at the N-7 position destabilizes the glycosidic bond and results in loss of the deoxyribose moiety. When the adduct is formed by reaction of CE-3,4—Q with DNA, it is re— leased from the DNA by spontaneous depurination. Reaction of E1(E2)-3,4-Q with dA produced no adducts; however, reaction of E1(E2)-3,4-Q with Ade resulted in the formation of 4-OHE1(E2)-l(a,B)-N3Ade (Fig. 8) (30). This adduct was ob- tained only with Ade because in dA the adjacent deoxyribose bonded to N-9 impedes the approach of the electrophile E1(E2)- 3,4-Q to N-3 (23,31). This interference is not present in DNA, as evidenced by formation of PAH—N3Ade adducts, which are rap- idly lost from the DNA by depurination (23,32). When E1-2,3-Q reacted with dG or dA, a profile of adducts totally different from those formed by E1—3,4-Q was obtained (Fig. 9) (26). Reaction of E,-2,3-Q with dG afforded 2-OHE1- 6—N2dG and with dA yielded 2—OHE1-6-N6dA. In this case, the E1-2,3-Q did not react as a quinone, but as its tautomer, the E1—2,3-Q methide. This electrophile reacts at G6 with the exocyclic amino group of dA or dG to yield the N6dA and NZdG adducts, which retain the deoxyribose and are referred to as “stable” adducts because they remain in DNA unless re- paired. 78 To determine whether these adducts are formed in biologic systems, E2-3,4-Q or enzymically activated 4-OHE2 was reacted with DNA for 2 hours at 37 °C (Fig. 10). The stable adducts were determined by the 32P-postlabeling method, and depurinating adducts were analyzed by high-performance liquid chromatog- raphy (HPLC) interfaced with an electrochemical detector (27). When E2-3,4-Q reacted with DNA, almost the same amount of 4-OHE2—l(a,B)-N7Gua and 4-OHE2—l(a,B)-N3Ade were ob- tained (Table 3). The amount of stable adducts was 0.02% of the depurinating adducts. Activation of 4-OHE2 with horseradish peroxidase gave similar results, whereas lactoperoxidase pro- duced a similar amount of N3Ade adduct but about 50% more N7Gua adduct (Table 3). The same two depurinating adducts were obtained in equal but smaller amounts when 4-OHE2 was activated with phenobarbital—induced rat liver microsomes (cy- tochrome P450) (27). When female Sprague—Dawley rats were treated by intramam- millary injection of 4-OHE2 or E2—3,4—Q, the 4-OHE2-1(0L,B)- N7Gua adduct was detected in the mammary tissue (27). The N3Ade adduct presumably was also present, but its synthetic standard was not available at the time of the study. These data clearly show that CBS are enzymatically oxidized to CE-Q and bind to DNA in vitro and in vivo. Several additional lines of evidence suggest that oxidation of 4-hydroxyestrogens is the pathway leading to estrogen-induced cancer. 4-Hydroxyestrogen formation has been observed to pre— dominate in hamster kidney (33,34) and other organs prone to estrogen-induced tumors, such as rat pituitary (35) and mouse uterus (36). In fact, 4-hydroxyestrogens induce kidney tumors in male Syrian golden hamsters, whereas the 2—hydroxyestrogens do not (4,37). Predominant 4-hydroxylase activity has also been found in human microsomes of uterine myometrium and benign uterine leiomyomas (38) as well as in microsomes of benign and malignant breast tumors (39,40). In tissues resistant to estrogen- induced tumors, such as the liver, formation of 2—hydroxyestro- gens predominates (39). Furthermore, 2,3,7,8-tetrachlorodibenzo— p-dioxin induces cytochrome P450 1B1, which predominantly catalyzes hydroxylation of E2 at the C-4 position (41—44). The importance of this finding is related to evidence that exposure to dioxin greatly increases the risk of developing cancer (42,45). Journal of the National Cancer Institute Monographs No. 27, 2000 1 13 EXONI 37 38 EXON 2 61 97 GGC CAA Normal —I ICCGI I'——I I GTTI I 111 321 500 Codon 13 l 13 37 38 97 GTC Mutation —|:l m l::_l————I l l 111 321 500 l 37 38 61 97 Codon 61 TA . — ]__..—__ c I Mutation I L I GAT 111 321 500 Fig. 5. Mouse Harvey—ras mutations. The normal Harvey-ras proto-oncogene (top) can be activated by mutation at codon 61 (CAA ——> CTA, middle) or codon 13 (GGC —> GTC, bottom). C E G A= r A = r PAH LPAH-N7Ade Adduct Abasic [C = $1 sute A = T lReplication [:31 FEE] lReplication [C=i] E$::] Fixed A=T A=T Mutant Fig. 6. Possible scheme for inducing A —> T mutations from depurinating Ade adducts. The combination of increased formation of the 4-hydroxylated CBS and their oxidation to CE-3,4-Q, which react with DNA to form the depurinating adducts associated with oncogenic muta— tions, suggests that the 4-hydroxyestrogen pathway producing CE-3,4-Q is responsible for the genotoxic effects leading to estrogen-induced initiation of cancer. The kidney of male Syrian golden hamsters is an established model for estrogen-induced tumorigenesis (46,47). Two hours after hamsters were given an injection intraperitoneally with 4—OHE2, the kidneys were removed and extracts were analyzed for the adducts formed by 4-OHE2 with DNA and GSH by using HPLC with electrochemical detection and confirmation by mass spectrometry (48). With DNA, the predominant adducts are 4-OHE2-1(0t,B)-N7Gua and 4-OHE2-1(a,B)-N3Ade; only the Journal of the National Cancer Institute Monographs No. 27, 2000 Table 2. Frequency of the Harvey—ras mutation in codon 61 (CAA —> CTA) after treatment of mouse skin with dibenzo[a,l]pyrene % of H-ras genes Time after treatment, days with codon 61 mutations 0 <0.001 1 0.1 2 1 3 5 6 0.5 9 <0.00l 35 0.1 63 (tumors) *Data taken from (25). N7Gua adduct was analyzed because different gradient condi- tions would have been needed for the N3Ade adduct. With GSH, 4-OHE2 forms the 4-OHE2-2-SG conjugate (49,50). This con— jugate is further metabolized to 4-OHE2-2-cysteine and 4-OHE2- 2-N—acetylcysteine by the mercapturic acid biosynthesis path- way. Therefore, all of these conjugates were searched for, along with the 4—OHE2-1(0L,B)—N7Gua adduct. Preliminary results in- dicate that 4-OHE2-1(a,B)-N7Gua and all of the GSH-derived conjugates are present in the kidney 2 hours after injection of 4-OHE2, with the cysteine conjugates being the most abundant. As hypothesized for the initiation of cancer by estrogens, these results demonstrate that, in the hamster kidney, 4-OHE2 is oxi- dized to E2-3,4-Q, which binds to GSH and to DNA, forming depurinating adducts. The nonsteroidal estrogen hexestrol, which is diethylstilbes- trol (DES) hydrogenated at the C-C double bond, is carcinogenic in Syrian golden hamsters (46,51). The major metabolite of hex- estrol and DES is their catechol (51—54), which can be meta- bolically converted to their catechol quinone. This hexestrol qui— none has chemical and biochemical properties similar to those of CE-3,4-Q, i.e., it specifically forms an N7Gua adduct after re- action with dG or DNA (55). These data suggest that the hex- estrol catechol quinone is the electrophile involved in tumor initiation by hexestrol. In turn, these results substantiate the hypothesis that CE—3,4-Q may be endogenous tumor initiators. In conclusion, the pathway of activation, i.e., oxidation of estrogens to CE and then to CE-Q, affords the ultimate carci- 79 R H300 Glutathione Conjugates HO 0: GSH S-transferase COMT R R non- enzymatic HO peroxidases peroxidases or P450 —'> Stable Adducts H0 or 2-OHE1(E2) P450 E1(E2)- -,2 3- so P450 reductase 5453-23-0 » Quinone Reductase NH2 N «,/J R N N E‘: R, = E2: H, -OH a non- enzymatlc —> HO peroxidases HO peroxudases or P450 0 4'0HE1(52)-‘1"(41J5) N3Ad6 OH or P450 4-OHE‘(E2) E1(E2)- -a 4- so P450 reductase E1(E2) 3.4- -o + H2N N / | \> H HN COMT R GSH S-transferase N o . H0 Glutathione Conjugates OH HO ocna 4-OHE1(E2)-1-(a,|3)-N7Gua Depurinating Adducts Fig. 7. Activating and deactivating (protecting) pathways of estrogen metabolism and formation of DNA adducts. nogenic metabolites that are CE-3,4-Q for endogenous estrogens and catechol quinones for synthetic estrogens. This competitive, oxidative pathway takes place only when excessive formation of 4—CE and/or their incomplete inactivation occur. The DNA dam- age by these reactive electrophiles consists of the formation of depurinating adducts and apurinic sites in DNA. Misrepair and/ or misreplication of the apurinic sites in DNA may generate the critical mutations that trigger induction of cancer by estrogens (21,25). ESTROGEN-MEDIATED FORMATION OF OXIDANTS AND OXIDATIVE DNA DAMAGE: THEIR ROLE IN CARCINOGENESIS Oxidants are continuously formed and degraded in normal cellular processes. They are necessary for a plethora of bio- chemical reactions, without which life itself could not be sus- tained. Cells are equipped with extensive multilayer antioxidant defenses to intercept excess oxidants. However, when ROS are generated at an inappropriate time, in excessive amounts, or when antioxidant defenses are overwhelmed, then the negative effects of oxidants become apparent. In the following presenta— tion, we will concentrate on the damaging effects of oxidants 80 induced endogenously by estrogens and their putative role in the carcinogenic process. One of the major types of damage that oxidants directly in- duce is oxidative modification of the genetic material. There are well over 30 different types of oxidized bases that can be formed in DNA; their levels exceed those of the stable carcinogen- induced adducts with DNA bases by about two orders of mag- nitude, being on average 1/105 versus 1/107 normal DNA bases, respectively (56,57). Some of these oxidized DNA bases are mutagenic (58—60) and induce DNA hypomethylation (61,62), a process known to increase gene expression (63). The oxidized base derivatives most frequently discussed in this presentation are 8-hydroxy-2’—deoxyguanosine (8-OHdG) and 5-hydroxymethyl-2’-deoxyuridine (HMdU) (for structures see Fig. 11). As oxidized bases are formed in DNA, various types of repair enzymes start removing them (64,65). There is always a background level of oxidized bases present in DNA, which seems to be tolerated by the cells. However, conditions leading to a continuous elevation of oxidized bases in DNA are the same as those that induce tumor promotion processes. The following examples illustrate that target sites for estrogen carcinogenesis invariably also contain elevated levels of oxi- dized bases in cellular DNA. Conversely, the presence of higher Journal of the National Cancer Institute Monographs No. 27. 2000 o {:l NA HzNHYNI CHZOH ””2 o 0 OH HO OH 4-OHE1(E2)-1(oc,B)-N7Gua R o o NH2 E1(E2)-3,4-Q N / NH2 H HO OH 4-OHE1(E2)-1(0t,B)-N3Ade Fig. 8. Reaction of E,(E2)-3,4-Q with dG or Ade. than normal amounts of oxidized DNA bases may be indicative of a carcinogenic process induced by estrogens. Oxidized DNA Bases as Evidence of Endogenous Oxidant Formation by Estrogens Animal models. The formation of either S-OHdG or HMdU in the target tissues for estrogen-mediated carcinogenesis has repeatedly been shown. The animal models often utilized are Syrian hamster kidney (66) and dorsolateral prostate in the Noble rat [(67); M. Bosland: unpublished data]. For example, the highest 8—OHdG increase (sevenfold) occurred in the peri- urethral section of the dorsolateral prostate isolated from the Noble rat treated with testosterone and E2 (Table 4; M. Bosland: unpublished data). This is the same part of the organ where F: {El HO CHZOH “2 HO HN E(-E2) 2 3 Q HNgN OH / N O Oi R 0% Q HOHZC _ 2-OHE1(E2)-6-N2dG o R HO E1(E2)-2,3-Q Methide HO NH2 E: R, :0 N \ “0 HOH c 52: H, -OH —OH \ N N | R 8-Hydroxy-2'-deoxyadenosine (8-OHdA) R = 2'-deoxyribose N | \>—OH N Fig. 11. Examples of oxidized DNA bases. adenocarcinoma growth (83% of animals), DNA adduct forma- tion (~fourfold increase), and lipid peroxidation (>threefold) oc- cur (Chapter 2). There was no cancer growth, DNA adducts, or 8—OHdG formation in the periphery. Thus, the oxidative DNA damage was detected at the same selected tissue site (10% 0f the whole prostate) where adenocarcinoma develops. 82 Formation of 8-OHdG and HMdU was also significantly el- evated and persisted in mouse skin topically treated with DMBA, a potent skin and mammary carcinogen, through the stages of tumor promotion and progression, and was evident at the time of tumor growth (Fig. 12) (68). The importance of oxidants and oxidative DNA damage in DMBA carcinogenesis Journal of the National Cancer Institute Monographs No. 27, 2000 Table 4. Testosterone and estradiol—induced changes in dorsolateral prostate* Periurethral Periphery Endpoint measured Control Treated % Change Control Treated % Change Adenocarcinoma 0 10/12 83 O 0 DNA adducts 2.7 102* 380 ND ND Lipid hydroperoxides 1.4 4.5 320 2.0 3.8 190 8—Ol-ldG 0.3 2.1+ 700 ND 0.] *From (67) and M. Bosland; unpublished data. TP 2 .02; ND 2 not detectable. is underscored by the long-known fact that pretreatment with antioxidants causes a suppression of DMBA-induced tumors without affecting levels of stable DNA adducts (69). DMBA treatment evoked sustained, long-lasting inflammatory re- sponses characterized by neutrophilic infiltration and edema (68). DMBA also enhances expression of IL—la messenger RNA (mRNA) and elevates the activity of IL-10L (70). Estrogens, even at a low physiologic dose, also increase formation of this in- flammatory cytokine (71), which, in turn, has a pronounced effect on a cascade of further inflammatory and carcinogenic responses (72—74). Cellular models. Both HMdU and 8-OHdG are present in MCF-lOA human breast epithelial cells (immortal but not tu- morigenic) and in MCF-7 breast cancer cells (75, 76). The basal levels of both modified bases are ~80% higher in MCF-7 cells, a finding that is expected because tumor cells produce substan— tial levels of hydrogen peroxide (H202), one of the major cel- lular oxidants (5 7, 77, 78). HMdU levels increased in response to H202 (75), and those of 8-OHdG, in response to the DMBA treatment ( 76 ). These increases are higher in MCF-lOA cells and reach levels prevalent in tumor cells. Carcinogen treatment of MCF-lOF cells (also immortal but otherwise a normal cell line) is known to lead to their malignant transformation (79). Hence, the increase in oxidized DNA bases likely occurs during the process of cellular transformation. Humans. More important, HMdU was shown to be present in white blood cell DNA of women at a high risk for breast cancer and those diagnosed with breast cancer (80,81). Of interest, a decrease in fat intake and presumed increase in vegetables and fruit consumption significantly decreased HMdU levels in women at high risk for breast cancer. In general, levels of oxi- dized purines were significantly elevated in human breast can- cer. However, they were also increased even at sites distal to the cancerous tissue, not only in the breast, but also at other sites of hormonal carcinogenesis, such as ovarian and, in men, prostatic cancers [(82,83); Chapter 9]. In fact, it has been proposed that the increased levels of oxidized bases in human DNA precede cancer development and may serve as biomarkers of cancer risk (81,83,84). There is extensive evidence that chemical carcino- gens generally induce formation of oxidized bases in DNA of target tissues in vivo. Estrogens are no exception and induce tumors by comparable mechanisms (57,66,85). Mechanisms of Estrogen-Mediated Oxidant Formation What are the sources of endogenous oxidants formed in re— sponse to estrogens? Estrogens, like other chemical carcinogens, are metabolized by cytochrome P450 enzymes and form hydrox- ylated products. The main metabolites of E2 include 2-, 4-, and l60t—hydroxyestradiol (Fig. 13) (33—36,44,86—88). The 2- and 4-hydroxylated catechols contain the hydroxyl groups in a vici- nal position, which predisposes them to further oxidation. Both can be oxidized to semiquinones, which in the presence of mo- lecular oxygen are oxidized to quinones with formation of su- peroxide anion radicals (023), as illustrated in Fig. 7 (66,89,90). These 023 readily dismutate to H202 either spontaneously or even faster when catalyzed by superoxide dismutase. H202 is neutral and rather nonreactive, except in the presence of the reduced transition metal ions (i.e., Fe2+ and Cu+), which cause formation of the most powerful and indiscriminate oxidants, hydroxyl radicals (OOH) (91,92). However, as a neutral mol- ecule, H202 can readily cross the cellular and nuclear mem— branes and reach DNA in neighboring cells, where it can cause site-specific oxidation of bases (57). Quinones and semiqui- nones are capable of redox cycling as long as there is molecular Fig. 12. Formation of HmdU and 8—OHdG in 7,12—dimethylbenz[a]anthra- cene (DMBA)—treated SENCAR mouse skin DNA. Adapted from (68). Per 104 Bases 0 1 2 7 14 21 35 8-OHdG ‘ Acetone Time after Last Tx (Days) Journal of the National Cancer Institute Monographs No. 27, 2000 83 OH P450 3A4 H0 173 - Estradiol (E2) 1A1, 1A2 O 3A4, 131 2 - Hydroxyestradiol (2-OHE2) OH .0. OH 160: - Hydroxyest-radiol OH 4 - Hydroxyestradiol OH (4-OHE2) Fig. 13. Oxidative metabolism of E2. Adapted from (44) and (87). oxygen available and, therefore, even a small amount of E2 may cause substantial ROS production and subsequent cellular damage. This ROS formation by redox cycling of semiquinones and quinones is mitigated by cellular quinone reductase, an enzyme that reduces quinones back to catechols by use of using reduced nicotinamide adenine dinucleotide (NADH) as a reducing co- factor. Moreover, COMT may prevent oxidation of CE to CE- SQ by methylating 2- or 4—hydroxyl groups (Fig. 14). However, it appears that the 4-hydroxyl group is not as readily methylated as is the 2-hydroxy1 substituent, which results in the predomi- nance of 4-OHE2 in redox cycling, while 2-OHE2 is virtually inactivated by methylation (93). Furthermore, methylated 2-OHE2 was shown to inhibit COMT-mediated methylation of 4-OHE2, which may allow for the accumulation of this carcino‘ genic metabolite in those organs where both metabolites are formed (93). The rapid methylation of 2-OHE2 may be one reason for its lack of carcinogenic activity (4,37), whereas the lesser methylation of 4-OHE2 contributes to its carcinogenic properties (4,37,39). Like PAH, estrogens can generate ROS by peroxidatic me- tabolism (94). For example, E2 was shown to produce phenoxyl radicals in the 1actoperoxidase—catalyzed reaction. These phenoxyl radicals rapidly react with the cellular-reducing agents GSH and NADH. However, instead of detoxification, other radi- cal species are formed (GS- and NADO, respectively), which reduce molecular oxygen to 02:, followed by H202 formation COMT —-—> HO HO HO OCH3 OH -OHE2 \ .— 0 Oz U 4_ HO O HO 0' 0 17B - Estradiol (E2) HO COMT H3CO ————> HO HO 2-OHE2 Fig. 14. Redox cycling of catechols. 84 Journal of the National Cancer Institute Monographs No. 27, 2000 (Fig. 15). The regenerated GSH and NADH can continue this process as long as 02 is available. Of the cellular reductants tested, only ascorbate radical does not further react with 02, thereby breaking the radical chain that leads to ROS formation. Lactoperoxidase and estrogens are ubiquitously present in milk ducts and in the mammary gland (94). Estradiol-Induced Lipid Peroxidation and DNA Damage ROS may also cause oxidation of cellular macromolecules other than DNA, which include proteins and lipids. For example, oxidation of cysteine residues at an active site of an enzyme would either inactivate or at least change the activity of that enzyme (95). Many biosynthetic and energy—producing antioxi- dants as well as repair enzymes have redox—sensitive centers, which are readily modified by prooxidant changes (96). Hence, E2-induced ROS may have a pronounced effect on cell mainte- nance and functioning. Lipids, particularly polyunsaturated lipids, are readily peroxi— dized, and the products often participate in a chain reaction propagating formation of various radical species (97). Lipid hydroperoxides are formed during prooxidant conditions gener- ated by different sources, which include inflammation and carcinogen exposures. Again, like other carcinogens, estro- gens induce lipid peroxidation during their metabolic activation (66). The insidiousness of this process is demonstrated by the fact that lipid hydroperoxides formed during E2 metabolism may serve as cofactors in further E2 (or other carcinogen) me- tabolism to hydroxylated products and in the oxidation of CE to quinone intermediates, which continuously amplifies the forma- tion of lipid hydroperoxides and cellular damage (Fig. 16). Fur- thermore, lipid hydroperoxide-derived aldehydes, such as malo- ndialdehyde and 4—hydroxynonenal, interact with bases in cellular DNA, thus increasing the burden of DNA modification (98,99). An interplay between estrogen metabolites and oxidants leads to at least three types of DNA base damage (Fig. 16): DNA base adducts produced by quinones, as described above, lipid hy- droperoxide-derived aldehyde DNA adducts, as well as a plethora of oxidized DNA bases. The importance of ROS for- mation during estrogen metabolism is underscored by the fact that H202 generated by the redox cycling of the semiquinone- quinone couple is readily reduced by cellular transition metal ions, such as Fe2+ and Cu’", to hydroxyl radicals (OOH), the most potent oxidants. Hydroxyl radicals not only oxidize bases in DNA, but also cause lipid peroxidation. Lipid hydroperoxides then serve as cofactors in further estrogen metabolism, which leads to additional semiquinone—quinone redox cycling, ROS production, and so on. Estradiol-Mediated Modulation of Immune Responses Although estrogens themselves can induce ROS production, they can also modulate immune responses and immune- mediated diseases, as indicated in Table 5, which can predispose to cancer (71,100). This process may take place under prooxi- dant conditions occurring in the course of inflammatory pro- cesses. In fact, at physiologic doses, E2 potently induces inter- leukin (IL)-lOL, a cytokine that can initiate a cascade of other cytokines, chemotactic and growth factors (71,101). Chemotac— tic factors cause infiltration of phagocytes, which may be acti- vated to secrete a plethora of other cytokines, ROS, and reactive nitrogen species (RNS) (72, 74,102—105). On the other hand, E2 inhibits IL-ld-induced IL-6 production. Therefore, by suppress- ing IL-6 formation, E2 increases human epithelial cell prolifera- tion, a process important in tumor growth, while it also inhibits the activity of natural killer cells, thus allowing tumor growth (101,105). E2 mediates macrophage proliferation and decreases cell differentiation (71). Each of the affected processes contrib— utes to the environment, allowing or encouraging tumor cell development and growth. Estrogen Effects on Macrophages and Their Production of ROS and RNS Macrophages have been found in normal human breast tissue. However, their numbers increase tremendously in breast tumors, providing up to 50% of the tumor mass (106,107). This increase in mass might be compounded by estrogen—stimulating macro- H202 H20 Lacto eroxidase E2 Peroxidase , . (Compound 11) D E2 [¢-0 l Compoundl (Compound II) E 2 . . 02 T ¢-o + NADH—>NAD —>o2 —>H202 0 _ $0 + GSH—> GS' —-—2> 02'—>H202 (b-O' + ascorbate—> ascorbate. —>X —>H202 Fig. 15. Futile estrogen metabolism. Adapted from ( 94). Journal of the National Cancer Institute Monographs No. 27, 2000 85 LPH-derived . . Aldehyde-DNA W Adducts Lipid FCZ+ .OH _ H202 \ DNA Adducts 02 02T LPH U E 4'OHE2 U Semiquinone _.> Quinone n NADPH NADPH NADH Fig. 16. Estradiol-induced formation of lipid hydroperoxides (LPH) and DNA damage. Adapted from (66). Table 5. Effects of estradiol (E2) on immune responses* E2 induces IL-lor O Chemotactic factors (IL-8, LTB4) O Infiltration of phagocytic cells Ez decreases IL-la-induced IL-6 O Human epithelial cell proliferation O Differentiation of T and B cells O Immunomodulation (IL-6) O Induction of aromatase E2 induces aromatase in macrophages E2 inhibits natural killer activity E2 increases immunoglobulins (IgM, IgG, etc.) *IL = interleukin; lg = immunoglobulin. phage proliferation (71). Substantial levels of estrogen are pro— duced from androgens within the female breast by the action of aromatase, a cytochrome P450 enzyme that is present in cells as well as in macrophages (88,108,109). Thus, estrogens cause macrophage proliferation and activation and, in turn, macro- phages produce estrogens, which may act on other phagocytic cells, and so on. On stimulation, macrophages produce oxidants such as 0;; and H202 by an reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase catalyzed reduction of molecu— lar oxygen. This process occurs rapidly by the activation of the constitutive NADPH oxidase. A few hours after macrophage stimulation, inducible nitric oxide synthase is synthesized by the cells and mediates production of L-arginine—derived nitric oxide (ONO), another radical species, which participates in signal transduction, numerous reactions, and cellular processes. 023 and ONO may rapidly interact, with the evolution of per- oxynitrite, a much more potent oxidant (110). Hence, macro— phage activation may lead to ROS and RNS, which include 023 , H202, OOH, and singlet oxygen [102, known to oxidize dG to 8-OHdG (111)], peroxynitrite (ONOO‘), nitrite, nitrate, as well as nitrating species. Among them, ROS and RNS may cause 86 DNA base hydroxylation, oxidation, nitration, and deamination (1 1 1—1 13). Estrogen Effects on PMN Function and ROS Formation Estrogens also affect the function of PMNs (polymorpho- nuclear leukocytes, neutrophils, granulocytes), which are an- other group of phagocytic cells that produce copious amounts of ROS on their stimulation. In addition to NADPH oxidase, PMNs express myeloperoxidase, an enzyme that catalyzes oxidation of chloride ions by H202 (generated by NADPH oxidase) to hy— pochlorite/hypochlorous acid (HOCl/OCI‘), one of the most po— tent oxidants (5 7). This reaction is catalyzed by myeloperoxi— dase released from PMNs during their activation process. HOCl/ OCl‘ is utilized as a bacteriocidal and tumoricidal agent within the organism. However, estrogens and some of their metabolites (i.e., E2, E1, 160t-OHE1, and estriol) may induce myeloperoxi— dase release from the resting (inactivated) cells and stimulate generation of oxidants in the absence of pathogens (114). Of interest, 2-hydr0xylated estrogens act as powerful inhibitors of PMNs activity, possibly one of the protective properties of the 2—hydroxylated CE. The estrogen-mediated action causes HOCl/OCI’ formation and ensuing oxidative cell damage, even in the absence of the proper targets. Moreover, estrogens can stimulate (by 10-fold) the activity of the released myeloperoxidase, thus compounding the damaging effects. The interaction of HOCl/OCl‘ with an excess of H202 causes regeneration of chloride ions as well as evolution of species that can chlorinate and oxidize DNA (Fig. 17) (112,114). Various interactions occur at physiological pH among reac- tive oxygen and nitrogen species generated by the phagocytic cells, macrophages and PMNs (Fig. 17). Those interactions lead to various types of DNA modification, many of which result in mutations. Therefore, estrogens, by modulating phagocytic cell Journal of the National Cancer Institute Monographs No. 27, 2000 4 - + Myeloperoxidase CM Nitric oxide synthase L -NO __—_, G, A + C L—arginine ._ \02 ONOO' —> . O _ O NADPH ox1dase —2> 02' —2+ H202 F—ZC’ ~OH—>Hydroxylation e, u OCl' Deamination H+ / NO'—> Nitration \ 102 —> Oxidation -—> 102 —-—> Oxidation / H+ —> C12 __, Chloramjne/ Chlorination Fig. 17. Oxygen-derived, enzyme-driven major cellular ROS at physiologic pH. Summary of literature by Khan AU, Frenkel K. proliferation and activation, have a pronounced effect on the integrity of DNA and mutagenesis. ROS generated by PMNs and macrophages cause not only DNA modification but also oxidation of proteins and lipid per- oxidation. As shown in Fig. 16, lipid hydroperoxides can now serve as cofactors of estrogen metabolism, during which ROS are produced, as well as other DNA-damaging species. There— fore, estrogens affect inflammatory responses and, in turn, their activities are affected by the inflammation products. Immunomodulation by Estrogens One of the more pronounced properties of E2 is its ability to differentiate T and B cells, increase immunoglobulin production, and aggravate immune complex-mediated diseases, such as sys- temic lupus erythematosus, which occur predominantly in fe- males (115—117). This disease is characterized by strong inflam- matory responses, which are thought to contribute to it, as well as to various types of cancer (57, 73,112,117). Women at high risk of breast cancer, those with benign breast diseases, and those who are diagnosed with breast cancer years later have significantly elevated anti—HMdU autoantibodies ( 84 ). As Fig. 18 shows, the levels of anti-HMdU autoantibodies are remarkably stable over a period of years. The presence of high anti-HMdU autoantibody levels attests to the prooxidant conditions that have led to oxidation of bases in cellular DNA and have evoked an autoimmune response. These results also suggest that the oxidative DNA base damage (HMdU) and the biologic responses it evokes (anti-HMdU autoantibodies) start occurring early in the carcinogenic process. Such a conclusion is strengthened by data obtained by other investigators, who showed that oxidative DNA base damage is evident in DNA of individuals at risk for hormone-dependent cancers (81,84,118). Effect of Estrogens 0n Oncogenes and Transcription Factors E2 was shown to induce macrophages to produce immediate early genes c-fos and c-jun, and the AP-l transcription factor, which is a heterodimer of these two genes ( 71 ). AP—l—binding Journal of the National Cancer Institute Monographs No. 27, 2000 120 Control BBD Future Breast_ 100 _ 0 Cancer _ a _ .. B 80 ' <> ‘ — \j 60- 1 AI - ox ' _ <5“ 40—0 3 _ _- _ if If f *3 — . , A ‘ o 6 fi Q ‘ ‘— 0 I I I ‘—JI T I I I I I I $ I I I I I 0 4 8 12 16 20 24 28 32 Individuals Fig. 18. Histogram of Anti—HMdU autoantibody titers in human sera (healthy controls, those with benign breast diseases [BBD], and those who were appar— ently healthy at the time of blood donation but were diagnosed with breast cancer 0.5—6 years later). Adapted from (84). sites are present in promoter regions of many growth factors as well as antioxidant enzymes (112,119,120). E2 also induces c- myc, an oncogene known to be important in tumor promotional processes, and bcl-2, a gene inhibiting apoptosis. Oxidants are known to induce all of these oncogenes, whereas antioxidants counteract their formation (57). Thus, E2 through its ROS- inducing capabilities affects processes leading to mutations, governing tumor growth, and tumor surveillance. Effects of estrogens on oxidant formation, oxidative DNA damage, and other cellular processes, which contribute to the carcinogenic properties of estrogens, are summarized in Table 6. Most of these effects are exactly the same as those induced by other chemical carcinogens. Tamoxifen as an Anticarcinogen Tamoxifen has been used as a drug that effectively prevents recurrence of breast cancer. It has also been recently shown to 87 Table 6. Selected properties of estrogens Carcinogenesis in susceptible organs Oxidative estrogen metabolism Oxidant formation and lipid peroxidation Genetic factors 0 Immediate early gene (c-fos, c-jun) - Transcription factors (AP—l) 0 Growth factors (TGF-B) o Oncogenes (c—myc, bcl»2) Epithelial cell growth (T) and cell differentiation (t) Increase of macrophage—mediated immune responses 0 Proliferation 0 Activation 0 Secretions (cytokines, ROS, RNS) 0 Function (Antigen processing and antigen presenting to T cells) Genetic damage 0 DNA adducts - Oxidized DNA bases 0 Lipid—derived aldehyde— DNA adducts decrease substantially breast cancer in women at high risk for this disease based on a strong family history. The controversy over its use as a preventive agent exists because tamoxifen in- creases the rate of endometrial cancer in susceptible individuals (121). Tamoxifen exerts its antiproliferative effects on the estrogen receptor-positive cells as well as on cells lacking that recep- tor (122,123). This suggests that the mode of action of tamoxi— fen in suppressing breast cancer relies not only on its antiestro— genic properties but must involve effects on other factors im— portant in the carcinogenic process. Tamoxifen has been shown to be a potent antitumor promoter in animal models, while es- trogens can act as tumor promoters. Many of the processes known to contribute to tumor promotion are inhibited by tamoxi- fen (Table 7) (124). These include inhibition of hyperplasia, inflammation, ROS production, oxidative DNA base damage, and lipid peroxidation. Estrogens induce all of these processes. Hence, the fact that tamoxifen, an antibreast cancer drug, sup— presses so many of the estrogen-induced prooxidant processes and factors strengthens the hypothesis of estrogen being a car— cinogen that acts through elevation of oxidative stress. The in- terrelationships among many factors contributing to estrogen— induced oxidative stress and carcinogenesis are summarized in Fig. 19. Is E2 A GENOTOXIC 0R EPIGENETIC CARCINOGEN? Pharmacologic levels of estrogens are known to produce toxic effects, such as embryotoxicity, teratogenicity, and carci— nogenicity (2,125,126). The molecular mechanism(s) by which estrogens cause such adverse effects are under investigation from several different angles. Recent epidemiologic and labora— tory findings have increased the growing concern that instability in the genome induced by estrogens may be involved in the induction of certain types of cancer in humans (127). Estrogen— induced genotoxicity is an important contributor to the induction Table 7. Effects of tamoxifen, an antitumor promoter Suppresses hyperplasia Antagonizes inflammatory responses Inhibits oxidant formation by neutrophils Decreases oxidative DNA base damage Decreases lipid peroxidation Reduces MDA serum levels in breast cancer patients Enhances apoptosis of tumor cells 88 OXIDATIV'E \ DNA DAMAGE TUMOR IMMUNE SYSTEM"--..._ Ami-HMdU ""7 antibody Fig. 19. Interaction among estrogen, immune system, tumor, and oxidative stress. of toxic effects, because estrogen receptor-mediated events by themselves cannot explain the carcinogenic and noncarcinogenic adverse properties of estrogens. The lack of mutagenic activity in bacterial and mammalian cell mutation assays (3—7) led to estrogens being categorized as nongenotoxic and nonmutagenic chemicals (8,9,128). However, more recent results, such as the arrest of DNA replication deriving from estrogen—DNA adducts, the enhancement of homologous recombination, and mutations in individual genes and in microsatellite repeat sequences clearly indicate that estrogens are able to induce multiple types of ge- netic insults in cells. This part of the chapter will focus on estrogen—mediated damage at the genome level leading to the development of mutations. Indirect Evidence of Mutations Induced by Estrogens Mutations may result from numeric changes or structural al- terations in the genome. These include extra or missing copies of microsatellite DNA, transcriptional silencing, chromosomal de- letions, frameshifts, amplifications, rearrangements, transloca- tions, and other changes that interfere with the integrity of the genome. More recent experiments have shown that some estro- gens are capable of producing such instability (126,129,130). For instance, estrogens induce numeric changes in chromosomes (genome mutation or aneuploidy) with and without apparent DNA damage (131 ). Both DES and E2 are potent inhibitors of mitosis in vitro and are capable of inducing genomic mutations in cultured cells (132). Potential targets for numeric changes in chromosomes are the spindle apparatus (microtubules and cen— trioles), DNA, regulating proteins, and centromere. Chromo- somal analysis of tissues of mice exposed perinatally to estrogen demonstrates that this treatment induces chromosomal aberra- tions in the same target tissues in which tumors subsequently develop (133,134). DES or E2 treatment of hamsters produces renal chromosomal aberrations, including deletions, inversions, and translocations (135,136). DNA damage either by free radicals or by genotoxic reactive metabolites is known to cause structural changes in chromo— somes. Formation of oxygen-free radicals by redox cycling of Journal of the National Cancer Institute Monographs No. 27, 2000 estrogens or DNA modification by reactive estrogen metabolites may explain some of the structural and numeric chromosomal changes observed in response to estrogen exposure (131,132). Damage to DNA by ROS generated by estrogen treatment, i.e., 8-OHdG, lipid-DNA adducts, and DNA strand breaks, may in- duce structural and numeric alterations in chromosomes and may be important lesions capable of producing mutagenic changes in the genome. Direct Evidence of Gene Mutations Induced by Estrogens The potential of estrogens to induce mutations has been highly controversial. Previous studies of the mutagenic potential of estrogens showed that neither they nor their reactive interme- diates induced mutations in the Ames bacterial reversion test or in Syrian hamster embryo cells (3—7). However, these results are not consistent with the covalent binding of reactive estrogen metabolites to bases of DNA or with the ROS—mediated changes to DNA bases, as discussed above. Both of these types of DNA lesions are capable of inducing mutations. Some earlier studies showed that DES and E2 can increase mutations leading to oua- bain resistance (137). Experiments demonstrate that DES qui- none increases homologous recombination in Escherichia coli (138). Both DES and E2 are mutagenic in the gpt+ Chinese hamster G12 cell line (7). Re-examination of the mutagenic potential of E2 at various concentrations demonstrated a weak, but detectable, mutagenic— ity of E2 at the lowest dose assayed (10—10 uM E2) in V79 Chinese hamster lung cells [(7); Albrecht T, Liehr JG: unpub- lished data]. Covalent DNA adducts formed by DES quinone and CE quinone arrest the progression of cytochrome oxidase III gene synthesis (139). The mRNA for the repair enzyme DNA polymerase [3 obtained from DES-induced kidney tumors carries several mutations in the catalytic domain compared to that of age-matched control kidney (140). Kidney tumors and premalignant kidney of E2-treated hamsters contain muta- tions in repeat sequences of microsatellite DNA (141,142). Re- cently, mutational changes in an unidentified gene have been observed in the stilbene-induced hamster kidney tumor (Singh LP, Roy D: unpublished data). A high frequency of genomic rearrangements have been observed in transformed 10T1/2 mouse cell subclones treated with E2, indicating that this and other natural hormones may accelerate the accumulation of mu- tations ( 143 ). Genetic instability manifested by somatic mutation of micro- satellite repeats occurs with high frequency in clear cell adeno— carcinomas of the vagina and cervix, with evidence of micro- satellite instability in all DES-associated tumors examined (144). Furthermore, mutations have been reported in the p53 gene at codons 274 and 140 in hepatocellular carcinoma asso- ciated with use of oral contraceptives (145). A good associa- tion has been shown between induction of aneuploidy, DNA adducts, and estrogendnduced cell transformation ([46). These findings indicate that several types of mutations induced by estrogens may not be detectable by the Salmonella reversion test or assay of gene mutations at specific narrowly defined loci in Syrian hamster embryo cells. Taken together, these findings suggest that estrogens can produce multiple types of genetic insults contributing to the induction of genomic instability. Sev- eral structural and numeric changes have already been demon- strated at the cellular level in response to DES or E2 exposure (147). Journal of the National Cancer Institute Monographs No. 27, 2000 CONCLUSIONS In this chapter, we have outlined a large body of evidence that estrogens, including the natural hormones E2 and E1, damage genetic material in various ways. Direct estrogen DNA adducts and/or indirect forms of covalent DNA alterations may be in- duced in experimental systems in vitro, cells in culture, and laboratory animals or may be detected in humans. Study of model carcinogenic PAH has led to the discovery that apurinic sites in DNA generated by depurinating adducts can lead to the mutations that trigger the cancer process. The CE metabolites, when oxidized to the electrophilic CE-Q, may react with DNA to form stable and depurinating adducts. The 4—CE that form pre- dominantly depurinating adducts are carcinogenic, whereas the noncarcinogenic 2-CE exclusively form stable DNA adducts. Oxidation of CE also leads to overwhelming amounts of ROS that generate extensive DNA damage and contribute to tumor promotion. Preliminary data also suggest that estrogens, including the natural hormones E2 and E], may induce gene mutations. The more recent gene mutation experiments reported in this chapter are compatible with the lack of mutagenic activity of synthetic and steroidal estrogens and their metabolites reported previously (3—7). Mutagenicity assays are designed to detect only the rela- tively high-mutation frequencies of potent carcinogenic com- pounds. In contrast, estrogens may be only weakly mutagenic, as is to be expected for endogenous compounds. Thus, mutagenic— ity assay conditions may have to be redesigned to detect low mutation frequencies at multiple gene loci with high accuracy and precision. The current data presented in this chapter lead to the conclu- sion that estrogens are genotoxic carcinogens. Oxidation of E1 and E2 to CE, with the 4-CE playing the major role, is the critical pathway of activation to form ultimate electrophilic metabolites and ROS. REFERENCES (1) Overall evaluations of carcinogenicity: an updating of IARC monographs volumes 1 to 42. IARC Monographs on the Evaluation of Carcinogenic Risk to Humans. Suppl 7. Lyon (France); 1987. p. 272—310. (2) Hormonal contraception and post—menopausal hormonal therapy. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Vol 72. Lyon (France); 1999. (3) Lang R, Redmann U. Non—mutagenicity of some sex hormones in the Ames salmonella/microsome mutagenecity test. Mutat Res 1979;67: 361—5. (4) Liehr JG, Fang WF, Sirbasku DA, Ari-Ulubelen A. Carcinogenicity of catechol estrogens in Syrian hamsters. J Steroid Biochem 1986;24:353—6. (5) Lang R, Reimann R. Studies for a genotoxic potential of some endog- enous and exogenous sex steroids. 1. Communication: examination for the induction of gene mutations using the Ames Salmonella/microsome test and the HGPRT test in V79 cells. Environ Mo] Mutagen 1993;21: 272—304. (6) Drevon C, Piccoli C, Montesano R. Mutagenicity assays of estrogenic hormones in mammalian cells. Mutat Res 1981;89:83—90. (7) Rajah TT, Pento IT. The mutagenic potential of antiestrogens at the HPRT locus in V79 cells. Res Comm Mol Pathol Pharmacol 1995;89:85—92. (8) Li J], Li SA. Estrogen carcinogenesis in hamster tissues: a critical review. Endocr Rev 1990;1 1:524—31. (9) Barrett JC, Tsutsui T. Mechanisms of estrogen—associated carcinogenesis. In: Huff J, Boyd J, Barrett JC, editors. Cellular and molecular mechanisms of hormonal carcinogenesis: environmental influences. New York (NY): Wiley-Liss; 1996. p. 105—12. (10) Feigelson HS, Henderson BE. Estrogens and breast cancer. Carcinogen— esis 1996;17:2279—84. 89 (1]) Clarke RB, Howell A, Potten CS, Anderson E. Dissociation between steroid receptor expression and cell proliferation in the human breast. Cancer Res 1997;57:4987—91. (12) Russo J, A0 X, Grill C, Russo IH. Pattern of distribution of cells positive for estrogen receptor a and progesterone receptor in relation to prolifer- ating cells in the mammary gland. Breast Cancer Res Treat 1999;53: 217—27. (13) Endo S, Metzler M, Hieber L. Nonrandom karyotypic changes in a spon- taneously immortalized and tumourigenic Syrian hamster embryo cell line. Carcinogenesis 1994;15:2387—90. (I4) Liehr JG, Gladek A, Macatee T, Randerath E, Randerath K. DNA adduct formation in liver and kidney of male Syrian hamsters treated with estro- gen and/or ot—naphthoflavone. Carcinogenesis 1991;12:385—9. (I5) Liehr JG, Wheeler WJ. Inhibition of estrogen—induced renal carcinoma in Syrian hamsters by vitamin C. Cancer Res 1983;43:4638—42. (16) Zhu BT, Roy D, Liehr JG. The carcinogenic activity of ethinyl estrogens is determined by both their hormonal characteristics and their conversion to catechol metabolites. Endocrinology 1993;132:577—83. (I7) Stalford AC, Maggs JL, Gilchrist TL, Park BK. Catecholestrogens as mediators of carcinogenesis: correlation of aromatic hydroxylation of es— tradiol and its fluorinated analogs with tumor induction in Syrian harn— sters. Mol Pharrnacol 1994;45:1259—67. (18) Weinberg RA. How cancer arises. Sci Am 1996;275:62—77. (19) Cavalieri EL, Rogan EG. The approach to understanding aromatic hydro— carbon carcinogenesis. The central role of radical cations in metabolic activation. Pharrnacol Ther 1992;55:183—99. (20) Cavalieri EL, Rogan EG. Mechanisms of tumor initiation by polycyclic aromatic hydrocarbons in mammals. In: Neilson AH, editor. The hand- book of environmental chemistry. Vol 31. PAHs and related compounds. Heidelberg (Germany): Springer-Verlag; 1998. p. 81—117. (21) Chakravarti D, Pelling JC, Cavalieri EL, Rogan EG. Relating aromatic hydrocarbon—induced DNA adducts and c—H-ras mutations in mouse skin papillomas: the role of apurinic sites. Proc Natl Acad Sci U S A 1995;92: 10422—6. (22) Devanesan PD, RamaKrishna NV, Padmavathi NS, Higginbotham S, Ro— gan EG, Cavalieri EL, et al. Identification and quantitation of 7,12- dimethylbenz[a]anthracene—DNA adducts formed in mouse skin. Chem Res Toxicol 1993;6:364—7l. (23) Chen L, Devanesan PD, Higginbotham S, Ariese F, Jankowiak R, Small GJ, et al. Expanded analysis of benzo[a]pyrene—DNA adducts formed in vitro and in mouse skin: their significance in tumor initiation. Chem Res Toxicol 1996;9z897—903. (24) Chakravarti D, Mailander P, Higginbotham S, Cavalieri E, Rogan E. Relationship of depurinating adducts with H—ras oncogenic mutations in mouse skin papillomas induced by benzo[a]pyrene, benzo[a]pyrene—7,8— dihydrodiol and 6-methylbenzo[a]pyrene. Proc Am Assoc Cancer Res 1999;40:507. (25) Chakravarti D, Mailander P, Franzen J, Higginbotham S, Cavalieri E, Rogan E. Detection of dibenzo[a,l]pyrene-induced H-ras codon 61 mutant genes in preneoplastic SENCAR mouse skin using a new PCR—RFLP method. Oncogene 1998;16:3203—10. (26) Stack DE, Byun J, Gross ML, Rogan EG, Cavalieri EL. Molecular char- acteristics of catechol estrogen quinones in reactions with deoxyribo- nucleosides. Chem Res Toxicol 1996;92851—9. (27) Cavalieri EL, Stack DE, Devanesan PD, Todorovic R, Dwivedy I, Hig— ginbotham S, et al. Molecular origin of cancer: catechol estrogen-3,4- quinones as endogenous tumor initiators. Proc Natl Acad Sci U S A 1997; 94:10937—42. (28) Ball P, Knuppen R. Catecholoestrogens (2- and 4-hydroxyestrogens): chemistry, biogenesis, metabolism, occurrence and physiological signifi— cance. Acta Endocrinol Suppl (Copenh)l980;232:l—127. (29) Martucci CP, Fishman J. P450 enzymes of estrogen metabolism. Phar- macol Ther 1993;57:237—57. (30) Li KM, Devanesan PD, Rogan EG, Cavalieri EL. Formation of the depu— rinating 4-hydroxyestradiol (4-OHE2)-1—N7Gua and 4-OHE2-l-N3Ade adducts by reaction of E2-3,4-quinone with DNA. Proc Am Assoc Cancer Res 1998;39:636. (31) Mulder PP, Chen L, Sekhar BC, George M, Gross ML, Rogan EG, et al. Synthesis and structure determination of the adducts formed by electro- chemical oxidation of 1,2,3,4-Tetrahydro-7,l2—dimethylbenz[a]anthra— 90 Gene in the presence of deoxyribonucleosides or adenine. Chem Res Toxi- col 1996;9z1264—77. (32) Li KM, Todorovic R, Rogan EG, Cavalieri EL, Ariese F, Suh M, et al. Identification and quantitation of dibenzo[a,l]pyrene—DNA adducts formed by rat liver microsomes in vitro: preponderance of depurinating adducts. Biochemistry 1995;34:8043—9. (33) Weisz J, Bui QD, Roy D, Liehr JG, Elevated 4-hydroxylation of estradiol by hamster kidney microsomes: a potential pathway of metabolic activa— tion of estrogens. Endocrinology 1992;131:655—61. (34) Zhu BT, Bui QD, Weisz J, Liehr JG. Conversion of estrone to 2— and 4—hydr0xyestrone by hamster kidney and liver microsomes: implications for the mechanism of estrogen-induced carcinogenesis. Endocrinology 1994;135:1772—9. (35) Bui Q, Weisz J. Identification of microsomal, organic hydroperoxide- dependent catechol estrogen formation: comparison with NADPH- dependent mechanism. Pharmacology 1988;36:356—64. (36) Bunyagidj C, McLachlan JA. Catechol estrogen formation in mouse uterus. J Steroid Biochem 1988;31:795—801. (37) Li JJ, Li SA. Estrogen carcinogenesis in Syrian hamster tissues: role of metabolism. Fed Proc 1987;46:1858-63. (38) Liehr JG, Ricci MJ, Jefcoate CR, Hannigan EV, Hokanson JA, Zhu BT. 4-Hydroxylation of estradiol by human uterine myometrium and myoma microsomes: implications for the mechanism of uterine tumorigenesis. Proc Natl Acad Sci U S A 1995;92:9220—4. (39) Liehr JG, Ricci MJ. 4-Hydroxylation of estrogens as marker of human mammary tumors. Proc Natl Acad Sci U S A 1996;93:3294—6. (40) Castagnetta LA, Granata OM, Arcuri FP, Polito LM, Rosati F, Cartoni GP. Gas chromatography/mass spectrometry of catechol estrogens. Ste- roids 1992;57:437—43. (41) Spink DC, Hayes CL, Young NR, Christou M, Sutter TR, Jefcoate CR, et al. The effects of 2,3,7,8—tetrachlorodibenzo-p-dioxin on estrogen metabo— lism in MCF-7 breast cancer cells: evidence for induction of a novel l7B-estradiol 4-hydroxylase. J Steroid Biochem Mol Biol 1994;51:251—8. (42) Hayes CL, Spink DC, Spink BC, Cao JQ, Walker NJ, Sutter TR. 178- estradiol hydroxylation catalyzed by human cytochrome P450 1B1. Proc Natl Acad Sci U S A 1996;93:9776—81. (43) Spink DC, Spink BC, Cao JQ, Gierthy JF, Hayes CL, Li Y, et a1. Induc- tion of cytochrome P450 1B] and catechol estrogen metabolism in ACHN human renal adenocarcinoma cells. J Steroid Biochem Mol Biol 1997,62: 223—32. (44) Spink DC, Spink BC, Cao JQ, DePasquale JA, Pentecost BT, Fasco MJ, et al. Differential expression of CYP1A1 and CYPlBl in human breast epithelial cells and breast tumor cells. Carcinogenesis 1998;19:291—8. (45 ) Wolff MS, Toniolo PG, Lee EW, Rivera M, Dubin N. Blood levels of organochlorine residues and risk of breast cancer. J Natl Cancer Inst 1993;85:648—52. (46) Li J], Li SA, Klicka JK, Parsons JA, Lam LK. Relative carcinogenic activity of various synthetic and natural estrogens in the Syrian hamster kidney. Cancer Res 1983;43:5200—4. (47) Kirkman H. Estrogen-induced tumors of the kidney in the Syrian hamster. III. Growth characteristics in the Syrian hamster. Monogr Natl Cancer Inst 1959;] :1—57. (48) Devanesan P, Todorovic R, Zhao J, Higginbotham S, Gross M, Rogan E, et al. Formation of catechol estrogen adducts in kidneys of male hamsters treated with 4—hydroxyestradiol. Proc Am Assoc Cancer Res 1999;40:46. (49) Cao K, Stack DE, Ramanathan R, Gross ML, Rogan ER, Cavalieri EL. Synthesis and structure elucidation of estrogen quinones conjugated with cysteine, N—acetylcysteine, and glutathione. Chem Res Toxicol 1998;11: 909—16. (50) Cao K, Devanesan PD, Ramanathan R, Gross ML, Rogan ER, Cavalieri EL. Covalent binding of catechol estrogens to glutathione catalyzed by horseradish peroxidase, lactoperoxidase, or rat liver microsomes. Chem Res Toxicol 1998;11:917—24. (51) Liehr JG, Ballatore AM, Dague BB, Ulubelen AA. Carcinogenicity and metabolic activation of hexestrol. Chem Biol Interact 1985;55:157—76. (52) Haaf H, Metzler M. In vitro metabolism of diethylstilbestrol by hepatic, renal and uterine microsomes of rats and hamsters. Effects of difference inducers. Biochem Pharmacol 1985;34:3107—15. (53) Blaich G, Gottlicher M, Cikryt P, Metzler M. Effects of various inducers on diethylstilbestrol metabolism, drug-metabolizing enzyme activities and Journal of the National Cancer Institute Monographs No. 27, 2000 the aromatic hydrocarbon (Ah) receptor in male Syrian golden harnster liver. J Steroid Biochem 1990;35:201—4. (54) Metzler M, McLachlan JA. Oxidative metabolism of the synthetic estro- gens hexestrol and dienestrol indicates reactive intermediates. Adv Exp Med Biol 1981;136 Pt A2829—37. (55) Jan ST, Devanesan PD, Stack DE, Ramanathan R, Byun J, Gross ML, et al. Metabolic activation and formation of DNA adducts of hexestrol, a synthetic nonsteroidal carcinogenic estrogen. Chem Res Toxicol 1998;11: 412—9. (56) Dizdaroglu M. Chemical determination of free radical-induced damage to DNA. Free Radic Biol Med 1991;10:225—42. (57) Frenkel K. Carcinogen—mediated oxidant formation and oxidative DNA damage. Pharmacol Ther 1992;53:127—66. (58) Shimame—More L, Rossman TG, Troll W, Teebor GW, Frenkel K. Ge— netic effects of 5-hydroxymethyl—2’—deoxyuridine, a product of ionizing radiation. Mutat Res 1987;178:177—86. (59) Tchou J, Grollman AP. Repair of DNA containing the oxidatively- damaged base, 8-oxoguanine. Mutat Res 1993;299:277—87. (60) Feig DI, Sowers LC, Loeb LA. Reverse chemical mutagenesis: identifi- cation of the mutagenic lesions resulting from reactive oxygen species- mediated damage to DNA. Proc Natl Acad Sci U S A 1994;91:6609—13. (61) Berkrier KL, Folk WR. EcoRI cleavage and methylation of DNAs con— taining modified pyrimidines in the recognition sequence. J Biol Chem 1977;252:3185—93. (62) Cerda S, Weitzman SA. Influence of oxygen radical injury on DNA methylation. Mutat Res 1997;386:141—52. (63) Zingg JM, Jones PA. Genetic and epigenetic aspects of DNA methylation on genome expression, evolution, mutation and carcinogenesis. Carcino- genesis 1997;18:869—82. (64) Teebor GW. Excision DNA repair. In: Vos JM, editor. DNA repair mechanisms: impact on human diseases and cancer. Austin (TX): RG Landes Co.; 1995. p. 99—123. (65) Demple B, Harrison L. Repair of oxidative damage to DNA: enzymology and biology. Annu Rev Biochem 1994;63:915—48. (66) Liehr JG. Hormone—associated cancer: mechanistic similarities between human breast cancer and estrogen—induced kidney carcinogenesis in ham— sters. Environ Health Perspect 1997;105 Suppl 3:565—9. (67) Han X, Liehr JG, Bosland MC. Induction of a DNA adduct detectable by 32P-postlabeling in the dorsolateral prostate of NBL/Cr rats treated with estradiol-l7B and testosterone. Carcinogenesis 1995;16:951—4. (68) Frenkel K, Wei L, Wei H. 7,l2—dimethylbenz[a]anthracene induces oxidative DNA modification in viva. Free Radic Biol Med 1995;19:373— 80. (69) Dipple A, Pigott MA, Bigger CA, Blake DM. 7,12—dimethylbenz[a]— anthracene—DNA binding in mouse skin: response of different mouse strains and effects of various modifiers of carcinogenesis. Carcinogenesis 1984;5z 1087—90. (70) Frenkel K, Li X, Li C. Effects of interleukin (IL)-10t on 7,12- dimethylbenz[a]anthracene (DMBA)-induced carcinogenesis in SENCAR mice. Proc Am Assoc Cancer Res 1998;39:487. (71) Cutolo M, Sulli A, Seriolo B, Accardo S, Masi AT. Estrogens, the im— mune response and autoimmunity. Clin Exp Rheumatol 1995;13:217— 26. (72) Kupper TS. Immune and inflammatory processes in cutaneous tissues. Mechanisms and speculations [published erratum appears in J Clin Invest 1991;87:753]. J Clin Invest 1990;86:1783—9. (73) Weitzman SA, Gordon LI. Inflammation and cancer: role of phagocyte- generated oxidants in carcinogenesis. Blood 1990;76:655—633. (74) Stadnyk AW. Cytokine production by epithelial cells. FASEB J 1994;8z 1041—7. (75) Djuric Z, Everett CK, Luongo DA. Toxicity, single-strand breaks, and 5—hydroxymethyl—2’-deoxyuridine formation in human breast epithelial cells treated with hydrogen peroxide. Free Radic Biol Med 1993:14: 541—7. (76) Huang X, Yusaf R, Shah R, Frenkel K. Pro-oxidant effects of 7,12— dimethylbenz[a]anthracene (DMBA) on human epithelial breast cells. Proc Am Assoc Cancer Res 1998;39:372. (77) Szatrowski TP, Nathan CF. Production of large amounts of hydrogen peroxide by human tumor cells. Cancer Res 1991;51:794—8. (78) Bhimani RS, Troll W, Grunberger D, Frenkel K. Inhibition of oxidative Journal of the National Cancer Institute Monographs No. 27, 2000 stress in HeLa cells by chemopreventive agents. Cancer Res 1993;53: 4528—33. (79) Calaf G, Russo J. Transformation of human breast epithelial cells by chemical carcinogens. Carcinogenesis 1993;14:483—92. (80) Djuric Z, Heilbrun LK, Reading BA, Boomer A, Valeriotta FA, Martino S. Effects of a low-fat diet on levels of oxidative damage to DNA in human peripheral nucleated blood cells. J Natl Cancer Inst 1991;83: 766—9. (81) Djuric Z, Heilbrun LK, Simon MS, Smith D, Luongo DA, LoRusso PM, et al. Levels of 5-hydroxymethyl-2’-deoxyuridine in DNA from blood as a marker of breast cancer risk. Cancer 1996;77:691—6. (82) Malins DC, Polissar NL, Gunselman SJ. Progression of human breast cancers to the metastatic state is linked to hydroxyl radical-induced DNA damage. Proc Natl Acad Sci U S A 1996;93:2557—63. (83) Malins DC, Polissar NL, Schaefer S, Su Y, Vinson M. A unified theory of carcinogenesis based on order-disorder transitions in DNA structure as studied in the human ovary and breast. Proc Natl Acad Sci U S A 1998; 95:7637—42. (84) Frenkel K, Karkoszka J, Glassman T, Dubin N, Toniolo P, Taioli E, et al. Serum autoantibodies recognizing 5-hydroxymethyl-2'-deoxyuridine, an oxidized DNA base, as biomarkers of cancer risk in women. Cancer Epidemiol Biomarkers Prev 1998;7:49—57. (85) Trush MA, Kensler TW. An overview of the relationship between oxida- tive stress and chemical carcinogenesis. Free Rad Biol Med 1991;10: 201—9. (86) Bni QD, Weisz J. Monooxygenase mediating catecholestrogen formation by rat anterior pituitary is an estrogen-4-hydroxylase. Endocrinology 1989;124:1085—7. (87) Shou M, Korzekwa KR, Brooks EN, Krausz KW, Gonzalez FJ, Gelboin HV. Role of human hepatic cytochrome P450 1A2 and 3A4 in the meta- bolic activation of estrone. Carcinogenesis 1997;18:207—14. (88) Zhu BT, Conney AH. Functional role of estrogen metabolism in target cells: review and perspectives. Carcinogenesis 1998;19:1—27. (89) Nutter LM, Wu YY, Ngo E0, Sierra EE, Gutierrez PL, Abul-Hajj YJ. An o-quinone form of estrogen produces free radicals in human breast cancer cells: correlation with DNA damage. Chem Res Toxicol l994;7: 23—8. (90) Tabakovic K, Gleason WB, Ojala WH, Abul-Hajj YJ. Oxidative trans- formation of 2-hydroxyestrone. Stability and reactivity of 2,3-estrone qui- none and its relationship to estrogen carcinogenicity. Chem Res Toxicol 1996;9z860—5. (91) Klein CB, Frenkel K, Costa M. The role of oxidative processes in metal carcinogenesis. Chem Res Toxicol 1991;42592—604. (92) Lloyd RV, Hanna PM, Mason RP. The origin of the hydroxyl radical oxygen in the Fenton reaction. Free Radic Biol Med 1997;22:885—8. (93) Weisz J, Clawson GA, Creveling CR. Biogenesis and inactivation of catecholestrogens. Adv Pharmacol 1998;42:828—33. (94) Sipe HJ Jr, Jordan SJ, Hanna PM, Mason RP. The metabolism of 17B- estradiol by lactoperoxidase: a possible source of oxidative stress in breast cancer. Carcinogenesis 1994;15:2637—43. (95) Pero RW, Sheng Y, Olsson A, Bryngelsson C, Lund—Pero M. Hypochlo- rous acid/N-chloramines are naturally produced DNA repair inhibitors. Carcinogenesis 1996;17:13—8. (96) Ding H, Demple B. In vivo kinetics of a redox-regulated transcriptional switch. Proc Natl Acad Sci U S A 1997;94:8445—9. (97) Aruoma 01. Free radicals, oxidants, and antioxidants: trend towards the year 2000 and beyond. In: Aruoma OI, Halliwell B, editors. Molecular biology of free radicals in human diseases. London (UK): OICA Inter- national; 1998. p. 1—28. (98) Wang MY, Liehr JG. Lipid hydroperoxide-induced endogenous DNA adducts in hamsters: possible mechanism of lipid hydroperoxide-mediated carcinogenesis. Arch Biochem Biophys 1995;316:38—46. (99) Nath RG, Chen H], Chang L, Amin S, Ocando JE, Chung FL. Detection of 4—hydroxynonenal-derived l,N2-6-propanodeoxyguanosine adducts in the liver DNA of untreated rat by a 32P-postlabeling method. Proc Am Assoc Cancer Res 1998;39:490. (100) Ansar Ahmed S, Dauphinee MJ, Montoya AI, Talal N. Estrogen induces normal murine CD5+ B cells to produce autoantibodies. J Immunol 1989; 142:2647—53. (101) Ansel J, Perry P, Brown J, Damm D, Phan T, Hart C, et a1. Cytokine 91 modulation of keratinocyte cytokines. J Invest Dermatol l990;94(6 Suppl):lOlS—107S. (102) Griffiths HR, Lunec J. Molecular aspects of free radical damage in in- flammatory autoimmune pathology. In: Aruoma OI, Halliwell B, editors. Molecular biology of free radicals in human diseases. London (U.K.): OICA International; 1998. p. 327—66. (103) Rosen GM, Pou S, Ramos CL, Cohen MS, Britigan BE. Free radicals and phagocytic cells. FASEB J 1995;9z200—9. (104) Tabibzadeh SS, Santhanam U, Sehgal PB, May LT. Cytokine-induced production of IFN—B2/IL—6 by freshly explanted human endometrial stro- mal cells. Modulation by estradiol-17B. J Immunol 1989;142:3134—9. (105) Mor G, Yue W, Santen RJ, Gutierrez L, Eliza M, Berstein LM, et al. Macrophages, estrogen, and the microenvironment of breast cancer. J Steroid Biochem Mol Biol 1998;67:403—1 1. (106) Lewis CE, Leek R, Harris A, McGee JO. Cytokine regulation of angio— genesis in breast cancer: the role of tumor-associated macrophages. J Leukoc Biol 1995;57:747c51. (107) Santner SJ, Pauley RJ, Tait L, Kaseta J, Santen RJ. Aromatase activity and expression in breast cancer and benign breast tissue stromal cells. J Clin Endocrinol Metab 1997;82:200—8. (108) Reed MJ, Topping L, Coldham NG, Purohit A, Ghilchik MW, James VH. Control of aromatase activity in breast cancer cells: the role of cytokines and growth factors. J Steroid Biochem Mol Biol 1993;44: 589—96. (109) Yue W, Wang JP, Hamilton CJ, Demers LM, Santen R]. In situ aroma— tization enhances breast tumor estradiol levels and cellular proliferation. Cancer Res 1998;58:927—32. (110) Esumi H, Tannenbaum SR. U.S.—Japan Cooperative Research Program: seminar on nitric oxide synthase and carcinogenesis. Cancer Res 1994; 54:297—301. (111) Floyd RA. The role of 8-hydroxyguanine in carcinogenesis. Carcinogen- esis 1990;11:1447—50. (112) Frenkel K. Carcinogenesis: the role of active oxygen species. In: Bertino JR, editor. Encyclopedia of cancer. San Diego (CA): Academic Press; 1997. p. 233—45. (113) Klein CB, Snow ET, Frenkel K. Molecular mechanisms in metal carci- nogenesis: role of oxidative stress. In: Aruoma OI, Halliwell B, editors. Molecular biology of free radicals in human diseases. London (U.K.): OICA International; 1998. p. 79—137. (114) Jansson G. Oestrogen-induced enhancement of myeloperoxidase activity in human polymorphonuclear leukocytes—a possible cause of oxidative stress in inflammatory cells. Free Radic Res Commun 1991;14:195—208. (115) Carlsten H, Holmdahl R, Tarkowski A, Nilsson LA. Oestradiol— and tes— tosterone—mediated effects on the immune system in normal and autoim— mune mice are genetically linked and inherited as dominant traits. Immu- nology 1989;68:209—14. (116) Carlsten H, Verdrengh M, Taube M. Additive effects of suboptimal doses of estrogen and cortisone on the suppression of T lymphocyte dependent inflammatory responses in mice. Inflamm Res 1996;45:26—30. (117) Cooke MS, Mistry N, Wood C, Herbert KE, Lunec J. Immunogenicity of DNA damaged by reactive oxygen species—implications for anti—DNA antibodies in lupus. Free Radic Biol Med 1997;22:151—9. (118) Malins DC, Polissar NL, Nishikida K, Holmes EH, Gardner HS, Gunsel- man SJ. The etiology and prediction of breast cancer. Fourier transform- infrared spectroscopy reveals progressive alterations in breast DNA lead— ing to a cancer-like phenotype in a high proportion of normal women. Cancer 1995;75:503—17. ([19) Arrigo AP, Kretz—Remy C. Regulation of mammalian gene expression by free radicals. In: Aruoma OI, Halliwell B, editors. Molecular biology of free radicals in human diseases. London (U.K.): OICA International; 1998. p. 183—223. (120) Suzuki K, Hei TK. Induction of heme oxygenase in mammalian cells by mineral fibers: distinctive effect of reactive oxygen species. Carcinogen- esis 1996;17:661—7. (121) O’Regan RM, Cisneros A, England GM, MacGregor JI, Muenzner HD, Assikis VJ, et al. Effects of the antiestrogens tamoxifen, toremifene, and ICI 182,780 on endometrial cancer growth. J Natl Cancer Inst 1998;90: 1552—8. (122) Wiseman H. Tamoxifen: Molecular basis of use in cancer treatment and prevention. Chichester (U.K.): John Wiley & Sons; 1994. p. 209. 92 (123) Jordan VC, editor. Long—term tamoxifen treatment for breast cancer. Madison (WI): The University of Wisconsin Press; 1994. p. 289. (124) Wei H, Frenkel K. Relationship of oxidative events and DNA oxidation in SENCAR mice to in viva promoting activity of phorbol ester-type tumor promoters. Carcinogenesis 1993;14:1 195—201. (125) Hertz R. The estrogen problem—retrospect and prospect. In: McLachlan 1A, editor. Estrogens in the environment. 11. Influences on development. New York (NY): Elsevier; 1985. p. 1—1 1. (126) Roy D, Palangat M, Chen CW, Thomas RT, Colerangle JC, Atkinson A, et a1. Biochemical and molecular changes at the cellular levels in response to exposure of environmental estrogen-like chemicals. J Toxicol Environ Health 1997;50:1—29. (127) Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human cancers. Nature 1998;396:643—9. (128) Barrett JC, Wong A, McLachlan JA. Diethylstilbestrol induces neoplastic transformation without measurable gene mutation at two loci. Science 1981;212:1402—4. (129) Roy D, Colerangle JC, Singh KP. Is exposure of environmental or indus- trial endocrine disrupting estrogen—like chemicals able to cause genomic instability? Frontiers Bioscience 1998;3zd913—21. (130) Roy D, Liehr JG. Estrogen. DNA damage and mutations. Mutat Res 1999;424:107—15. (131) Metzler M, Pfeiffer E, Schuler M, Rosenberg B. Effects of estrogen on microtubule assembly: significance for aneuploidy. In: Li JJ, Nandi S, Li SA, Gustafsson J, Sekely L, editors. Hormonal carcinogenesis II. New York (NY): Springer-Verlag; 1996. p. 193—9. (132) Metzler M, Pfeiffer E, Kohl W, Schnitzler R. Interactions of carcinogenic estrogens with microtubular proteins. In: Li JJ, Nandi S, Li SA, editors. Hormonal carcinogenesis. New York (NY): Springer—Verlag; 1992. p. 86—94. (133) Jones LA, Hajek RA. Effects of estrogenic chemicals on development. Environ Health Perspect 1995;103:63—7. (134) Hajek RA, Pathak S, Boddie AK, Jones LA. Aneuploidy of mouse cer— vicovaginal epithelium induced by prenatal estrogen treatment. Proc Am Assoc Cancer Res 1989;30:299. (135) Banerjee SH, Banerjee S, Li SA, Li J]. Cytogenetic changes in renal neoplasms and during estrogen-induced carcinogenesis. In: Li JJ, Nandi S, Li SA, editors. Hormonal carcinogenesis. New York (NY): Springer— Verlag; 1992. p. 247—50. (136) Banerjee SK, Banerjee S, Li SA, Li J]. Induction of chromosome aber— rations in Syrian hamster renal cortical cells by various estrogens. Mutat Res 1994;311:191—7. (137) Tsutsui T, Suzuki N, Maizumi H, Barrett JC. Aneuploidy induction in human fibroblasts: comparison with results in Syrian hamster fibroblasts. Mutat Res 1990;240:241—9. (138) Korah RM, Humayun MZ. Mutagenic and recombinagenic effects of diethylstilbestrol quinone. Mutat Res 1993;289:205—14. (139) Roy D, Abul-Hajj YJ. Estrogen-nucleic acid adducts: guanine is major site for interaction between 3,4-estrone quinone and COIII gene. Carcinogen- esis 1997;18:1247—9. ( 140) Yan 21, Roy D. Mutations of DNA polymerase [3 mRNA of stilbene estrogen—induced kidney tumors in Syrian hamster. Biochem Mol Biol Int 1995;37:175—83. (141) Hodgson AV, Ayala—Torres S, Thompson EB, Liehr JG. Estrogen-induced microsatellite DNA alterations are associated with Syrian hamster kidney tumorigenesis. Carcinogenesis 1998;19:2169—72. (142) Hodgson AV, Warnke PH, Liehr JG. Estrogen-induced microsatellite DNA instability during Syrian hamster kidney tumorigenesis. Proc Am Assoc Cancer Res 1999;40:510. (143) Paquette B. Enhancement of genomic instability by l7B—estradiol in mini- satellite sequences of X-ray—transformed mouse 10T1/2 cells. Carcino— genesis 1996;17:1221—5. (144) Boyd J, Takahashi H, Waggoner SE, Jones LA, Hajek RA, Wharton JT, et al. Molecular genetic analysis of clear cell adenocarcinomas 0f the vagina and cervix associated and unassociated with diethylstilbestrol ex- posure in utero. Cancer 1996;77:507—13. (145) De Benedetti VM, Welsh JA, Yu MC, Bennett WP. p53 mutations in hepatocellular carcinoma related to oral contraceptive use. Carcinogenesis 1996;17:145—9. (146) Tsutsui T, Barrett JC. Neoplastic transformation of cultured mammalian Journal of the National Cancer Institute Monographs No. 27, 2000 cells by estrogens and estrogen—like chemicals. Environ Health Perspect 1997;105 Supp 32619—24. (147) Jackson AL, Chen R, Loeb LA. Induction of microsatellite instability by oxidative DNA damage. Proc Natl Acad Sci U S A 1998;95:12468—73. NOTES Supported by Public Health Service grants R01CA25176, R01CA49917, and POlCA49210 (to E. Cavalieri and E. Rogan) (National Cancer Institute Journal of the National Cancer Institute Monographs N0. 27, 2000 [NCI]), R01CA37858 (NCI) and R01AG14587 (National Institute on Aging) (to K. Frenkel), CA63129 and CA74971 (to J. Liehr) (NCI), and CA52584 (to D. Roy) (NCI), National Institutes of Health, Department of Health and Human Services. E. Cavalieri and E. Rogan thank the key contributors to this research: D. Chakravarti, P. Devanesan, K.-M. Li, D. Stack, and R. Todorovic. K, Fren— kel acknowledges the contributions of M. Bosland and R. Yasuf from New York University School of Medicine (NY) for sharing their unpublished data. 93 3:,» A xv a. r »/ :Ybne ,‘miwnpxiwwu. 1, i _ . _ . c rs » g _ _ , . try», $15.36 V . , L 4:..wa , . . , , V , . . . . .m Chapter 5: Tissue-Specific Synthesis and Oxidative Metabolism of Estrogens Colin R. Jefcoate, Joachim G. Liehr, Richard J. Santen, Thomas R. Sutter, James D. Yager, Wei Yue, Steven J. Santner, Rajeshwar Tekmal, Laurence Demers, Robert Pauley, Frederick Naftolin, Gil Mor, Lev Berstein Estrogen exposure represents the major known risk factor for development of breast cancer in women and is implicated in the development of prostate cancer in men. Human breast tissue has been shown to be a site of oxidative metabolism of estrogen due to the presence of specific cytochrome P450 enzymes. The oxidative metabolism of 17B-estradiol (E2) to E2—3,4-quinone metabolites by an E2-4-hydroxylase in breast tissue provides a rational hypothesis to explain the mam- mary carcinogenic effects of estrogen in women because this metabolite is directly genotoxic and can undergo redox cy- cling to form genotoxic reactive oxygen species. In this chap- ter, evidence in support of this hypothesis and of the role of P4501B1 as the 4-hydroxylase expressed in human breast tissue is reviewed. However, the plausibility of this hypoth- esis has been questioned on the grounds that insufficient E2 is present in breast tissue to be converted to biologically significant amounts of metabolite. This critique is based on the assumption that plasma and tissue E2 levels are concor- dant. However, breast cancer tissue E2 levels are 10-fold to 50-fold higher in postmenopausal women than predicted from plasma levels. Consequently, factors must be present to alter breast tissue E2 levels independently of plasma concen- trations. One such factor may be the local production of E2 in breast tissue through the enzyme aromatase, and the evi- dence supporting the expression of aromatase in breast tis- sue is also reviewed in this chapter. If correct, mutations or environmental factors enhancing aromatase activity might result in high tissue concentrations of E2 that would likely be sufficient to serve as substrates for CYPlBl, given its high affinity for E2. This concept, if verified experimentally, would provide plausibility to the hypothesis that sufficient E2 may be present in tissue for formation of catechol metabo- lites that are estrogenic and which, upon further oxidative metabolism, form genotoxic species at levels that may con- tribute to estrogen carcinogenesis. [J Natl Cancer Inst Monogr 2000;27:95—112] The major known risk factors for development of breast can- cer in women are associated with prolonged exposure to in- creased levels of estrogen. Estrogen and testosterone are also thought to be involved in the development of prostate cancer in men (see Chapter 2). From information presented in Chapters 3 and 4, it is clear that evidence is building for a role of oxidative metabolites of 17B-estradiol (E2) and/or estrone (E1), particu- larly the catechols, in breast cancer. New information implicates the catechols as signaling molecules with relative binding affini- ties for the human estrogen receptor that are equal to or greater than E2 itself ( I ). It is also clear that, upon further oxidative metabolism, the catechol metabolites can form quinones that can directly form adducts with glutathione and guanine and adenine Journal of the National Cancer Institute Monographs No. 27, 2000 bases in DNA (see Chapter 4). In particular, the 3,4-quinone forms a depurinating adduct with guanine and adenine, leaving an abasic site with mutagenic potential. In addition, as will be discussed in more detail below, the catechols are capable of redox cycling, a process accompanied by the generation of re- active oxygen species able to cause oxidative damage to lipids, proteins, and DNA. A critical issue in relation to estrogen and the potential con- tribution of the catechol estrogen (CE) metabolites to breast cancer is their source. Estrogens themselves and their oxidative metabolites are formed by the activities of various cytochromes P450 (CYPs). These enzymes have dual functions, the biosyn— thesis and/or inactivation of physiologic regulators on the one hand and the metabolism of environmental chemicals on the other. Natural processes in which they participate include the synthesis of estrogen from cholesterol, which involves multiple, very specific CYP enzymes, typically compartmentalized into different organelles, cells, or organs. The final process in the synthesis of estrogens involves a three-step oxygenation reaction catalyzed by a single P450, CYP19 (aromatase), which converts an androgen (testosterone or androstenedione) to an estrogen (E2 or E1, respectively). Numerous low-specificity CYPs are in- volved in the oxidative inactivation and clearance of these same steroids, drugs, and environmental pollutants. The low specific- ity of these P450 cytochromes, which are most abundant in liver but are also found in most cells, results in the conversion of these chemicals to multiple products with increased hydrophilicity and functional groups for subsequent metabolism. For some chemi— cals, particularly those that contain olefinic double bonds or aromatic rings, these products may include Chemically reactive metabolites that can cause DNA damage and thereby cause er- rors during replication, which result in mutations. The metabo- lism of steroids, in conjunction with conjugation reactions (sul- fation, glucuronidation, and methylation), may contribute to lowering serum and cellular levels of steroidal parent hormones Affiliations of authors: C. R. Jefcoate, Department of Pharmacology, Univer- sity of Wisconsin—Madison; J. G. Liehr, Stehlin Foundation for Cancer Re— search, Houston, TX; R. J. Santen, W. Yue, Division of Hematology, Oncology, and Endocrinology, Cancer Center, University of Virginia Health Science Cen— ter, Charlottesville; T. R. Sutter, W. Harry Feinstone Center for Genomic Re— search, University of Memphis, TN; J. D. Yager, Division of Toxicological Sciences, Department of Health Sciences, The Johns Hopkins University School of Hygiene and Public Health, Baltimore, MD; S. J. Santner; R. Pauley, Wayne State University, Detroit, MI; R. Tekmal; Emory University, Atlanta, GA; L. Demers, Pennsylvania State University, Hershey; F. Naftolin, G. Mor, Yale University, New Haven, CT; L. Berstein, Petrov Institute, St. Petersburg, Russia. Correspondence to: James D. Yager, Ph.D., Division of Toxicological Sci— ences, Department of Environmental Health Sciences, The Johns Hopkins Uni- versity School of Hygiene and Public Health, 615 North Wolfe St., Baltimore, MD 21205 (e—mail: jyager@jhsph.edu). See “Notes” following “References.” © Oxford University Press 95 and, hence, to altered regulation of biologic activity. Identifica- tion of any new P450 cytochrome raises the question of which of these functions underlie its pattern of expression. It is most important to appreciate that the various P450 enzymes show tissue-specific expression; i.e., forms expressed in liver may not reflect those expressed in breast or prostate tissue. The implica- tions of this are potentially enormous, since metabolites (wheth- er of exogenous or endogenous compounds) will show tissue specificity. The mammary gland is involved in steroid synthesis (cer- tainly aromatase and possibly other steps), steroid metabolism (hydroxylation and conjugation), and xenobiotic metabolism (li- pophilic molecules stored in fat and excreted in the milk). As will be discussed in detail below, an investigation (2) indicated that substantial levels of estrogen arise from aromatase activity localized in breast tissue. This is present in both breast epithelia and fibroblasts. Thus, estrogen biosynthesis occurs in the target tissue. Furthermore, the estrogen oxidative metabolites are formed by various CYPs. Human breast and breast tumor tissue express various CYPs. Among these is CYPlBl. As will be discussed in detail below, CYPlBl is a catalytically efficient estrogen 4—hydroxylase. CYPlBl is present in fibroblasts, as well as in epithelial cells (3—6); thus, it is co-localized with aromatase, the enzyme producing estrogens. This co-localization of CYPlBl and aromatase means that the effective concentra- tion of the 4-hydroxylated product is much higher than that indicated by the circulating levels. Human breast fibroblasts contain estrogen receptors that regulate cell growth in vitro, but estrogen receptors are present in only 5%—10% of the epithelial cells. A study (7) indicated that aromatase is elevated by pros- taglandin E2 (via cyclic adenosine monophosphate [cAMP]) and by various cytokines. However, very little is known about the regulation of CYPlBl and estrogen receptors in these cells. Expression of each of these genes is also dependent on the extracellular matrix and growth factor milieu that surrounds these cells. Several presentations in this monograph raise the issue of oxidative stress (see Chapters 3 and 4), mediated by the 4-CE/0-quinone redox cycling. One possibility is that this may also contribute to signaling that promotes aromatase expression. The remainder of this chapter will provide greater detail regard- ing the expression and potential roles of aromatase, CYPlBl, and CE, particularly 4-hydroxy—E2 (4—OHE2), in breast cancer. POTENTIAL ROLE OF AROMATASE OVEREXPRESSION IN THE GENESIS 0F BREAST CARCINOMA Estrogens and Risk of Breast Cancer Evidence from multiple sources suggests that estrogens are involved in the genesis of breast carcinoma. Administration of exogenous estrogens causes breast cancer in rodents (8). Breast tumors induced by the potent carcinogen 7,12-dimethyl— benz[a]anthracene can be delayed or prevented by castration and administration of anti—estrogens (9). Aromatase inhibitors pre- vent the Spontaneous development of breast tumors in aging Sprague-Dawley rats (10). By the age of 2 years, approximately 50% of these animals develop either benign or malignant breast lesions (Fig. l) (10). Increasing doses of the aromatase inhibitor fadrozole inhibit these tumors in a dose-dependent fashion. No- tably, 1.25 mg of fadrozole/kg of body weight daily causes 100% inhibition. Taken together, these data provide strong evi- dence that estrogens are involved in the carcinogenic process, resulting in breast neoplasms in animals. 96 60 50- 4o- 30— 20— 10- o- . 0 Fig. 1. Inhibition of Spontaneous benign and malignant breast tumors in aging Sprague—Dawley rats with administration of the aromatase inhibitor fadrozole. No tumors appeared in animals receiving the 1.25 mg/kg dose. Figure adapted from the data of Gunson et al. (10). % with tumor I l l 0,25 1.25 0.05 Fadrozole (mg/kg) Indirect evidence in women also supports a role for estrogen in the genesis of breast cancer. Early menarche and late meno- pause are associated with an increased risk of breast cancer. These factors result in prolonged exposure of the breast to es- trogen. Prospective epidemiologic Studies (1 1—14) detected an increase in the plasma levels of E2 in women developing breast cancer 5 or more years later. Obesity is also associated with a greater risk of breast cancer ( I 5). This relationship might also be explained by increased estrogen production, since the degree of obesity correlates linearly with total—body aromatase activity (16). Aromatase catalyzes the rate—limiting step in estrogen bio— synthesis, the conversion Of androgens to estrogens. Additional indirect evidence regarding estrogens and breast cancer derives from analyses of the rate of breast cancer in women receiving estrogen replacement therapy during the menopause. More than 50 studies are now available to examine this relationship [reviewed in (17)]. Six meta-analyses have pooled this information, and a critical review of these data al- lows several points to be made. The first is that no randomized, controlled, double-blind studies have been conducted to demon- strate conclusively that estrogen replacement therapy during the menopause increases the risk of breast cancer. Only ObSCI‘Va— tional studies are available. Secondly, while susceptible to sev- eral biases, most of these Studies show an increased risk of breast cancer with the use of estrogen replacement therapy for a period of more than 5—10 years. The relative risk of breast cancer under these circumstances increases by about 30%. Thirdly, the abso- lute increased risk is small, approximating one additional breast cancer case in 100 women of age 50 years who have taken estrogen for at least 10 years. The conclusion that these data prove that estrogen replace- ment therapy is associated with an increased risk of breast cancer is controversial. However, as current studies evolve, the evi— dence increases. For example, the Nurses’ Health Study (18) now involves more than 725 550 patient-years of observation over a 10-year period. In this study, the risk of breast cancer increased with more than 5 years of current estrogen use in women of all ages over 50 years. This cohort study allows ad- justment for most, but not all, confounding factors and provides the most convincing evidence of a relationship between estrogen Journal of the National Cancer Institute Monographs No. 27, 2000 ingestion and breast cancer. The definitive study, conducted as part of the Women’s Health Initiative, involves a randomized, placebo-controlled trial. The results of this study will not be available until midway into the first decade of the 21St century. Correlations between estrogen levels and subsequent risk of breast cancer have not, until recently, been positive. Earlier stud- ies measured levels of urinary estrogens, plasma estrogens, or the free fraction of E2 with insensitive or relatively nonspecific methods. Conflicting results were reported, and the common View was that estrogens were not elevated in women who would later develop breast cancer (19,20). Recently, however, Toniolo et al. (11) and three other groups (12—14) demonstrated in- creased levels of E2 and its precursor, testosterone, in women found to develop breast cancer prospectively 5 or more years later. While not every group could replicate these results (19,20), the majority of data support such a relationship. Further evidence regarding estrogen production and breast cancer risk has been provided by experiments demonstrating that a reduction in estrogen production in women reduces the inci- dence of breast cancer (21,22). Data from two classic studies (21,22) are consistent with such an effect (Fig. 2). One of these studies (22) examined the incidence of breast cancer in a group of women who had undergone bilateral oophorectomy before A . Breast Cancer after Surglcal Menopause 1.4 1:.» 1.2 (D S 1.0 (D $0.8 E '8 0.6 5 0.4 ‘8 0.2 I I 0.0 <3 13-? 8-12 13-17 18-22 23+ years after surgery B . Breast Cancer after Surglcal Menopause 1 g 0.8 ‘6 3 0.6 1.: E 0.4 3 0.2 .0 O 0 unilateral bilateral oophorectomy at < age 40 Fig. 2. A) Observed over expected rates of breast cancer over time after surgical menopause. Adapted from the data of Trichopoulos et al. (21). B) Observed over expected rates of breast cancer in women undergoing unilateral and bilateral oophorectomy. Adapted from the data of Feinleib et a1. (22 ). Journal of the National Cancer Institute Monographs No. 27, 2000 age 35 years. The control group consisted of women subjected to a unilateral oophorectomy at the same ages. The end point of the study was the ratio between observed and expected breast can- cers in these two groups of women. After a period of 20 years, the women undergoing bilateral oophorectomy had a 75% re- duction in the incidence of breast carcinoma (Fig. 2, A). In the other study (21), with a similar design, the decrease in breast cancer incidence over that expected gradually declined as a func~ tion of time after oophorectomy (Fig. 2, B). Although these studies were also subject to bias, they provide compelling evi- dence that ovarian factors, and presumably E2, are involved in the genesis of breast carcinoma. Sources of Estrogen The sources of E2 production in women are important to consider, since overproduction may result from altered regula- tion at any site. Estrogen can be made in several tissues. Aro- matase, the enzyme catalyzing the rate-limiting step in estrogen biosynthesis, is widely present throughout the body. The pre- menopausal ovary, which contains the highest level of aro- matase, except for the placenta, is the major source of E2 during the premenopausal years. Peripheral adipose tissue also contains aromatase and is a major source of this enzyme, since the mass of adipose tissue (particularly in obese women) is substantial. Breast tissue itself contains aromatase, both in its fatty compo— nents and in its epithelial cells, and can synthesize estrogen in sztu. Importance of In Situ Aromatase in Breast Tissue Emerging evidence suggests that estrogen produced in situ, as opposed to E2 made in other tissues and delivered to the breast via an endocrine mechanism, plays a major biologic role in breast physiology. Several lines of evidence support this con— cept. Demonstration of the aromatase enzyme and its messenger RNA in breast tissue by immunohistochemical and molecular biologic techniques, studies in nude mice to show that the amount of estrogen made locally causes biologic effects, and clinical studies of aromatase inhibitors in patients provide proof of the importance of in situ production of estrogen in breast tissue. Several groups of investigators (23—25) over the past few years have demonstrated the aromatase enzyme in breast tissue. Both radiometric and product isolation methods demonstrated that tritiated androgens could be converted to estrogens in hu- man breast cancer tissue as evidence of aromatase activity. The biologic significance of this finding was initially questioned, since the amount of enzyme present, compared with that in placenta or ovary, was low (26). In this regard, other investiga- tors postulated that focal concentrations of aromatase in selected cell populations might be high, but overall activity in breast cancer tissue might be low. This might occur because of the presence of fibrous or other tissues in the tumor, which would dilute the concentration of enzyme in the tumor overall. To examine this possibility, immunohistochemical studies were used to detect aromatase in individual breast cancer cells. Re- sulting data (27—29) demonstrated high levels of aromatase staining in individual cells, supporting the concept that aro- matase might act in an autocrine or paracrine fashion in breast tissue (Fig. 3). Controversy exists at present whether aromatase activity is 97 Fig. 3. Immunohistochemical staining for aromatase in human breast cancer. A) Section of human placental villus stained with the Harada antibody. B) Control section of human placenta stained with an irrelevant antibody, anti-neuropeptide Y. C) Section of human breast tumor stained with the anti-aromatase antibody. D) Control section of human breast tumor stained with an irrelevant antibody, predominantly in epithelial cancer cells or in the surrounding stromal cells (30,31). Certain monoclonal antibodies, used in conjunction with antigen retrieval techniques, suggest that the majority of aromatase is in epithelial cells (30). Monospecific polyclonal antibodies, utilized on the same tissue sections, show a preponderance of aromatase activity in stromal cells (27,32,33). To provide additional data regarding stromal aro- matase, Santen and his group (33) grew stromal cells isolated from breast cancer lesions as well as from benign tissue sur- rounding the tumors. They found that isolated stromal cells from breast cancer tissue contain high levels of aromatase enzyme when stimulated by dexamethasone, phorbol esters, and CAMP in combination. Aromatase enzyme activity, assessed by a ra— diometric assay, increases by nearly four logs in response to this combination of enhancers, and message increases 30-fold (Fig. 4). Other investigators have found that aromatase message levels are higher in areas of breast cancer with high stromal cell content than in areas with low content. Quantitative assessment using a histologic scoring system and immunohistochemistry detected average H-scores of 13 (range, 0—45) for stromal cells and 4 (range, 1.4—16) for stroma in 26 human breast tumors (27). This technique takes into account the relative abundance of stromal 98 anti-neuropeptide Y. E) Section of human breast tumor stained with the anti- aromatase antibody showing stromal spindle cells positively stained. F) Photo- micrograph of greater magnification from the periphery of a group of tumor epithelial cells showing positive staining of stromal spindle cells. Reprinted from (27) with the permission of authors and publisher. and epithelial cells in the tumor (27). Considered together, these data support the biologic importance of aromatase in breast can- cer tissue and suggest that stroma may contribute to a greater extent to this process than epithelial cells. Further support for the importance of aromatase in breast tissue itself derives from studies in a nude mouse model devel- oped by Yue et a1. (34). Using this model, these investigators (34) examined the relative importance of uptake from plasma versus local E2 synthesis in breast tissue. This model involves the use of MCF-7 breast cancer cells transfected stably with aromatase (A+) that are implanted on one side of castrated nude mice. On the other side, sham-transfected MCF-7 cells (A—) are implanted. Administration of the aromatase substrate andro— stenedione causes no growth stimulation of aromatase-negative cells (Fig. 5). This important control demonstrates that no aro- matase activity is present in non-breast tissue in the mouse. Aromatase-positive cells implanted on the other side of the same animals are stimulated to grow by androstenedione, providing evidence of the biologic effect of aromatase present locally in the breast. The aromatase inhibitor 4-hydroxyandrostenedione blocked this growth effect. One can conclude that the growth stimulation cannot have occurred as a result of peripheral con- Joumal of the National Cancer Institute Monographs N0. 27, 2000 A. Benign breast 4000 - ' 3000 - " 2000 - ‘ 1000 - ' O 8. Breast tumor 8000 6000 Aromatase Activity (pmol/g protein/hr) 4000 2000 O ‘ Control PDA+cAMP DEX DEX+ DEX+ PDA+CAMP PDA+cAMP 44:65 20267 Fig. 4. Regulation of aromatase enzyme activity in benign breast and breast tumor stromal cell cultures. A) Enzyme activity in myofibroblasts isolated from human breast tissue. Basal activity was compared with phorbol ester plus cyclic adenosine monophosphate (cAMP) (100 nmol/L phorbol diacetate [PDA] and l nmol/L CAMP), 100 nmol/L dexamethasone (DEX) alone, the combination of DEX, PDA, and CAMP, and the addition of the aromatase inhibitor letrozole (l umol/L) to the combination treatment (Ciba Geigy Substance 20267). The re- sults of cultures from three patients were used. Data reprinted from (33) with the permission of authors and publisher. 700 (D E 2 600 — S /A A+ ‘6 500 — E / 3 T, 400 — f 25 .s 300 ~ / 0 A/ U, E 200 A x/ 4: 4/ o X 4 A7 =,\° 100 “+_._..——.—+—./ l l l | l | O 10 20 30 40 50 60 70 Days of androstenedione treatment Fig. 5. Change in tumor volumes in aromatase—transfected MCF—7 cells (A+) and in sham—transfected cells (A—). Data reprinted from (34) with the permission of authors and publisher. version to E2, which would have stimulated the aromatase— negative cells as well. As expected, the levels of E2 in aro- matase-positive tumors markedly exceeded those in aromatase- negative tumors (Fig. 6). The relative importance of in situ production versus uptake of E2 from plasma was then examined. Silastic implants designed Journal of the National Cancer Institute Monographs No. 27, 2000 A 1500 UJ U) 2' 1200— m a.) E {:3 900a B) e E 5004 2 ma 5 300— % E .3 o A— Fig. 6. Tumor estradiol (E2) concentrations in aromatase-transfected (A+) and sham-transfected cells (A—) implanted into and grown in nude mice. Data re— printed from (34) with the permission of authors and publisher. SE = standard error. to produce plasma estrogen levels ranging from 5 to 20 pg/mL were implanted into castrate animals to evaluate the effect of E2 uptake. Androstenedione was administered to others to examine in situ production. With this experimental system, tissue E2 lev- els and tumor growth were higher as a result of in situ aroma- tization than from plasma delivery of estrogen [data shown in (34)]. This series of experiments in mice supports the hypothesis that an important determinant of tissue E2 levels is local pro- duction in the breast. If this hypothesis is correct, the level of E2 produced in breast tissue may be the most important determinant of EZ-induced carcinogenesis. This conclusion is supported by the direct measurement of aromatase activity with elegant iso- topic techniques in human breast tumors by Reed et al. (35). These investigators demonstrated that 83% i- 9% (standard de- viation) of tumor estrogen levels resulted from in Situ aromatase in four of six tumors. In the other two tumors, no estrogen synthesis in the breast itself could be demonstrated. Although speculative, the hypothesis of in situ E2 synthesis could explain the relatively poor correlations between use of estrogens for menopausal replacement therapy and breast cancer risk. This might also explain why it has been difficult to demonstrate higher estrogen levels in women who will later develop breast cancer. If the local synthesis hypothesis is correct, measurement of the concentration of E2 in breast tissue itself would be the most precise predictor of later development of breast cancer. This concept is supported by the fact that plasma and tissue E2 levels are not well correlated. The ratio of tissue—to—plasma E2 levels in premenopausal women approximates l : 1, whereas in postmenopausal women the ratio is 10—50:] (36). Taken to- gether, these data suggest that certain factors present in breast tissue can influence local production of estrogen and that these may be the prime determinants of tissue estrogen concentrations. If these concepts are correct, elevated plasma levels of estrogen would be associated with high tissue concentrations in some, but not in all, patients and breast cancer risk might only be increased in those with high tissue levels. Mechanism of Carcinogenesis After considering the sources of E2 available for stimulating breast tissue, one must consider how E2 causes breast cancer. The predominant theory at the present time relates to effects of estrogen on cell growth (37). Enhanced cell proliferation, in— 99 duced either by endogenous or by exogenous estrogens, in- creases the number of cell divisions and, by inference, the pro- portionate number of mutations. With an enhanced rate of proliferation, the time available for DNA repair is reduced. In addition, single-stranded DNA, present during cell division, is more susceptible to damage than double-stranded DNA, and gene duplication can occur. Another current theory, discussed in Chapters 3 and 4 and in the following section, is that estrogens can be metabolized to genotoxic products. These two current theories of enhanced cell proliferation and genotoxic metabolites are not mutually exclu- sive but could act in an additive or even synergistic fashion. For example, DNA damage originating from CBS would be propa- gated more rapidly by increased cellular proliferation, and in- sufficient time might be available for DNA repair. Additional data will be required to determine the precise interactions be- tween these two pathways of carcinogenesis. The major critique of the genotoxic metabolite theory is that estrogen levels are not sufficiently high to produce biologically relevant amounts of these metabolites. This critique, however, is based on an analysis of plasma E2 levels and not tissue levels. If E2 is synthesized locally in breast tissue, the levels would be higher than expected from plasma concentrations. This concept is supported by the fact that E2 concentrations in breast cancers from postmenopausal women are as high as those from pre- menopausal women. This is surprising, since the levels of E2 in the plasma of premenopausal women are 10- to 50-fold higher than those in the plasma of postmenopausal women (36). Hypothesis of Aromatase Overexpression As a means of integrating these data, Santen et al. (personal communication) have postulated that aromatase overexpression in breast tissue may be a cause of breast cancer. Through aro- matase overexpression, tissue levels of E2 would be sufficiently high to undergo metabolism to biologically important quantities of genotoxic metabolites. Four separate models of aromatase overexpression and breast cancer have been well characterized and provide strong support for this hypothesis. Three involve the hyperplastic alveolar nodule (HAN) model systems. Zhang and Medina (38) have developed a series of transplantable breast explants that grow in the mammary fat pads of highly inbred strains of mice. Two of these are induced by carcinogens and are called the C4 and C5 HAN. One is induced by hormonal stimu- lation and is called the D2 HAN. Upon serial passage in mam- mary fat pads, these lesions develop frank cancer with an inci- dence that approaches 90% under certain conditions. Each of these HAN models has been shown to have an insertional mu— tation called Int 5. This mutation has now been characterized and involves the insertion of a portion of the long terminal repeat of the mouse mammary tumor virus into genomic DNA (39—41). Of great interest is the fact that, in each of these models, the insertion is into the 3’ untranslated region of the tenth exon of the aromatase gene. This results in overexpression of the aro- matase gene and, by inference, in tumor development. The fourth model is a transgenic mouse model in which aromatase is overexpressed, predominantly in mammary tissue (42). These animals develop atypical ductal hyperplasia, a type of lesion that predicts an increased rate of breast cancer development when found in patients. They also develop fibroadenomas, typical duc- tal hyperplasia, and dysplasia, lesions that are not associated with an increased risk of breast cancer in women. Rarely do the aromatase-transfected animals develop frank breast cancer. 100 Evidence of Aromatase in Benign Breast Tissue Several laboratories have obtained evidence that malignant breast tissue contains both aromatase message and enzyme ac- tivity. However, if aromatase overexpression is important in the genesis of breast cancer, this enzyme must be present in benign breast tissue as well. To assess this possibility, core and exci- sional biopsy specimens containing atypical ductal hyperplasia were evaluated with an immunohistochemical method using a monospecific polyclonal antibody (43 ). This technique revealed aromatase immunohistochemical staining in both stromal and epithelial cells contained in the hyperplastic lesions. In the sur- rounding normal tissue, aromatase staining was present pre- dominantly in glandular epithelial cells but to a lesser extent in stroma. Surprisingly, macrophages with substantial aromatase activ- ity were also detected (43). This finding led to an extensive series of experiments to verify that macrophages indeed express aromatase. THP-1 cells, a malignant cell line that can be differ- entiated into macrophages upon exposure to phorbol esters (44 ), were used. These cells contained aromatase enzyme activity with levels close to those found in human placenta. The aro- matase inhibitor letrozole completely inhibited this activity. Conditioned media from these cells, exposed to the aromatase substrate testosterone, stimulated the growth of EZ-responsive MCF-7 indicator cells. As evidence of specificity, growth of indicator cells could be blocked with letrozole or with the pure anti-estrogen ICI 182780. As further evidence of aromatase expression in macrophages, human monocytes were examined, basally and after differentia- tion into macrophages, in tissue culture with the addition of phorbol esters. These cells contained aromatase message when differentiated into macrophages but not under control condi- tions. Finally, monoclonal antibodies specific for macrophages were used to demonstrate, by double labeling, that the cells in the breast that contained aromatase were in fact macrophages. The production of E2 by macrophages is of further interest because chemokines regulating tissue invasion by macrophages are also controlled by estrogen. The levels of macrophage che- mokine 1 (MCP-l) are lowered by E2 in MCF—7 breast cancer cells and in other tissues (see Chapter 8). It is interesting that a classical negative feedback loop could exist whereby E2 lowers MCP-l, which would result in a decrease in invasion of tissue by macrophages. This would result in a lower production of E2 in that tissue. As a consequence, the levels of MCP-l would in- crease and the macrophages would again be stimulated to invade the tissue. Since breast tumors contain a substantial number of macrophages, their contribution to local E2 production could be biologically important. These data suggest that breast tissue can make E2 from epi— thelial cells, from stroma, and from macrophages that infiltrate normal tissue. Potentially, either one of these three cell types could overexpress aromatase and provide sufficient amounts of E2 locally to allow conversion to genotoxic quinone metabolites. Several examples of aromatase overexpression are known to exist. A rat Leydig cell tumor overexpresses aromatase through activation of a CAMP-dependent enhancer of aromatase (45). The breast tissue of goats is the major source of aromatase prior to parturition, and bilateral mastectomy delays the time of par- turition (46). The Sebright bantam syndrome is caused by aro- matase overexpression, which feminizes the feather pattern of roosters and gives them the phenotypic appearance of chickens Journal of the National Cancer Institute Monographs No. 27, 2000 (47). Familial causes of aromatase overexpression occur in pa- tients, resulting in prepubertal gynecomastia in boys and preco- cious thelarche and/or macromastia in girls. The Peutz-Jeghers syndrome is characterized by aromatase overproduction and leads to testicular tumors in boys and ovarian tumors in girls. Each of these examples provides evidence that aromatase over- expression is somewhat common in animals and in patients (48). Mechanism of Aromatase Overexpression A variety of potential mechanisms could result in aromatase overexpression. Aromatase transcription is regulated by multiple enhancers, including CAMP, phorbol esters, dexamethasone, prostaglandin E2, transforming growth factor-B and interferon gamma among others (49). Fibroblasts isolated from breast tu— mors as well as from benign tissue surrounding the tumors con— tain aromatase. The activity of this enzyme and its message can be stimulated up to lOOOO—fold in cell culture with addition of phorbol esters, CAMP, and dexamethasone (Fig. 4) (41). Acti- vating mutations involving any of these or other steps could result in aromatase overexpression in breast tissue. Simpson and co-workers (50) have postulated that prostaglandin E2 may be important in this process and have pointed out that use of non- steroidal anti-inflammatory agents is associated with a decreased incidence of breast cancer in women. These agents are known to block prostaglandin E2 production and putatively could decrease breast cancer through this mechanism. Prevention of Breast Cancer If aromatase overexpression were an etiologic factor for breast cancer, third-generation aromatase inhibitors might be used for prevention. In premenopausal women, it would be pos— sible to block estrogen production in breast tissue without af- fecting ovarian E2 synthesis. The ovary is relatively resistant to aromatase inhibitors because of the extremely high levels of androstenedione present as substrate in the ovary (51). While these concepts are speculative, further evaluation of aromatase expression in various premalignant breast lesions is warranted. METABOLIC ACTIVATION OF ESTROGENS BY 4-HYDROXYLATION Role of Metabolism in Estrogen-Induced Cancer The oxidative metabolism of estrogens has been studied in detail as part of investigations of the regulation and control of hormonal action and has been reviewed by Zhu and Conney (52). In this section, estrogen metabolism will be discussed only to the extent that it affects induction of tumors by this hormone. The metabolic activation of estrogens has increasingly been sus— pected to play a role in the carcinogenic process (53—55) because the modulation of metabolic oxidation affects Ez—induced tumor incidence in a rodent model, the kidney tumor in male Syrian hamsters, in a way that is not consistent with previous hypoth- eses. Estrogens have previously been postulated to act in hor- mone-associated cancers, including breast cancer, primarily by estrogen receptor—mediated proliferation of cells mutated by spontaneous replication errors [reviewed by Feigelson and Hen- derson (56)]. According to this hypothesis, inhibition of estrogen metabolism should enhance tumor formation, since estrogen me- tabolites are less hormonally active than the parent E2 or E1. In contrast, OL-naphthoflavone, an inhibitor of estrogen hydroxyl- ation (57,58), and ascorbic acid (vitamin C), a reducing agent Journal of the National Cancer Institute Monographs No. 27, 2000 known to reduce estrogen quinone intermediates to their hydro- quinones (59), either completely or partially inhibited kidney tumor induction in hamsters by E2 (Table 1) (59—61). Neither of these chemicals is known to have estrogen agonist or antagonist activities and, thus, could not have interfered with the estrogen receptor-mediated proliferation of mutated cells. In addition, 170L-ethinylestradiol, a poor carcinogen in the hamster kidney tumor model (62 ), is converted to catechol metabolites at much lower rates than E2 (63). All of these data are consistent with the conclusion that metabolic activation of estrogens is necessary for the induction of tumors by estrogenic hormones (53—55). This metabolic activation of estrogens, as a necessary part of the tumorigenesis process, has been proposed by analogy to meta- bolic activation of carcinogenic chemicals such as benzo[a]py- rene or other carcinogenic hydrocarbons (64). Moreover, the conversion to catechol metabolites and their further metabolism to quinone and semiquinone intermediates were explored, since such metabolites have been shown to play a role in DNA damage induced by other carcinogens (65). Formation of CEs CEs, 2- or 4-hydroxylated E2 or E], are the focus of metabolic research in the context of estrogen-induced cancer because these compounds are major oxidized metabolites of estrogenic hor- mones in most mammalian species (66,67) and are precursors to reactive intermediates (53—55). For instance, catechols including CEs may be oxidized chemically or by enzymatic processes to semiquinones, which are free radicals, and further to reactive quinone intermediates (68—71). These semiquinone/quinone spe- cies react with nucleophilic sulfur- or nitrogen-rich endogenous chemicals including nucleic acids (71—74). Therefore, the for— mation and metabolic activation of CBS, specifically 4-hydrox- ylated estrogens, are described below. Catechol Formation by Aromatic Oxidation of Estrogens 2-Hydroxylation of steroidal estrogens is the major metabolic oxidation of estrogenic hormones in most mammalian species (Fig. 7) (66,67). For instance, in human or hamster livers, the 2-hydroxylation is catalyzed by CYP 3A isoforms, whereas CYP 1A isoforms represent the predominant estrogen-2—hydroxylase activity in extrahepatic tissues (57, 75—77). These estrogen-2- hydroxylases convert E2 to approximately 80%—85% 2-hy- droxyestradiol (2-OHE2) and, because of a lack of specificity of the enzyme(s), 15%—20% of 4-OHE2 (78), as shown, for in- stance, for liver in Table 2. In contrast, specific estrogen-4- hydroxylase(s) that convert E2 mainly to 4-OHE2 (3) have been identified (78—80) in those organs of rodents in which chronic estrogen exposure induces malignant or benign tumors, such as in hamster kidney (62), mouse uterus (81), or rat pituitary (82). Table 1. Reduced incidence of estrogen-induced kidney tumors in male Syrian golden hamsters by modulation of estrogen metabolism Kidney tumors, Reference Treatment % tumor—bearing animals No. None 0 (62) 17B—Estradiol 100 ( 62 ) 17B-Estradiol + oc-naphthoflavone 0 (61) 17B-Estradiol + ascorbic acid 50 (60) (vitamin C) 170L-Ethinylestradiol 10 ( 62 ) 2-F1uoroestradiol 0 (84) 101 Fig. 7. Formation and catabolism of cat- echol estrogens (CEs). Estradiol is con- verted to 2-hydroxyestradiol (2-OHE2) by reduced nicotinamide adenine di- nucleotide phosphate (NADPH)—depen~ dent estrogen—Z—hydroxylases, which form a small amount of 4—hydroxyestra— diol (4-OHE2) by a lack of specificity of the enzymes. Conversely, NADPH- dependent estrogen—4—hydroxylases form preferentially 4-OHE2 and a small amount of 2—OHE2. In addition, organic hydroperoxide (OHP)-dependent estro— gen-2/4-hydroxylase(s) form approxi— E2 mately equal amounts of both catechol products. Both 2—OHE2 and 4—OHE2 may be converted to the corresponding methoxyestrogens (2-OC1-I3E2 and 4-OCH3E2, respectively) catalyzed by catechol—O—methyltransferase or other conjugate metabolites catalyzed by uri— dine diphosphate (UDP) glucuronyl transferases or other conjugating en— zymes. These phase II metabolites, in turn, may be converted back to CEs by demethylating enzyme activities, gluc— uronidase, or other enzyme activities. NADPH NADPH Other Conjugates Table 2. Metabolic conversion of 10 uM 17B-estradiol (E2) to catechol estrogens by microsomal preparations of target organs of estrogen—induced cancer 4—Hydroxy—E2 Tissue 2-Hydroxy—E2* 4—Hydroxy—E2* 2-Hydroxy—E2 Reference No. Hamster liver: control 365 52.4 0.14 (78) Hamster kidney Control 13.4 7.2 0.5 Ez—treated 2.2 6.1 2.8 CD—l mouse uterus: control 0.1 1.3 13.0 (80) Sprague—Dawley rat pituitary Control 0.03 0.29 9.7 (79) Ez-treated 0.08 0.24 3.0 *Values in column = picomoles per milligram protein per minute. The specific formation of 4-hydroxylated estrogens is important because, in the hamster kidney tumor model, 4-OHE2 is as car- cinogenic as E2, whereas 2-hydroxylated estrogens did not in- duce any tumors (83—85). In addition to the specific reduced nicotinamide adenine dinucleotide phosphate (NADPH)- dependent estrogen 2- or 4-hydroxylation, an organic hydroper- oxide-dependent estrogen 2- and 4-hydroxy1ase activity has been detected, which produces both catechols in roughly equal amounts (78,86). In humans, the predominant conversion of E2 to 4—OHE2 has been detected in microsomes of uterine myometrium and fi— broids—i.e., benign uterine myomas (87)—and in benign and malignant mammary tumors ( 88). In addition, a specific estro- gen-4—hydroxylase activity has been identified in MCF-7 breast cancer cells, which can be induced by 2,3,7,8—tetrachlorodi- benzo-p-dioxin (TCDD), a common environmental pollutant (89,90). As described in the introduction to this chapter, this human estrogen-4-hydroxy1ase activity was designated as cyto- chrome P4501B1 (CYPlBl), a novel extrahepatic isozyme de- tected in mammary tissue, ovary, adrenal gland, uterus, and elsewhere (89—91). In one reported measurement of estrogen metabolite concentrations in a human breast cancer extract (92), 102 the ratio of 4-OHE2 to 2-OHE2 was determined to be 4: 1. The same 4: 1 ratio was detected for the rates of formation of these CEs by breast cancer microsomes (88). It was concluded from these studies that, in rodent organs prone to estrogen-induced cancer or in human uterus or breast, which are targets of hor- mone-associated cancers, the predominant formation of 4-OHE2 may result in elevated concentrations of this carcinogenic estro- gen metabolite in these tissues. Catechol Formation by Deconjugation of Conjugated Estrogens De nova formation of CEs by hydroxylation of E1 or E2 by CYP enzymes (66,67) is only one of several pathways of CE formation. Other pathways include the deconjugation of conju- gated estrogen metabolites. The demethylation of methoxyestro- gens has been investigated with the use of liver and kidney microsomes of male Syrian hamsters (93). Rates of demethyl— ation of 2- and 4-methoxyestradi01 by kidney microsomes are comparable, whereas the rate of demethylation of 2-methoxyes- tradiol by liver microsomes is approximately fivefold higher than that of 4-methoxyestradiol (Table 3) (93 ). The rates of renal demethylation of methoxyestrogens are comparable to the rates Journal of the National Cancer Institute Monographs No. 27, 2000 Table 3. Kinetic parameters for catechol estrogen formation by hamster microsomes catalyzed either by aromatic hydroxylation of 17B-estradiol (E2) or by demethylation of methoxyestrogens* Demethylation of Aromatic hydroxylation methoxyestrogens Km, Vmax, pmol/mg Km, Vmax, pmol/mg ProductT MM protein per min MM protein per min Liver 2-OHE2 28 1573 15.4 606 4—OHE2 26 453 1 1.5 109 Kidney 2—OHE2 11.8 24.3 5.2 24 4-OHE2 8.7 31.9 15.5 30 *2-OHE2 = 2—hydroxyestradiol; 4—OHE2 = 4-hydroxyestradiol; Km = Mi— chaelis constant; Vmax = maximum initial velocity of an enzyme reaction. ”(Data are taken from Weisz et al. (78) and Zhu et a1. (93). of direct 2- and 4-hydroxylation of E2 by kidney microsomes, whereas the rates of hepatic demethylation are approximately one fifth of the corresponding de novo hydroxylation rates (78). These data demonstrate that metabolic demethylation of me thoxyestrogens is an important source of CE metabolites in the hamster kidney, a target of estrogen-induced carcinogenesis (62), whereas CE formation by direct aromatic hydroxylation of E2 predominates in liver, where this hormone does not induce tumors under these conditions. The identities of the CYP en- zymes catalyzing demethylation of methoxyestrogens are not known. CYP 3A enzymes have been identified as one of the activities capable of catalyzing the N—demethylation of a xeno- biotic substrate (94). Increased demethylase activity for me- thoxyestrogens in liver microsomes from phenobarbital-treated rats also indicates participation of CYP 2B isoforms in this reaction (95). However, specific identification of the enzymes catalyzing this reaction requires further research. Other phase II estrogen metabolites, such as Ez- and E1-3-D- glucuronides, are also deconjugated to the parent hormones by lysosomal glucuronidases (93). For instance, lysosomes from male Syrian hamster kidney catalyze the deconjugation of es- trogen glucuronides at rates that are one-third to two-thirds greater than corresponding rates by liver lysosomes. Treatment of hamsters with E2 implants for 9 days increases lysosomal glucuronidase activities for these estrogen glucuronides by 15%—25% in kidney and doubles the activities in liver, but it does not alter their corresponding Km (i.e., Michaelis constant) values (93). Microsomal glucuronidase activities are approxi- mately 10%—20% of lysosomal activities and do not appear to contribute appreciably to deconjugation of glucuronide metabo- lites. These results have been obtained with E2- and El-3-D- glucuronides as substrates. However, CE-glucuronide metabo- lites exist (96) and may also be deconjugated by these enzymes. Therefore, conjugates of both CE metabolites and of parent hor- mones formed in the liver may circulate and be deconjugated in extrahepatic tissues, including in organs prone to hormonal can- cer. For instance, comparable concentrations of 2- and 4—OHE2 have been found in the kidney and liver of male Syrian hamsters treated with E2, despite the fact that renal activities of estrogen- 2— and 4-hydroxylases are only one-tenth to one-twentieth the hepatic values (97). These comparable estrogen concentrations suggest that, in addition to direct aromatic hydroxylation of par- ent hormones, other metabolic pathways of CE formation exist in the kidney, such as deconjugation of methoxyestrogens and glucuronide metabolites. In extrahepatic tissue, such as in the hamster kidney, deconjugation reactions may be as important a source of CEs as aromatic hydroxylation of parent hormone. Metabolic Activation by Redox Cycling CEs, including 4-OHE2, are capable of metabolic redox cy- cling. This process consists of the organic hydroperoxide- dependent oxidation of the CEs (the hydroquinone) to the qui- none and the NADPH-dependent CYP reductase-catalyzed reduction of the quinone intermediate back to the hydroquinone (Fig. 8) (68). The quinones may react with nucleic acids and OH HO OCH: OH I OH o.‘ Conjugates HO H0 H0 0 OH OH O. 0 Fig. 8. Metabolic redox cycling between hydroquinone or quinone forms of 4-hydroxyestradiol (as shown) or 2-hydroxyestradiol (not shown). Either cat- echol estrogen (CE) may be oxidized to its respective quinone by organic hy— droperoxide—dependent P4501A enzymes. The quinones, in turn, may be reduced by reduced nicotinamide adenine dinucleotide phosphate (NADPH)—dependent cytochrome P450 reductase or other reductases. The semiquinone form is an intermediate in both oxidation and reduction reactions and may react with mo- Joumal of the National Cancer Institute Monographs No. 27, 2000 lecular oxygen to form superoxide radicals and quinone. Superoxide radicals may be reduced to hydrogen peroxide and further to hydroxy radical in the presence of metal ions. Although both CEs have the ability to undergo metabolic redox cycling, hydroxy radical damage has been observed mainly with 4-hy— droxylated estrogens, presumably because elevated concentrations of these me- tabolite exist in target tissues where estrogens induce tumors. 103 form stable or unstable DNA adducts, as described in Chapter 4. The semiquinone free radical is an intermediate in each of these conversions. The estrogen semiquinone may react with molecu- lar oxygen and form quinone and superoxide radicals (98). Thus, metabolic redox cycling is a mechanism of metabolic activation resulting in the continuous formation of free radical species from possibly small amounts of CE substrate (Fig. 8). Superoxide radicals may be reduced enzymatically or nonen- zymatically to hydrogen peroxide and further to hydroxy radi- cals, which are capable of initiating lipid peroxidation. Lipid hydroperoxides are, thus, formed by metabolic redox cycling of CBS and, in turn, support the formation of CBS and their redox cycling. This is achieved by lipid hydroperoxides functioning as cofactors for the organic hydroperoxide-dependent estrogen—2/ 4—hydroxylase (78,86) and for CYP1A1, which catalyzes the oxidation of hydroquinones, including CEs, to corresponding quinones (99). This shift toward lipid hydroperoxides as cofac- tors for enzyme-catalyzed reactions represents a loss of meta- bolic control in the cell, which is exerted in part by the bioavail- ability of NADPH as cofactor for biosynthetic reactions. Therefore, metabolic redox cycling of CEs, including 4-OHE2, may persist indefinitely in an unregulated fashion. Moreover, this redox cycling process amplifies damage because relatively small amounts of CEs may repeatedly undergo this process and generate greater than stoichiometric amounts of various radical species. Conclusions 4-OHE2 is a carcinogenic metabolite and may be formed by four different processes. 4-Hydroxyestrogens may be generated: (a) as a minor byproduct of NADPH-dependent 2—hydroxylation of either E1 or E2 due to a lack of specificity of the estrogen-2- hydroxylases; (b) by specific NADPH-dependent 4-hydroxyl— ation of E1 or E2 catalyzed by CYPlBl and possibly other 4-hydroxylases; (c) by organic hydroperoxide-dependent estro- gen-2- and 4-hydroxylase activity, which forms these two cat- echols in approximately equal amounts [little is known about the location and regulation of this enzyme(s)]; and (d) by deconju- gation of 4-methoxyestrogens and other CE conjugates. Tumors are expected to arise in cells or tissues that experience high concentrations of 4—OHE2 or 4—OHE1 formed by one or several of the processes cited above. High concentrations of 4-hy- droxyestrogens may be found in cells with high formation of CEs by aromatic hydroxylation of parent estrogens, inadequate phase II metabolism of these catechols, or high rates of decon- jugation of phase II metabolites back to CEs. Thus, CE metabo- lite concentrations are the result of a balance of several processes of formation and catabolism and may be unique for a given cell type, depending on its specific profile of metabolizing and ca- tabolizing enzyme activities. Elevated concentrations of 4-hydroxy1ated estrogens may be formed as a result of a loss of regulatory control of cells because organic hydroperoxides may be cofactors for the formation and metabolic redox cycling of CEs. In other cell types, 4-OHE2 may be formed by specific 4—hydroxylation of E2. The unique distri- bution of estrogen-4-hydroxylases in tissues, such as the uterus, breast, and others, points to 4—OHE2 as a hormone with charac- teristics distinct and different from those of E2. Specific 4-OHE2-induced processes, such as blastocyst implantation in the uterus of mice (100), support this view. Therefore, the organ- specific distribution of E2_4-hydroxylases may be a function of the role of this catechol in physiologic processes. 104 Tumors may be formed in organs and cells, which experience a loss of regulatory control as metabolism is converted to an organic hydroperoxide-dependent process from an NADPH— dependent process. Tumors may also arise in cells that express estrogen-4-hydroxylase activity and limited phase II metabo- lism, as may occur in mammary or uterine cells. More research is needed to examine the metabolic characteristics of each cell type in these organs and, thus, their potential for high CE con- centrations and for elevated tumor risk. ESTROGEN 4-HYDROXYLATION CATALYZED BY HUMAN CYTOCHROME P4501B1: OXIDATIVE METABOLISM OF ESTROGENS The oxidative metabolism of estrogens in vivo, especially E2 and E1, is known to occur at several positions, including carbons C-1, C-2, C-4, C-6, C-7, C-ll, C-14, C-lS, C-16, and C-18 [reviewed in (52,67)]. Studies of the routes of formation of these estrogen metabolites comprise an active area of research in sev- eral scientific fields, including endocrinology, pharmacology, environmental health sciences, oncology, and epidemiology. Additional knowledge of the rates of formation, metabolic fate (including further metabolism and routes of elimination), and activities of these hydroxylated steroids will support the scien- tific assessment of the relative importance of such metabolic pathways to hormonal carcinogenesis and other hormone-related diseases. The mechanism of carcinogenesis of estrogens has been pri- marily attributed to their specific action as steroid hormone re- ceptor agonists, controlling cellular growth and differentiation in estrogen-responsive tissues through concerted gene regulation. As discussed more fully in other sections of this monograph (see Chapters 3 and 4), increasing evidence of an additional mecha- nism of carcinogenesis has focused attention on the CE metabo- lites, which are less potent estrogens than E2 yet are biologically reactive, capable of directly or indirectly damaging protein, lipid, and DNA (101—103). Worthy of reiteration, our present knowledge of the mechanisms of estrogen—related disease pro- cesses indicates a strong estrogen receptor-mediated action. While the activities and potencies of the various estrogen me- tabolites are less understood, it is reasonable to predict, and in certain cases known, that the activities of these metabolites range from inactive to toxic to protective. Regarding estrogen— related cancers, the routes and rates of formation of certain es- trogen metabolites, specifically the 2—, 16a—, and more recently the 4-hydroxyestrogens, are being investigated as etiologic fac- tors. Of particular interest to this research focus group (The Cancer Cube) and a major topic of the conference preceding this monograph are the local production and activity of 4—OHE2. 4—Hydroxylated metabolites represent only a small percent- age of the total amount of estrogens detected in the urine, and 4-hydroxylation was previously considered a minor metabolic route (104). However, tissue 4-hydroxylation of E2 may play an important role in estrogen homeostasis. In human (87) and mouse (80) uteri, rat pituitary (79), and hamster kidney (105), the rate of E2—4-hydroxylation equals or exceeds the rate of 2-hydroxy1ation, and, notably, these organs are sites of estrogen- induced tumors (62, 81,106,107). In comparison to normal tissue, elevated E2-4—hydroxylase activity has been observed in human tissue samples prepared from breast (88,108) and uterine ( 87) tumors. Furthermore, in male hamster kidney, the carcinogenic and DNA-damaging activity of 4-OHE2 and the lack of activity Journal of the National Cancer Institute Monographs No. 27, 2000 of 2-OHE2 (83,85,109) implicate the 4—hydroxylated metabolites in estrogen carcinogenesis (11). Evidence of the toxicity of CEs via the generation of free radicals through lipid hydroperoxide- supported (110) reductive-oxidative cycling mechanisms contin- ues to accumulate (111—113). Furthermore, several laboratories (71,114,115) have demonstrated that estrogen-3,4—quinones can form DNA adducts, indicating the direct genotoxic potential of these compounds. Requisite to elucidating the contributions of the 4-hydroxyestrogens to estrogen carcinogenicity are the iden- tification and characterization of the enzymes that produce these metabolites. IDENTIFICATION AND CHARACTERIZATION OF HUMAN CYPlBl A novel human cytochrome P450 was first isolated by dif- ferential hybridization as a TCDD-responsive complementary DNA (cDNA) clone from a human keratinocyte cell line treated with TCDD. Levels of this messenger RNA (mRNA) (P4501B 1) were shown to be increased 50-fold by treatment with 10 nM TCDD, in part as the result of increased rates of gene transcrip- tion (116). Analysis of the complete cDNA sequence of this 5.1-kilobase (kb) TCDD—inducible mRNA identified a new gene subfamily of cytochrome P450, CYPlBl, based on 40% se- quence homology to other polycyclic aromatic hydrocarbon (PAH)-inducible isoforms, CYP1A1 and CYP1A2 (90). The CYPlBl gene was mapped to human chromosome 2 by poly- merase chain reaction (PCR) amplification of human/rodent so- matic cell hybrid panels using specific primers to the 3’- untranslated region of the CYPlBl cDNA. Segregation analysis of the data showed 100% concordance between the presence of the PCR product specific for CYPlBl and chromosome 2 and greater than 8% discordance for all other chromosomes. South- ern blot analysis of human genomic DNA, using a single cDNA probe corresponding to the 5’—portion of the CYPlBl open read- ing frame, suggested the presence of only a single gene in this subfamily (90). In the 19805 through studies on the metabolism of PAH, Jefcoate and co-workers (117) found that several rodent tissues, including embryo fibroblasts, adrenals, and the mammary gland, produced anomalous product ratios suggesting the presence of a novel P450 cytochrome. This was established 10 years later by the purification of a novel P450 that was found in each of these tissues. Subsequently, Jefcoate and co-workers tried to address the function of this form, notably by characterizing its expres- sion pattern and regulation. They used antibodies raised against the purified protein to identify expression of equivalent human forms in human breast cells and keratinocytes (118), as well as in a variety of rodent steroidogenic cells (adrenal, testis, ovary) (119). Expression in drug-metabolizing organs, like the liver, was low. This suggested that the function was physiologic in these nonhepatic tissues (i.e., the mammary gland). This was supported by the strong hormonal regulation demonstrated in steroidogenic cells. The induction of this P450 cytochrome by PAH was established, suggesting involvement of the aryl hy- drocarbon (Ah) receptor. Cloning of the mouse and rat cDNAs from mouse embryo fibroblasts and rat adrenals, respectively (91,120), completed the initial characterization of this P450. This characterization allowed Jefcoate’s laboratory to confirm that this rodent P450 had the same characteristics of and high sequence homology (82%) to human CYPlBl. In human tissues, Northern blot analysis of CYPlBl expres- Journal of the National Cancer Institute Monographs No. 27, 2000 sion showed that CYPlBl mRNA could be detected in each of 15 different tissue RNA samples prepared from heart, brain, placenta, lung, liver, skeletal muscle, kidney, spleen, thymus, prostate, testis, ovary, small intestine, colon, and peripheral blood leukocytes. In primary cultures of normal human epider- mal keratinocytes treated for 24 hours with 10 nM TCDD, levels of CYPlBl mRNA were increased more than 70-fold (90), dem- onstrating that TCDD induction of CYPlBl expression in ke- ratinocytes was not restricted to the transformed cell line SCC— 12F (116). Isolation and initial characterization of the human P4501B1 gene (121) described the DNA sequence of a 12-kb genomic clone corresponding to the entire 5.1-kb CYPlBl cDNA (90) and containing 3.0 kb of upstream (5’- of the ATG) DNA (123). Comparison of these sequences revealed the presence and posi— tions of three exons (371, 1044, and 3707 base pairs [bp]) and two introns (390 and 3032 bp), with the CYP1B1 open reading frame spanning exons 2 and 3. Southern blot analysis using cDNA probes corresponding to each of the three exons of CYPlBl supported the presence of only a single CYPlBl gene and excluded the existence of pseudogenes (121). High- resolution chromosome mapping confirmed the previous so- matic cell hybrid analysis (90) and further mapped the CYPlBl gene to human chromosome 2 at 2p21—22 ( 121 ). A single tran- scription initiation site was identified in this CYPlBl gene, which lacks a consensus TATA box sequence. Deletion analysis of CYPlBl-promoter reporter—gene constructs identified a spe- cific region (—1022 to —835) containing three TCDD-responsive enhancer—core-binding motifs (5’-GCGTG-3’) contributing to the TCDD-inducible expression of CYPlBl in the human kera- tinocyte cell line SCC—12F (121 ). Comparison of the human CYPlBl genomic and cDNA se- quences, obtained independently from two cell lines derived from different individuals, revealed three sequence differences (allelic variants, potential polymorphisms). Concurrent human genetic studies to identify one of two loci for primary congenital glaucoma (PCG) led to the mapping of the PCG locus GCL3A to human chromosome 2 at the 2p21 region [reviewed in (122)]. Subsequent investigations (123—125) have led to independent reports that identify distinct CYPlBl gene mutations that seg— regate with the GCL3A phenotype in PCG families. GCL3A- linked PCG has an autosomal recessive mode of transmission, and it appears that the observed mutations will result in the absence of P4501B1 protein and/or activity. Nonetheless, further understanding of this disease will require additional knowledge of the physiologic function of CYPlBl. As an added benefit of the extensive DNA sequence analysis of the translated regions of the CYPlBl gene in 22 PCG families and in 100 randomly selected normal individuals (124), the Val 432—Leu CYPlBl polymorphism (90,121) was confirmed. Three additional polymorphisms predicting variant amino acid sequences were identified: Arg 48—Gly, Ala 119—Ser, and Asn 453—Ser. The frequency of each wild—type allele, calculated for the reported sample of 100 normal individuals ( 124), is as fol- lows: Arg 48, 0.71; Ala 119, 0.71; Val 432, 0.28; and Asn 453, 0.76. The identification of several frequently occurring polymor- phisms in the human CYPlBl gene demonstrates the strength of direct genomic sequence analysis as a method to identify single nucleotide polymorphisms. At this time, the appropriate proce- dures to determine the functional significance of such variant proteins are less clear, making this an active area of research. 105 Nonetheless, epidemiologic studies, similar to those reported for catechol-0-methyltransferase polymorphisms and the risk for breast cancer (126,127), are under way for CYPlBl. Such stud- ies should further clarify the role of CYPlBl, specifically, and hydroxylated estrogen metabolites, generally, in estrogen- related diseases. ESTROGEN HYDROXYLATION CATALYZED BY HUMAN P4501B1 The widespread clinical use of antiestrogens in the adjuvant treatment of breast cancer provided a strong rationale for studies to determine the mechanism(s) of the antiestrogenic activity of ligands for the Ah receptor (128,129). In the MCF-7 breast tumor cell line, a sensitive gas chromatography/mass spectrom- etry method of analysis was used to show that treatment with TCDD resulted in large, more than lO—fold, increases in the rates of hydroxylation of E2 at positions C—2, C-4, C-150L, and C-60L (129). These studies support the role of increased E2 metabolism in the observed antiestrogenic effects of TCDD, which include the inhibition of estrogen-mediated expression of tissue plasmin- ogen—activator activity and the formation of multicellular foci (130,131). Spink et al. (130) convincingly demonstrated that P4501Al catalyzed the hydroxylation of E2 at the C-2, C-150L, and G601 positions in MCF-7 cells treated with TCDD. How— ever, the metabolism at the C-4 position was shown to be dis— tinct, catalyzed by an unknown TCDD-inducible CYP, best characterized as a low-Km E2-4—hydroxylase. Isolation and characterization of the TCDD-inducible CYPlBl from humans (90,116) and from rodent species (91,120,131) and the observation that human CYPlBl mRNA was elevated by treatment of human breast cell lines with TCDD led to studies of CYPlBl expression and activity in MCF-7 cells (89). Antibodies raised against mouse CYPlBl (anti-P450—EF) and cross-reactive with human CYPlBl (117,118) were used to investigate E2 hydroxylation catalyzed by microsomes prepared from TCDD-treated MCF-7 cells. This antibody preparation was shown to selectively inhibit E2-4-hydroxylation in a concentra~ tion-dependent manner. Furthermore, the elevated expression of P4501Bl protein and mRNA in these TCDD~treated cells closely paralleled the observed rates of E2 4-hydroxylation ( 89). Collectively, these results provided strong support for the as- signment of human CYPlBl as the low-Km E2-4—hydroxylase that is induced in MCF—7 human breast cancer cells by TCDD treatment. To establish the specific relationship between the P4501Bl gene product and E2 metabolism, human P450 protein was ex- pressed in Saccharomyces cerevisiae (3). Two recombinant P4501Bl expression plasmids were described. One construct, identified as CYPlBlAO, was designed to express a protein of 543 amino acids, corresponding to the entire deduced amino acid sequence of P4501B1. A second expression construct, identified as CYP1B1A3, was designed to produce a protein that did not contain the three amino acid residues after the initial methionine (deletion of amino acids 2—4). The design of the second plasmid was based on the reported amino acid sequence of rat P4501B1, showing that this protein did not contain the first four amino acid residues of its corresponding amino acid sequence (120). Levels of P4501B1 protein present in the microsomal fraction of the CYP1B1A3 construct (340 pmol/mg protein) were about 10-fold greater than the levels of P450 of CYPlBlAO preparations (3 ). This difference in expression of P450 may be important, since a 106 similar construct containing this same deletion of amino acids 2—4 also resulted in the highest expression levels of human P4501B1 in Escherichia coli (132). Further studies of the trans- lation, processing, and subcellular localization of this protein are warranted. Yeast CYPlBlAO microsomal preparations, contain- ing the lower P4501Bl-specific content, were determined to have the greatest E2 hydroxylation turnover numbers; these rates were not increased by the addition of NADPH cytochrome P450 reductase to the reaction mixture. Reactions containing these microsomes were shown to catalyze the 4- and 2—hydroxylation of E2, with Km values of 0.71 and 0.78 p.M and turnover num- bers of 1.39 and 0.27 nmol product/minute per nmol P450, re- spectively (3 ). The major CEs detected in serum and urine are the 2-hydrox— ylated metabolites. The liver is the primary site of estrogen metabolism, where rates of 2-hydroxylation, catalyzed by P4501A2, 3A3, and 3A4, greatly exceed the rate of 4-hydr0x- ylation. The reported apparent Km values for E2 hydroxylation catalyzed by the human forms of these enzymes range from 20 to 156 (LM, which are considerably higher that the apparent Km values of P4501B1, less than 1 MM (Table 4) (3,132—137). The turnover number for the formation of 4—OHE2 by P4501Bl is similar to the turnover numbers for the formation of 2-OHE2 by human P4501A2, 3A3, and 3A4. The E2—4-hydroxylase activity of P4501B1 has the highest catalytic efficiency (turnover/Km) of any reported estrogen hydroxylase, and the apparent Km values of E2-4- and 2-hydroxylase activities are the lowest reported values for estrogen hydroxylation (Table 4) (3 ). Comparisons of the values presented in Table 4 indicate a potential role of P4501Bl hydroxylase activity in estrogen homeostasis, espe- cially in extrahepatic organs, which express much lower levels of P4501A2 and 3A4 than does the liver (138). In further experiments to explore the significance of the low- Km 4—hydroxylase activity of P4501B1, cellular E2 metabolism was studied in MCF—7 cells treated with the potent Ah receptor ligand, indolo[3,2b]carbazole (ICZ) (139). MCF-7 cells exhib- ited ICZ concentration-dependent increases in P4501B1 and 1A1 mRNA levels. In parallel experiments to determine E2 me- tabolism, treatment of MCF-7 cells with 10 p.M ICZ for 72 hours, followed by replacement of medium containing 1 (LM E2 for 6 hours, resulted in a TCDD-like profile of E2 metabolites (129), with increased rates of hydroxylation occurring at the C-2, C-4, C-60L, and C-15a positions of E2; the rates of 2-hy- droxylation were the greatest, approximately threefold to seven- fold greater than the rates of hydroxylation at the other positions. When the E2 concentration was decreased to 10 nM, the rates of 2- and 4-hydroxylation in ICZ-treated cells were detectable and approximately equal, demonstrating that the E2 4- and 2-hydrox- ylation activities of P4501B1 are significant at low, physiologi- cally relevant concentrations of E2 (3 ). In summary, these stud- ies demonstrate that human P4501B1 is a catalytically efficient E2-4—hydroxylase that is likely to participate in endocrine regu- lation and estrogen—related disease. Tissue Expression of Human P4501B1 Large differences are demonstrated between species in CYPlBl metabolism of PAH (5,140). Thus, the question arose whether the expression of CYPlBl in breast cells gives any indication of physiologic or pathologic functions linked to such activity. The relative expression levels of CYP1A1 and CYPlBl will determine the conversion of E2 to the 2- and 4-hydroxyl- Journal of the National Cancer Institute Monographs No. 27, 2000 Table 4. Activites, kinetic parameters, and catalytic efficiencies of human estrogen hydroxylases* 2-Hydroxylation 4-Hydroxylation 160L—Hydroxylati0n Turnover, ce, k/Krn Turnover, ce, k/Km, Turnover, ce, k/Km, Reference Km, uM k (min—1) (sec‘1 ~M'l) Km, (1M k (min—l) (sec—1 -M‘1) Km, (1M k (min—1) (sec'l -M") No. Estradiol hydroxylation CYPlBl 0.78 0.27% 2.1 x 107 0.71 1.391 1.2 x 108 ND (3) CYP1A2 20 11 3.3 ><107 28 0.9 1.9 x106 58 0.7 7.2 ><105 (I35) 32 3.31 6.2 x 106 NR NR (136) CYP3A4 54 0.8 8.9 x 105 111 0.3 1.6 x 105 75 0.4 3.2 x 105 (135) 156 3.31 1.3 x 106 NR NR (136) CYP19 1.58 0.45 1.7 x 107 NR NR (137) Estrone hydroxylation CYP1B1 NR NR NR CYP1A2 19 9.2 2.9 x 107 17 2.0 7.1 x 106 <01 (135) 14 5.43 2.3 x 107 <0.1 <0.1 (138) CYP3A4 102 0.7 4.1 x 105 78 0.6 4.6 x 105 64 0.5 4.7 x 105 (135) <0.2 <02 95 0.68 4.3 x 105 (138) *Activities and kinetic parameters were determined as described in the referenced manuscripts. Catalytic efficiencies (turnover/km) were calculated with the use of these reported values. ND = not detected; NR = not reported; Km = Michaelis constant; ce = catalytic efficiency; k = turnover. TShimada et al. ([32) reported similar turnover numbers for recombinant P4501B1 purified from Escherichia coli (2-hydroxy1ation, 0.13 min—1; 4—hydroxylation, 1.4 min”). iAoyama et al. (137) reported similar turnover numbers for CYP1A2 (2—hydroxy1ation, 2.74 min—1; 4—hydroxylation, 0.27 min—1) and for CYP3A4 (2— hydroxylation, 1.30 min”; 4-hydroxy1ation, 0.31 min’l). ation products, respectively. CYPlBl is expressed constitutively in cultured breast luminal epithelial cells, which are the source of most breast tumors (5). This implies that there will be a low basal formation of 4-CE in all tumors. Typically, CYP1A1 is essentially undetectable in breast tissue from most donors. En- vironmental activators of the Ah receptor, such as polychloro- biphenyls, which have been detected at considerable levels in breast fat, could potentially induce both forms. This certainly is the case for these chemicals in cultured breast epithelial cells ( 5 ). However, analysis of CYP1A1 mRNA in human breast tissue by reverse transcription—PCR (RT—PCR) indicated that this isoform is typically undetectable. Low levels are occasionally present in cultured epithelial cells, but this is retained independent of me— dium changes and is almost certainly constitutive. These data indicate that relatively little 2-CE will be produced in vivo. A central question to be addressed and resolved in future research is whether this very low formation of 4-CE is physiologically or pathologically significant. The relatively high affinity of CYPlBl for E2 allows this activity to be retained even at low physiologic hormone levels. While it might seem that this ac— tivity is too low to remove E2, a substantial proportion of CYPlBl is localized in nuclear and perinuclear membranes. This may exert a relatively greater effect on nuclear levels of E2 and 4-CE production. This constriction to depletion of estrogen will depend on the relative activity of other metabolic processes acting on E2. CYPlBl seems to be the major source of 4-CE formation, and the measurement of this steroid in vivo points to substantial accumulated activity. Information on the “basal” and inducible expression of P4501B1 in human cells and tissues continues to be an interest- ing, yet less developed, aspect of the knowledge of this recently identified CYP. Additional knowledge of tissue P4501B1- specific content and activity will strengthen scientific assess- ments of the relative importance of estrogen metabolic pathways to estrogen-related diseases and their prevention. Journal of the National Cancer Institute Monographs No. 27, 2000 As stated above, the initial characterization of human P4501B1 revealed that levels of this 5.1-kb transcript were detectable in multiple adult tissue samples, the kidney sample exhibiting the greatest apparent signal relative to the other samples tested (90). In a subsequent study (4), these results were extended to include analyses of additional adult tissue samples representing five additional organs and a second kid- ney sample, comparative analyses of the relative expression of P4501A1 and 1A2 in these same tissue samples, and analyses of tissue samples from five fetal organs. P4501A1 mRNA was detected in 12 of the 21 adult tissue RNA samples but in none of the five fetal tissue RNA samples. The most intense hybridization signals occurred in the prostate and mammary tissues. P4501A2 mRNA was detected only in the adult liver sample. P4501B1 mRNA was detected in 20 of the 21 adult RNA samples (heart, brain, placenta, lung, liver, skeletal muscle, kidney 1, kidney 2, spleen, thymus, prostate, testis, ovary, small intestine, colon, peripheral blood leukocytes, adre- nal, pituitary, uterus, and mammary tissue) and in fetal heart, brain, lung, and kidney. The levels of P4501B1 RNA in adult kidney of both sexes and fetal kidney, as well as in prostate, uterus, and mammary tissue, were apparently greater than those in other tissues (4). Many of these findings were confirmed in a report describing an RT—PCR-based analysis of P4501B1 ex- pression showing that human P4501B1 is expressed mainly in extrahepatic tissues of adults and fetuses (14]). Contrary to the earlier report (4 ), P4501B1 RNA was not detected in samples of adult lung (141). While the basis of this discrepancy is unlikely to be identified, other investigators (142) have reported the de- tectable expression of this mRNA in lung tissue samples from smokers. In summary, these preliminary analyses of the organ- specific RNA distribution of the P4501 gene family have re- vealed that, while P4501A2 is expressed primarily in the liver, P4501A1 and 1B1 are expressed widely and found in many of the same tissues. Furthermore, P4501B1 appears to be the pre- 107 Fig. 9.1mmunohistochemical analysis of P4501B1 pro- tein in breast tissue of a nonpregnant woman of repro— ductive age. Photonricrographs are at a magnification of x100 (A and B) or x400 (C). Panel A: preimmune immunoglobulin G (IgG). Panels B and C: anti- P4501B1 IgG. The analysis was performed as previ- ously described [(146), except that the concentration of the primary antibody was 0.3 ug/mL]. The secondary antibody was a goat anti-rabbit streptavidin-conjugated antibody, and the staining was detected using biotin- conjugated horseradish peroxidase (HRP) complex and subsequent development using diaminobenzidine as the chromogen. The slides were lightly counterstained with hematoxylin. dominant family 1 P450 expressed in human fetal tissues (4,141). Regarding the expression of P4501B1 in organs such as breast and uterus, where strong associations between estrogen exposure and the risk for cancer are known, accumulating evi- dence indicates the presence and activity of this enzyme in these tissues. Elevated CE production has been associated with tumors of the breast (88,108,143,145), and P4501B1 RNA (4,145,146) and protein (147) have been detected in both normal (4,146) and tumor (145—147) breast tissue samples. Furthermore, in human uterine tissue, where 4-hydroxylation of E2 is increased in myo- mas compared with surrounding myometrium, the 4-hydroxy- lase activity of myoma microsomes was strongly inhibited by an antimouse P4501B1 antibody, suggesting that, in human uterine tissue, E2 4-hydroxylation‘is catalyzed by P4501B1 (87). In addition, Larsen et al. (5) demonstrated the expression of P4501B1 protein and activity in early-passage human mammary epithelial cells isolated from reduction mammoplasty tissue of seven individual donors. Specific contents of CYP1B1 and CYP1A1 protein in microsomal preparations (day 6 of culture) were quantitated by immunoblot analysis. Levels of constitutive CYPlBl protein ranged from 0.01 to 1.4 pmol/mg microsomal protein; exposure to TCDD increased these levels, ranging from 2.3 to 16.6 pmol/mg microsomal protein. Levels of constitutive CYP1A1 protein were much lower than those of CYPlBl. How- ever, the inductive response of CYP1A1 to TCDD treatment was very strong, resulting in comparable specific contents of the two P450 cytochromes in microsomes prepared from treated cells. This study indicates that human mammary epithelia constitu- tively expresses variable levels of functional CYP1B1 protein (11), which may contribute to the oxidative metabolism of E2 and other estrogens. To aid in the analysis of CYPlBl protein expression, our laboratory has produced polyclonal rabbit anti-P4501Bl anti- bodies that were shown to be both sensitive and CYP1B1 spe- cific, detecting this protein by both immunoblot and immuno— histochemical analyses (147,148). Recently, we (149) have used these antibodies to investigate the constitutive expression and cellular localization of P4501Bl in normal human tissue samples. A representative immunohistochemical analysis of P4501B1 protein expression in normal human breast tissue is presented in Fig. 9 (Sutter TR, Kim J, Sherman M: manuscript in preparation). As best seen in Fig. 9, panels B and C, P4501B1 is expressed ubiquitously in the ducts of this breast lobule, con- spicuously present in the epithelia, as well as the myoepithelia, stromal fibroblast, and endothelial cells. P4501B1 staining is 108 predominantly cytoplasmic, yet some nuclear staining is evident. Like the previous study ( 5 ), these results are consistent with the concept that human breast epithelial cells constitutively express significant amounts of CYPlBl protein, which may contribute to the oxidative metabolism of E2 and other estrogens in this tissue. Conclusions The kinetic parameters determined for the E2-4-hydr0xylase activity of human P4501B1 establish this enzyme as the most catalytically efficient estrogen-hydroxylase described to date. This observation is important because it suggests that this en- zyme is responsible for the E2-4-hydroxylase activity that has been observed in several tissues, such as human uterus and breast. The specific expression of P4501Bl and formation of CEs have been independently associated with estrogen-related tumors in multiple tissues and species. Further elucidation of the physiologic and/or pathologic significance of the E2-4- and 2-hydroxylase activities of P4501B1 will require additional knowledge of the tissue, cellular, and developmental expression (and regulation thereof) of this gene, and its protein product in humans and other animals. Despite the present knowledge gaps, our understanding of human P4501B1 as a low-Km E2-4- and 2-hydroxylase and a major extrahepatic form of CYP firmly supports the hypothesis of the role of CE metabolites in estro- gen-related carcinogenesis. REFERENCES (1) Anstead GM, Carlson KE, Katzenellenbogen IA. The estradiol pharma- cophore: ligand structure—estrogen receptor binding affinity relationships and a model for the receptor binding site. Steroids 1997;63:268-303. (2) Zhao Y, Nichols JE, Bulun SE, Mendelson CR, Simpson ER. Aromatase P450 gene expression in human adipose tissue. Role of a Jak/STAT path- way in regulation of the adipose-specific promoter. J Biol Chem 1995; 270:16449—57. (3) Hayes CL, Spink DC, Spink BC, Cao JQ, Walker NJ, Sutter TR. 17B- Estradiol hydroxylation catalyzed by human cytochrome P450 1B1. Proc Natl Acad Sci U S A 1996;93:9776—81. (4) Shimada T, Hayes CL, Yamazaki H, Amin S, Hecht SS, Guengerich FP, et al. Activation of chemically diverse procarcinogens by human cyto- chrome P-450 1B1. Cancer Res 1996;56:2979—84. (5 ) Larsen MC, Angus WG, Brake PB, Eltom SE, Sukow KA, Jefcoate CR. Characterization of CYPlBl and CYP1A1 expression in human mam- mary epithelial cells: role of the aryl hydrocarbon receptor in polycyclic aromatic hydrocarbon metabolism. Cancer Res 1998;58:2366—74. (6) Eltom SE, Larsen MC, Jefcoate CR. Expression of CYPlBl but not CYP1A1 by primary cultured human mammary stromal fibroblasts con- Joumal of the National Cancer Institute Monographs No. 27, 2000 stitutively and in response to dioxin exposure: role of the Ah receptor. Carcinogenesis 1998;19:1437—44. (7) Zhao Y, Agarwal VR, Mendelson CR, Simpson ER. Estrogen biosynthe- sis proximal to a breast tumor is stimulated by PGEZ via cyclic AMP, leading to activation of promoter II of the CYP19 (aromatase) gene. Endocrinology 1996;137:5739—41. (8) Furth J. Hormones as etiological agents in neoplasia, In: Becker FF, editor. Cancer: a comprehensive treatise, etiology: chemical and physical carcinogenesis. 2‘1d ed. New York (NY): Plenum Press; 1982. p. 89—134. (9) Hollingsworth AB, Lerner MR, Lightfoot SA, Wilkerson KB, Hanas JS, McKay PB, et a1. Prevention of DMBA-induced rat mammary carcinomas comparing leuprolide, oophorectomy, and tamoxifen. Breast Cancer Res Treat 1998;47:63—70. (10) Gunson DE, Steele RE, Chau RY. Prevention of spontaneous tumours in female rats by fadrozole hydrochloride, an aromatase inhibitor. Br J Can- cer 1995;72:72—5. (11) Toniolo PG, Levitz M, Zeleniuch-Jacquotte A, Banerjee S, Koenig KL, Shore RE, et al. A prospective study of endogenous estrogens and breast cancer in postmenopausal women. J Natl Cancer Inst 1995;87:190—7. ([2) Berrino F, Muti P, Micheli A, Bolelli G, Krogh V, Sciajno R, et al. Serum sex hormone levels after menopause and subsequent breast cancer. J Natl Cancer Inst 1996;88:291—6. , (13) Wysowski DK, Comstock GW, Helsing KJ, Lau HL. Sex hormone levels in serum in relation to the development of breast cancer. Am J Epidemiol 1987;125:791—9. (I4) Dorgan JF, Longcope C, Stephenson HE Jr, Falk RT, Miller R, Franz C, et al. Serum sex hormone levels are related to breast cancer risk in post- menopausal women. Environ Health Perspect 1997;105 Suppl 3:583—5. (15) Magnusson C, Baron J, Persson I, Wolk A, Bergstrom R, Trichopoulos D, et a1. Body size in different periods of life and breast cancer risk in post-menopausal women. Int J Cancer 1998;76:29—34. (16) Longcope C, Baker R, Johnston CC Jr. Androgen and estrogen metabo- lism: relationship to obesity. Metabolism 1986;35:235—7. (1 7) Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,411 women without breast cancer. Lancet 1997;350:1047—59. (18) Colditz GA, Hankinson SE, Hunter DJ, Willett WC, Manson JE, Stampfer MJ, et al. The use of estrogens and progestins and the risk of breast cancer in postmenopausal women. N Engl J Med 1995;332:1589—93. (19) Garland CF, Friedlander NJ, Barrett-Connor E, Khaw KT. Sex hormones and postmenopausal breast cancer: a prospective study in an adult com- munity. Am J Epidemiol 1992;135:1220—30. (20) Helzlsouer KJ, Alberg AJ, Bush TL, Longcope C, Gordon GB, Comstock GW. A prospective study of endogenous hormones and breast cancer. Cancer Detect Prevent 1994;18:79—85. (21) Trichopoulos D, MacMahon B, Cole P. Menopause and breast cancer risk. J Natl Cancer Inst 1972;48:605—13. (22) Feinleib M. Breast cancer and artificial menopause: a cohort study. J Natl Cancer Inst 1968;41:315—29. (23) Miller WR, O‘Neill J. The importance of local synthesis of estrogen within the breast. Steroids 1987;50:537—48. (24 ) Bezwoda WR, Mansoor N, Dansey R. Correlation of breast tumour aro- matase activity and response to aromatase inhibition with aminogluteth— imide. Oncology 1987;44:345—9. (25) Lipton A, Santen RJ, Santner SJ, Harvey HA, Sanders SI, Matthews YL. Prognostic value of breast cancer aromatase. Cancer 1992;70:1951—5. (26) Bradlow HL. A reassessment of the role of breast tumor aromatization. Cancer Res 1982;42(8 Suppl):3382s—3386s. (27) Santen RJ, Martel J, Hoagland M, Naftolin F, Roa L, Harada N, et al. Stromal spindle cells contain aromatase in human breast tumors. J Clin Endocrinol Metab 1994;79:627—322. (28) Esteban JM, Warsi Z, Haniu M, Hall P, Shively JE, Chen S. Detection of intratumoral aromatase in breast carcinomas. An immunohistochemical study with clinicopathologic correlation. Am J Pathol 1992;140:337—43. (29) Berstein LM, Larionov AA, Kyshtoobaeva ASh, Pozharisski KM, Semi— glazov VF, Ivanova OA. Aromatase in breast cancer tissue—localization and relationship with reproductive status of patients. J Cancer Res Clin Oncol 1996;122:495—8. Journal of the National Cancer Institute Monographs N0. 27, 2000 (30) Dowsett M. Theoretical considerations for the ideal aromatase inhibitor. Breast Cancer Res Treat 1998;49 Suppl 1:839—44. (3]) Lu Q, Nakamura J, Savinov A, Yue W, Weisz J, Dabbs DJ, et al. Ex- pression of aromatase protein and messenger ribonucleic acid in tumor epithelial cells and evidence of functional significance of locally produced estrogen in human breast cancers. Endocrinology 1996;137:3061—8. (32) Sasano H, Frost AR, Saitoh R, Harada N, Poutanen M, Vihko R, et al. Aromatase and 17 beta—hydroxysteroid dehydrogenase type 1 in human breast carcinoma. J Clin Endocrinol Metab 1996;81:4042—6. (33) Santner SJ, Pauley RJ, Tait L, Kaseta J, Santen RJ. Aromatase activity and expression in breast cancer and benign breast tissue Stromal cells. J Clin Endocrinol Metab 1997;82:200—8. (34 ) Yue W, Wang JP, Hamilton C], Demers LM, Santen R]. In situ aroma— tization enhances breast tumor estradiol levels and cellular proliferation. Cancer Res 1998;58:927—32. ( 35 ) Reed MJ, Owen AM, Lai LC, Coldham NG, Ghilchik MW, Shaik NA, et al. In situ oestrone synthesis in normal breast and breast tumour tissues: effect of treatment with 4-hydroxyandrostenedione. Int J Cancer 1989;44: 233—7. (36) van Landeghem AA, Poortman J, Nabuurs M, Thijssen JH. Endogenous concentration and subcellular distribution of estrogens in normal and ma- lignant human breast tissue. Cancer Res 1985;45:2900—6. (37) Preston—Martin S, Pike MC, Ross RK, Henderson BE. Epidemiologic evidence for the increased cell proliferation model of carcinogenesis. En- viron Health Perspect 1993;101 Suppl 5:137—8. (38) Zhang L, Medina D. Gene expression screening for specific genes asso- ciated with mouse mammary tumor development. M01 Carcinog 1993;8: 123—6. ( 39) Tekmal RR, Durgam VR. The overexpression of int-S/Aromatase, a novel MMTV integration locus gene, is responsible for D2 mammary tumor cell proliferation. Cancer Lett 1995;88:147—55. (40) Morris VL, Rao TR, Kozak DA, Gray CA, Lee Chan EC, Cornell TJ, et al. Characterization of Int-5, a locus associated with early events in mam- mary carcinogenesis. Oncogene Res 1991;6253—63. (41) Gray DA, McGrath CM, Jones RF, Morris VL. A common mouse mam- mary tumor virus integration site in chemically induced precancerous mammary hyperplasia. Virology 1986;148:360—8. (42) Tekmal RR, Ramachandra N, Gubba S, Durgam VR, Mantione J, Toda K, et al. Overexpression of int-5/aromatase in mammary glands of transgenic mice results in the induction of hyperplasia and nuclear abnormalities. Cancer Res 1996;56:3180—5. (43) Mor G, Yue W, Santen RJ, Gutierrez L, Eliza M, Berstein LM, et a1. Macrophages, estrogen and the microenvironment of breast cancer. J Ste- roid Biochem Mol Biol 1998;67:403—11. (44) Jakob F, Homann D, Adamski J. Expression and regulation of aromatase and 17 beta-hydroxysteroid dehydrogenase type 4 in human THP 1 leu- kemia cells. J Steroid Biochem Mol Biol 1995;55:555—63. (45) Fitzpatrick SL, Richards JS. Identification of a cyclic adenosine 3’,5’ monophosphate-response element in the rat aromatase promoter that is required for transcriptional activation in rat granulosa cells and R2C ley- dig cells. Mol Endocrinol 1994;8zl309—19. (46) Peaker M, Taylor E. Oestrogen production by the goat mammary gland: transient aromatase activity during late pregnancy. J Endocrinol 1990; 125:R1—3. (47) Wilson JD, Leshin M, George FW. The Sebright bantam chicken and the genetic control of extraglandular aromatase. Endocr Rev 1987;8:363—76. ( 48 ) Coen P, Kulin H, Ballantine T, Zaino R, Frauenhoffer E, Boal D, et al. An aromatase-producing sex-cord tumor resulting in prepubertal gynecomas— tia. N Engl J Med 1991;324:317—22. (49) Simpson ER, Zhao Y, Agarwal VR, Michael MD, Bulun SE, Hin- shelwood MM, et al. Aromatase expression in health and disease. Recent Prog Horm Res 1997;52:185—213. (50) Zhao Y, Agarwal VR, Mendelson CR, Simpson ER. Estrogen biosynthe- sis proximal to a breast tumor is stimulated by PGE2 via cyclic AMP, leading to activation of promoter 11 of the CYP19 (aromatase) gene. Endocrinology 1996;137:5739—42. (51) Santen RJ, Samojlik E, Wells SA. Resistance of the ovary to blockade of aromatization with aminoglutethimide. J Clin Endocrinol Metab 1980;51: 473—7. 109 (52) Zhu BT, Conney AH. Functional role of estrogen metabolism in target cells: review and perspectives. Carcinogenesis 1998;19:1—27. (53) Yager JD, Liehr JG. Molecular mechanisms of estrogen carcinogenesis. Annu Rev Pharmacol Toxicol 1996;36:203—32. (54) Liehr JG. Hormone-associated cancer: mechanistic similarities between human breast cancer and estrogen—induced kidney carcinogenesis in ham— sters. Environ Health Perspect 1997;105:565—9. (55) Liehr JG. Dual role of estrogens as hormones and pro—carcinogens: tu- mour initiation by metabolic activation of oestrogens. Eur J Cancer Prev 1997;6z3—10. (56) Feigelson HS, Henderson BE. Estrogens and breast cancer. Carcinogen— esis 1996;17:2279—84. (57) Hammond DK, Zhu BT, Wang MY, Ricci MJ, Liehr JG. Cytochrome P450 metabolism of estradiol in hamster liver and kidney. Toxicol Appl Pharmacol 1997;145:54—60. (58) Blaich G, Raabe H, Metzler M. Modification of 7,8—benzoflav0ne me- tabolism in hamster liver and kidney microsomes by hepatic tumor in— ducing treatments. Carcinogenesis 1990;11:95—101. (59) Liehr JG, Wheeler WJ, Ballatore AM. Influence of vitamin C on estrogen induced renal carcinogenesis in Syrian hamster. In: Meyskens FL, Prasad KN, editors. Modulation and mediation of cancer by vitamins. Basel (Switzerland): Karger Scientific Publishers; 1983. p. 132—9. (60) Liehr JG, Wheeler WJ. Inhibition of estrogen-induced renal carcinoma in Syrian hamsters by vitamin C. Cancer Res 1983;43:4638—42. (61) Liehr JG, Gladek A, Macatee T, Randerath E, Randerath K. DNA adduct formation in liver and kidney of male Syrian hamsters treated with estro- gen and/or oc-naphthoflavone. Carcinogenesis 1991;12:385—9. ( 62 ) Kirkman H. Estrogen—induced tumors of the kidney in the Syrian hamster. III. Growth characteristics in the Syrian hamster. Natl Cancer Inst Monogr 1959;121—57. ( 63 ) Zhu BT, Roy D, Liehr JG. The carcinogenic activity of ethinyl estrogens is determined by both their hormonal characteristics and their conversion to catechol metabolites. Endocrinology 1993;132:577—83. (64) Cavalieri E, Rogan E. Mechanisms of tumor initiation by polycyclic aro- matic hydrocarbons in mammals. In: Neilson AH, editor. The handbook of environmental chemistry, vol 3J. PAHs and related compounds. Heidel- berg (Germany): Springer—Verlag', 1998. p. 81—117. (65) Flowers L, Bleczinski WF, Burczynski ME, Harvey RG, Penning TM. Disposition and biological activity of benzo[a]pyrene—7,8—dione. A geno- toxic metabolite generated by dihydrodiol dehydrogenase. Biochemistry 1996;35:13664—72. (66) Slaunwhite WR, Kirdani RY, Sandberg AA. Metabolic aspects of estro- gens in man. In: Greep O, Astwood EB, Geiger SR, editors. Handbook of physiology, section 7: endocrinology, vol 11, part 1. Washington (DC): American Physiological Society; 1973. p. 485—523. (67) Manucci C, Fishman J. P450 enzymes of estrogen metabolism. Pharmacol Ther 1993;57:237—57. (68) Liehr JG, Ulubelen AA, Strobe] HW. Cytochrome P-450-mediated redox cycling of estrogens in Syrian hamsters. J Biol Chem 1986;261:16865—70. (69) Liehr JG, DaGue BB, Ballatore AM. Reactivity of 4’—4’ ’—diethylstilbestrol quinone, a metabolic intermediate of diethylstilbestrol. Carcinogenesis 1985;6z829—36. (70) Dwivedy I, Devanesan P, Cremonesi P, Rogan E, Cavalieri E. Synthesis and characterization of estrogen 2,3- and 3,4-quinones. Comparison of DNA adducts formed by the quinones versus horseradish peroxidase— activated catechol estrogens. Chem Res Toxicol 1992;5z828—33. (71) Cavalieri EL, Stack DE, Devanesan PD, T odorovic R, Dwivedy I, Hig- ginbotham S, et al. Molecular origin of cancer: catechol estrogen—3,4— quinones as endogenous tumor initiators. Proc Natl Acad Sci U S A 1997; 94:10937—42. (72) Abul—Hajj YJ, Tabakovic K, Tabakovic I. An estrogen—nucleic acid ad- duct. Electroreductive intermolecular coupling of 3,4—estrone-o—quinone and adenine. J Am Chem Soc 1995;117:6144—5. (73) Stack DE, Byun J, Gross ML, Rogan EG, Cavalieri EL. Molecular char— acteristics of catechol estrogen quinones in reactions with deoxyribo- nucleosides. Chem Res Toxicol 1996;9285l—9. (74) Cao K, Stack DE, Ramanathan R, Gross ML, Rogan ER, Cavalieri EL. Synthesis and structure elucidation of estrogen quinones conjugated with cysteine, N—acetylcysteine, and glutathione. Chem Res Toxicol 1998,11: 909—16. 110 (75) Guengerich FP. Characterization of human microsomal cytochrome P-450 enzymes. Annu Rev Pharmacol Toxicol 1989;29:241—64. (76) Aoyama T, Korzekwa K, Nagata K, Gillette J, Gelboin HV, Gonzalez FJ. Estradiol metabolism by complementary deoxyribonucleic acid-expressed human cytochrome P4505. Endocrinology 1990;126:3101—6. (77) Kerlan V, Dreano Y, Bercovici JP, Beaune PH, Floch HH, Berthou F. Nature of cytochrome P450 involved in the 2—/4-hydroxylations of estra- diol in human liver microsomes. Biochem Pharmacol 1992;44:1745—56. (78) Weisz J, Bui QD, Roy D, Liehr JG. Elevated 4-hydroxylation of estradiol by hamster kidney microsomes: a potential pathway of metabolic activa— tion estrogens. Endocrinology 1992;131:655—61. (79) Bui QD, Weisz J. Monooxygenase mediating catecholestrogen formation by rat anterior pituitary is an estrogen-4-hydroxylase. Endocrinology 1989;124:1085—7. (80) Paria BC, Chakraborty C, Dey SK. Catechol estrogen formation in mouse uterus and its role in implantation. Mol Cell Endocrinol 1990;69:25—32. (81) Newbold RR, Bullock BC, McLachlan JA. Uterine adenocarcinoma in mice following developmental treatment with estrogens: a model for hor— monal carcinogenesis. Cancer Res 1990;50:7677—81. (82) Clifton KH, Meyer RK. Mechanism of anterior pituitary tumor induction by oestrogen. Anat Rec 1956;125:65—81. (83) Liehr JG, Fang WF, Sirbasku DA, Ulubelen AA. Carcinogenicity of cat— echol estrogens in Syrian hamsters. J Steroid Biochem 1986;24:353—6. (84) Liehr JG, Sirbasku DA. Estrogen—dependent kidney tumors. In: Taub M, editor. Tissue culture of epithelial cells. New York (NY): Plenum Press; 1985. p. 205—34. (85) Li J], Li SA. Estrogen carcinogenesis in Syrian hamster tissues: role of metabolism. Fed Proc 1987;46:1858-63. (86) Weisz J. Biogenesis of catecholestrogens: a mechanism of metabolic ac- tivation of estrogens. Polycyclic Arom Compounds 1994;62241—51. (87) Liehr JG, Ricci MJ, Jefcoate CR, Hannigan EV, Hokanson JA, Zhu BT. 4-Hydroxylation of estradiol by human uterine myometrium and myoma microsomes: implications for the mechanism of uterine tumorigenesis. Proc Natl Acad Sci U S A 1995;92:9220—4. (88) Liehr JG, Ricci MJ. 4-Hydroxylation of estrogens as marker of human mammary tumors. Proc Natl Acad Sci U S A 1996;93:3294—6. (89) Spink DC, Hayes CL, Young NR, Christou M, Sutter TR, Jefcoate CR, et al. The effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on estrogen metabo- lism in MCF—7 breast cancer cells: evidence for induction of a novel l7B—estradiol 4-hydroxylase. J Steroid Biochem Mol Biol 1994;51:251—8. (90) Sutter TR, Tang YM, Hayes CL, Wo YY, Jabs EW, Li X, et a]. Complete cDNA sequence of a human dioxin-inducible mRNA identifies a new gene subfamily of cytochrome P450 that maps to chromosome 2. J Biol Chem 1994;269:13092—9. (91) Savas U, Bhattacharyya KK, Christou M, Alexander DL, Jefcoate CR. Mouse cytochrome P—450EF, representative of a new 1B subfamily of cytochrome P—450s. J Biol Chem 1994;269:14905—11. (92) Castagnetta LA, Granata OM, Arcuri FF, Polito LM, Rosati F, Cartoni GP. Gas chromatography/mass spectrometry of catechol estrogens. Ste— roids 1992;57:437—43. (93) Zhu BT, Evaristus EN, Antoniak SK, Sarabia SF, Ricci MJ, Liehr JG. Metabolic deglucuronidation and demethylation of estrogen conjugates as a source of parent estrogens and catecholestrogen metabolites in Syrian hamster kidney, a target organ of estrogen—induced tumorigenesis. Toxicol Appl Pharmacol 1996;136:186—93. (94) Jacolot F, Simon 1, Dreano Y, Beaune P, Riche C, Berthou F. Identifica- tion of the cytochrome P450 IIIA family as the enzymes involved in the N-demethylation of tamoxifen in human liver microsomes. Biochem Phar— macol 1991;41:1911—9. (95) Hoffman AR, Paul SM, Axelrod J. The enzymatic formation of catecho- lestrogens from 2—methoxyestrogens by rat liver microsomes. Endocrinol- ogy 1980;107:1192—7. (96) Ritter JK, Chen F, Sheen YY, Lubet RA, Owens IS. Two human liver cDNAs encode UDP-glucuronosyltransferases with 2 log differences in activity toward parallel substrates including hyodeoxycholic acid and cer- tain estrogen derivatives. Biochemistry 1992;31:3409—14. (97) Zhu BT, Liehr JG. Inhibition of catechol 0-methyltransferase-catalyzed 0-methylation 0f 2- and 4—hydr0xyestradiol by quercetin. Possible role in estradiol-induced tumorigenesis. J Biol Chem 1996;271:1357—63. (98) Roy D, Liehr JG. Temporary decrease in renal quinone reductase activity Journal of the National Cancer Institute Monographs No. 27, 2000 induced by chronic administration of estradiol to male Syrian hamsters. Increased superoxide formation by redox cycling of estrogen. J Biol Chem 1988;263:3646—51. (99) Roy D, Bernhardt A, Strobel HW, Liehr JG. Catalysis of the oxidation of steroid and stilbene estrogens to estrogen quinone metabolites by the B—naphthoflavone-inducible cytochrome P450 IA family. Arch Biochem Biophys 1992;296:450—6. (100) Das SK, Taylor JA, Korach KS, Paria BC, Dey SK, Lubahn DB. Estro- genic responses in estrogen receptor-0L deficient mice reveal a distinct estrogen signaling pathway. Proc Natl Acad Sci U S A 1997;94: 12786—91. (101) Liehr JG, Roy D. Free radical generation by redox cycling of estrogens. Free Radic Biol Med 1990;82415—23. (102) Dwivedy I, Devanesan P, Cremonesi P, Rogan E, Cavalieri E. Synthesis and characterization of estrogen 2,3- and 3,4-quinones. Comparison of DNA adducts formed by the quinones versus horseradish peroxidase- activated catechol estrogens. Chem Res Toxicol 1992;52828—33. (103) Nutter LM, Wu YY, Ngo EO, Sierra EE, Gutierrez, PL, Abul-Hajj YJ. An o—quinone form of estrogen produces free radicals in human breast cancer cells: correlation with DNA damage. Chem Res Toxicol 1994;7z23—8. (104) Fishman J. The catechol estrogens. Neuroendocrinology 1976;22:363—74. (105) Weisz J , Bui QD, Roy D, Liehr JG. Elevated 4-hydroxylation of estradiol by hamster kidney microsomes: a potential pathway of metabolic activa— tion of estrogens. Endocrinology 1992;131:655—61. (106) Henderson BE, Ross RK, Pike MC. Toward the primary prevention of cancer. Science 1991;254:1131—8. (107) Furth J, Cliftonk H, Gadsden EL, Buffet RF. Dependent and autonomous mammotropic pituitary tumors in rats; their somatropic features. Cancer Res 1956;16:608—16. (108) Lemon HM, Heidel JR, Rodriguez-Sierra JF. Increased catechol estrogen metabolism as a risk factor for nonfamilial breast cancer. Cancer Res 1992;69:457—65. ([09) Han X, Liehr JG. DNA single-strand breaks in kidneys of Syrian hamsters treated with steroidal estrogens: hormone-induced free radical damage preceding renal malignancy. Carcinogenesis 1994;15:997—1000. (110) Wang MY, Liehr JG. Identification of fatty acid hydroperoxide cofactors in the cytochrome P450-mediated oxidation of estrogens to quinone me— tabolites. Role and balance of lipid peroxides during estrogen—induced carcinogenesis. J Biol Chem 1994;269:284—91. (111) Seacat AM, Kuppusamy P, Zweier JL, Yager JD. ESR identification of free radicals formed from the oxidation of catechol estrogens by Cu“. Arch Biochem Biophys 1997;347:45—52. (112) Yoshie Y, Ohshima H. Synergistic induction of DNA strand breakage by catechol-estrogen and nitric oxide: implications for hormonal carcinogen— esis. Free Radic Biol Med 1998;24:341—8. (113) Chen Y, Shen L, Zhang F, Lau SS, van Breemen RB, Nikolic D, et al. The equine estrogen metabolite 4-hydroxyequilenin causes DNA single-strand breaks and oxidation of DNA bases in vitro. Chem Res Toxicol 1998;11: 1105—1 1. ([14) Roy D, Abul—Hajj YJ. Estrogen—nucleic acid adducts: guanine is major site for interaction between 3,4—estrone quinone and COIII gene. Carci- nogenesis 1997;18:1247—9. (115) Tsutsui T, Taguchi S, Tanaka Y, Barrett JC. 17Beta-estradiol, diethylstil— bestrol, tamoxifen, toremifene and ICI 164,384 induce morphological transformation and aneuploidy in cultured Syrian hamster embryo cells. Int J Cancer 1997;70:188—93. (116) Sutter TR, Guzman K, Dold KM, Greenlee WF. Targets for dioxin: genes for plasminogen activator inhibitor—2 and interleukin»1[3. Science 1991; 254:415—8. (117) Pottenger LH, Christou M, Jefcoate CR. Purification and immunological characterization of a novel cytochrome P450 from C3H/10T1/2 cells. Arch Biochem Biophys 1991;286:488—97. (118) Christou M, Savas U, Spink DC, Gierthy JF, Jefcoate CR. Co—expression of human CYP1A1 and a human analog of cytochrome P450—BF in re— sponse to 2,3,7,8-tetrachlor0-dibenzo-p-dioxin in the human mammary carcinoma—derived MCF—7 cells. Carcinogenesis 1994;15:725—32. (119) Otto S, Marcus C, Pidgeon C, Jefcoate C. A novel adrenoconicotropin— inducible cytochrome P450 from rat adrenal microsomes catalyzes poly— cyclic aromatic hydrocarbon metabolism. Endocrinology 1991;129: 970—82. Journal of the National Cancer Institute Monographs No. 27, 2000 (120) Bhattacharyya KK, Brake PB, Eltom SE, Otto SA, Jefcoate CR. Identi- fication of a rat adrenal cytochrome P450 active in polycyclic hydrocar- bon metabolism as rat CYPlBl. Demonstration of unique tissue-specific pattern of hormonal and aryl hydrocarbon receptor-linked regulation. J Biol Chem 1995;270:11595—602. (121) Tang YM, W0 YY, Stewart J, Hawkins AL, Griffin CA, Sutter TR, et al. Isolation and characterization of the human cytochrome P450 CYPlBl gene. J Biol Chem 1996;271:28324—30. (122) Safarazi M. Recent advances in molecular genetics of glaucomas. Hum Mol Genet 1997;621667—77. (123) Stoilov I, Akarsu AN, Sarfarazi M. Identification of three different trun— cating mutations in cytochrome P450181 (CYPlBl) as the principal cause of primary congenital glaucoma (Buphthalmos) in families linked to the GLC3A locus on chromosome 2p21. Hum M01 Genet 1997;6z64l—7. (124) Stoilov I, Akarsu AN, Alozie I, Child A, Barsoum—Homsy M, Turacli ME, et al. Sequence analysis and homology modeling suggest that primary congenital glaucoma 0n 2p21 results from mutations disrupting either the hinge region or the conserved core structures of cytochrome P4501Bl. Am J Hum Genet 1998;62:573—84. (125) Bejjani BA, Lewis RA, Tomey KF, Anderson KL, Dueker DK, Jabak M, et al. Mutations in CYPlBl, the gene for cytochrome P4501B1, are the predominant cause of primary congenital glaucoma in Saudi Arabia. Am J Hum Genet 1998;62:325—33. (126) Lavigne JA, Helzlsouer KJ, Huang HY, Strickland PT, Bell DA, Selmin O, et al. An association between the allele coding for a low activity variant of catechol—0-methyltransferase and the risk for breast cancer. Cancer Res 1997;57:5493—7. (127) Thompson PA, Shields PG, Freudenheim JL, Stone A, Vena JE, Marshall JR, et al. Genetic polymorphisms in catechol-0-methyltransferase, meno— pausal status, and breast cancer risk. Cancer Res 1998;58:2107—10. (128) Gierthy JF, Lincoln DW 2d, Kampcik SJ, Dickerman HW, Bradlow HL, Niwa T, et al. Enhancement of 2- and 16a-estradiol hydroxylation in MCF-7 human breast cancer cells by 2,3,7,8-tetrachlorodibenzo—P—dioxin. Biochem Biophys Res Commun 1988;157:515-20. (129) Spink DC, Lincoln DW 2d, Dickerman HW, Gierthy JF. 2,3,7,8- Tetrachlorodibenzo-p-dioxin causes an extensive alteration of 1713- estradiol metabolism in MCF—7 breast tumor cells. Proc Natl Acad Sci U S A 1990;87:6917—21. (130) Spink DC, Eugster HP, Lincoln DW 2nd, Schuetz JD, Schuetz EG, Johnson JA, et al. l7B-Estradiol hydroxylation catalyzed by human cy— tochrome P450 1A1: a comparison of the activities induced by 2,3,7,8- tetrachlorodibenzo-p-dioxin in MCF-7 cell with those from heterologous expression of the cDNA. Arch Biochem Biophys 1992;293:342—8. (131) Walker NJ, Gastel JA, Costa LT, Clark GC, Lucier GW, Sutter TR. Rat CYP1B1: an adrenal cytochrome P450 that exhibits sex-dependent ex- pression in livers and kidneys of TCDD-treated animals. Carcinogenesis 1995;16:1319—27. (132) Shimada T, Wunsch RM, Hanna IH, Sutter TR, Guengerich FP, Gillam EM. Recombinant human cytochrome P450 lBl expression in Esch- erichia coli. Arch Biochem Biophys 1998;357:111—20. (133) Yamazaki H, Shaw PM, Guengerich FP, Shimada T. Roles of cyto— chromes P450 1A2 and 3A4 in the oxidation of estradiol and estrone in human liver microsomes. Chem Res Toxicol 1998;11:659—65. (134) Waxman DJ, Lapenson DP, Aoyama T, Gelboin HV, Gonzalez FJ, Kor— zekwa K. Steroid hormone hydroxylase specificities of eleven cDNA— expressed human cytochrome P4505. Arch Biochem Biophys 1991;290: 160—6. (135) Osawa Y, Higashiyama T, Shimizu Y, Yarborough C. Multiple functions of aromatase and the active site structure; aromatase is the placental estrogen 2-hydroxylase. J Steroid Biochem Mol Biol 1993;44:469—80. (136) Shou M, Korzekwa KR, Brooks EN, Krausz KW, Gonzalez FJ, Gelboin HV, Role of human hepatic cytochrome P450 1A2 and 3A4 in the meta- bolic activation of estrone. Carcinogenesis 1997;18:207—14. (137) Aoyama T, Korzekwa K, Nagata K, Gillette J, Gelboin HV, Gonzalez FJ. Estradiol metabolism by complementary deoxyribonucleic acid-expressed human cytochrome P4505. Endocrinology 1990;126:3101—6. (138) de Waziers I, Cugnenc PH, Yang CS, Leroux JP, Beaune PH. Cytochrome P 450 isoenzymes, epoxide hydrolase and glutathione transferase in rat and human hepatic and extrahepatic tissues. J Pharmacol Exp Ther 1990; 253:387—94. lll (I39) Bjeldanes LF, Kim JY, Grose KR, Bartholomew JC, Bradfield CA. Aro— matic hydrocarbon responsiveness-receptor agonists generated from in— dole-3-carbinol in vitro and in viva: comparisons with 2,3,7,8- tetrach]orodibenzo-p-dioxin. Proc Natl Acad Sci U S A 1991;88:9543—7. (140) Christou M, Savas U, Schroeder S, Shen X, Thompson T, Gould MN, et al. Cytochromes CYP1A1 and CYPlBl in the rat mammary gland: cell- specific expression and regulation by polycyclic aromatic hydrocarbons and hormones. Mol Cell Endocrinol 1995;115:41—50. (141) Hakkola J, Pasanen M, Pelkonen O, Hukkanen J, Evisalmi S, Anttila S, et al. Expression of CYP1B1 in human adult and fetal tissues and differential inducibility of CYPlBl and CYP1A1 by Ah receptor ligands in human placenta and cultured cells. Carcinogenesis 1997;18:391—7. (142) Willey JC, Coy EL, Frarnpton MW, Torres A, Apostolakos MJ, Hoehn G, et al. Quantitative RT—PCR measurement of cytochromes p450 1A1, 1B1, and 2B7, microsomal epoxide hydrolase, and NADPH oxidoreductase expression in lung cells of smokers and nonsmokers. Am J Respir Cell Mol Biol 1997;14:114—24. ( I 43) Abul-Hajj YJ, Thij ssen JH, Blankenstein MA. Metabolism of estradiol by human breast cancer. Eur J Cancer Clin Oncol 1988;24:1171—8. (144) Hoffman AR, Paul SM, Axelrod J. Catecholestrogen synthesis and me— tabolism by human breast tumors in vitro. Cancer Res 1979;39:4584—7. (I45) McKay JA, Melvin WT, Ah-See AK, Ewen SW, Greenlee WF, Marcus CB, et al. Expression of cytochrome P450 CYPlBl in breast cancer. FEBS Lett 1995;374:270—2. (I46) Huang Z, Fasco MJ, Figge HL, Keyomarsi K, Kaminsky LS. Expression of cytochromes P450 in human breast tissue and tumors. Drug Metab Dispos 1996;24:899—905. (147) Murray GI, Taylor MC, McFadyen MC, McKay JA, Greenlee WF, Burke MD, et al. Tumor—specific expression of cytochrome P450 CYPlB 1. Can- cer Res 1997;57:3026—31. 112 (148) Walker NJ, Crofts FG, Li Y, Lax SF, Hayes CL, Strickland PT, et a1. Induction and localization of cytochrome P450 1B1 (CYPlBl) protein in the livers of TCDD—treated rats: detection using polyclonal antibodies raised to histidine-tagged fusion proteins produced and purified from bac— teria. Carcinogenesis 1998;19:395—402. (149) Kim JH, Sherman ME, Strickland PT, Guengerich FP, Sutter TR. Differ— ential expression of cytochrome P450 1A1 and 1B1 in human lung from smokers, non-smokers and ex-smokers. The Toxicologist, Toxicological Science 1999;48(1-S):96—7. NOTES R. J. Santen is on the advisory board of Lilly Pharmaceuticals, Indianapolis, IN. Supported by Public Health Service (PHS) grants CA63129 and CA74971 (to J. G. Liehr) from the National Cancer Institute, National Institutes of Health (NIH), Department of Health and Human Services, and by PHS grants E508148, ESO3819, and E50714] (to T. R. Sutter) from the National Institute of Envi- ronmental Health Sciences, NIH. J. G. Liehr acknowledges the contributions of M. J. Ricci and Drs. B. T. Zhu, X. Han, D. Roy, A. Gladek, M. Y. Wang, and H. Bhat who contributed to these studies while at the University of Texas Medical Branch, Galveston. T. R. Sutter acknowledges and thanks all of the students, postdoctoral fellows, collaborators, colleagues, and friends who have contributed to the research described in this review; especially the contributions of Dr. William Greenlee (Chemical Industry Institute of Toxicology, Research Triangle Park, NC), Dr. David Spink (Wads- worth Center, Albany, NY), and Dr. Carrie Hayes (University of Memphis, TN). He thanks M. Sherman and J. Kim (The Johns Hopkins University, Baltimore, MD) for their contributions to Fig. 9. Journal of the National Cancer Institute Monographs No. 27, 2000 Chapter 6: Estrogen Metabolism by Conjugation Rebecca Raftogianis, Cyrus Creveling, Richard Weinshilboum, Judith Weisz The involvement of estrogens in carcinogenic processes within estrogen-responsive tissues has been recognized for a number of years. Classically, mitogenicity associated with estrogen receptor-mediated cellular events was believed to be the mechanism by which estrogens contributed to carci- nogenesis. Recently, the possibility that estrogens might con- tribute directly to mutagenesis resulting from DNA damage has been investigated. That damage is apparently a result of the formation of catechol estrogens that can be further oxi- dized to semiquinones and quinones. Those molecules rep- resent reactive oxygen species and electrophilic molecules that can form depurinating DNA adducts, thus having the potential to result in permanent nucleotide mutation. Con- jugation of parent estrogens to sulfate and glucuronide moi- eties; of catechol estrogens to methyl, sulfate, and glucuro- nide conjugates; and of catechol estrogen quinones to glutathione conjugates all represent potential “detoxifica- tion” reactions that may protect the cell from estrogen- mediated mitogenicity and mutagenesis. In this chapter, the biochemistry and molecular genetics of those conjugative reaction pathways are discussed. When applicable, the in- volvement of specific enzymatic isoforms is presented. Fi- nally, the activity of many of these conjugative biotransfor- mation reactions is subject to large interindividual variation—often due to the presence of common nucleotide polymorphisms within the genes encoding those enzymes. Functionally significant genetic polymorphisms that might contribute to variable conjugation of estrogens and catechol estrogens are also discussed. [J Natl Cancer Inst Monogr 2000;27:113—24] The involvement of estrogens in carcinogenic processes within the breast has been appreciated for a number of years (1—3). The classical concept of estrogens as carcinogens recog- nizes the mitogenicity of estrogens via estrogen receptor (ER)- mediated cellular events (1). More recently, as has been detailed throughout this monograph (Chapters 3—5), the role of catechol estrogens (CEs) as genotoxic chemical procarcinogens, indepen- dent of ER mediation, has been recognized (2—4). Although estrogens and CBS differ with regard to the role of the ER in mediating their carcinogenicity, they have in common the po- tential for “detoxification” Via enzyme-mediated conjugation to glucuronide, glutathione (GSH), methyl, and/0r sulfate moieties (2). In this chapter, we will discuss primary estrogen and CE conjugation reactions, with particular emphasis on the biochem— istry and molecular genetics of the human enzymes that catalyze those reactions. Estrogens exert biologic responses in steroid hormone— responsive cells largely via interaction with ERs, members of a superfamily of nuclear hormone receptors that act as ligand- activated transcription factors (5). There are two known ER subtypes, EROL and ERB , which share similar estrogen affinities but have dissimilar expression patterns and response to anties- trogens (5—7). The two most potent endogenous estrogens, es- Journal of the National Cancer Institute Monographs No. 27, 2000 trone and 17B-estradiol, are both ligands for the ERs, although those receptors have higher affinity for l7B-estradiol than for estrone and it is 17B-estradiol that is believed to be the predomi- nant endogenous activator of ER-mediated cellular processes (5 ). The most abundant circulating estrogen, however, is the sulfate conjugate of estrone (8,9). The process by which estro- gens, synthesized and secreted predominantly by the ovaries, are transported to and exert their biologic effects in steroid hormone target tissues is not completely understood. As will be discussed in this chapter, estrogen conjugates, particularly estrone sulfates, are believed to play an important role in that process (9—11). Chemical carcinogenesis emerged as a scientific discipline approximately 50 years ago (12,13). One of the principles of that discipline is that compounds often require metabolic “activa— tion” to form genotoxic and carcinogenic metabolites (12,13). That process involves the establishment of a balance between “activating” and “inactivating” metabolic pathways. The hy- pothesis that estrogens might contribute to the pathophysiology of breast cancer as direct genotoxins (3,4) has raised the possi- bility of just such a balance between estrogen activation and inactivation in those hypothetical genotoxic effects. Specifically, oxidative reactions, often catalyzed by isoforms of the cyto- chromes P450, can result in the formation of CEs from parent estrogens and, subsequently, semiquinones and quinones de- rived from CBS that are capable of forming either stable or depurinating DNA adducts (14—16). Countering the effects of these pathways of metabolic activation are enzymatic reactions that inactivate the parent estrogens, the CBS, and quinones. In— activation pathways involving conjugation reactions, such as methylation, sulfation, glucuronidation, or conjugation with GSH, will be detailed in this chapter. It is important to note that, although a number of animal models have been developed to facilitate the study of CE-mediated carcinogenesis, the focus of this chapter will be primarily on the conjugation of estrogens and CBS in humans. Although it has been hypothesized that conjugated CEs may exhibit biologic activity (2), the focus of this chapter is on conjugation as a detoxification mechanism. Finally, conjugation pathways of both estrogens and CBS dis- play large variations among individuals—often as a result of common genetic polymorphisms. Therefore, the possibility arises that common, inherited variations in enzymatic pathways for estrogen bioactivation or in the inactivation of either the parent compound or downstream metabolites might represent individual risk factors for the occurrence of breast cancer. The Afiiliations of authors: R. Raftogianis, Department of Pharmacology, Fox Chase Cancer Center, Philadelphia, PA; C. Creveling, National Institute of Dia- betes and Digestive and Kidney Diseases, Bethesda, MD; R. Weinshilboum, Department of Pharmacology, Mayo Medical School/Mayo Clinic/Mayo Foun- dation, Rochester, MN; J. Weisz, Department of Obstetrics and Gynecology, Milton S. Hershey Medical Center, Pennsylvania State University. Correspondence to: Rebecca Raftogianis, Ph.D., Department of Pharmacol- ogy, Fox Chase Cancer Center, 7701 Burholme Ave., Philadelphia, PA 1911] (e—mail: RB_Raftogianis@FCCC.EDU). See “Notes" following “References.” © Oxford University Press 113 molecular epidemiology of estrogen carcinogenesis is detailed in Chapter 7. That chapter focuses on genetic polymorphisms that have been studied as risk factors for estrogen-mediated carcino- genesis. This chapter will present the current state of knowledge with regard to functionally significant genetic polymorphisms in human genes encoding estrogen—conjugating enzymes, many of them as yet untested as breast cancer risk factors. ESTROGEN CONJUGATION Biologic Role of Estrogen Conjugation The endogenous formation of estrogen conjugates has long been recognized as a major route of estrogen metabolism (17). Both endogenous and synthetic exogenous estrogens are exten- sively biotransformed to estrogen conjugates in humans (Fig. l) (2,18). The most abundant circulating estrogen conjugates are the sulfates, followed by the glucuronides. It is important to note that conjugated estrogens are not appreciable ligands for the ERs; thus, they do not promote ER—mediated activity (2). Intu- itively, it was initially assumed that sulfate and glucuronide conjugation of estrogens represented a pathway resulting in less active, more polar, and more readily excreted estrogenic com- pounds. It is now appreciated, however, that estrogen sulfates actually exhibit a much longer half—life than do the parent es- trogens (2, 8, 1 1). Estrone sulfate is the most abundant circulating estrogen, at concentrations approximately 10—fold higher than unconjugated estrone (8). That finding, as well as increasing knowledge about the transport and subsequent desulfation of estrogen sulfates, has led to a widely held belief that sulfated steroid hormones serve an important biologic role as steroid hormone precursors, particularly for steroid hormone-responsive tissues (2,9). An increasing body of scientific data supports the hypothesis that sulfation and desulfation of estrogens may well represent an endogenous system important in the regulation of biologically active steroid hormones in target tissues (10,11). Specifically, it is currently hypothesized that inactive estrone sulfate is transported to target tissues via the circulatory system, taken into target cells, most likely by organic anion transporters, enzymatically hydrolyzed to estrone by intracellular membrane- bound steroid sulfatase (arylsulfatase C), and hydroxylated to active l7B-estradiol via catalysis by l7B-hydroxysteroid dehy- drogenases (2,11,18,19). l7B—Estradiol activates the ER via li- gand binding and initiates a number of downstream ER— mediated events—most notably related to transcriptional activation of those genes that contain DNA sequences that bind and respond to activated ERs (5,18). The transport of estrone sulfate into steroid hormone- responsive cells is not well understood; however, some studies (19,20) have shown that a human organic anion transporter (Oatpl) has high affinity for both sulfate and glucuronide estro- gen conjugates. Furthermore, this transporter is typically respon— sible for intracellular import of organic ions rather than the efflux of these compounds out of the cell. The level of expres— sion or activity of this transporter in human breast tissues has not yet been reported. Many target tissues including the breast ex- hibit estrogen sulfation activities in addition to the ability to desulfate estrogen sulfates (8,9). This “cycling” has been dem- onstrated in mammalian cells and, like phosphorylation and de- phosphorylation of proteins during cell—signaling processes, sulfation and desulfation of steroid hormones possibly represent an intracellular regulatory mechanism for estrogenic activity 114 (Fig. 1) (811,21). Recognition of the importance of steroid sul— fatase activity in the formation of intratumoral estrogens has resulted in the development of a number of steroid sulfatase inhibitors for the treatment of steroid hormone—responsive tu- mors (22,23). Enzymes responsible for the glucuronidation and degluc— uronidation of estrogens are also expressed in a variety of human tissues, including the breast ( 24,25 ). Estrogen glucuronides have received much less attention, however, than have the sulfate conjugates as steroid hormone precursors, most likely because they are less abundant and more readily cleared from the body (2). Breast tumors and breast cancer cell lines express high lev- els of B-glucuronidase, the enzyme that catalyzes the hydrolysis of estrogen glucuronides; however, appreciable estrogen gluc- uronide cycling in breast tissue has not been demonstrated (24). Although the concept of estrogen glucuronides as steroid pre- cursors has been underinvestigated, it is generally accepted that glucuronidation of estrogens serves primarily a classical excre- tory role. Estrogen glucuronide conjugates are readily excreted in both urine and bile (26). Biochemistry of Estrogen Conjugation Sulfation. Sulfate conjugation of estrogens is catalyzed by several members of a superfamily of cytosolic sulfotransferase (SULT) enzymes (27,28). SULT enzymes catalyze the transfer of S03" from 3’-phosphoadenosine—5’-phosphosulfate, the en— zymatic cofactor, to, in the case of estrogens, phenolic acceptor groups (28). Cytosolic SULTs are active as homodimers. Sulfa- tion of estrone and l7B-estradiol occurs at the 3-phenolic O // // H03 5— 0 glucuronide —0 Estrone Sulfate B-gluc- Estrone Glucumnide uronidase Vows / Sterold\ O UGTS Sulfatase // H Estrone l7fl-hydmxysteroid» dehydrogenase OH H Estradiol SULTs B-gluc- Steroid Vomdase Sulfatase UG'I\ OH OH Ho3 s—O glucuronide—o Estradiol Sulfate Estradiol Glucuronide Fig. 1. Conjugative pathways for estrone and l7B—estradiol. Sulfation, desulfa- tion, glucuronidation, and deglucuronidation reactions catalyzed by sulfotrans- ferases (SULTs), steroid sulfatase, UDP—glucuronosyltransferase (UGTs), and B-glucuronidase, respectively, are shown. Journal of the National Cancer Institute Monographs No. 27, 2000 group of the steroidal A ring (Fig. 1). Estrogen SULT activity has been demonstrated in a variety of human tissues, including liver, small intestine, kidney, placenta, uterus, adrenal gland, and breast (29—33). The level of estrogen SULT activity in the hu- man liver is high, and this activity is believed to contribute significantly to the high circulating levels of estrone-3-sulfate (29,30). Although from a quantitative perspective, sulfation of estrogens in the liver is probably the most important overall estrogen—conjugating activity in the body, sulfation of estrogens in steroid target tissues, including the breast, has also been dem- onstrated and may well be important in affecting the biologic activity of estrogens within those tissues (2,] 1). Study of the association of estrogen SULT activity with breast cancer has been an active area of research. Expression of estrogen SULT activity within breast tumors has been reported to correlate with the ER status of the tumor as well as with the response of tumors to estrogens and adrenalectomy (33,34). However, other studies have shown no such correlation or even an inverse correlation (35). Such contradictory findings are in- dicative of the difficulty investigators have encountered in the study of estrogen sulfation in steroid target tissues. As will be discussed shortly, the reason for those difficulties has recently been appreciated in that we now know that multiple SULT en- zymes contribute to estrogen SULT activity and, importantly, there is significant interindividual variation in the level of ac- tivity of the enzymes catalyzing the sulfation of estrogens (27,30,36). Furthermore, SULTs are subject to profound sub— strate inhibition (32). The concentration of substrate at which inhibition occurs differs among estrogen-sulfating isoforms such that slight differences in experimental conditions would have important implications in the interpretation of resulting data. Glucuronidation. Estrogen glucuronidation is catalyzed by several members of a superfamily of microsomal UDP- glucuronosyltransferase (UGT) enzymes (25,37). UGTs catalyze the conjugation of UDP-glucuronic acid, the UGT cosubstrate, to a variety of endogenous and exogenous aglycones, including steroid hormones (38). Whereas estrogens are sulfated predomi- nantly at the 3 position, glucuronidation can occur at either the 3 or 178 hydroxyl group of steroidal hormones, with the 178 position being the apparent predominant site of glucuronidation for l7B-estradiol (Fig. 1). Glucuronidation of estrogens renders those molecules less lipophilic and more readily excreted in both urine and bile. l7B-Glucuronides of estradiol are known to in- duce cholestasis, putatively via interaction with hepatocyte can— alicular membrane efflux transporters such as MDRl and MRPZ/CMOAT (39,40). Steroid hormone glucuronidation has been observed in hu— man liver, biliary epithelium, kidney, gut, prostate, ovary, and breast (25,26,38). In a study comparing UGT activity in matched breast ductal carcinoma and peritumoral tissues, the authors (24) reported activity in tissues from only four of the 12 individuals studied. Furthermore, in those four sample pairs, the level of activity was fivefold lower in tumor tissue than in the peritu- moral tissue. However, those studies were conducted with the use of 4-methylumbelliferone as substrate (as opposed to an estrogen), and it is not clear whether that activity correlates with estrogen glucuronidation in the breast. Glucuronidation is a ma— jor route of androgen metabolism, and the study of this pathway has received much attention in terms of its role in the patho- physiology of androgen-dependent diseases (41). However, the role of estrogen glucuronides in breast cancer has received little Journal of the National Cancer Institute Monographs N0. 27, 2000 attention compared with sulfate conjugation. This is most likely due to the perception that estrogen glucuronidation serves a predominantly excretory role, secondary to sulfate conjugation. It is clear that much further study of the glucuronidation of estrogens is required before we can fully understand the bio- chemistry of this pathway and its role in affecting estrogen ac- t1v1ty. Molecular and Cellular Aspects of Estrogen Conjugation Biochemical studies of estrogen conjugation provided much knowledge about these important metabolic pathways. However, there were also many questions left unanswered by these studies, and we now have begun to be able to answer some of those questions using the tools and further knowledge gained with the advent of molecular biology. There are a surprising number of SULT and UGT isoforms capable of contributing to the conju- gation of estrogens. Those isoforms are often expressed in a tissue—specific manner and are often under specific regulatory control. Furthermore, a number of those conjugative enzymes are encoded by genes known to harbor common genetic poly- morphisms. These factors help explain many of the complexities of estrogen conjugation—and this knowledge allows us to probe estrogen conjugation reactions in a systematic fashion. Sulfotransferases. The cloning of SULT genes and comple— mentary DNAs (cDNAs) is a very active area of research (27). Currently, there are at least 10 unique cytosolic SULT enzymes known to be expressed in human tissue (27,42—45). On the basis of amino acid sequence identity, those 10 human SULTs fall within two families, SULTl and SULTZ. Subfamilies include SULTs 1A, 1B, 1C, 1E, 2A, and 2B. The 1A, 1C, and 2B families each have multiple members. Although amino acid identity allows the classification of these enzymes into families and subfamilies, members exhibit overlapping substrate affini- ties even across families. Estrone and l7B-estradiol are sub- strates for SULTlAl, SULTlEl, and SULT2Al, although the affinity of these enzymes for estrogens varies (Table l) (27). Overlapping substrate specificity of SULTs toward estrogens complicates the study of estrogen sulfation. For example, the high affinity of SULTlEl for estrogen substrates suggests that this enzyme plays a major role in the endogenous sulfation of estrogens, and the activity of this enzyme in the liver likely contributes significantly to the quantitatively large pool of cir— culating estrogen sulfates (29). It would be logical to hypoth- esize that this enzyme activity might be important in regulating estrogen activity in breast tumors. However, studies have sug- gested that, although SULTlEl appears to be expressed in nor- mal breast epithelial cells, it is not highly expressed in breast tumors or cell lines derived from breast tumors (46). SULTlAl and, to a lesser extent, SULT2Al appear to be the SULT iso- forms primarily responsible for estrogen sulfation in breast tu- mors (46—48). These findings suggest that a specific SULT iso- form may play a variable role in endogenous steroid hormone sulfation, depending on the tissue and the disease of interest. Relatively little is known about the regulation of SULT genes. Although genes have been cloned for most of the human SULT cDNAs and enzymes identified to date, the DNA se- quences contributing to the promotion or regulation of transcrip- tion of those genes have not been well defined. Of the human SULT genes cloned thus far, only SULTlEl contains a canoni- cal TATA box element, and experimentally determined sites of transcription initiation appear to correspond to the location of 115 Table 1. Specificity of sulfotransferase (SULT) and UDP—glucuronosyltransferase (UGT) isoforms with various estrogen and catechol estrogen substrates Substrate* Isoform Estrone Estradiol 2-OH-estrone 4—OH—estrone 2-OH-estradiol 4-OH-estradiol SULTs 1A1 X X ND X 1E1 X X ND ND ND ND 2A1 X X ND — — X UGTs 1A1 — X X X X X 1A3 X ND X X X X 1A4 — X — - X X 1A7 — — ND — X ND 1A8 X ND X X X X 1A9 X X X X X X 1A10 X X ND X X ND 2B4 — — ND X X ND 2B7 — — X X X X *X = isoform has been shown to conjugate indicated substrate; — = isoform has been shown not to conjugate the indicated substrate; ND = interaction of the indicated isoform/substrate pair has not been determined. See text for details and references. that element (49). The level of estrogen SULT activity in human tissues has been reported to be under the influence of steroid hormones (28). In concordance with that finding, the 5’-flanking region of the SULTlEl gene contained half palindrome gluco- corticoid and thyroid hormone response elements (49). How- ever, the functional significance of those elements has not yet been studied experimentally. Additional evidence of SULT regulation includes the identification of alternative sites of tran- scription initiation for the SULT1A1 gene (50). The regulation or tissue selectivity of alternative transcriptional initiation of SULT1A1 has not been well studied. Finally, conjugation of estrogens is known to vary signifi— cantly among individuals (29,51). That observation raises the possibility that genetic variation in the genes contributing to estrogen conjugation (pharmacogenetics) may contribute to in- terindividual variation in estrogen metabolism. As will be dis- cussed in Chapter 7, a number of genetic variants in genes con- tributing to estrogen metabolism have been reported to represent risk factors for the development of breast cancer. The study of the pharrnacogenetics of SULT enzymes is currently an active endeavor. A common, functionally significant genetic polymor- phism has been described for SULTlAl (52,53). That SULTlAl polymorphism results in an Arg2l3His amino acid substitution. Correlation of the level of SULT activity in human blood platelet samples and SULTlAl genotype suggests that individuals homozygous for the His allozyme exhibit a signifi- cantly diminished capacity to sulfate prototypic phenolic mol- ecules (52). The contribution of this polymorphism to interindi- vidual variability in the conjugation of estrogens or as a risk modifier for breast cancer has not yet been reported. Similarly, large interindividual variations in the level of SULT2A1 activity in human liver and the level of immunore- active protein in intestinal tissues have also been reported (30,54). Genetic polymorphisms resulting in Met57Thr and Glu186Val amino acid changes in SULT2A1 have been reported (55). Functional studies of the recombinant SULT2A1 allo— zymes (55) have shown that those amino acid changes, particu— larly when coexpressed, result in a diminished level of recom- binant enzyme activity. However, SULT2A1 genotype does not appear to correlate significantly with the level of apparent SULT2A1 activity in human tissues (55 ). Finally, the presence 116 of large differences in the level of immunoreactive SULTlEl protein in samples of human small intestines raises the possibil— ity that genetic polymorphisms might also exist for that enzyme (30). That possibility is currently the subject of active study, but no polymorphisms in the SULT1E1 gene have yet been reported. Glucuronosyltransferases. As with the SULTs, a number of UGT isoforms are now known to contribute to the conjugation of estrogens (Table 1). The degree of contribution of individual UGTs to that activity is not yet well understood, and it is likely that specific isoforms will contribute differently, depending on the tissue and the disease of interest. There are currently at least 12 functional UGT isoforms known to be expressed in human tissues (37). Like the SULTs, those 12 enzymes fall into two families within the human UGT superfamily of microsomal en- zymes. Glucuronidation of estrone and 17B—estradiol appears to be catalyzed by several members of the UGTl family. Thus far, recombinant human UGTs 1A1, 1A3, 1A4, 1A8, 1A9, and 1A10 have all been shown to catalyze the glucuronidation of estrone and/or 17 B—estradiol (Table 1) (25,26,56—60). It is interesting to note that, for some human recombinant UGT isoforms, there appears to be selective affinity for estrone or l7B-estradiol as substrate. For example, UGTs 1A1 and 1A4 exhibit activity toward l7B-estradiol but not toward estrone, whereas UGTs 1A9 and 1A10 have been reported to catalyze the glucuronida- tion of both of these estrogens (25,26,56,58,60). Activity for UGTs 1A8 and 1A3 toward estrone has been reported, but the activity of those isoforms toward 17B-estradiol has apparently not been evaluated (57,59). There is much known about the expression of human UGTlA isoforms in various tissues. It should be noted, however, that the tissue distribution profile of some UGT isoforms has not been as extensively characterized as others. UGT1A1 is expressed in human liver, colon, and biliary epithelium (gallbladder) but not in stomach (60). UGT1A3 has been reported to be expressed in human colon, biliary epithelium, and liver, but the level of ex- pression in liver varied significantly between individuals and was fivefold to 10-fold less than the level of UGTs 1A1 and 1A4 (57, 60). UGT1A4 is expressed in human liver, colon, and biliary epithelium but not in stomach (60). UGT1A8 appears to be expressed specifically in human intestinal tissues (59,60). UGT1A9 is expressed in human prostate, testis, breast, ovary, Journal of the National Cancer Institute Monographs No. 27, 2000 skin, skeletal muscle, stomach, small intestine, colon, liver, and kidney but not in biliary epithelium or stomach (25,60). UGTlAlO is expressed in colon, biliary epithelium, and stom- ach but not in liver (60). It is important to note that, although only the UGT1A9 isoform has been reported to be expressed in human breast to date, that is likely a reflection of the lack of evaluation of the level of expression of various UGT isoforms in that tissue. The regulation of the UGTl family is currently not well characterized but is an active area of study. The most notable feature of this gene family is that all of the UGTlA isoforms disseminate from a single “nested” gene structure (37). There are six coding exons in the human UGTlA genes, and each isoform is encoded by the same exons 2 through 5. The only differen- tiation between isoforms is that each exon 1, encoding the N- terminal half of the protein, is unique, and isoform specificity results from alternative transcription initiation and usage of unique exons 1 (37). Therefore, each isoform is under the control of individual promoter sequences, and isoform-specific regula- tion has been observed. For example, as noted in the previous paragraph, UGTlA isoforms are differentially expressed in hu- man tissues. As previously noted, the capacity for estrogen conjugation and, specifically, glucuronidation is known to vary widely in the human population (51). That observation raises the possibility that genetic variation may exist in the UGT isoforms that con— tribute significantly to estrogen glucuronidation. A functionally significant common polymorphism in the promoter sequence of the UGTlAl gene has been described and well characterized (61—63). That polymorphism is a variable length (TA)n TAA repeat in the functional TATA box upstream of exon 1 of the UGTlAl gene. The wild-type allele is defined as n = 6. Allelic variants identified to date include n = 5, 7, and 8 (63). In vitro studies utilizing reporter constructs driven by allelic variants of the UGTlAl promoter (63) have shown that promoter activity appears to decrease with increasing n. Furthermore, clinical as- sociation of the most common variant (n = 7) with a relatively poor ability to glucuronidate bilirubin (Gilbert’s syndrome), as well as the chemotherapeutic agent SN-38, has been observed (61,64). Studies determining the association of UGTlAl alleles with estrogen metabolism and risk modification of breast cancer have not yet been reported, but they will be of great interest. CE CONJUGATION Biologic Role of CE Conjugation The putative role of CBS in the mediation of breast carcino— genesis has been described in Chapters 3—5 of this monograph. The biotransformation of estrone and estradiol to CEs involves hydroxylation at the 2 or 4 position of the steroidal A ring of these parent estrogens (3,14). Those reactions are catalyzed by multiple cytochrome P450 isoforms. Both the 2- and the 4-hy- droxy CEs can be further oxidized to CE quinones (CE-Qs) or semiquinones (Fig. 2) (16). The 2-hydroxy CE-Qs have been shown to form stable DNA adducts, whereas the 4-hydroxy CE- Qs have been shown to form depurinating adducts (16,65,66). There is good evidence suggesting that those depurinating ad- ducts can lead to apurinic DNA sites and permanent mutations that, when inflicted upon critical DNA sequences, can lead to tumorigenesis (16,66). CEs can also enter into redox cycling Journal of the National Cancer Institute Monographs No. 27, 2000 and, thereby, become a source of reactive oxygen species (3). Hence, unless CEs are inactivated, they may contribute to car- cinogenesis by causing DNA damage mediated by reactive oxy- gen species and by direct interaction of CE-Qs with DNA to form depurinating adducts (65,66). Fortunately, our cells are fortified with an armament of conjugative pathways that result in the biotransformation of toxic estrogen metabolites to relatively nontoxic moieties (2 ). Generally, the reactive CEs are detoxified by biotransformation to predominantly methyl conjugates, to a lesser extent glucuronides, and possibly sulfate conjugates (Fig. 2) (2). A further conjugative safeguard lies in the detoxification of CE-Qs via conjugation to glutathione (16,67). Therefore, the actual risk of CBS in causing DNA damage may well depend on the ability of individual cells to conjugate CBS and CE-Qs rela— tive to the rate of formation of these toxic estrogen metabolites. In the rest of this chapter, we will focus on CE conjugative pathways and those conjugative enzymes responsible for the detoxification of CBS and CE—Qs. The most well-studied CE conjugation reaction is that of methylation. CE methylation is catalyzed by catechol-0- methyltransferase (COMT), an enzyme that exists in both “soluble” (S—COMT) and membrane-bound (M-COMT) forms, as discussed in detail below (16,68). Studies in the hamster kidney (69) provided the first example linking an estrogen- induced cancer with the induction of COMT. The localization of COMT in the epithelial cells of the proximal convoluted tubules of the hamster kidney is similar to its localization in the rat kidney reported earlier (70). Hamsters treated with primary es- trogens, such as estrone and estradiol, develop tumors in the renal cortex. There is evidence that the carcinogenicity of estro- gens for hamster kidney results from a combination of factors: 1) an increase in the catechol load, 2) the presence of high levels of 2- and 4-hydroxylated CEs subject to oxidative metabolism in the renal cortex, and 3) a relative insufficiency of COMT (71). In control hamsters, COMT was localized in the cytoplasm of proximal convoluted tubules, predominantly in the juxtamedul- lary region where estrogen-induced tumors arise. After 2 or 4 weeks of treatment with estrogen, COMT was seen in epithelial cells of the proximal convoluted tubules throughout the cortex. Moreover, many cells showed intense nuclear COMT immuno- reactivity (Fig. 3) (69). The estrogen-induced cancers were COMT negative but were surrounded by tubules with epithelial cells with intense cytoplasmic and nuclear immunostaining. Im- munoblot analysis indicated that the nuclear COMT, shown in Fig. 3, was S-COMT. This translocation to the nucleus was shown by sequencing of hamster kidney COMT messenger RNA to occur in the absence of a nuclear localization signal. This pattern of induction of COMT in hamster kidney in re- sponse to estrogen treatment, in particular in the nucleus, has been interpreted as a possible response to “a threat” to the ge— nome by products of oxidative metabolism of CEs. It is of interest that nuclear localization of COMT is not unique to hamster kidney but also can be seen in some normal, as well as neoplastic, mammary epithelial cells (72). Human breast tissues have the capacity to synthesize both 2- and 4-hy— droxyestrogens (71,73). A cytochrome P450 that catalyzes the 2- and 4-hydroxylation of estrogen has been identified by immu— nocytochemistry in human ductal epithelial cells (74). COMT has also been identified in those cells (75). High levels of oxi- datively damaged DNA have been found in breast tissue from women in the United States (76,77). It is reasonable to propose 117 COMT SULTs R GSTs R UGTs / CYPs HO peroxidases O H Estrogen Sulfate and Glucuronide Co 'u ates Fig. 2. Conjugation of estrogens, n] g HO 0 catechol estrogens, and estrogen SULTs Steroid Sulfatase 2-0H-Estradiol(Estrone) Estradiol(Estrone)—2,3-Quin0ne quinones. Reaction pathways and UGTs B—Glucuronidase the enzymes involved are shown. R For s1mpl1c1ty, the formanon of es— CYPs trogen semiquinones is not de- picted. (See Chapter 5 for details.) COMT = catechol-O—methyltrans— ferase; SULTs = sulfotransfer- H ases; UGTs = UDP-glucuronosyl— EstradiolzR, -OH CYIPS transferases; CYPs = cyto— EmmeR =0 peroxidases . , —_) chromes P450; GSTs = glutathi- one S-transferases. Catechol Estrogen Methyl, Sulfate and Glucuronide Conjugates R HO w HO 4-OH-Estradiol (Estrone) COMT SULTs UGTs Catechol Estrogen Methyl, Sulfate and Glucuronide Conjugates Estrogen Quinone Glutathione Conjugates R ;| ./ O o Estradiol (Estrone)-3,4—Quinone lGSTs Estrogen Quinone Glutathione Conjugates that, in human breast tissue, like the kidneys of hamsters treated with estrogen, oxidative metabolism of CBS might contribute to this oxidative damage. Biochemistry of CE Conjugation Methylation. Quantitatively, the most active CE conjugative pathway is methylation. CE methylation is catalyzed by COMT, a member of a superfamily of methyltransferase enzymes (68). COMT, a classical phase II enzyme, catalyzes the transfer of methyl groups from S—adenosyl methionine, the enzyme cofac- tor, to hydroxyl groups of a number of catechol substrates, in- cluding the CEs. Under normal circumstances, CEs are, for the most part, promptly O-methylated by COMT to form 2- and 4-0-methylethers, which are then excreted (78). While virtually all catechols are substrates for COMT, the highest affinities for the enzyme are exhibited by the CEs ( 78). The existence of this metabolic pathway helps to explain the extremely short half-life of CBS and the predominance of O-methylethers of CBS, in particular of 2-methylethers, as the major metabolites of estrone and estradiol in urine (79). However, under circumstances dur- ing which the capacity for O-methylation is reduced or inhibited by an excess catechol load, the half—life of CEs may be extended. This phenomenon could have special importance for specific cellular sites, such as breast epithelial cells, where CEs are formed. COMT might play an important role in protecting the genome from damage that could be caused by the metabolism of estrogens through activation of the CE-Q pathway. A number of investigators are now studying the involvement of this enzyme as well as the interindividual variability of COMT enzyme ac- tivity in detoxification of CEs specifically in the context of breast carcinogenesis. 118 The hypothesis that COMT provides a protective mechanism against cytotoxicity and genotoxicity by preventing the oxida- tion of catechols is in its infancy. At present, we know enough to consider O-methylation an important mechanism for prevent- ing cytotoxic and genotoxic damage caused by products of the oxidative metabolism of catechols. This knowledge may gener- ate avenues for therapeutic intervention where a deficit in the capacity for O-methylation appears to be a risk factor in carci— nogenesis (80,81). Sulfation and glucuronidation. While methylation of CEs has been well studied, very little is known about the role of sulfation and glucuronidation in the detoxification of CEs. The excretion of both sulfate and glucuronide conjugates of CEs has been observed in rats (82), and it is clear from a number of in vitro studies that UGTs and SULTs are able to catalyze the conjugation of CBS. We also know that those enzymes are ex- pressed in the liver and estrogen-responsive tissues, such as the breast epithelium (25,27). Therefore, it is plausible to suggest that these reactions may play a biologic role in the detoxification of CEs. A number of studies [reviewed in (2 )] have reported that the major urinary metabolites of CBS are the methyl conjugates. From a quantitative perspective, therefore, the formation of sul- fate and glucuronide CEs does not appear to represent major pathways in the overall metabolism and excretion of CBS. How- ever, because the reactivity and toxicity of the CEs are intracel- lular phenomena, it has been suggested that local metabolism of CBS within target cells will be just as important as the overall detoxification of CBS in the liver (2). For that reason, a number of investigators are now studying the role of sulfate and gluc— uronide conjugation of CBS in the intracellular detoxification of these carcinogens. COMT, UGTs, and SULTs often share affin- Joumal of the National Cancer Institute Monographs No. 27, 2000 Fig. 3. Immunoreactive catechol—0-methyltransferase (COMT) (left) and CuZn—superoxide dismutase (CuZnSOD) (right) in proximal convoluted tubules from the same region in adjacent tissue sections from the kidney of a hamster treated for 2 weeks with estradiol (original magnification X500). Intense immunostaining for COMT is seen with many nuclei in contrast to the perinuclear immunostaining for CuZnSOD. Arrows point to some of the cells with distinct immunostaining for COMT and perinuclear staining for CuZnSOD. This figure is used by permission of Oxford University Press from Carcinogenesis (69). ity for the same substrates. An indirect role that has been sug— gested for the relevance of SULTs and UGTs in the detoxifica— tion of CBS is that those enzymes might represent pathways for the conjugation of other catechol substrates that may compete for, and thus inhibit, the capacity of COMT to detoxify CBS (2 ). Those hypotheses have yet to be rigorously tested at the experi- mental level. GSH conjugation. The reactivity of CE—Qs relates to their ability to undergo redox cycling, creating oxidative stress, and/ or to react directly with cellular nucleophiles (such as DNA) (3,16). Conjugation of quinones to GSH, a major cellular sulfhydryl tripeptide, is generally considered a detoxifica- tion mechanism (16,83). GSH conjugation of CE-Qs has been shown to occur both in vivo and in vitro (84). GSH-conjugated CE-Qs are then rapidly converted to mercapturic acid meta- bolites that are readily excreted from the cell. It is primarily this excretory role of GSH conjugation that is believed to contribute to the detoxification of CE-Qs. However, the actual degree of detoxification of CE-Qs that is imparted by GSH conjugation is unclear because GSH-conjugated quinones are capable of undergoing the same redox cycling reactions as are the parent quinones and semiquinones (83,84). Those reac- tions result in the formation of reactive oxygen species that can themselves cause DNA damage. Therefore, the “net” protective effect of conjugation of CE-Qs by GSH depends on the relative balance between GSH-mediated CE excre- tion and the GSH-mediated formation of reactive oxygen species. Most studies appear to confirm that conjugation of CE-Qs with GSH results in a net decrease in DNA damage (67,85). Molecular and Cellular Aspects of CE Conjugation Catechol-0-methyltransferase. A single gene encoding COMT is expressed at the protein level in two forms as a con- sequence of the existence of alternative transcription initiation sites (86). The two transcriptional products result in the trans- lation of an S-COMT and an M-COMT enzyme with Mr values of 23 000 and 26 000 daltons, respectively. M—COMT includes Journal of the National Cancer Institute Monographs No. 27, 2000 an additional 50 amino acid residues at the N-terminus of the protein that are not present in S-COMT (86,87). Of the two forms, the cytosolic S-COMT has a lower affinity but a higher capacity for catecholamines than M-COMT (68). The relative expression of the two COMT enzymes varies with different tissues, but S-COMT appears to be the dominant form in most cell populations (87—89). COMT is widely distributed, with high levels of activity being reported in the liver and kidney epithe- lium, as well as in the ependymal and glial cells. In breast tissue, immunoreactive COMT has been observed in both normal and neoplastic epithelial cells (75). In neoplastic cells of rodent and human breast, COMT enzyme activity, expressed as units per milligram of protein, has been reported to be elevated (75,90). However, this apparent increase may be due to an increase in cell numbers in neoplastic breast parenchyme. Extensive cytochemical studies of the localization of COMT both at the cellular and subcellular levels (91) support the hy- pothesis that COMT plays a critical role in the local regulation of catechols at specific target sites. Regulation of COMT ex- pression appears to be tissue selective and site specific. In liver and possibly in red blood cells, COMT functions in the 0— methylation of circulating endogenous and xenobiotic catechols (92). In addition, in liver, quantitatively the most important site for the metabolism of estrogens via 2-hydroxylation, COMT serves to inactivate 2—OH CEs close to the site where they are formed. In many other tissues in which COMT is expressed, it appears to have a critical role in restricting the passage of cat— echols between tissue compartments (93). An example is the dense concentration of COMT in the epithelial cells of the cho- roid plexus that separate the vascular system from the spinal fluid. Another example is the presence of COMT in ependymal cells lining brain ventricles separating the spinal fluid from the brain parenchyma. The presence of COMT in astrocytes, oligo- dendrocytes, and microglia may well restrict the movement of catechols to “fields” in the central nervous system. In certain tissues, the expression of COMT has been shown to be under hormonal control. Studies of the expression of COMT in the rat uterus provide 119 an example of a precise spatial and temporal expression of COMT and of its hormonal regulation in relation to a critical physiologic event, implantation (94). Immunoreactive COMT becomes evident in the luminal epithelium of the uterus at the site of decidualization just before implantation on day 3 of preg— nancy. The role of progesterone in the induction of COMT was demonstrated by the effective blockade of enzyme expression by RU—486 (95). Since there is evidence that CBS generated in the uterus may have an important role in the process of implantation, the induction of COMT by progesterone could serve to delimit the action of CEs to the implantation site (96). Finally, levels of COMT activity in humans were shown more than 20 years ago to be controlled, in part, by a common genetic polymorphism (97). The phenotypic trait of low COMT activity was found in approximately 25% of a Caucasian popu- lation. Molecular pharmacogenetic studies (98) have identified a single nucleotide polymorphism in the COMT gene that results in a Va1108Met (amino acid 108 in S-COMT) amino acid sub- stitution. This amino acid change is of great functional signifi- cance, since the methionine substitution results in a protein with low enzyme activity, and correlation of low COMT activity with COMT genotype has been reported in human tissues. It is no— table that this COMT genetic variant represents a truly “bal— anced” polymorphism, in that the frequency of occurrence of each allele is approximately 50%. The description of the mo- lecular genetic basis for low COMT activity made possible ge- netic epidemiologic studies and, as pointed out in Chapter 7, COMT has been a focus for studies of the genetic epidemiology of breast cancer. Unfortunately, the results of those studies (80,81,99) are conflicting. Therefore, these two complementary issues serve to illustrate—in both a broad and a highly focused fashion—the promise and limitations of this overall research strategy. This approach almost certainly will be applied with increasing frequency to help elucidate the possible contribution of direct estrogen genotoxicity to the pathophysiology of breast cancer and other neoplasia. Sulfotransferases and UDP-glucuronosyltransferases. The role of sulfation and glucuronidation as detoxification pathways for CBS is underinvestigated. However, catalysis of CE conju- gation by human recombinant SULT and UGT enzymes has been reported. Those results will be presented here, but it should be cautioned that the relevance these studies have to in vivo CE conjugation is not yet clear. SULTlAl, in addition to catalyzing the sulfation of estrone and estradiol, also catalyzes the sulfation of 4—hydroxyestrone, as well as of 2- and 4-hydroxyestradiol (Table 1) (100). In that same study, SULT2Al was reported not to catalyze the sulfation of 2-hydroxyestradiol or 4-hy- droxyestrone but to have marginal activity toward 4-hydroxyes- tradiol. There have apparently been no other reports of specific SULTs catalyzing the sulfation of CEs. It is tempting to speculate that SULT1A3, a catechol- preferring SULT, or SULTlEl, an estrogen SULT, might par— ticipate in the sulfation of CEs. Reverse transcription— polymerase chain reaction studies have suggested that SULT1A3 was highly expressed in human breast tumors and cell lines relative to the expression of SULTlAl (Raftogianis R: unpublished data). However, it is not yet known whether SULTlA3 contributes to the sulfation of CEs. It has also been suggested that SULT] A3 might indirectly contribute to the regu- lation of CE conjugation by sulfating other catechols that would otherwise compete for COMT (2), thus inhibiting CE methyl— 120 ation. Hypotheses involving the role of SULT1A3 in CE con- jugation have yet to be rigorously tested experimentally. SULTlEl is expressed in normal breast epithelium, but it is not known whether that enzyme catalyzes the sulfation of CBS (46). A common polymorphism has been described for SULTlAl (52,53), and a number of laboratories are currently testing the hypothesis that this polymorphism may represent a risk factor for breast cancer. SULTlAl polymorphisms are hypothesized to modify susceptibility to estrogen-mediated carcinogenesis via both sulfation of parent estrogens and variable detoxification of CEs (Figs. 1 and 2). Finally, biochemical pharmacogenetic stud- ies (101) have shown that a common genetic polymorphism results in interindividual variation in the activity of SULTlA3. However, there have been no reports on the molecular basis for this polymorphism. Should SULT1A3 be involved in the detoxi- fication of CEs, it is possible that polymorphisms in this gene might represent risk factors for susceptibility to CE-mediated breast cancer. A large number of human recombinant UGTs, from both the UGTl and UGTZ families, catalyze the glucuronidation of CEs (Table 1). Although there is much substrate overlap among these isoforms, there does appear to be some selectivity of isoforms toward specific CEs. UGT1A1 and UGT1A3 both catalyzed the conjugation of 2~ and 4-hydroxy CEs, with particularly high activity toward the 2—hydroxy CEs ( 102). UGT1A4 exhibited low levels of activity toward 2- and 4-hydroxyestradiol and no activity with estrone CBS (58). UGT1A7 has been shown to catalyze the glucuronidation of 2-hydroxyestradiol (60). UGTlAS and UGT1A9 have also been reported to conjugate all four CEs, but with particularly high activity toward the 4~hy- droxy CEs (25,59). In a separate publication (60), however, UGT1A8 was reported not to catalyze the conjugation of 2-hy- droxyestradiol or 4-hydroxyestrone. UGT1A10 catalyzed the conjugation of 2-hydroxyestradiol and 4-hydroxyestrone (60). In the UGT2 family, the recombinant enzymes for both UGT2B4 and UGT2B7 catalyzed the glucuronidation of CEs (102—104). UGT2B4 (previously referred to as 2B1 1) catalyzed the conju- gation of 4—hydroxyestrone and 2—hydroxyestradiol (104). UGT2B7 exhibited activity toward the 2- and 4—hydroxy CEs, with particularly high activity toward the 4-hydroxy CEs (102,103). In addition to the functional variable repeat polymorphism in the TATA box already discussed for UGTlAl, common poly- morphisms exist in both UGT2B4 and UGT2B7. The UGT2B4 polymorphism is defined by an Asp458Glu amino acid substi- tution that results in a protein with diminished UGT activity (105). The UGT2B7 polymorphism causes a His268Tyr amino acid change that apparently does not alter the function of UGT2B7 (102). Whether the UGTlAl or 2B4 polymorphisms result in clinically significant variation in the in vivo conjugation of CBS is not yet known. Glutathione S-transferases. Members of a superfamily of cytosolic GSTs catalyze the conjugation of GSH, the reactive cosubstrate, to a variety of electrophiles (106). Although GSH conjugation can occur independent of GST-mediated catalysis, GSTs likely play a role in the catalysis of GSH conjugation of CE—Qs. GSTs are a major class of detoxification enzymes. There are estimated to be at least 20 human GST isoforms (106). Their activity has been associated with the inactivation of a large number of xenobiotics, including many drugs. The ability of many tumors to exhibit increased levels of intracellular GST Journal of the National Cancer Institute Monographs No. 27, 2000 expression has been implicated as a mechanism of chemothera- peutic drug resistance (107). GST enzymes are encoded by a superfamily of GST genes (106). The nomenclature adopted for this superfamily is quite different from that for the cytochromes P450, SULTs, or UGTS. The five families of GSTs have been designated GST alpha (or), mu (pt), pi (1T), sigma (0'), and theta (0). Humans possess a single functional GSTTr, but each of the other families contains multiple family members. GST enzymes are active as either homodimers or heterodimers. The frequent occurrence of functional GST heterodimers has made the study of substrate specificity for particular GST isoforms difficult. Perhaps it is for this reason that there is a lack of reports re- garding the specific GST isoforms that contribute to the forma- tion of CE—Q—GSH conjugates. There is much known about the molecular genetics of human GSTS (106). Many GSTor isoforms are expressed in human liver and skin, while some are ubiquitously expressed. Some mem- bers of the GSTp. family are expressed in human liver, while others are expressed in muscle, testis, brain, and heart. GSTTr is ubiquitously expressed, and GSTG has been reported in human liver and red blood cells. There have been a number of reports indicating the high inducibility of GSTS by a variety of agents. Polycyclic aromatic hydrocarbons, phenolic antioxidants, reac- tive oxygen species, barbiturates, and synthetic glucocorticoids have been shown to induce GSTs. Induction of GSTTr has been of particular interest because of its putative role in drug resis— tance (107). The mechanisms by which GSTS are inducible are apparently diverse (106). The regulation of GST expression ap- pears to be quite complex. A number of genetic response ele— ments have been characterized in GST genes, including xeno- biotic, antioxidant, and glucocorticoid—responsive elements. Furthermore, GST subunit expression is quite tissue specific, and regulatory elements contributing to tissue specificity are beginning to be defined. An NF-KB-like repressor element has recently been described in the human GSTrr gene. Expression of GSTS also appears to undergo sex— and age-specific regula- tions. A number of genetic polymorphisms have been described for human GSTS, including variations in the GSTu, GSTn, and GSTB genes (106,108). The most notorious GST polymorphism is the null gene for GSTp. (106). This polymorphism is defined by a deletion of the GSTM1 ((1.) gene. The frequency of homo- zygosity of this deletion varies with ethnicity, from approxi- mately 22% in Nigerians to 58% in Chinese populations. Epi- demiologic studies have suggested that individuals who are null for the GSTM1 gene may be at increased risk for a variety of neoplastic diseases. The epidemiology of this polymorphism in breast cancer is discussed in Chapter 7 of this monograph. A common single nucleotide polymorphism in the human GSTTI' gene resulting in an IlelOSVal amino acid substitution has been identified, and the VallO4 variant is associated with low GSTn activity (108). In addition, the VallO4 allele has been associated with increased risk for prostate cancer. Epidemiologic studies of the role of this polymorphism in breast cancer are discussed in Chapter 7. An additional null allele for a GSTB gene, GSTTl, has also been reported (106). The frequency of the homozygous GSTTl null genotype has been reported to vary from 16% in British Caucasians to 38% in Nigerians. The biologic conse- quences of the GSTTI null genotype are not yet clear, but stud- ies of this polymorphism and breast cancer susceptibility are also discussed in Chapter 7. Journal of the National Cancer Institute Monographs No. 27, 2000 CONCLUDING REMARKS AND FUTURE DIRECTIONS Conjugation is clearly a major biotransformation pathway for estrogens in humans. Recognition of the contribution of estrogen conjugation and deconjugation in breast cancer has been a rela- tively recent event. Increasing evidence suggests that the role of estrogen conjugation, particularly sulfation, goes beyond that of an excretory function and is perhaps even a major regulator of biologically active estrogens. Much less is known about conju- gation of CES, but the role that these conjugative pathways play in the biotransformation of CBS is an emerging story. Methyl- ation of CEs appears to be an important detoxification mecha- nism, and some evidence suggests that variation in the capacity of cells to methylate CES may represent a risk factor for sus- ceptibility to breast cancer. Clearly, more investigative effort will be required to fully understand which, if any, of these conjugative pathways modify cancer susceptibility or progression. As described in the next chapter, the study of low-penetrance, risk—modifying genes is very active, and we are beginning to see the inclusion of genes, such as COMT, that contribute to estrogen and CE conjugation among those being studied. As more genetic polymorphisms in estrogen- and CE-conjugating enzymes are identified, even larger epidemiologic studies will be necessary to delineate which of these variations or—more likely—which set of these genetic variants, represent cancer risk factors. The identification of novel genes encoding conjugating enzymes and “functionally significant” polymorphisms within those genes is occurring at a rapid pace. New molecular information arising from this era of “functional genomics” will require careful biochemical and large-scale epidemiologic studies before we can understand the biologic interplay and ultimate cellular consequence of the ap- parent myriad biotransforrnations that estrogens undergo and how these reactions contribute to carcinogenesis. REFERENCES (1) Safe SH. Interactions between hormones and chemicals in breast cancer. Annu Rev Pharmacol Toxicol 1998;38:121—58. (2) Zhu BT, Conney AH. Functional role of estrogen metabolism in target cells: review and perspectives. Carcinogenesis 1998;19:1—27. (3) Yager JD, Liehr JG. Molecular mechanisms of estrogen carcinogenesis. Annu Rev Pharrnacol Toxicol 1996;36:203—32. (4) Service RF. New role for estrogen in cancer? Science 1998;279:1631—3. (5) MacGregor JI, Jordan VC. Basic guide to mechanisms of antiestrogen action. Pharmacol Rev 1998;50:151—96. (6) Paech K, Webb P, Kuiper GG, Nilsson S, Gustafsson J, Kushner PJ, et al. Differential ligand activation of estrogen receptors ERor and ERB at APl sites. Science 1997;277:1508—10. (7) Barkhem T, Carlsson B, Nilsson Y, Enmark E, Gustafsson J, Nilsson S. Differential response of estrogen receptor or and estrogen receptor [3 to partial estrogen agonists/antagonists. Mol Pharmacol l998;54;105—12. (8) Pasqualini JR, Gelly C, Nguyen BL, Vella C. Importance of estrogen sulfates in breast cancer. J Steroid Biochem 1989;34:155—63. (9) Hobkirk R. Steroid sulfotransferases and steroid sulfate sulfatases: char— acteristics and biological roles. Can J Biochem Cell Biol 1985;63: 1127—44. (10) Hobkirk R. Steroid sulfation: current concepts. Trends Endocrinol Metab 1993;4269—74. (11) Coughtrie MW, Sharp S, Maxwell K, Innes NP. Biology and function of the reversible sulfation pathway catalysed by human sulfotransferases and sulfatases. Chem Biol Interact 1998;109:3—27. (12) Miller JA. Research in chemical carcinogenesis with Elizabeth Miller—a trail of discovery with our associates. Drug Metab Rev 1994;26:1—36. (13) Miller JA. Brief history of chemical carcinogenesis. Cancer Lett 1994; 8329—14. 12] (I4) Liehr JG, Ulubelen AA, Stroben HW. Cytochrome P-450-mediated redox cycling of estrogens. J Biol Chem 1986;261:16865—70. (15) Roy D, Strobel HW, Liehr JG. Cytochrome bS-mediated redox cycling of estrogen. Arch Biochem Biophys 199l;285:331~8. (16) Cavalieri EL, Stack DE, Devanesan PD, Todorovic R, Dwivedy I, Hig— ginbothan S, et al. Molecular origin of cancer: catechol estrogen—3,4— quinones as endogenous tumor initiators. Proc Natl Acad Sci U S A 1997; 94:10937—42. (17) Guengerich FP. Metabolism of 170t-ethynylestradiol in humans. Life Sci 1990;47:1981—8. ([8) Williams CL, Stancel GM. Estrogens and progestins, In: Hardman JG, Limbird LE, editors. Goodman & Gilman’s the pharmacological basis of therapeutics. 9'h ed. New York (NY): McGraw—Hill; 1996. p. 1411—40. (19) Eckhardt U, Schroeder A, Stieger B, Hochli M, Landmann L, Tynes R, et al. Polyspecific substrate uptake by the hepatic organic anion transporter Oatpl in stably transfected CHO cells. Am] Physiol 1999;276:G1037—42. (20) Kanai N, Lu R, Bao Y, Wolkoff AW, Vore M, Schuster VL. Estradiol 17B-D-glucuronide is a high-affinity substrate for oatp organic anion transporter. Am J Physiol 1996;270(2 Pt 2):F326—31. (21) Kauffman FC, Sharp S, Allan BB, Burchell A, Coughtrie MW. Micro- somal steroid sulfatase: interactions with cytosolic steroid sulfotransfer— ases. Chem Biol Interact 1998;109:169—82. (22) Li PK, Chu GH, Guo JP, Peters A, Selcer KW. Development of potent non-estrogenic estrone sulfatase inhibitors. Steroids 1998;63:425—32. (23) W00 LW, Howarth NM, Purohit A, Hejaz HA, Reed MJ, Potter BV. Steroidal and nonsteroidal sulfamates as potent inhibitors of steroid sul- fatase. J Med Chem 1998;41:1068—83. (24) Albin N, Massaad L, Toussaint C, Mathieu MC, Morizet J, Parise O, et a]. Main drug—metabolizing enzyme systems in human breast tumors and peritumoral tissues. Cancer Res 1993;53:3541—6. (25) Albert C, Vallee M, Beaudry G, Belanger A, Hum DW. The monkey and human uridine diphosphate—glucur0nosyltransferase UGTlA9, expressed in steroid target tissues, are estrogen-conjugating enzymes. Endocrinology 1999;140:3292—3102. (26) Ebner T, Remmel RP, Burchell B. Human bilirubin UDP—glucuronosyl- transferase catalyzes the glucuronidation of ethinylestradiol. Mol Phar— macol 1993;43:649—54. (27) Weinshilboum RM, Otterness DM, Aksoy IA, Wood TC, Her C, Rafto- gianis RB. Sulfation and sulfotransferases 1: sulfotransferase molecular biology: cDNAs and genes. FASEB J 1997 ,1 123-14. (28) Falany CN. Enzymology of human cytosolic sulfotransferases. FASEB J 1997;11:206—16. (29) Hernandez JS, Watson RW, Wood TC, Weinshilboum RM. Sulfation of estrone and l7B-estradiol in human liver: Catalysis by thermostable phe- nol sulfotransferase and by dehydroepiandrosterone sulfotransferase. Drug Metab Dispos 1992;20:413—22. (30) Her C, Szumlanski C, Aksoy IA, Weinshilboum RM. Human jejunal estrogen sulfotransferase and dehydroepiandrosterone sulfotransferase: immunochemical characterization of individual variation. Drug Metab Dispos 1996;24:1328—35. (31) Adams JB, Phillips NS. Properties of estrogen and hydroxysteroid sul- photransferases in human mammary cancer. J Steroid Biochem 1990,36: 695—701. (32) Weinshilboum R, Otterness D. Sulfotransferase enzymes. In: Kauffman FC, editor. Conjugation—deconjugation reactions in drug metabolism and toxicity. New York (NY): Springer-Verlag; 1994. p. 45—78. (33) Adams LB, Phillips NS, Pewnim T. Expression of hydroxysteroid sul— photransferase is related to estrogen receptor status in human mammary cancer. J Steroid Biochem 1989;33:637—42. (34) Adams JB, Pewnim T, Chandra DP, Archibald L, Foo MS. A correlation between estrogen sulfotransferase levels and estrogen receptor status in human primary breast carcinoma. Cancer Res 1979;39:5124—6. (35) Raju U, Blaustein A, Levitz M. Conjugation of androgens and estrogens by human breast tumors in vitro. Steroids 1980;35:685—95. (36) Price RA, Spielman RS, Lucena AL, Van Loon JA, Maidak BL, Wein— shilboum RM. Genetic polymorphism for human platelet thermostable phenol sulfotransferase (TS PST) activity. Genetics 1989;122:905—14. (37) Meech R, Mackenzie PI. Structure and function of uridine diphos- phate glucuronosyltransferases. Clin Exp Pharmacol Physio] 1997;24: 907-15. 122 (38) Clarke DJ, Burchell B. The uridine diphosphate glucuronosyltransferase multigene family: function and regulation. In: Kauffman FC, editor. Con— jugation—deconjugation reactions in drug metabolism and toxicity. New York (NY): Springer-Verlag; 1994. p. 3—43. (39) Huang L, Hoffman T, Vore M. Adenosine triphosphate-dependent trans- port of estradiol—17B (B—D-glucuronide) in membrane vesicles by MDRl expressed in insect cells. Hepatology 1998;28:1371—7. (40) Loe DW, Almquist KC, Cole SP, Deeley RG. ATP—dependent 17B— estradiol 17-(B—D-glucuronide) transport by multidrug resistance pro- tein (MRP). Inhibition by cholestatic steroids. J Biol Chem 1996;271: 9683—9. (41) Belanger A, Brochu M, Lacoste D, Noel C, Labrie F, Dupont A, et al. Steroid glucuronides: human circulatory levels and formation by LNCaP cells. J Steroid Biochem 1991;40:593—8. (42) Her C, Wood TC, Eichler EE, Mohrenweiser HW, Ramagli LS, Siciliano MJ, et al. Human hydroxysteroid sulfotransferase SULT2B1: two en— zymes encoded by a single chromosome 19 gene. Genomics 1998;53: 284—95. (43) Fujita K, Nagata K, Ozawa S, Sasano H, Yamazoe Y. Molecular cloning and characterization of rat ST1B1 and human ST1B2 cDNAs, encoding thyroid hormone sulfotransferases. J Biochem (Tokyo) 1997;122: 1052—61. (44) Her C, Kaur GP, Athwal RS, Weinshilboum RM. Human sulfotransferase SULTlCl: cDNA cloning, tissue-specific expression, and chromosomal localization. Genomics 1997;41:467-—70. (45) Sakakibara Y, Yanagisawa K, Katafuchi J, Ringer DP, Takami Y, Na— kayama T, et al. Molecular cloning, expression, and characterization of novel human SULTlC sulfotransferases that catalyze the sulfonation of N—hydroxy—2—acetylaminofluorene. J Biol Chem 1998;273:33929—35. (46) Falany JL, Falany CN. Expression of cytosolic sulfotransferases in normal mammary epithelial cells and breast cancer cell lines. Cancer Res 1996; 56:1551—5. (47) Falany JL, Lawing L, Falany CN. Identification and characterization of cytosolic sulfotransferase activities in MCF-7 human breast carcinoma cells. J Steroid Biochem Mol Biol 1993;46:481—7. (48) Sharp S, Anderson JM, Coughtrie MW. Immunohistochemical localisa— tion of hydroxysteroid sulphotransferase in human breast carcinoma tis- sue: a preliminary study. Eur J Cancer 1994;30A21654—9. (49) Her C, Aksoy IA, Kimura S, Brandriff BF, Wasmuth JJ, Weinshilboum RM. Human estrogen sulfotransferase gene (STE): cloning, structure, and chromosomal localization. Genomics 1995;29:16—23. (50) Raftogianis RB, Her C, Weinshilboum RM. Human phenol sulfotransfer— ase pharmacogenetics: STPl gene cloning and structural characterization. Pharmacogenetics 1996;6z473—87. (51) Helton ED, Simmons R, Meltz ML, Goldzieher JW. Variability in con— jugates of ethinylestradiol produced by in vitro incubation of human liver tissue. Contraception 1977;16:257—60. (52) Raftogianis RB, Wood TC, Otterness DM, Van Loon JA, Weinshilboum RM. Phenol sulfotransferase pharmacogenetics in humans: association of common SULTlAl alleles with TS PST phenotype. Biochem Biophys Res Commun 1997;239:298—304. (53) Raftogianis RB, Wood TC, Weinshilboum RM. Human phenol sulfotrans- ferases SULT1A2 and SULTlAl: genetic polymorphisms, allozyme properties, and human liver genotype—phenotype correlations. Biochem Pharmacol 1999;58:605—16. (54) Aksoy IA, Sochorova V, Weinshilboum RM. Human liver dehydroepi— androsterone sulfotransferase: nature and extent of individual variation. Clin Pharmacol Ther 1993;54:498—506. (55) Wood TC, Her C, Aksoy I, Otterness DM, Weinshilboum RM. Human dehydroepiandrosterone sulfotransferase pharmacogenetics: quantitative Western analysis and gene sequence polymorphisms. J Steroid Biochem Mol Biol 1996;59:467—78. (56) King CD, Green MD, Rios GR, Coffman BL, Owens IS, Bishop WP, et al. The glucuronidation of exogenous and endogenous compounds by stably expressed rat and human UDP—glucuronosyltransferase 1.1. Arch Biochem Biophys 1996;332:92—100. (57) Mojarrabi B, Butler R, Mackenzie PI. cDNA cloning and characterization of the human UDP glucuronosyltransferase, UGT1A3. Biochem Biophys Res Commun 1996;225:785—90. (58) Green MD, Tephly TR. Glucuronidation of amines and hydroxylated Journal of the National Cancer Institute Monographs N0. 27, 2000 xenobiotics and endobiotics by expressed human UGT1.4 protein. Drug Metab Dispos 1996;24:356—63. (59) Cheng Z, Radaminska-Pandya A, Tephly TR. Cloning and expression of human UDP—glucuronosyltransferase (UGT) 1A8. Arch Biochem Biophys 1998;356:301—5. (60) Strassburg CP, Manns MP, Tukey RH. Expression of the UDP- glucuronosyltransferase 1A locus in human colon: identification and char— acterization of the novel extrahepatic UGT1A8. J Biol Chem 1998;273: 8719—26. (6]) Bosma PJ, Chowdhury JR, Bakker C, Gantla S, de Boer A, Oostra BA, et al. The genetic basis of the reduced expression of bilirubin UDP- glucuronosyltransferase 1 in Gilbert’s syndrome. N Engl J Med 1995;333: 1171—5. (62) Aono S, Adachi Y, Uyama E, Yamada Y, Keino H, Nanno T, et a1. Analysis of genes for bilirubin UDP-glucuronosyltransferase in Gilbert’s syndrome. Lancet 1995;345:958—9. (63) Beutler E, Gelbart T, Demina A. Racial variability in the UDP— glucuronosyltransferase 1 (UGTlAl) promoter: a balanced polymorphism for regulation of bilirubin metabolism? Proc Natl Acad Sci U S A 1998; 95:8170—4. (64) Iyer L, Hall D, Das S, Mortell MA, Ramirez J, Kim S, et a1. Pheno- type—genotype correlation of in vitro SN-38 (active metabolite of irino— tecan) and bilirubin glucuronidation in human liver tissue with UGTlAl promoter polymorphism. Clin Pharmacol Ther 1999;65:576—82. (65) Stack D, Byun J, Gross ML, Rogan EG, Cavalieri EL. Molecular char— acteristics of catechol estrogen quinones in reactions with deoxyribo- nucleosides. Chem Res Toxicol 1996;92851—9. (66) Li KM, Devanesan PD, Rogan EG, Cavalieri EL. Formation of the depu— rinating 4-hydroxyestradiol (4-OHE2)-1-N7Gua and 4-OHE2—1—N3Ade adducts by reaction of E2-3,4—quinone with DNA [abstract]. Proc Am Assoc Cancer Res 1998;39:636. (67) Mobley J A, Bhat AS, Brueggemeier RW. Measurement of oxidative DNA damage by catechol estrogens and analogues in vitro. Chem Res Toxicol 1999;12:270—7. (68) Lennard L. Methyl transferases. In: Guengerich FP, editor. Comprehen— sive toxicology, Vol 3: biotransformation. London (U.K.): Pergamon; 1997. p. 437—54. (69) Wiesz J, Fritz-Wolz G, Clawson GA, Benedict CM, Abendroth C, Crev- eling CR. Induction of nuclear catechol-O-methyltransferase by estrogens in hamster kidney: implications for estrogen—induced renal cancer. Carci— nogenesis 1998;19:1307—12. (70) Inoue K, Tice LW, Creveling CR. Immunocytochemical localization of catecholvO-methyltransferase. In: Usdin E, Weiner N, Youdim MB, edi- tors. Biochemistry and function of monoamine enzymes. New York (NY): Marcel Dekker; 1977. p. 835—59. (71) Liehr J. Mechanism of metabolic activation and inactivation of catechol estrogens: the basis of genotoxicity. Polycyclic Aromat Comp 1994;62 232—40. (72) Weisz J, Fitz-W011 G, Gestl S, Clawson GA, Creveling CR, Liehr JG, et a]. Nuclear localization of catechol-O-methyltransferase in non-neoplastic as well as neoplastic mammary epithelial cells in human breast paren- chyma. Am J Pathol. In press 2000. (73) Weisz J. Metabolism of estrogens in target cells. Diversification and amplification of hormone action and the catechol estrogen hypothesis. In: Hochberg R, Naftolin F, editors. New biology of steroid hormones. New York (NY): Raven Press; 1991. p. 102—12. (74) Gestl SA, Dabbs DJ, Clawson GA, Liehr JG, Creveling CR, Weisz J. Nuclear localization of catechol—O—methyltransferase (COMT) in human breast parenchyma, in normal, hyperplastic and neoplastic epithelial cells: implications for catecholestrogens’ (CEs), role in carcinogenesis [ab— stract]. Proc Am Assoc Cancer Res 1998;39:636. (75) Amin AM, Creveling CR, Lowe MC. Immunohistochemical localization of catechol methyltransferase in normal and cancerous breast tissues of mice and rats. J Natl Cancer Inst 1983;70:337—42. (76) Malins DC, Polissar NL, Gunselman SJ. Progression of human breast cancers to the metastatic state is linked to hydroxyl radical—induced DNA damage. Proc Natl Acad Sci U S A 1996;93:2557—63. (77) Malins DC, Polissar NL, Nishikida K, Holmes EH, Gardner HS, Gunse1< man SJ. The etiology and prediction of breast cancer. Fourier transfonn- infrared spectroscopy reveals progressive alterations in breast DNA lead- Joumal of the National Cancer Institute Monographs No. 27, 2000 ing to a cancer—like phenotype in a high proportion of normal women. Cancer 1995;75:503—17. (78) Ball P, Haupt M, Knuppen R. Biogenesis and metabolism of catechol estrogens in vitro. In: Merriam GR, Lipsett MB, editors. Catechol estro- gens. New York (NY): Raven Press; 1983. p. 91—103. (79) Lipsett M, Merriam G, Kono S, Brandon D, Pheiffer D, Loriaux, D. Metabolic clearance of catechol estrogens, In: Merriam GR, Lipsett MB, editors. Catechol estrogens. New York (NY): Raven Press; 1983. p. 105—14. (80) Lavigne JA, Helzlsouer KJ, Huang HY, Strickland PT, Bell DA, Selmin O, et al. An association between the allele coding for a low activity variant of catechol-O-methyltransferase and the risk for breast cancer. Cancer Res 1997;57:5493—7. (81) Thompson PA, Shields PG, Freudenheim JL, Stone A, Vena JE, Marshall JR, et a1. Genetic polymorphisms in catechol—O-methyltransferase, meno- pausal status, and breast cancer risk. Cancer Res 1998;58:2107—10. (82) Brueggemeier RW, Kimball JG, Kraft F. Estrogen metabolism in rat liver microsomal and isolated hepatocyte preparations—I. Metabolite forma- tion and irreversible binding to cellular macromolecules. Biochem Phar- macol 1984;33:3853—9. (83) Monks TJ, Lau SS. Glutathione conjugate-mediated toxicities. In: Kauff- man FC, editor. Conjugation—deconjugation reactions in drug metabolism and toxicity. New York (NY): Springer—Verlag; 1994. p. 459—509. (84) Butterworth M, Lau SS, Monks TJ. Formation of catechol estrogen glu- tathione conjugates and y-glutamyl transpeptidase-dependent nephrotox- icity of 17B—estradiol in the golden Syrian hamster. Carcinogenesis 1997; 18:561—7. (85) Butterworth M, Lau SS, Monks TJ. 17B—Estradiol metabolism by hamster hepatic microsomes. Implication for the catcchol—O-methyl transferase— mediated detoxication of catechol estrogens. Drug Metab Dispos 1996; 24:588-94. (86 ) Tenhunen J, Salminen M, Lundstrom K, Kiviluoto T, Savolainen R, U1- manen I. Genomic organization of the human catechol O-methyltrans- ferase gene and its expression from two distinct promoters. Eur J Biochem 1994;223:1049—59. ( 87) Lundstrom K, Salminen M, J alanko A, Savolainen R, Ulmanen I. Cloning and characterization of human placental catechol-0-methyltransferase cDNA. DNA Cell Biol 1991;10:181—9. (88) Bertocci B, Miggiano V, Da Prada M, Dembic Z, Lahm HW, Malherbe P. Human catechol—O-methyltransferase: cloning and expression of the membrane-associated form. Proc Natl Acad Sci U S A 1991;88:1416—20. (89) Ulmanen I, Peranen J, Tenhunen J, Tilgmann C, Karhunen T, Panula P, et a1. Expression and intracellular localization of catechol-O-methyltrans- ferase in transfected mammalian cells. Eur J Biochem 1997;243:452—9. (90) Assicot M, Contesso G, Bohuon C. Catechol-0-methyltransferase in hu— man breast cancers. Eur J Cancer 1977;13:961—6. (9]) Thakker DR, Creveling CR. O-Methylation. In: Mulder GJ, editor. Con- jugation reactions in drug metabolism: an integrated approach. London (U.K.): Taylor and Francis Ltd.; 1990. p. 193—231. (92) Kopin IJ. Catecholamine metabolism; basic aspects and clinical signifi— cance. Pharmacol Rev 1985;37:333—64. (93) Creveling CR, Hartman BK. Relationship between the cellular localiza- tion and physiological function of catechol—O-methyltransferase. In: Us- din E, Borchardt RT, Creveling CR, editors. Biochemistry of S—adenosyl— L-methionine and related compounds. London (U.K.): Macmillan Press; 1983. p. 479—86. (94) Inoue K, Creveling CR. Induction of catechol—0—methyltransferase in the luminal epithelium of rat uterus by progesterone. J Histochem Cytochem 1991;39:823—8. (95) Inoue K, Creveling CR. Induction of catechol-0-methyltransferase in the luminal epithelium of rat uterus by progesterone: inhibition by RU-486. Drug Metab Dispos 1995;23:430—2. (96) Paria B, Das S, Dey S. Embryo implantation requires estrogen-directed uterine preparation and catechol estrogen-mediated embryonic activation. In: Goldel DS, editor. Catecholamines: bridging basic science with clini- cal medicine. New York (NY): Marcel Dekker; 1977. p. 835—59. (97) Weinshilboum RM, Raymond FA. Inheritance of low erythrocyte cat- echol—0—methyltransferase activity in man. Am J Hum Genet 1977;29: 125—35. (98) Lachman HM, Papolos DF, Saito T, Yu YM, Szumlanski CL, Weinshil- 123 boum RM. Human catechol-0-methyltransferase pharmacogenetics; de- scription of a functional polymorphism and its potential application to neuropsychiatric disorders. Pharmacogenetics 1996;6z243—50. (99) Millikan RC, Pittman GS, Tse CK, Duell E, Newman B, Savitz D, et a1. Catechol-0-methyltransferase and breast cancer risk. Carcinogenesis 1998;19:1943—7. (100) Falany CN, Wheeler J, Oh TS, Falany JL. Steroid sulfation by expressed human cytosolic sulfotransferases. J Steroid Biochem Mol Biol 1994;48: 369—75. (10]) Price RA, Cox NJ, Spielman RS, Van Loon IA, Maidak BL, Weinshil- boum RM. Inheritance of human platelet thermolabile phenol sulfotrans- ferase (TL PST) activity. Genet Epidemiol 1988;5:1—15. (102) Cheng Z, Rios GR, King CD, Coffman BL, Green MD, Mojarrabi B, et al. Glucuronidation of catechol estrogens by expressed human UDP— glucuronosyltransferases (UGTs) 1A1, 1A3, and 2B7. Toxicol Sci 1998; 45252—7. (103) Ritter JK, Sheen YY, Owens IS. Cloning and expression of human liver UDP—glucuron0syltransferase in COS-1 cells. 3,4—catechol estrogens and estriol as primary substrates. J Biol Chem 1990;265:7900—6. (104) Jin CJ, Miners JO, Lillywhite KI, Mackenzie PI. cDNA cloning and expression of two new members of the human liver UDP- glucuronosyltransferase 2B subfamily. Biochem Biophys Res Commun 1993;194:496—503. ( 105) Levesque E, Beaulieu M, Hum DW, Belanger A. Characterization and 124 substrate specificity of UGTZB4 (E458): a UDP-glucuronosyltransferase encoded by a polymorphic gene. Pharrnacogenetics 1999;92207—16. (106) Hayes JD, Pulford DJ. The glutathione S-transferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer che- moprotection and drug resistance. Crit Rev Biochem Mol Biol 1995;30: 445—600. (107) Farber E. Clonal adaptation during carcinogenesis. Biochem Pharrnacol 1990;39:1837—46. (108) Harries LW, Stubbins MJ, Forman D, Howard GC, Wolf CR. Identifica— tion of genetic polymorphisms at the glutathione S—transferase Pi locus and association with susceptibility to bladder, testicular and prostate can— cer. Carcinogenesis 1997;18:641—4. NOTES Supported by a Mary L. Smith Charitable Lead Tmst Award and a Louise and Gustavus Pfeiffer Research Foundation Award (to R. Raftogianis); and by Public Health Service grants CA65532 (National Cancer Institute) (to J. Weisz) and grants GM28157 and GM3572O (National Institute of General Medical Sciences) (to R. Weinshilboum), National Institutes of Health, Department of Health and Human Services. We thank Ms. Kathleen Buchheit for her assistance in the preparation of this manuscript. Journal of the National Cancer Institute Monographs No. 27, 2000 Chapter 7: Molecular Epidemiology of Genetic Polymorphisms in Estrogen Metabolizing Enzymes in Human Breast Cancer Patricia A. Thompson, Christine Ambrosone Epidemiologic studies indicate that most risk factors for breast cancer are related to reproductive and hormonal fac- tors. For a number of years, the mechanism for estrogens in carcinogenesis was thought to be that of mitotic stimulation, with the growth promotion of ductal epithelial cells harbor- ing precursor mutations in the breast. However, evidence is now available that estrogens may act as initiators of cellular alterations and tumorigenesis. Investigation and measure- ment of serum levels of estrogens in epidemiologic studies may, therefore, be misleading, because they may reflect lev- els quite different from those of hormone metabolites to which the target tissue is exposed. Proportions of hormone metabolites may be estimated by evaluation of associations between breast cancer risk and genetic polymorphisms in enzymes involved in hormone metabolism. A number of mo- lecular epidemiologic studies have been conducted to evalu- ate associations between polymorphic genes involved in ste- roid hormone metabolism (i.e., CYP17, COMT, CYP1A1, CYP19, GST, and MnSOD) that may account for a propor- tion of enzymatic variability, and results are discussed in this review. There are strengths and limitations to such an ap- proach, foremost of which may be the lack of insight into the extent to which individual variability in estrogen exposure may be explained by allelic variation. Variability in other endogenous and exogenous factors that impact parent hor- mones and their metabolites along activation and conjuga- tion pathways may also affect associations in case—control comparisons. This and other possible reasons for inconsis- tencies in results of molecular epidemiologic studies are dis- cussed. Contributions from population-based studies and those from the laboratory may together move this field ahead and more clearly elucidate the basis of hormonally related cancers, identifying etiologic factors and susceptible populations for preventive strategies. [J Natl Cancer Inst Monogr 2000;27:125—34] Breast cancer is the most commonly occurring cancer among women in the United States, representing 29% of all newly diagnosed cancers in women and is second only to lung cancer as cause of cancer death in women (1). Currently estimated to affect 175 000 women in 1999 (I), a number of putative risk factors for breast cancer have been identified. Of those factors examined in epidemiologic studies, aside from a family history of breast cancer, the majority of risk factors are related to repro- ductive history and are widely thought to reflect longer lifetime exposures to the endogenous steroid hormones (2). Currently, the hypotheses proposed to explain the role of reproductive risk factors in breast cancer etiology are controversial and are based on the dual effects of estrogens as both promoters of ductal epithelial cell growth (Chapter 8) and as precursors for muta- genic estrogen metabolites [(3—5); Chapters 4 and 5]. Journal of the National Cancer Institute Monographs No. 27, 2000 Because epidemiologic and animal studies indicate that breast cancer risk may be related to endogenous exposures to steroid hormones, intensive epidemiologic research has been targeted at serum and urinary measurement of parent hormones and their metabolites in both case—control and cohort studies, yielding inconsistent results (3,6). Multiple factors complicate measure— ment of urinary and serum hormones, including intraindividual variation of hormones as a result of menstrual and diurnal tim- ing, disease status (case—control studies), laboratory variability, and hormone degradation in transport and storage. Furthermore, Zhu and Conney (7) suggest that the measurement of serum estrogens may reflect levels associated with hepatic-specific me- tabolism that may differ from the local milieu of estrogen me- tabolites. This supposition about tissue specific differences, as well as the development of approaches to assess tissue-specific exposures in relation to risk require further investigation and are beyond the scope of this discussion (for a discussion of tissue specific metabolism, see Chapter 5). According to the paradigms that have been developed for studies of bladder and lung cancers with respect to individual susceptibility to chemical carcinogens, metabolic variability that affects xenobiotic metabolism has been suggested by us, and others, to increase the risk of breast cancer among defined subsets of women [reviewed recently in (8,9)]. As observed in drug and chemical metabolism, there is considerable interin- dividual genetic variability in the metabolic and biosynthe- tic pathways in steroidogenesis. These person-to-person differ- ences, which are, in part, attributed to allelic variability or gene polymorphisms, might define subpopulations of women with higher lifetime exposures to hormone-dependent growth promo- tion or to cellular damage from particular estrogens and estrogen metabolites. Such variation could explain a portion of the can- cer susceptibility associated with reproductive events and hor- mone exposure. Currently, the evaluation of associations be- tween breast cancer risk and genetic polymorphisms in enzymes involved in hormone metabolism may be the most effective manner in which to evaluate metabolic variability, until techni— cal and epidemiologic methodologies have been developed to accurately quantify specific estrogens and their metabolites. In this chapter, we will discuss the use of genetic markers in population studies to determine the impact of metabolic di— versity in estrogen hormone biosynthesis and metabolism on Afiiliations of authors: P. A. Thompson, Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston; C. Ambrosone, Division of Molecular Epidemiology, National Center for Toxicological Re— search, Jefferson, AK. Correspondence to: Patricia Thompson, Ph.D., Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030. © Oxford University Press 125 breast cancer risk. More important, we will discuss the strengths and limitations of this approach in considering the multigenic nature and complexity of the problem. The powerful introduc- tion of molecular epidemiology as an approach to directly test the role of common allelic variants as risk factors for disease may allow us to determine which steroid hormone pathways are critical determinants of excessive hormonal exposures, and per- haps cancer risk, and ultimately identify targets for preventive strategies. ESTROGEN BIOSYNTHESIS Six isoforms of cytochrome—dependent monooxygenase (CYP) are involved in the biosynthesis of various steroid hor- mones, starting from cholesterol—CYPllA, CYP17, CYP19, CYPl 1B1, CYP21B, and CYP11B2 (10). As presented in Fig. 1, several enzymes are important in the biosynthesis of estrogens starting from cholesterol. The rate-limiting step in all steroid hormone biosynthesis is the cleavage of the side chain of cho- lesterol by CYPl 1A to form the C21 steroids, pregnenolone and progesterone. Hydroxylation and subsequent cleavage of the two-carbon side chain of the C21 steroids by the CYP17 (l7-a hydroxylase activity/C17—20 lyase activity) yields the C19 ste- roids, androstenedione and dehydroepiandrosterone. Andro- stenedione is the immediate precursor for the formation of tes— tosterone (11). Estrogens are ultimately formed by aromatization of androstenedione and testosterone, catalyzed by the CYP19 (aromatase) (12). In addition to the cytochrome P450, a series of hydroxysteroid dehydrogenases (3B~HSD and 17B-HSD) par- ticipate in the biosynthesis of the steroid hormones (7, 13,14). As depicted in Fig. l, the 3B-HSD converts pregnenolone to pro- gesterone, whereas the l7B-HSD converts androstenedione to testosterone. There are several forms of l7B-HSD isoforms, and current data indicate that each form has a distinct role in the metabolism of steroid hormones. l7B-HSD type 1 and CYP19 catalyze the end steps in l7B-estradiol (E2) biosynthesis through androstenedione, supporting a role for this particular isoform in the biosynthesis of E2. ESTROGEN METABOLISM Once formed, estrogens are extensively metabolized by a number of oxidative and conjugative reactions that can lead to their deactivation and subsequent elimination [(7,15,16); Chap- ters 5 and 6]. Alternatively, oxidation and conjugation reac- tions of estrogens may generate metabolites that have distinct biologic activities, including altered hormonal properties; geno- toxicity, through the formation of reactive species that modify cellular DNA and protein; and/or chemotherapeutic properties, by forming derivatives that are antagonistic at the estrogen receptor (ER) or potentially antiangiogenic (7). Oxidative me— tabolism of estrogens, largely by hydroxylation, is mediated by the same CYPs that metabolize therapeutic agents and xenobi- otics (15). Nicotinamide adenine dinucleotide phosphate (NADPH)-dependent oxidative metabolism of estrogens has been demonstrated for some, but not all, cytochrome P450. These CYPs include the major hepatic CYP 1A2 and 3A4 and the extrahepatic CYP 1A1 and 1B1 as well as members of the 3A family, which are expressed in a number of steroid hormone- responsive tissues, including brain, breast, ovary, kidney, and prostate (7,15). The biologic significance of tissue—specific ex- pression of enzymes involved in extrahepatic estrogen metabo- lism remains largely unknown. However, it has been postulated by a number of investigators that local estrogen metabolism may generate important biologically active estrogen species or, in the Cholesterol P450”C Pregnenolone Hydroxylase Lyase Activity Activity P450", I Progesterone ‘ ’ ‘ ‘ Dehydroepiandrosterone (DHEA) Androstenedione P450mm Estrone -l Testosterone l- “ Estradiol P450mm Fig. 1. Biosynthesis of estrogens from cholesterol. Participation of six key forms of cytochrome P450 and two hydroxysteroid dehydrogenases in the generation of estrogens are shown. The boxes containing enzymes with known polymorphisms are bolded. 126 Journal of the National Cancer Institute Monographs No. 27, 2000 case of carcinogenesis, may predispose certain tissues to expo- sures from highly reactive and genotoxic metabolites of E2 [( 7); Chapter 5]. Aromatic hydroxylation of E2 occurs primarily at C-2 and to a lesser extent at C-4, to form the 2,3- and 3,4—catechol estrogens (CEs), respectively. These metabolites are intermediates for the generation of the reactive semiquinones and quinones. Both semiquinones and quinones have the ability to damage DNA and protein directly or indirectly through redox cycling and genera— tion of reactive oxygen species (ROS), and they have been im- plicated as potential initiators of tumor formation (4, 7,16,17). P450-mediated hydroxylation of estrone (E) at C—16 forms the “suspect” carcinogen l6a-hydroxy E1, an estrogen metabo— lite thought to bind covalently to DNA and ERs (I8). Consid- erable controversy remains regarding which pathways and me- tabolites, if any, are most important in estrogen-induced carcinogenesis. The secondary metabolism of E2 involves 0-methylation by catechol-0-methyltransferase (COMT), conjugation to glucuronides and sulfates, and clearance of reactive semiqui— nones and quinones, reported to involve catechol oxidation coupled to glutathione conjugation [( 7); Chapter 6]. The dispo— sition of the hydroxylated estrogens remains unclear and some- what controversial. Conjugation of E2 and E1 to glucuronides and sulfates by specific enzymes in liver and target cells de- creases their bioavailability by facilitating their excretion. How- ever, there is increasing interest in these compounds as slow- clearing precursor compounds for biologically active estrogens reactivated, following cleavage of conjugated groups in target tissues (i.e., by sulfatases). Enzymatic 0-methylation of CE by COMT leads to more lipophilic estrogen metabolites with longer half-lives than the other conjugated forms. As suggested by Zhu and Conney (7), studies indicate that these metabolites may have biologic activities atypical of those associated with classic ER occupation, some of which may be potential inhibitors of estrogen-dependent carcinogenesis. Because of the complexity of the biology of hormonal carcinogenesis and the experimental design issues for human population studies, we refer the reader to Chapters 3—5 and 8 for a comprehensive review of the current competing hypotheses in estrogen-induced carcinogenesis and for the specific role of estrogen metabolites in breast carcino- genesis. GENETIC POLYMORPHISMS IN ESTROGEN BIOTRANSEORMATION AND BREAST CANCER RISK Many of the enzymes involved in estrogen metabolism/ biosynthesis are polymorphically distributed within the human population (i.e., CYP17, CYP19, CYP3A4, CYP1A2, and COMT) (7,16,19,20), and, for some, there are known gene vari— ants (i.e., CYP17, COMT, CYP1A1, CYP19, and glutathione S -transferase [GST]) that may account for a proportion of enzy- matic variability (Table 1). For others with currently no defined genetic basis (e.g., CYP1A2 and CYP3A4) and, to some extent, for those with a defined variant allele, some portion of individual variability is determined by such factors as age, sex, smoking status, alcohol consumption, and dietary effects on the regula- tory factors that control gene transcription and protein expres- sion. Several research groups have begun measuring, in popu- lation—based studies, the distribution of some of the known allelic variants that have been predicted to alter estrogen me- tabolism and biosynthesis to determine the impact of these vari- ants on breast cancer risk. Unlike measuring estrogen metabo- lites, an approach such as measuring genotype presents a more stable measure or biomarker of hormone exposure (i.e., cumu— lative high versus low), but it should be approached with caution because only a proportion, the extent of which is currently un- known, of individual variability in estrogen exposure may be explained by allelic variation. Investigating the distribution of functionally relevant genetic polymorphisms that alter the bio— availability of steroid hormones among people with disease and people without disease may provide more direct evidence for estrogen and estrogen metabolites as modifiers of human dis- eases, including breast cancer. The molecular epidemiologic studies and their results performed, to date, on genetic polymor- phisms that alter estrogen metabolism/biosynthesis and breast cancer risk are reviewed below. Table 1. Enzymes in the biosynthesis and metabolism of estrogen with known genetic polymorphisms analyzed to date in epidemiologic studies of breast cancer Role in estrogen biosynthesis and metabolism Allelic variants/effect of genotype (reference No.) CYP1A1 2-hydroxylase, extrahepatic (generates 2—OH CE) CYP17 17a hydroxylase/Cl7—20 lyase; catalyzes the rate-limiting step in ovaIian and adrenal biosynthesis pathways for androstenedione, the immediate precursor of testosterone CYP19 Aromatase/estrogen synthetase; converts testosterone and androstenedione to E2 and E1, respectively GST Glutathione sulfotransferases; conjugate glutathione with reactive oxygen species decreasing oxidative stress generated during estrogen metabolism COMT Methyltransferase; methylates and inactivates CE MnSOD Manganese superoxide dismutase, mitochondrial superoxide dismutase that converts two superoxide radicals to H202 and 02 M1, a Thr to Cys substitution 3’ noncoding region; m2, an lle to Val substitution in exon 7 (23); m3, an MspI RFLP that is specific to African-Americans (24); and m4, a Thr to Asp substitution in codon 461, which is adjacent to m2 (25) Base pair change in the 5’-untra.nslated region creates an Spl—type (CCACC box) promoter site 34 bp upstream from the initiation but downstream from the transcription start site, predicted to introduce an additional Spl binding site and enhanced promoter activity (32) Polymorphic (I I IA)n repeat in intron 5 close localization to the exon/intron border of exon 4, may alter splice site (43) GSTMl, deletion of the entire gene; ~50% of Caucasians inherit the null allele (60) GSTTI, deletion of the entire gene; ~30% of Caucasians inherit the null allele (65) GSTPl, Ile to Val change at amino acid position 105, reduced activity (30) Val t0 Met at amino acid position 158/108, alters heat stability, decreased methylation activity (47) Val to Ala change in the 9 amino acid position in signal peptide sequence, altered protein trafficking (59) Ala = alanine; Asp = aspartic acid; Cys = cysteine; CE = catechol estrogen(s); COMT = catechol-0-methyltransferase; CYP = cytochrome P450; E1 = estrone; E2 = l70-estradiol; GST = glutathione-S—transferase; Ile = isoleucine; Met = methionine; RFLP = restriction fragment—length polymorphism; Thr = threonine; Val = valine. Journal of the National Cancer Institute Monographs No. 27, 2000 127 MOLECULAR EPIDEMIOLOGY OF HORMONAL CARCINOGENESIS CYP1A1 Early studies of genetic polymorphisms in CYP1A1 focused primarily on its role in lung cancer risk because it activates polycyclic aromatic hydrocarbons, potent tobacco smoke car- cinogens. However, CYP1A1 also catalyzes the hydroxylation of E2 at the C-2, C-6or, and C-15a positions in several extrahe- patic tissues, including epithelial cells (15,21). To date, four polymorphisms have been identified within this gene; ml, which is a threonine to cysteine substitution that results in a MspI restriction site in the 3’ noncoding region (22); m2, an amino acid substitution of isoleucine to valine in exon 7 of the gene (23); m3, an A-T to G-C transition mutation in the 3’ noncoding region 300 base pairs (bp) from the polyadenylation site introducing an MspI restriction fragment-length polymor- phism (RFLP) that is specific to African-Americans (24); and m4, an amino acid substitution of threonine to asparagine in codon 461 adjacent to m2 (25). Studies evaluating the role of CYP1A1 genotypes on breast cancer risk are summarized in Table 2. Rebbeck et al. (26) investigated the m1 polymorphism in a small sample of case patients (n = 96) and “convenience sample” control subjects (n = 146). In the control subjects, the polymorphic allele had a low frequency (3% homozygotes and 6% heterozygotes) that was similar to that observed among the case patients. In the Western New York Diet Study, Ambrosone et al. (27) evaluated the m2 polymorphism in a case—control study with 404 postmenopausal women. The exon 7 substitution was associated with a nonsignificant increase in risk that was most pronounced among women who were moderate smokers (<20 pack years). Taioli et a1. ( 28) hypothesized that CYP1A1 polymorphisms could affect breast cancer risk through their me- diating effect on estrogen metabolism. In a case—control study of African—American and European-American women (51 case pa- tients and 269 control subjects), they noted that, among African— American women, the m1 polymorphism significantly increased breast cancer risk (odds ratio [OR] = 9.7; confidence interval [CI], 2.0—47.9). Numbers in these stratified analyses, however, were quite small. With the use of data from the Nurses’ Health Study, Ishibe et al. (29) studied the effects of CYP1A1 poly- morphisms (ml and m2) and cigarette smoking on breast cancer risk in a European-American population. Although women with variant alleles for either polymorphism were not at increased risk overall, those women with variant alleles who began smok- ing before the age of 16 years were at significantly increased risk of breast cancer. Recently, Bailey et al. (30) evaluated all four known CYP1A1 polymorphisms in relation to breast cancer risk in a casewcontrol study with approximately 164 Caucasian and 59 African-American women with breast cancer and equal num- bers of age-matched control subjects. None of these polymor- phisms, including those specific to African-Americans, was as- sociated with increased risk; risk was not modified by smoking status “ever/never.” CYP17 Another cytochrome P450 enzyme that has received much attention of late is CYP17. As discussed above, CYP17 func- tions at key branch points in human steroidogenesis, catalyzing the ovarian and adrenal biosynthesis pathways for androstene- dione, the immediate precursor of testosterone (Fig. 1) (31). A single base-pair change in the upstream promoter site creates an additional MspI recognition site in the 5’-untranslated region and creates a Spl-type (CCACC box) promoter site 34—bp up- stream from the initiation, but downstream from the transcrip— tion start site. This variant is referred to as the A2 allele and is predicted to introduce an additional Spl binding site and en— hanced promoter activity (32). Feigelson et al. (32) hypothesized that the polymorphism (A2 allele) could result in an increased rate of transcription and, thus, an increase in estrogen levels and perhaps increased risk of breast cancer in carriers of the variant allele. In a multiethnic cohort (Latino, Asian, and African-American) in Los Angeles and Hawaii, the researchers genotyped DNA from 174 case pa- tients and 285 control subjects. Allele differences were not dif- ferent between groups, so all were analyzed together. Overall, risk was not significantly increased for women with the A2 allele, but when women were stratified by stage Of disease, it was observed that the A2 allele conferred more than a twofold Table 2. Associations between breast cancer and CYP1A1 variant alleles* Study (reference No.) Polymorphism Case patients Control subjects OR (95% CI) Modifying factor OR (95% CI) Rebbeck et al., 1994 (26) ml 96 146 Not calculated, no association Ambrosone et al., 1995 (27) m2 (heterozygote 176 228 1.61 (0.9—2.8) Smoking, <20 pack yr 5.2 (1.2—23.7) & homozygote) Taioli et al., 1995 (28) ml 51 269 African-American women 9.7 (2.0—47.9) only (20 case patients, 81 control subjects) Ishibe et al., 1998 (29) m1 466 466 1.5 (0.7—1.5) Smoking 5.7 (1.5—21.3) m2 0.9 (0.6—1.3) Smoking < age 16 yr 3.6(1.1—11.7) Bailey et al., 1998 (30) ml 164, Caucasian 162, Caucasian 1.4 (0.8—2.4) m2 1.4 (0.6-3.1) m3 — m4 0.8 (0.4—1.9) ml 59, African- 59, African- 0.5 (0.2—1.1) m2 American American 1.0 (098—1 .1) m3 0.8 (0.3—2.0) m4 1.0(1.0—1.1) *OR = odds ratio; CI = confidence interval. 128 Journal of the National Cancer Institute Monographs NO. 27, 2000 increase in risk among those with advanced disease. The re- searchers also noted that late age at menarche was protective only among women who were homozygous for the A1 allele. Because it is unknown whether or not this polymorphism is functional, Feigelson et al. (33) followed this report with a con— trolled study of serum hormone levels throughout the menstrual cycle in nulliparous healthy women. Among women with A2 alleles, levels of E2 and progesterone were consistently higher on days 11 and 22, respectively. Further studies, however, have failed to confirm these findings. In a large case—control study in the UK. (835 case patients and 591 control subjects), Dunning et al. (34) found no evidence of increased risk for women with variant alleles. CYP17 did not modify associations between age at menarche and risk and there was no effect observed among women with advanced disease. Similar null associations were noted in studies by Helzlsouer et a1. (35) and by Weston et al. (36), and our own analyses in the Western New York Diet Study (unpublished data) also support the null hypothesis of no association between CYP17 genotype and breast cancer risk. Most recently, Haiman et al. (37) as- sessed the association between the A2 variant of CYP17 and breast cancer risk in a prospective nested case—control study in the Nurses’ Health Study Cohort. Within the Nurses’ Health Study, women with the A2 allele were not at increased risk for incident breast cancer or for advanced disease. However, like the findings of Feigelson et al. (32), the protective effect of later age at menarche (>13 years) was only observed among women with the A1 allele and not among women carrying the A2 alleles, adding further support to the hypothesis that the A2 variant allele in CYP17 may act as a modifier of breast cancer risk but is not an independent risk factor. CYP19 Aromatase or estrogen synthetase, encoded by the CYP19 gene, converts androstenedione to El and testosterone to E2 and completes the pathway for estrogen biosynthesis from choles- terol (38). The majority of circulating estrogens in premeno- pausal women is in the form of E2 and is produced cyclically by the granulosa cells of the ovarian follicles (39). However, ex- tensive extragonadal production of estrogens also occurs in liver, muscle, and adipose tissue by aromatization of adrenal andro— gens (40). After menopause, the majority of estrogen is derived from fat by the aromatization of adrenal androstenedione to E, a weaker estrogen than E2. The extent of androstenedione con— version to E1 is associated with higher body fat content and with increasing age, indicating that a major source of estrogen expo— sure in older women is the continuous production of E1 in adi— pose tissue (41). Several studies [reviewed in (42)] have sug- gested a relationship between increased adiposity, elevated El levels, and postmenopausal breast cancer risk. A polymorphic tetranucleotide repeat (TTTA)n has been identified in intron 5 about 80 nucleotides downstream of exon 4 in the CYP19 gene near the intron/exon border. This close proximity to the intron/exon suggests a possible role for these tetranucleotide repeats in the determination of splicing sites. In a study performed by Kristensen et a1. (43), five different alleles containing 7, 8, 9, 11, and 12 TTTA repeats were iden- tified. Although relatively rare, Kristensen et al. (43) noted a significant association with breast cancer risk in carriers of the longest repeat variant (TTTA)12, designated the A1 allele, in a case—control study with 366 case patients and 252 control sub- Journal of the National Cancer Institute Monographs N0. 27, 2000 jects. The A1 allele was present in less than 2% 0f the control population but in almost 4% of case patients. Siegelmann- Danieli et al. (44) also evaluated this association and found increased risk with the variant A1 allele. These data suggest that polymorphisms in the CYP19 gene may be involved as a low- penetrance gene in breast cancer susceptibility. Other polymor— phisms have been identified in the coding region of CYP19, but they do not appear to alter enzyme function or expression (45) and have not been analyzed in population studies in association with risk. COMT COMT is one of several phase II enzymes involved in the conjugation and inactivation of CE (46). Because there is evi— dence that CE, particularly the 4-hydroxyCE, may bind to DNA and result in DNA damage (4,17), the possible role of lower activity in the enzyme in relation to breast cancer risk is impor- tant. An amino acid change (valine to methionine) at position 158/108 in the membrane-bound/cytosolic form of the protein has been linked to decreased methylation activity of COMT (47). The allelic variation at amino acid position 158/108 is believed to be closely associated with the observed trimodal distribution of COMT enzyme activity in the human population. The genotypes designated in relation to the predicted enzymatic activity of the protein are high (COMTval/Val), intermediate (COMTVQVME‘), and low (COMTMWMEI) (48,49). Three groups to date, all with conflicting results (Table 3), have evaluated the role of the COMT genetic polymorphism in relation to breast cancer risk. In a nested case—control study, Lavigne et al. (50) evaluated COMT low-activity alleles in 113 women with breast cancer and an equal number of control subjects. In the entire sample, overall associations with heterozygosity and homozygosity for the “low-activity allele” were 1.30 (95% CI, 0.66—2.58) and 1.45 (95% CI, 0.69—2.58), respectively. When women were stratified by menopausal status, women who were postmenopausal had a greater than twofold increase in risk with the COMTMe"MCt ge- notype or two low—activity alleles, but inverse associations were noted for premenopausal women with the same genotype. Be— cause in postmenopausal women, most estrogens are produced by the conversion of androgens in adipose tissue, associations were also evaluated among women stratified on body mass in- dex (BMI). Among postmenopausal women, associations were noted only among those whose BMI was greater than 24.47 kg/mz. Thompson et a1. (51) performed similar analyses in a study of 281 case patients and 289 control subjects in western New York. In the overall dataset, no relationship was observed between variant COMT alleles and breast cancer risk, but marked differ- ences were noted when data were stratified by menopausal sta- tus. Among premenopausal women with breast cancer, those with at least one low-activity allele showed significantly in— creased risk (OR = 2.4; 95% C1 = 14—43). In contrast to premenopausal women, there was an inverse association be- tween low-activity alleles and postmenopausal breast can- cer, which was most pronounced among those who were COMTMWMet (OR = 0.4; 95% CI = 0.2—0.7). When COMTMWval individuals were combined with individuals who were COMTMeUMe‘, having one or two low-activity alleles significantly decreased risk (OR = 0.5; 95% CI = 0.3—0.9). When data were stratified by menopausal status, the low-activity 129 Table 3. Associations of a low—activity allelic variant of COMT with breast cancer risk according to menopausal status* Study (reference No.) Study size case patients/control subjects COMT low-activity allele alone OR (95% CI) Interaction Exposure Lavigne et al., 1997 (50) All subjects 113/114 No association with breast cancer risk Premenopausal 24/25 Decreased risk of breast cancer Postmenopausal 89/89 Increased risk of breast cancer Thompson et al., 1998 (5]) All subjects 281/289 No association with breast cancer risk Premenopausal 141/134 Increased risk of breast cancer Postmenopausal 140/155 Decreased risk of breast cancer Millikan et al., 1998 (54) All subjects 654/642 No association with breast cancer risk European—American 389/379 No association with breast cancer risk African-American 265/263 No association with breast cancer risk Premenopausal (combined 331/297 No association with European-American and breast cancer African-American) risk; RR = 0.7 (0.4—1.2) Postmenopausal (combined 323/344 No association with European—American and African-American) breast cancer risk; RR = 0.8 1.4 (0.7—2.9) 0.2 (0.0415) 2.2 (0.9—5.1) Highest risk for breast cancer in women with BMI >24.47 g/mz; OR = 3.6 (95% CI = 1.1—12) 2.1 (1.0—4.4) Highest risk for breast cancer Increased risk for breast strongest in heaviest women cancer only among ever with BMI >27 kg/mz; OR = smokers 5.7 (95% CI = 1.1—30.1) 0.4 (0.2—0.7) Decreased risk for breast cancer Decreased risk only strongest in leanest women significant among never with BMI <23 kg/mz; OR = smokers 0.3 (95% CI = 0.1—0.7) 0.7 (0.5—1.1) 0.8 (0.4—.5) Decreased risk for breast cancer in heaviest women with BMI >27.8 kg/mz; RR = 0.5 (0.3—1.1) Decreased risk for breast cancer in women who were physically inactive; RR = 0.5 (95% CI = 0.2—0.9) Decreased risk for breast cancer in women who (0.5—1.4) were physically inactive; RR = 0.5 (95% CI = 0.3—0.9) *BMI = body mass index; CI = confidence interval; COMT = catechol—O—methyltransferase; OR = odds ratio; R = relative risk. COMTMm variant was most strongly associated with risk among the heaviest premenopausal women (OR = 5.7; 95% CI = 1.1—30.l); whereas, in postmenopausal women, an inverse as- sociation with COMT and risk was strongest in the leanest women with at least one low-activity allele (OR = 0.3; 95% CI = 0.1—0.7). The authors hypothesized that there may be an opposing role of CE metabolism in breast cancer etiology, de- pending on the hormonal environment, and that differing bio- logic effects of the CE reported in the literature (i.e., DNA damaging versus growth inhibition) may depend on the levels of circulating estrogens. They further suggested that, in a high- estrogen environment, such as in premenopausal and the heav- iest postmenopausal women, higher circulating levels of the cat- echol compounds (2-OH and 4-OH) may result in higher circulating levels of potentially mutagenic compounds (17). In a low—estrogen environment, as in leaner postmenopausal women, higher circulating levels of the unmethylated catechols in a low COMT background may elevate the levels of the putative anti— carcinogenic 2-OHE1 (52). Because the CEs are products of estrogen metabolism by CYP1A1 and CYP1A2, which are both induced by smoking, Ambrosone et al. (53) have also presented data evaluating the role of COMT on breast cancer in smoking and nonsmoking women. It is interesting that increased risk was observed only among premenopausal women who smoked and that inverse 130 associations were significant only among postmenopausal non- smokers. Millikan et a1. (54) also evaluated these possible rela- tionships in the Carolina Breast Cancer Study (654 case patients and 642 control subjects), within a population of European- American and African-American women. Neither low- nor high- activity alleles were associated with increased breast cancer risk for premenopausal or postmenopausal women, European- American women, or African—American women. Among pre- menopausal women (European-American and African- American women combined), there were inverse associations between low-activity alleles and breast cancer among women with BMI greater than 27.8 and among both premenopausal and postmenopausal women who were physically inactive. Smoking, hormone replacement therapy, and oral contraceptive use did not modify associations. These discrepancies may be due to small sample sizes in the studies by Lavigne et al. (50) and Thompson et al. (51), or there may be biologic factors that differentially impact risk associations; these issues will be addressed below. GST and MnSOD ROS may be generated through a number of mechanisms, including those related to metabolism of E2 (see Chapters 4 and 5). For example, ROS are produced via CE-mediated redox cy— cling of quinones and semiquinones (16). Substantial data indi- cate that oxidative stress is related to breast cancer risk (55—57). Journal of the National Cancer Institute Monographs No. 27, 2000 The glutathione—dependent peroxidases (e.g., GST and seleni— um-dependent glutathione peroxidases) are involved in detoxi- fication of products of oxidative damage, by catalyzing conju- gation of glutathione with ROS (58). Similarly, superoxide dismutase (Mn, Cu, and ZnSOD) catalyzes the dismutation of two superOxide radicals, producing hydrogen peroxide and oxy- gen. Genetic polymorphisms are known to affect enzyme activ- ity in GST M1, T1, and P1, and a recently identified polymor- phism in MnSOD apparently alters the structure of the enzyme, affecting its ability to enter the mitochondrion (59). Polymor— phisms in these enzymes could impact the relationship between oxidative stress and breast cancer etiology. The GSTMl genetic polymorphism is a deletion of the entire gene; approximately 50% of European-Americans inherit the null allele (60). Because GSTMl enzyme is present in human breast tissue (6]), it is plausible that lack of this isozyme could increase breast cancer risk. A number of research groups have evaluated possible associations between GSTMl and breast can- cer risk and, for the most part, have found no effect on risk (see Table 3). In a study of 197 women with breast cancer and 225 control subjects, Zhong et al. (62) found no increased risk with the null allele. Similarly, Ambrosone et al. (27,63) found no association between GSTMl genetic polymorphisms and breast cancer in a study with 212 premenopausal and 410 postmeno— pausal women. Risk relationships were not affected by smoking status or by high or low consumption of dietary sources of antioxidants. In the Nurses’ Health Study, Kelsey et a1. (64) also observed no overall associations between the GSTM1 null allele and breast cancer risk; however, the deletion appeared to be associated with better survival, perhaps because of the role of GST in metabolism of and conjugation with chemotherapeutic agents. Supporting these negative findings, in a study with 164 European-American and 59 African-American women with breast cancer and an equal number of matched control subjects, Bailey et al. (30) also found no associations between GSTMl genotype and breast cancer risk. Associations were observed, however, with not only GSTMl polymorphisms but also with GSTTl and GSTPl, in a nested case—control study evaluated by Lavigne et al. (50) for COMT. Helzlsouer et al. (65) noted that the null genotype was associated with a twofold increase in breast cancer risk, primarily among postmenopausal women. Although not significant, the GSTPl polymorphism also ap— peared to increase risk. When combining putative “high-risk” alleles for the three genes, women who were null for GSTTl and GSTMl, and heterozygous or homozygous for the GSTPl valine substitution, had an almost fourfold increase in breast cancer risk (OR = 3.77; 95% CI = l.lO—l2.88). Bailey et al. (30) also evaluated GSTTl in relation to breast cancer in the previously noted study, but observed no association with increased risk. More recently, Garcia-Closas et al. (66) reported no evidence for an association between GSTTl null (OR = 0.86; 95% CI = 0.61—1.21) or GSTM1 null (OR = 1.05; 95% CI = 0.80—1.37) in 466 women with incident breast cancer compared with an equal number of control subjects in the Nurses’ Health Study. Furthermore, when GST genotypes were considered in combi- nation or together with cigarette smoking, no associations with an increased risk of breast cancer were observed. Until recently, the MnSOD polymorphism had been evalu- ated only in relation to Parkinson’s disease in a study in Japan (59). Because ROS, including those generated by estrogens and their metabolites (Chapter 4), may be involved in breast carci- Journal of the National Cancer Institute Monographs No. 27, 2000 nogenesis and because MnSOD is a major enzyme involved in the scavenging of free radicals, Ambrosone et a1. (67) hypoth~ esized that the MnSOD alanine allele could be related to breast cancer risk by having an altered capacity to reduce oxidative stress. In the western New York study, the variant allele was associated with an almost twofold increase in risk in postmeno- pausal women and, in premenopausal women, heterozygosity or homozygosity for the alanine allele conferred a 3.5 risk (95% CI = 1.8—6.8). Risk appeared to be the greatest among women who were in the lower median for consumption of dietary sources of antioxidants, fruits, and vegetables overall. METHODOLOGIC ISSUES IN MOLECULAR EPIDEMIOLOGY The molecular epidemiologic study of susceptibility inferred by genetic polymorphisms and the elucidation of gene— environment interactions should advance our understanding of carcinogenic mechanisms. But many studies of polymorphisms in xenobiotic and steroid hormone-metabolizing enzymes and cancer risk, and those of gene—environment interactions have yielded conflicting results. The molecular epidemiologic litera— ture is rife with inconclusive data. There is a clear need for the molecular epidemiologic community to explore areas of bias and flaws in study design and analyses that may result in inconsistent study results. A number of plausible explanations are available for conflict- ing results in gene~environment interaction studies, the most obvious of which would be related to small sample size and misclassification of exposure, which can result in both Type I and Type II errors. A Type 11 error, which represents the inabil— ity to detect a true effect, results from inadequate power. Statis— tical power depends on sample size, the size of the effect to be detected, and the variability within the study population. Small sample sizes are common to molecular epidemiologic studies, not only because population-based molecular studies are expen— sive and, thus, restrict the number of subjects recruited but also because the means of analysis automatically cut the population in half or, in many instances, well below that in stratified analy- ses. For studies of gene—environment interaction, case and con- trol subjects are stratified by genotype, and associations between the risk factor and disease status are evaluated separately within each stratum. Theoretically, one would stratify by genotype and calculate relative risks or ORs to estimate risk associated with an exposure, adjusting for possible confounding factors or other known disease risk factors. Clearly, even in large studies, the numbers of subjects in each cell will be drastically reduced, greatly decreasing power and the likelihood of being able to detect an effect. Less often considered, however, is the likeli— hood of a Type I error (i.e., a false-positive result). As pointed out by Greenland and Rothman (68), when stratification has exceeded the limits of the data, the exposure effect estimates begin to get further and further from the null, and the OR be- comes inflated. Bias in the analysis may be introduced in the process of applying large-sample methods to sparse data. Addi- tionally, even modest errors in exposure measurements or ge- netic tests can confer a significant bias in epidemiologic Studies, requiring even larger-than-expected sample size to compensate for misclassification (69). As recently demonstrated and dis- cussed by Garcia—Closas et al. (69), the importance for improved accuracy and validation of exposure assessment and establish— ment of good quality control for genetic testing should be con- 131 sidered areas of high priority in gene—environment interaction studies. However, despite issues related to small sample sizes, results may also vary because of the nature of the individual study designs and the populations being studied. For the most part, investigators have used three approaches to study gene— environment interactions: 1) to evaluate effects of a polymor- phism alone on cancer risk, 2) to investigate the effects of the polymorphism on risk among persons exposed and unexposed to an environmental factor, and 3) to examine the effects of an exposure on risk within individuals with varying genotypes. This approach however, is inherently limited by the amount of infor— mation that genotype confers on enzyme activity. Genotype alone is an inexact or incomplete surrogate for the knowledge of enzyme activity that one would want and does not take into account the induction or inhibition of enzyme activity. And, other than a handful of null and inactive allelic variants (i.e., GST M1 and T1), genotype alone explains only a proportion of the phenotype. Levels of enzyme activity may rise or decline depending on the effects of any number of endogenous and exogenous factors. For example, enzyme induction may occur through the effects of steroid hormones or by exposure to drugs, pesticides, industrial chemicals, tobacco smoke, ethanol, and food sources. Conversely, enzyme activity can be inhibited through competitive binding at the active site, decreased biosyn- thesis, or increased breakdown of the enzyme by a number of chemicals and foods. An example of the impact of these multiple variables on predicted associations is demonstrated in the unexpected obser— vation that GST Ml may increase the risk of colorectal adeno- mas in a diet high in broccoli (70). The mechanism proposed by these investigators was that the GSTs conjugate the cancer che- mopreventive isothiocyanate, sulforaphane, rendering it inac- tive. Thus, although one would intuitively presume that an ab— sence of GST would increase cancer risk through a decreased ability to conjugate with ROS and chemical carcinogens as dis— cussed above, lack of GST may actually be protective by not negating the anticarcinogenic effects of isothiocyanates. Be- cause of the many factors that can impact enzyme activity, in- consistencies in results of studies measuring genotype as an independent marker of enzyme activity could be related to dif- ferential exposures of study populations to known or unknown factors that affect enzyme levels. Thus, molecular epidemiologists face formidable challenges in study design to investigate accurately the role of hormone metabolites in breast cancer risk, particularly through assess- ment of genetic polymorphisms in enzymes involved in hor— mone metabolism. First, a strong understanding of biologic mechanism of disease will be essential in defining and accu- rately monitoring risk-related exposures in subsets of people defined by genetic sensitivity. Second, large studies are needed, as well as creative analytic approaches, to be able to evaluate accurately the effects of multiple genes and multiple exposures on breast cancer risk. FUTURE DIRECTIONS The first approach to be applied to the study of susceptibility to breast cancer would be a reassessment of breast cancer disease and the related risk factors. Several intriguing but unconfirmed observations suggest that tumor phenotypes clinically subtyped by (ER)/progesterone receptor (PR) status, p53 mutations, or 132 HER-2/neu may not only be valuable as prognostic indicators for patient outcome but may also reflect distinct etiologic path- ways, providing a means to mirror past exposures (71—75). For example, breast tumors that overexpress the oncogene HER—2/ neu have been suggested to represent a breast cancer subtype with a common etiologic origin that may have evolved from a pathway separate from other breast cancers (74, 76). The hypoth- esis that separate breast tumor subtypes exist has led to the corollary that these tumor subsets may have distinctive sets of risk factors relevant to differing causative events. This concept merits serious consideration in future study designs. For ex— ample, amplification of the oncogene HER-2/neu, associated with a poor prognostic tumor phenotype, has been positively associated with early use of oral contraceptives (71,72) and in- versely associated with breast-feeding (73,74). The overexpres- sion of p53 has been positively associated with current smoking (77) and with oral contraceptive usage (78). In the Iowa Wom— en’s Health Study, Potter et al. (75) evaluated classic epidemio- logic risk factor data to determine the interaction between breast cancer risk factors (i.e., family history, BMI, reproductive fac- tors, hormone use, and alcohol use) and ER/PR status. In this study of 939 incident breast cancers, three patterns of association appeared in relation to epidemiologic risk factors. PR-positive breast cancers were associated with endogenous hormone expo- sure variables, whereas ER-negative/PR-negative and ER- positive/PR-positive breast cancers were inversely associated with a number of the widely accepted reproductive risks. Al— though the data are not presented but referred to, Huang et al. (79) also indicate significant differences for reproductive risk factors among case subjects stratified by ER/PR status in the Carolina Breast Cancer Study. These observations suggest that tumor phenotypes subtyped by ER/PR status, HER-2/neu, or p53 may reflect distinct clinical entities. Their etiologies may be related to host environmental and lifestyle factors that ultimately impact the natural history of the disease. Overall, these data are provocative and suggest that etiologic heterogeneity in subsets of breast cancers may have masked the impact of relevant risk factors. Incorporation of tumor phenotyping in disease stratifi- cation may yield stronger and more consistent associations with etiologic and genetic factors, such as reproductive variables, smoking, exogenous hormone usage, and metabolic variability among subsets of breast tumors. A second important area for future studies is the continued identification of functionally significant gene polymorphisms and the incorporation of research platforms to rapidly interrogate the associations between allele variability, exposure, and risk. Unfortunately, the identification of functionally distinct genetic variants that alter activity or gene expression in a modest fashion (i.e., changes of activity or expression), as opposed to rare mu— tations that result in overt metabolic deficiencies, is difficult and has been slow. The identification of widely distributed mutations or polymorphisms has been greatly facilitated by the ongoing efforts to map the more than 80000 genes in man. This explo— sion in information on the genetic heterogeneity of man will be fundamental to understanding complex disease risk. However, the incorporation of sequence-based information in population studies will only be practical with the validation of high- throughput, DNA-based methodologies, as promised by a num- ber of newly emerging platforms, such as DNA microarrays, multiplex analysis on fluorescent microspheres, or multiplex mass-tag strategies (80). The introduction of technologies that Journal of the National Cancer Institute Monographs No. 27, 2000 provide simultaneous assessment of tens to hundreds of allelic variants in large numbers of samples promises rapid, accurate, and cost-effective iterative hypothesis testing for gene—based complex diseases. With the tools to more clearly understand cancer etiology and to identify susceptible subgroups of individuals also comes the responsibility to devise strategies to produce results that are valid. Because of the pitfalls and sources of bias in molecular epidemiologic studies, it is imperative that we seriously critique our own data, establish working consortia to develop and to capitalize on large cohort studies, and introduce efforts to in- crease public and professional awareness of the need to perform human population studies. Presentation of data from small stud- ies, even though they may conflict with each other, will move the field forward. As insights are gained from epidemiologic studies and taken to the laboratory, they will help us to under- stand mechanisms of carcinogenesis and should be encouraged but interpreted with caution. At present, this remains an area of fundamental research with potential to significantly impact our understanding of individual variability and disease risk within the human population. Alternatively, this research should pro- vide much needed information for identifying subpopulations that would most benefit from aggressive prevention strategies. REFERENCES (I) Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1999. CA Cancer J Clin 1999;49:8—31. (2) Kelsey JL, Bernstein L. Epidemiology and prevention of breast cancer. Annu Rev Public Health 1996;17:47—67. (3) Pike MC, Spicer DV, Dahmoush L, Press MF. Estrogens, progestogens, normal breast cell proliferation, and breast cancer risk. Epidemiol Rev 1993;15:17—35. Cavalieri EL, Stack DE, Devanesan PD, Todorovic R, Dwivedy I, Higgin— bothan S, et al. Molecular origin of cancer: catechol estrogen-3,4-quinones as endogenous tumor initiators. Proc Natl Acad Sci US A 1997;94: 10937—42. Liehr JG. Dual role of oestrogens as hormones and pro-carcinogens: tu— mour initiation by metabolic activation of oestrogens. Eur J Cancer Prev 1997;6:3—10. Dorgan JF, Longcope C, Stephenson HE Jr, Falk RT, Miller R, Franz C, et al. Serum sex hormone levels are related to breast cancer risk in postmeno— pausal women. Environ Health Perspect 1997;105(suppl 3):583—5. (7) Zhu BT, Conney AH. Functional role of estrogen metabolism in target cells: review and perspectives. Carcinogenesis 1998:1921—27. (8) Dunning AM, Healey CS, Pharoah P, Teare MD, Ponder BA, Easton DF. A systematic review of genetic polymorphisms and breast cancer risk. Cancer Epidemiol Biomarkers Prev 1999;82843—54. (9) Coughlin SS, Piper M. Genetic polymorphisms and risk of breast cancer. Cancer Epidemiol Biomarkers Prev 1999;8:1023—32. (10) Omura T, Morohashi K. Gene regulation of steroidogenesis. J Steroid Biochem Mol Biol 1995;53:19—25. (11) Payne A. Hormonal regulation of cytochrome P450 enzymes, cholesterol side-chain cleavage and 17or-hydroxylase/Cl7-2O lyase in Leydig cells. Biol Reprod 1990;42:399—404. (12) Kelloff GJ, Lubet RA, Lieberman R, Eisenhauer K, Steele VE, Crowell J A, et al. Aromatase inhibitors as potential cancer chemopreventives. Cancer Epidemiol Biomarkers Prev 1998;7265—78. (13) Poutanen M, Isomaa V, Peltoketo H, Vihko R. Role of l7B—hydroxysteroid dehydrogenase type 1 in endocrine and intracrine estradiol biosynthesis. J Steroid Biochem Mol Biol 1995;55:525—32. (14) Dufort I, Rheault P, Huang XF, Soucy P, Luu-The V. Characteristics of a highly labile human type 5 l7beta—hydroxysteroid dehydrogenase. Endo- crinology 1999;140:568—74. (15) Manucci CP, Fishman J. P450 enzymes of estrogen metabolism. Pharma- col Ther 1993;57:237—57. (16) Yager JD, Liehr JG. Molecular mechanisms of estrogen carcinogenesis. Annu Rev Pharrnacol Toxicol 1996;36:203—32. (4 x, (5) (6) Journal of the National Cancer Institute Monographs No. 27, 2000 (I7) Liehr JG. Hormone—associated cancer: mechanistic similarities between human breast cancer and estrogen—induced kidney carcinogenesis in ham— sters. Environ Health Perspect 1997;105(suppl 3):565—9. (18) Telang NT, Katdare M, Bradlow HL, Osborne MP. Estradiol metabolism: an endocrine biomarker for modulation of human mammary carcinogene- sis. Environ Health Perspect 1997:105(suppl 3):559—64. (19) Feigelson HS, Ross RK, Yu MC, Coetzee GA, Reichardt JK, Henderson BE. Sex steroid hormones and genetic susceptibility to breast and prostate cancer. Drug Metab Rev 1998;30:421—34. (20) Daly AK, Cholerton S, Gregory W, Idle JR. Metabolic polymorphisms. Pharmacol Ther 1993;57:129—60. (21) Spink DC, Eugster HP, Lincoln DW 2nd, Schuetz JD, Schuetz EG, Johnson JA, et a1. l7B-Estradiol hydroxylation catalyzed by human cytochrome P450 1A1: a comparison of the activities induced by 2,3,7,8- tetrachlorodibenzo-p-dioxin in MCF-7 cells with those from heterologous expression of the cDNA. Arch Biochem Biophys 1992;293:342—8. Kawajiri K, Nakachi K, Imai K, Yoshii A, Shinoda N, Watanabe J. Iden» tification of genetically high risk individuals to lung cancer by DNA poly» morphisms of the cytochrome P4501A1 gene. FEBS Lett 1990;263:131—3. Hayashi S, Watanabe J, Nakachi K, Kawajiri K. Genetic linkage of lung cancer—associated MspI polymorphism with amino acid replacement in the heme binding region of the human cytochrome P4501A1 gene. J Biochem 1991;110:407—11. Crofts F, Cosma GN, Currie D, Taioli E, Toniolo P, Garte SJ. A novel CYP1A1 gene polymorphism in African Americans [published erratum appears in Carcinogenesis 1993;14:2652]. Carcinogenesis 1993;14: 1729—31. (25) Cascorbi I, Brockmoller J, Roots l. A C4887A polymorphism in exon 7 of human CYP1A1: population frequency, mutation linkages, and impact on lung cancer susceptibility. Cancer Res 1996;56:4965—9. (26) Rebbeck TR, Rosvold EA, Duggan DJ, Zhang J, Buetow KH. Genetics of CYP1A1: coamplification of specific alleles by polymerase chain reaction and association with breast cancer. Cancer Epidemiol Biomarkers Prev 1994;32511—4. (27) Ambrosone CB, Freudenheim JL, Graham S, Marshall JR, Vena JE, Bra— sure JR, et al. Cytochrome P4501A1 and glutathione S—transferase (Ml) genetic polymorphisms and postmenopausal breast cancer risk. Cancer Res 1995;55:3483—5. (28) Taioli E, Trachman J, Chen X, Toniolo P, Garte SJ. A CYP1A1 restriction fragment length polymorphism is associated with breast cancer in African— American women. Cancer Res 1995;55:3757—8. (29) lshibe N, Hankinson SE, Colditz GA, Spiegelman DL, Willett WC, Speizer FE, et a1. Cigarette smoking, cytochrome P4501A1 polymorphisms, and breast cancer risk in the Nurses’ Health Study. Cancer Res 1998;58:667—71. (30) Bailey LR, Roodi N, Verrier CS, Yee CJ, Dupont WD, Parl FF. Breast cancer and CYP1A1, GSTMl, and GSTTl polymorphisms: evidence of a lack of association in Caucasians and African Americans. Cancer Res 1998; 58:65—70. (31) Brentano ST, Picado—Leonard J, Mellon SH, Moore CC, Miller WL. Tis- sue-specific, cyclic adenosine 3',5’—monophosphate—induced, and phorbol ester-repressed transcription from the human P450c17 promoter in mouse cells. Mol Endocrinol 1990;4z1972-9. (32) Feigelson HS, Coetzee GA, Kolonel LN, Ross RK, Henderson BE. A polymorphism in the CYP17 gene increases the risk of breast cancer. Can- cer Res 1997;57:1063—5. Feigelson HS, Shames LS, Pike MC, Coetzee GA, Stanczyk FZ, Henderson BE. Cytochrome P450c17or gene (CYP17) polymorphism is associated with serum estrogen and progesterone concentrations. Cancer Res 1998; 58:585—7. Dunning AM, Healey CS, Pharoah PD, Foster NA, Lipscombe JM, Red— man KL, et al. No association between a polymorphism in the steroid metabolism gene CYP17 and risk of breast cancer. Br J Cancer 1998;77: 2045—7. Helzlsouer KJ, Huang HY, Strickland PT, Hoffman S, Alberg AJ, Com— stock GW, et a1. Association between CYP17 polymorphisms and the de- velopment of breast cancer. Cancer Epidemiol Biomarkers Prev 1998;72 945—9. Weston A, Pan CF, Bleiweiss, IJ, Ksieski HB, Roy N, Maloney N, et al. CYP17 genotype and breast cancer risk. Cancer Epidemiol Biomarkers Prev 1998;7:94l—4. (22) (23) (24) (33) (34) (35) (36) 133 (37) Haiman CA, Hankinson SE, Spiegelman D, Colditz GA, Willet WC, Speizer FE, et al. The relationship between a polymorphism in CYP17 with plasma hormone levels and breast cancer. Cancer Res 1999;59:1015-20. (38) Sasano H, Nagura H, Harada N, Goukon Y, Kimura M. Immunolocaliza— tion of aromatase and other steroidogenic enzymes in human breast disor- ders. Hum Pathol 1994;25:530-5. (39) Cole PA, Robinson CH. Mechanism and inhibition of cytochrome P-450 aromatase. J Med Chem 1990;33:2933—42. (40) Longcope C, Pratt JH, Schneider SH, Fineberg SE. Aromatization of an— drogens by muscle and adipose tissue in vivo. J Clin Endocrinol Metab 1978;46:146—52. (4]) Toniolo PG. Endogenous estrogens and breast cancer risk: the case for pro- spective cohort studies. Environ Health Perspect 1997;105(supp1 3):587—92. (42) Key TJ, Pike MC. The role of oestrogens and progestagens in the epide— miology and prevention of breast cancer. Eur J Cancer Clin Oncol 1988; 24:29—43. (43) Kristensen VN, Andersen TI, Lindblom A, Erikstein B, Magnus P, Bor- reson-Dale AL. A rare CYP19 (aromatase) variant may increase the risk of breast cancer. Pharmacogenetics 1998;8z43—8. (44) Siegelmann—Danieli N, Buetow KH. Constitutional genetic variation at the human aromatase gene (Cyp19) and breast cancer risk. Br J Cancer 1999; 79:456—63. (45) Watanabe J, Harada N, Suemasu K, Higashi Y, Gotoh O, Kawajiri K. Arginine-cysteine polymorphism at codon 264 of the human CYP19 gene does not affect aromatase activity. Pharmacogenetics 1997;7z4l9—24. (46) Guldberg HC, Marsden CA. Catechol—O—methyl transferase: pharmacologi— cal aspects and physiological role. Pharmacol Rev 1975;27:135—206. (47) Lachman HM, Papolos DF, Saito T, Yu YM, Szumlanski CL, Weinshil— boum RM. Human catechol-0—methyltransferase pharrnacogenetics: de- scription of a functional polymorphism and its potential application to neuropsychiatric disorders. Pharmacogenetics 1996;6z243—50. (48) Goldstein M, Lieberman A. The role of the regulatory enzymes of catechol— amine synthesis in Parkinson’s disease. Neurology l992;42(4 suppl 4):8—l2. (49) Price-Evans DA. Genetic factors in drug therapy: clinical and molecular pharmacogenetics. Cambridge (U.K.): Cambridge University Press; 1993. (50) Lavigne J A, Helzlsouer KJ, Huang HY, Strickland PT, Bell DA, Selmin O, et al. An association between the allele coding for a low activity variant of catechol-0-methyltransferase and the risk for breast cancer. Cancer Res 1997;57:5493—7. (51) Thompson PA, Shields PG, Freudenheim JL, Stone A, Vena JE, Marshall JR, et al. Genetic polymorphisms in catechol—0-methyltransferase, meno- pausal status, and breast cancer risk. Cancer Res 1998;58:2107—10. (52) Bradlow HL, Telang NT, Sepkovic DW, Osborne MP. 2-Hydroxyestrone: the “good” estrogen. J Endocrin 1996;150(Suppl):SZ59—65. (53) Ambrosone CB, Thompson PA, Shields PG, Freudenheim JL, Stone A, Marshall JR, et a1. Catechol-0-methyltransferase (COMT) genetic poly morphisms, smoking and breast cancer risk. Am J Epidemiol 1998;147:565. (54) Millikan RC, Pittman GS, Tse CK, Duell E, Newman B, Savitz D, et a1. Catechol—0—methyltransferase and breast cancer risk. Carcinogenesis 1998; 19:1943—7. (55) Schwartz JL, Antoniades DZ, Zhao S. Molecular and biochemical repro— gramming of oncogenesis through the activity of prooxidants and antioxi- dants. Ann N Y Acad Sci 1993;686:262—78. (56) Ames BN. Dietary carcinogens and anticarcinogens. Oxygen radicals and degenerative diseases. Science 1983;221:1256—64. (57) Djuric Z, Depper JB, Uhley V, Smith D, Lababidi S, Martino S, Heilbrun L. Oxidative DNA damage levels in blood from women at high risk for breast cancer are associated with dietary intakes of meats, vegetables, and fruits. J Am Diet Assoc 1998;98:524—8. (58) Hayes JD, Pulford DJ. The glutathione S-transferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer che— moprotection and drug resistance. Crit Rev Biochem Mol Biol 1995;30: 445—600. (59) Shimoda—Matsubayashi S, Matsumine H, Kobayashi T, Nakagawa-Hattori Y, Shimizu Y, Mizuno Y. Structural dimorphism in the mitochondrial targeting sequence in the human manganese superoxide dismutase gene. A predictive evidence for conformational change to influence mitochondrial transport and a study of allelic association in Parkinson's disease. Biochem Biophys Res Commun 1996;226:561—5. (60) Brockmoller J, Gross D, Kerb R, Drakoulis N, Roots 1. Correlation between 134 trans-stilbene oxide-glutathione conjugation activity and the deletion mu- tation in the glutathione S—transferase class mu gene detected by polymer- ase chain reaction. Biochem Pharmacol 1992;43:647—50. (61) Forrester LM, Hayes JD, Millis R, Barnes D, Harris AL, Schlager JJ, et al. Expression of glutathione S—transferases and cytochrome P450 in normal and tumor breast tissue. Carcinogenesis 1990;11:2163-70. (62) Zhong S, Wyllie AH, Barnes D, Wolf CR, Spurr NK. Relationship between the GSTM1 genetic polymorphism and susceptibility to bladder, breast and colon cancer. Carcinogenesis 1993;14:1821—4. (63) Ambrosone CB, Coles BF, Freudenheim JL, Shields PG. Glutathione S- transferase (GSTMl) genetic polymorphisms do not affect human breast cancer risk, regardless of dietary antioxidants. J Nutr 1999;129(2S suppl): 5655-88. (64) Kelsey KT, Hankinson SE, Colditz GA, Springer K, Garcia—Closas M, Spiegelman D, et a1. Glutathione S-transferase class mu deletion polymor- phism and breast cancer: results from prevalent versus incident cases. Can- cer Epidemiol Biomarkers Prev 1997;6:51 1—5. (65) Helzlsouer KJ, Selmin 0, Huang H, Strickland PT, Hoffman S, Alberg AJ, et al. Association between glutathione S-transferase M 1, P1, and T1 genetic polymorphisms and development of breast cancer. J Natl Cancer Inst 1998; 90:512—8. (66) Garcia—Closas M, Kelsey KT, Hankinson SE, Spiegelman D, Springer K, Willett WC, et a1. Glutathione S—transferase mu and theta polymorphisms and breast cancer susceptibility. J Natl Cancer Inst 1999;91:1960—4. (67) Ambrosone CB, Freudenheim JL, Thompson PA, Bowman E, Vena JE, Marshall JR, et a1. Manganese superoxide dismutase (MnSOD) genetic polymorphisms, dietary antioxidants, and risk of breast cancer. Cancer Res 1999;59:602—6. (68) Greenland S, Rothman K. Introduction to stratified analysis. In: Modern epidemiology. Philadelphia (PA): Lippincott-Raven; 1998. p. 258. (69) Garcia-Closas M, Rothman N, Lubin J. Misclassification in case—control studies of gene—environment interactions: assessment of bias and sample size. Cancer Epidemiol Biomarkers Prev 1999;8:1043—50. (70) Lin HJ, Probst-Hensch NM, Louie AD, Kau IH, Witte JS, Ingles SA, et a1. Glutathione transferase null genotype, broccoli, and lower prevalence of colorectal adenomas. Cancer Epidemiol Biomarkers Prev 1998;7:647—52. (71) Olsson H, Borg A, Femo M, Ranstam J, Sirgurdsson H. Her-2/neu and INT2 proto—oncogene amplification in malignant breast tumors in relation to reproductive factors and exposure to exogenous hormones. J Natl Cancer Inst l991;83;1483—7. (72) Treumiet HF, Rookus MA, Peterse JL, Hart AA, van Leeuwen FE. Dif— ferences in breast cancer risk factors to neu (c-erbB-2) protein overexpres— sion in breast tumor. Cancer Res 1992;52:2344—5. (73) Gammon MD, Hibshoosh H, Terry MB, Bose S, Schoenberg JB, Brinton LA, et a1. Oral contraceptive use and other risk factors in relation to HER- 2/neu overexpression in breast cancer among young women. Cancer Epi— demiol Biomarkers Prev 1999;8z413—9. (74) Huang WY, Newman B, Millikan RC, Conway K, Hulka BS, Schell MJ, et al. Risk of breast cancer according to the status of HER-2/neu oncogene amplification. Cancer Epidemiol Biomarkers Prev 2000;9:65—7l. (75) Potter JD, Cerhan JR, Sellers TA, McGovern PG, Drinkard C, Kushi LR, et al. Progesterone and estrogen receptors and mammary neoplasia in the Iowa Women’s Health Study: how many kinds of breast cancer are there? Cancer Epidemiol Biomarkers Prev l995;4:3l9—26. (76) De Potter CR. The neu-oncogene: more than a prognostic indicator? Hum Pathol 1994;25:1264—8. (77) Gammon MD, Hibshoosh H, Terry MB, Bose S, Schoenberg JB, Brinton LA, et al. Cigarette smoking and other risk factors in relation to p53 expression in breast cancer among young women. Cancer Epidemiol Bio— markers Prev 1999;8z255—63. (78) van der Kooy K, Rookus MA, Peterse HL, van Leeuwen FE. p53 protein overexpression in relation to risk factors for breast cancer. Am J Epidemiol 1996;144:924—33. (79) Huang WY, Newman B, Millikan RC, Schell MJ, Hulka BS, Moorrnan PG. Hormone—related factors and risk of breast cancer in relation to estrogen receptor and progesterone receptor status. Am J Epidemiol 2000;151: 703—14. (80) Schaid DJ, Buetow K, Weeks DE, Wijsman E, Guo SW, Ott J, et al. Discovery of cancer susceptibility genes: study designs, analytic ap proaches, and trends in technology. J Natl Cancer Inst Monogr 1999;26:1—16. Journal of the National Cancer Institute Monographs N0. 27, 2000 Chapter 8: Estrogen Receptor-Mediated Processes in Normal and Cancer Cells Robert B. Dickson, George M. Stance] The role of estrogens in breast and other cancers has been extensively investigated for many years, and historically most of these studies have focused on the hormonal regula- tion of cell proliferation. The most recent work in this area has focused on the expression of genes likely to mediate pro- liferation (e.g., growth factors, proto-oncogenes, etc.) and their regulation by the classic nuclear estrogen receptor, ER- oc. In this chapter, we present a synopsis of several new developments in this area of ER-regulated gene expression. These developments include the following: 1) the selective activation of ER domains by partial estrogen antagonists, such as tamoxifen and other ligands; 2) the effects of ER-a overexpression and gene knockout on the development of breast and uterine cancers in experimental animal models; 3) mechanisms by which steroid hormones regulate pro- grammed cell death, cell cycle progression, cell-substratum interactions, and genomic instability in cancer cells; 4) iden- tification of nuclear proteins that interact with the ER in the presence of agonists and antagonists, the effect of ligand binding on the receptor structure, and the interactions of liganded and nonliganded receptors with coactivators, core- pressors, and other regulatory proteins; and 5) the biochemi- cal properties, cellular distribution, and potential biologic roles for the newly discovered ER-B. Although there is an increasing interest in understanding the role of estrogens as endogenous carcinogens, it remains clear that ER-mediated regulation of gene expression plays many significant roles in normal and cancer cells, and increased knowledge of the mechanisms involved will improve our overall understand- ing of hormonal carcinogenesis. [J Natl Cancer Inst Monogr 2000;27:135—45] BACKGROUND AND RATIONALE FOR SESSION The primary focus of this meeting was “Estrogens as Endog- enous Carcinogens in the Breast and Prostate,” and most speak- ers thus discussed the actions of estrogens that are potentially related to initiation events (Chapters 3—5). This focus is some- what different from most historic studies on estrogens and can- cer, which focused largely on the role of estrogens in the process of proliferation. The View was that estrogens increased prolif— eration of target cells in the breast and other tissues, and this proliferation contributed to breast cancer by one of two major mechanisms. First, an increase in cell proliferation would be expected to cause an increase in spontaneous errors associated with DNA replication. Second, after mutations were introduced into a target cell by this or other mechanisms, estrogens would enhance the replication of clones of cells carrying such genetic errors. Much of the focus on estrogens in cancer to date has thus been on the mechanisms by which estrogens increase cell pro- liferation. The general view has been that estrogens regulate prolifera- tion of target cells by transcriptional mechanisms involving the Journal of the National Cancer Institute Monographs No. 27, 1999 classic estrogen receptor (ER), initially discovered in the labo- ratories of Elwood Jensen at the University of Chicago, IL, and Jack Gorski at the University of Illinois (Champaign-Urbana). Work from their laboratories, along with metabolic inhibitor studies of Gerald Mueller, then at the University of Wisconsin (Madison), led to the concept that estrogens increased prolifera- tion by stimulating RNA synthesis in target cells. This concept led to a search for target genes for which transcription was regulated by estrogenic hormones, and many laboratories iden— tified a number of growth factors, proto—oncogenes, and other regulatory molecules that were likely candidates for such genes. In recent years the emphasis of these studies has progressed to investigating the transcriptional regulation of such target genes by the ER, with special emphasis on identifying the regu- latory factors involved and their molecular mechanism of action, differences in the activity Of various ER ligands, the identifica- tion of new ER subtypes, and the types of gene families regu— lated by estrogens during hormonally induced increases in pro- liferation. It was thus felt important to have a session on ER mediated processes in normal and target cells. Because it was impossible to present a comprehensive review of this body of work in a single session, the goal was to invite a group of speakers who would address some of the most rapidly emerging paradigms of estrogen action that the Organizing Committee felt were particularly relevant to understanding the actions of estro- gens in cancer cells. Readers interested in additional information on this aspect of estrogen action are referred to a number of recent reviews and articles and references therein, in addition to the references pro- vided throughout the body of this article. These references in- clude information on the structure and function of the nuclear ERs (1—9), the roles of nuclear receptor coactivators and core- pressors in steroid hormone action (10—12), recent advances in the development of selective ER modulator(s) (SERM) (13—17), and the various phenotypes Observed in ER knockout mice, which indicate biologic actions mediated by these nuclear re- ceptors (18—21). OVERVIEW OF SPEAKERS AND TOPICS Until quite recently, the stimulation of transcription by estro- gens was viewed in terms of a relatively straightforward set of interactions initiated by estrogen binding to a Single type of ER that was thought to be identical in all target tissues. Ligand Afliliations of authors: R. B. Dickson, Lombardi Cancer Research Center, Georgetown University, Washington, DC; G. M. Stancel, Department of Inte— grative Biology, Pharmacology, and Physiology, University of Texas Medical School at Houston. Correspondence to: George M. Stancel, Ph.D., Department of Integrative Biology, Pharmacology, and Physiology, University of Texas Medical School at Houston, 6431 Fannin, Houston, TX 77030. See “Note” following “References.” © Oxford University Press 135 binding was viewed as a “trigger” that activated the receptor from an “off” state to an “on” state, and this activation enabled the receptor to activate or repress transcription of target genes. The activated receptor was then thought to interact with an es- trogen response element (ERE) in the 5’-flanking region of re- sponsive genes. It was thought that EREs of most endogenous hormone-regulated genes would have sequences similar to the palindromic sequence, GGTCAnnnTGACC, originally identi- fied in the vitellogenin gene and generally referred to as the consensus ERE. This scheme of estrogen action is illustrated in Fig. l, and, although highly schematized and oversimplified, it represents, to a good approximation, the state of our basic knowledge of estrogen-regulated transcription about 5 years ago. In terms of cancer, it has been known for many years that breast cancer cells contain steroid receptors, and the content of ERs and progesterone receptors (PRs) in individual tumors is a valuable predictor of whether an individual patient will respond to endocrine therapy. However, the correlations between recep- tor content and responses to endocrine therapy are far from perfect, and many tumors progress to states of hormone inde- pendence. Antihormones, such as tamoxifen, were known to compete with estrogenic agonists for receptor binding, but little else was known about the specific biochemical mechanisms by which these important drugs produced their actions in experi— mental or therapeutic settings. In addition, their use is compli— cated because most breast tumors eventually become refractory to antiestrogen treatment. Paradoxically, drugs such as tamoxi— fen also display strong agonist activity in the endometrium, which is highly problematic for their therapeutic use. Such ob- servations were difficult to reconcile with a view of the ER as a simple “on/off” switch that interacted with the same regulatory sequence in all target genes. Within recent years, significant advances in our understand- ing of ER-mediated events have occurred at the conceptual level, and major new experimental approaches to the study of hormone action have become available. Many of these approaches will be discussed in this session, and several key points are enumerated below. ESTROGEN TARGET CELL . Estrogen Receptor a E2 \/ D “Transcription” ERE Estrogen Responsive Gene E2 GGTCAnnnTGACC (Estradiol) Fig. 1. Model of estrogen action circa 1990. Estrogens such as estradiol (E2) enter target cells by diffusion and bind the classic estrogen receptor—(x (filled ovals). The receptor—hormone complex stimulates transcription of target genes via interactions with an estrogen response element similar to the sequence 5’- GGTCAnnnTGACC—3' identified in the vitellogenin A2 gene. 136 1) It is now known that the ER contains several “domains” that are involved in transcriptional regulation and that different ligands may selectively activate these functions. This knowl- edge raises the exciting prospect of developing estrogens that can be used to selectively produce desired therapeutic actions while minimizing untoward side effects; several such agents have already been discovered. Such functional studies on the actions of different estrogens and antiestrogens are being accompanied by structural studies of the molecular interac- tions between ligands and the receptor, and this combination will almost certainly lead to the discovery of even more selective agents. A related question of special importance to understand breast cancer etiology is whether structurally di- verse estrogens differentially stimulate proliferation of breast cancer cells. Dr. McDonnell discusses the role of the ER transcriptional activation functions (AFs) in ligand selective responses. 2) A major experimental advance has been the production of experimental animals that overexpress the ER or knockout animals that do not express the receptor. These experimental animals provide heretofore unavailable approaches to define unequivocally the role of ER in estrogen-mediated events, to identify redundant signaling pathways that compensate for changes in ER levels, and to identify previously unrecog— nized actions of estrogens. Dr. Couse describes the genera— tion and phenotypes of ER-a knockout (ERaKO) and over— expressing mice as well as the effect that receptor levels have on the development of breast and uterine cancers in experi- mental models. In the past, a major focus of study has been the regulation of growth factor and proto—oncogene expression. More recently, attention has increased to other ways by which estrogens might affect breast cancer. This attention includes the study of mechanisms that regulate cell death, the factors that con— trol cell—cycle progression, and the mechanisms that contrib- ute to genomic instability of cancer cells. In addition, interest has increased in processes, such as angiogenesis and cell— substratum interactions, that can affect tumor growth and metastases, and understanding these processes may also im- prove our understanding of the etiology of breast cancer and potential therapeutic targets. Dr. Dickson addresses several mechanisms regulating these pathways in mammary cancer cells. 4) It has been known for some time that cross talk exists be— tween ER-mediated events and other signaling pathways (e.g., those regulated by peptide growth factors and their second messenger systems), and the ER itself undergoes phosphorylation/dephosphorylation events that could alter its activity. More recent studies have also identified a number of other nuclear factors, including coactivators, corepressors, and integrator proteins that play important roles in ER- mediated transcriptional events. A key observation is that these factors can alter the magnitude of cellular responses to estrogens and other steroids. Identification of these factors and the mechanisms by which they operate are likely to provide additional indices that can be used in conjunction with ER/PR levels to classify breast tumors and predict the efficacy of current hormonal therapies and to develop new therapeutic targets. A related advance has been the recogni- tion that substantial diversity is found in the location and sequence of EREs in endogenous hormone responsive genes, 3 v Journal of the National Cancer Institute Monographs No. 27, 1999 and these differences may also increase our understanding of mechanisms by which ER-mediated processes affect breast cancer. Dr. Greene’s talk discusses the interaction of several factors with the ER and illustrates that different ligands pro- duce different structural states of the receptor that could in- teract differentially with other regulatory molecules, such as coactivators and corepressors. 5) In addition to the classic ER, now referred to as ER-a, a second receptor termed ER-B has been identified in humans and in animals. The two receptors show different tissue dis- tributions, and, although they have generally similar ligand binding patterns, at least several differences appear to exist. An exciting era of endocrine research will be to define further the properties, distribution, and regulation of these receptors and to identify the biologic responses that they mediate. This new receptor is discussed by Dr. Gustafsson, whose labora— tory has been the leader in the identification and character- ization of ER—B. Cellular Components That Distinguish Between Agonist- and Antagonist-Activated Steroid Receptors Research in Dr. McDonnell’s laboratory has been driven in large part by two key issues in estrogen and antiestrogen phar- macology. First is the issue of how to obtain tissue selectivity with estrogens used for hormone replacement therapy. It is clearly established that estrogens diminish vasomotor instability (“hot flashes”), preserve bone mass, and have beneficial effects on cardiovascular health. An emerging view is also based on epidemiologic evidence that they may also benefit cognitive function and delay the onset of Alzheimer’s disease. However, it is highly problematic that estrogens used for these desirable purposes produce proliferative effects on the breast and endo- metrium. It is the fear of breast cancer, in particular, that greatly limits the use of estrogen replacement therapy by many women. The second pharmacologic issue is the use of tamoxifen as an adjuvant treatment of breast cancer. The drug has established efficacy in the treatment of the disease, but tamoxifen treatment generally fails after a period of time, and use for prolonged periods (e.g., 10 years) may actually be less beneficial than use for shorter times (e.g., 5 years). These observations were very difficult to reconcile with a simple mechanism of estrogen action in which “all estrogens are alike” in that they simply activate the receptor, and antiestrogens simply act by “freezing” the ER in an inactive state akin to that of an unliganded receptor protein. Roughly 5 years ago, a number of studies began appearng that were inconsistent with this simple View of ER activation. One study was a clinical paper published by Love et a1. (22) in 1992. These workers examined the effect of tamoxifen on bone mineral density of the lumbar spine in women receiving the drug for the treatment of breast cancer, and their data indicated that tamoxifen increased bone mass. In other words, the drug acts as an estrogen agonist in bone, in contrast to its antiestrogen action in the breast. This study was one of the first well-documented clinical studies of a SERM and clearly indicated that an ER ligand could have opposite effects in different target tissues. Shortly thereafter, studies in Dr. McDonnell’s own laboratory demonstrated that the binding of different ligands caused the ER to assume different conformations (23). In these studies, he used protease digestion to probe subtle differences in ER conforma— tion. When trypsin was incubated with the unliganded ER, the 66-kd molecular weight native receptor was degraded to very Journal of the National Cancer Institute Monographs No. 27, 1999 low-molecular-weight fragments. When estradiol was bound to the receptor, however, the receptor assumed a conformation less susceptible to protease digestion because a relatively large re- ceptor fragment (32 kd) remained after prolonged digestion. When tamoxifen was bound to the receptor, the protein was not degraded to very small fragments, indicating that tamoxifen did not simply hold the receptor in an inactive conformation similar to that of the unliganded protein. Rather, tamoxifen binding produced a conformational change that protected a relatively large protein fragment (28 kd) from trypsin digestion. This find- ing indicated that tamoxifen actually put the ER in a conforma- tion that was distinct from either that produced by the endog- enous hormone (estradiol) or that of the unliganded receptor, which was previously presumed to represent its inactive confor- mation. This study provided physical evidence that different ligands caused the receptor to assume different conformations. These laboratory investigations suggested a molecular expla- nation for clinical findings such as those reported by Love et a1. (22) in different tissues (i.e., bone versus breast cancer cells). Because tamoxifen and estradiol put the receptor into different conformations, this investigation suggested that the different tis— sues had ways to functionally “distinguish” structural difference in the receptor (i.e., the conformation of the ER—tamoxifen com- plex could function as an agonist in bone but not in breast cancer cells). It was known at this time that the ER had a modular structure and that two different regions of the protein could function to activate transcription. One such region, termed tran- scription-activating function 1 (referred to as either TAF-l or AF-l in the literature) was present in the N-terminal region of the receptor, and a second (AF-2) was present in its carboxyl- terminal region. This knowledge raised the possibility that dif- ferent ligands (e.g., estradiol versus tamoxifen) might put the receptor into conformations in which the two AFs were differ- entially active. To test this hypothesis, Dr. McDonnell’s group performed a series of co—transfection studies with the use of wild-type ERs that contained both AFs and ER mutants in which only one of the AFs was active (24). A series of such studies indicated that most cultured cells (approximately 90% of those tested) required both AF-l and AF—2 functions for transcriptional activity when stimulated by estradiol, but the hormone could stimulate transcription in some cells via receptors with only an active AF-l or an active AF—2 function. Tamoxifen failed to activate transcription in all cases in which both the AF-l and AF-2 functions were required and in cases in which the AF-2 function alone could mediate estradiol— induced transcription. In these cell types, tamoxifen functioned as a pure estrogen antagonist to block the action of estradiol. In contrast, in those cells in which estradiol could activate tran- scription from receptors with only a functional AF-l, tamoxifen could act as a partial agonist with substantial estrogen—like ac— tivity. These interactions are illustrated schematically in Fig. 2. These studies were also important because they established that tamoxifen could function both as a partial estrogen agonist and as a pure estrogen antagonist via the same receptor system. This finding ruled out the possibility that the antagonist and partial agonist activities of antiestrogens were mediated by different receptor systems. In the early 1990s, the concept was also emerging that the role of the receptor AF was to serve as “contact” points for the interaction with other cellular proteins involved in transcription control, the so—called coactivators and corepressors. The idea 137 ESTROGEN RECEPTOR DNA & Ligand AF-2 — COOH NH — AF-l Estradiol: $2.1? Tamoxifen: {if at: o Fig. 2. Agonists and antagonists have differential effects on transcription-activating functions (AF) of the estrogen receptor (ER). The ER has a modular structure involving a N-terminal transcription activation function, termed AF-l, a DNA-binding domain, a ligand-binding domain, and a more C-terminal activation function, termed AF-2. Both estradiol and tamoxifen bind to a similar site in the ligand—binding domain of the receptor. Estradiol is able to activate both AFs. In contrast, tamoxifen prevents activation of AF—2, although the drug can activate the AF—l function of the receptor. was that the steroid receptor would bind to EREs in the regula- tory regions of target genes and that AF functions (activated by bound hormone) would then recruit these factors, which would alter transcription. This idea raised the possibility that different cells required different AFs in the ER for transcription activation because they contained different complements of coactivators, corepressors, or other regulator proteins. Such differences could be qualitative (i.e., different cells would express different types of proteins) or quantitative (different cells would contain differ- ent levels of coactivators/corepressors), or both. To investigate this possibility, Dr. McDonnell’s laboratory performed a series of studies in which they co-transfected one such coactivator (termed GRIP) into cells along with either the wild-type or mutated ER (25). In the cell line used, estradiol produced a full response (100%) of transfected reporter genes with the wild-type receptor, but only a 50% response was pro- duced with receptors in which only AF—l was active and only 20% in receptors in which only AF-2 was active. However, when vectors expressing GRIP were used to raise cellular levels of this protein, receptors with only AF-l or AF-2 function pro- duced the same transcriptional response as wild—type receptors. This finding established the principle that coactivators present in some cells are sufficient to enable the ER to activate transcrip- tion when only one of its two AFs is activated by ligand binding. Collectively, studies such as these indicated that tamoxifen could function as a partial agonist if only the AF-l function of the receptor was required, but it always functioned as an antago- nist if the AF-2 function was required, either alone or in com- bination with the AF-l function. A related question that Dr. McDonnell also considered was whether all antiestrogens would display this same type of behavior, and he thus began a series of studies to investigate the ability of different antiestrogens to activate AF-l activity but block AF-2 activity. These studies led to the recent discovery of an antiestrogen, GW—5638, with an activity profile different from that of tamoxifen. GW-5638 is a triphenylethylene antiestrogen that appears to- tally devoid of either the AF-l or AF-2 type of activity (26). This lack of AF activity is not due to poor entry into target cells, because the drug can block the transcriptional effects of estradiol and tamoxifen in cultured cell systems. This drug is thus a pure antiestrogen in the breast, but it retains the ability to maintain bone mass without producing any uterotrophic action in rats. This ability indicates that ER ligands without AF-l or AF-2 138 activity can function as estrogen agonists in bone. This function suggests that ER-mediated actions in target cells are even more complex than previously recognized and that other factors be- sides AF—l and AF-2 functions are likely to be involved in estrogen actions in some cell types. Role of ER-a in Carcinogenesis With the Use of Transgenic Mouse Models The estrogen signaling system has long been implicated as a possible factor in the induction and/or promotion of carcinogen- esis, especially in the tissues of the female reproductive tract and of the breast. The proliferative effects of the natural ligand, 17B-estradiol, as well as the synthetic estrogen diethylstilbestrol (DES) in the uterus, vagina, and mammary gland have been well studied. The majority of the cellular effects of estrogens are thought to be mediated by the ER, now known to exist in two types, the well—characterized ER-(x and the newly discovered ER—B. Although it has been established that the ER must be present for most estrogen-induced mechanisms, the relationship between the levels of ER and the extent to which a tissue is estrogen responsive is less understood. Furthermore, the influ- ence of varied ER levels in carcinogenesis is even less well known. Efforts to understand further the role of the ER-a in carcinogenesis have led to the generation and characterization of a series of transgenic mouse models that possess altered levels of ER—a expression. The MT-mER mice possess a transgene that results in overexpression of the ER-a protein, whereas the ERaKO mice are homozygous for a targeted disruption of the ER-or gene and, therefore, possess no functional levels of ER—a (27). By using these models, studies have been conducted to elucidate further the role of ER-a in the induction and promotion of hormonally (DES)-induced tumors of the reproductive tract (28) and of oncogene-induced tumors of the mammary gland (29). In utero exposure to DES, a potent synthetic estrogen, has been linked to a significantly higher risk of a rare form of vagi- nal cancer, as well as other reproductive abnormalities in hu- mans. The effects of neonatal DES exposure in the female mouse include structural abnormalities in the uterus, oviduct, and bone; uterine tumors; and vaginal adenosis and adenocarci- noma, whereas, in the male, increased incidence of retained testes and hypoplasia of the accessory sex organs have been reported. However, the exact mechanisms by which develop- Journal of the National Cancer Institute Monographs No. 27, 1999 mental exposure to DES leads to such abnormalities remain unknown. DES is able to bind ER-a and to mimic the prolif— erative effects of the natural hormone 17B-estradiol in the uterus and the vagina. However, DES and its metabolites are also able to directly bind DNA and tubulin, reportedly increasing the in- cidence of aneuploidy and of nondisjunction in dividing cells. Therefore, it is possible that the developmental and carcinogenic effects of DES may be a direct result of its ER-or-mediated activity, its nonreceptor-mediated genotoxic effects, or both. The role of ER-or in the induction and promotion of DES- induced tumors was first investigated by using the transgenic MT-mER mice. The uteri of adult MT-mER mice possessed approximately 25% more ER-a than their wild-type littermates (29). It was hypothesized that, because of this abnormal expres- sion of ER-a, the reproductive tract tissues of the MT-mER mice may be more susceptible to tumors after neonatal exposure to DES. Wild-type and MT-mER littermates were treated with DES on days 1—5 at 2 mg/pup per day and then killed at 4, 8, 12, and 18 months of age. At 8 months of age, DES-treated MT- mER mice demonstrated a significantly higher incidence of uter- ine adenocarcinoma at 73% compared with 46% in the DES- treated wild-type mice (Table 1). These tumors were also preceded at 4 months by a significantly higher incidence of the preneoplastic lesion, atypical hyperplasia, in the MT-mER mice at 26% compared with 0% in the wild-type mice. These data indicate that the level of ER—or present in a tissue may be a determining factor in the progression of estrogen—responsive tu- mors. Further studies (30,31) designed to possibly segregate the estrogenic and genotoxic effects of DES have utilized the ERaKO mice, which possess no functional levels of the ER-a protein. Wild—type, heterozygous, and EROLKO littermates were treated as described above in the MT-mER study and also killed at 4, 8, 12, and 18 months. At all time points, uterine weight was significantly reduced in DES-treated wild-type and heterozy- gous females, whereas no difference was observed in the ERorKO females. Furthermore, the persistent cornification and hyperplasia of the vaginal epithelium as well as the progressive proliferative lesions of the oviduct that are characteristic of neo- natal exposure to DES were observed in the wild-type and het- erozygous mice but absent in the EROLKO females. At 4 months of age, squamous metaplasia was occasionally observed in the Table 1. Effect of increased ER—ot on the progression of DES-induced tumors in the mouse reproductive tract No. of affected mice/total No. of mice Control DES Age, Wild Observation mo type MT-mER Wild type MT-mER Squamous metaplasia 4 0/15 0/14 2/19 (11%) 12/19 (63%)* 8 0/11 0/10 6/24 (25%) 1/26 (4%) Atypical hyperplasia 4 0/ 15 0/14 0/ 19 (0%) 5/19 (26%)* 8 0/11 0/10 3/24 (12%) 1/26 (4%) Adenocarcinoma 4 0/ 15 0/ 14 0/ 19 (0%) 0/ 19 (0%) 8 0/11 0/10 11/24 (46%) 19/26 (73%)* 12 0/15 0/15 11/15 (73%) 13/15 (87%) 18 0/19 0/19 13/14 (93%) 12/13 (92%) *P<.05 as calculated by the Fisher exact probability test, comparing DES— treated MT—mER to DES—treated wild type. Adapted from (28). DES = diethyl- stilbestrol; ER = estrogen receptor. Journal of the National Cancer Institute Monographs No. 27, 1999 uteri of DES-treated wild—type females but not in the DES- treated EROLKO mice. In the males, significant atrophy of the seminal vesicle was observed at all time points in both DES- treated wild-type and heterozygous mice, whereas no difference was observed between control and DES-treated EROLKO males. The incidence of tumors, as well as possible altered gene ex- pression in reproductive tract tissues of the different genotype/ treatment groups, is currently being assessed. These results thus far indicate that certain developmental effects of DES are, in- deed, ER-OL mediated. Finally, the group of Couse and Korach has utilized the ERorKO to study the influence of ER-OL in mammary tumors induced by the ectopic expression of the Wnt-l oncogene. Mice possessing the MMTV—LTR-driven Wnt-l transgenic construct are known to develop hyperplastic ductal and alveolar epithe- lium and eventually mammary adenocarcinoma during adult- hood (32). Therefore, they have crossed Wnt-l transgenic mice with the ERaKO mice to generate mice that possess the Wnt—l transgene on a background of altered ER-a levels (33). The adult female ERaKO mammary gland is completely undeveloped, ex- hibiting only a rudimentary ductal structure and lacking any terminal end or alveolar buds. However, ectopic expression of the Wnt-l gene in the EROLKO mammary gland did result in hyperplasia of the existing ductal structure, but it did not lead to further ductal branching or the development of terminal end buds as exhibited by the Wnt-l/wild-type ER-a females. In ad- dition, the average time of tumor onset in the Wnt-l/EROLKO females was much delayed (50% at 48 weeks) compared with the Wnt-l/Wild—type ER-a littermates (50% at 24 weeks), even though the serum levels of estradiol in the ERorKO females are approximately 10-fold higher than normal. Postpubertal ovari— ectomy, as well as pregnancy, had no effect on the growth rate of the mammary tumors in the Wnt-l/wild-type ER-or females. However, prepubertal ovariectomy did result in a delayed aver— age time to tumor onset in the Wnt-l/wild-type ER—a as well as the Wnt-l/EROLKO females compared with that of their respec- tive intact study groups. The results of these studies indicated that Wnt—l—induced mammary tumors can arise in the absence of functional ER—or, as well as ovarian hormones. However, their results have demonstrated that estrogen actions, as mediated by the ER-or, do act to promote the growth of Wnt-l-induced tu- mors. Regulation of Cell Cycle and Cell Death in Mammary Cancer Physiologic levels of estrogens and progestins are well known to promote both onset and malignant progression of breast cancer. A number of investigators in the field believe that an imbalance of mammary epithelial proliferation and death (ap— optosis) contributes to tumor formation, genomic instability, and metastasis. They have hypothesized that imbalanced expression of steroid-regulated genes triggers this diverse cascade of pro— cesses (34,35). Both estrogenic and progestational steroids are known to regulate expression of genes encoding several poly- peptide growth factors, growth factor—binding proteins, and growth factor receptors. In the case of the epidermal growth factor (EGF) family of ligands and receptors (including trans- forming growth factor-(x [TGF-a], amphiregulin, EGF receptor, and c-ersz), their pathologic overexpression and functional rel- evance for breast cancer has received experimental support in vitro, in vivo, and in ongoing clinical studies. Conversely, 139 growth factors have been shown to regulate expression and func- tion of steroid receptors (34,35). Sex steroids and growth factors appear to exert their principal influences on the cell cycle and cell survival through regulation of cyclin D1 and Bcl-Z/BchL, respectively ( 35—3 7). The c—myc gene and the bcl-2 gene family have been shown to be important downstream mediators, both of the actions of steroids and of the EGF ligand/receptor family on cellular proliferation and survival; of particular interest, the c- myc gene is amplified in 20%—30% of breast cancer cases and aberrantly expressed in a much higher proportion of cases (38,39). A recent meta—analysis of published clinical pathologic studies (39) has demonstrated that amplification of the c-myc gene is associated with increased lymph node metastases and poorer survival, irrespective of expression of the ER. Myc ap- pears to exert its principal effect on the cell cycle through acti- vation of CDK-2; however, its overexpression sensitizes cells to apoptosis coincident with induction of the proapoptotic p53 and bax genes (38). As a model system to examine the consequences in viva of disregulated expression of two important, but functionally quite distinct, mediators of the action of estrogen, Dr. Dickson’s group has carried out a cross of transgenic Myc- and TGF-a- overexpressing mouse strains. Bitransgenic progeny exhibited a remarkable synergy between the two genes for mammary tu- morigenesis, independent of sex, parity, and reproductive hor- monal status, indicating that disregulated expression of these two estrogen-inducible genes can entirely supplant an etiologic role for estrogen in malignancy (40). They observed that the mechanism of interaction of Myc and TGFOL involved a coordi- nated stimulation of the cell cycle and suppression of apoptosis (Fig. 3) (41). First, the two gene products interacted such that TGF-a-induced BchL allowed cellular survival in the presence of Myc-induced p53 and Bax (36). The EGF receptor-mediated effect on survival appears to depend on signal transduction, both through the Erkl/Erk2 and the PISK pathways (42). An inde- pendent survival effect is also mediated in these cells by colla- gen IV acting through a B] integrin-PI3K mechanism (43). Sec- ond, through their concordant activation of CDK-4 (by induction of cyclin D1) and CDK-2 (by destruction of the CDK inhibitor p27), the two gene products markedly stimulated aberrant cell cycles and promoted the appearance of multiple chromosomal aberrations (44,45). The results of decreased modulation of p27 by c-myc appear to be sufficiently potent to allow abrogation of the anchorage-independent G1/S cell cycle checkpoint (46). The appearance of dicentric chromosomes and of a wide array of other chromosomal abnormalities was suggestive of a bridge— break—fusion cycle mechanism at work. The p53 gene was ob— served not to be mutated in Myc-expressing mammary tumors; it played no obvious role in surveillance of the chromosomal defects observed (47,48). Future studies must further address these mechanisms of cell cycle disregulation, apoptosis suppres- sion, genomic instability, their relevance to sex steroid action, and their roles in human breast cancer. ER Structure, Modulators, and Targets in Hormone-Responsive Tissues and Cancers Dr. Greene emphasized that the ER does not function in a vacuum but that it interacts with many other proteins. For ex- ample, it is well established that the ER and other steroid recep- tors interact with the heat-shock proteins (49) during the initial synthesis of the receptor to ensure its proper folding and traf- ficking, and, in turn, dissociation of heat-shock proteins seems to be required for the ligand-occupied receptor to activate tran- scription. He also emphasized that one of the major functions of ligand binding is to change the nature of protein—protein inter— actions between steroid receptors and other proteins and, con- versely, that other proteins can alter the state of the ER inde— pendent of ligand binding (e. g., by receptor phosphorylation). In Growth Inhibitors Apoptosis 4— G2 myc Overexpression Cell Cycle Progression Growth or Survival Factors G1 Accelerated Proliferation Fig. 3. Model for the dual action of c—myc overexpression in mammary epithelial cells. Deregulated c-myc expression promotes cell—cycle progression through a mechanism that is currently under investigation. The end result of such inap- propriate cell—cycle stimulation depends on a number of factors, such as cell genotype and environment. For example, in the presence of certain growth or survival factors (such as activators of the epidermal growth factor receptor or 140 integrin-mediated adhesion), c—myc expression is proposed to accelerate cell proliferation and promote cell survival. In the absence of such factors or in the presence of certain growth inhibitors (such as transforming growth factor—B), constitutive expression of c—myc is more likely to induce apoptosis [adapted from ( 38)]. Journal of the National Cancer Institute Monographs No. 27, 1999 his talk, Dr. Greene presented data from his laboratory on the following three related topics: 1) the identification of gene tar- gets for the ER-ligand complex, 2) the factors that serve to modulate the actions of the ER in target cells, and 3) the struc- tural changes produced in the ER by the binding of different estrogenic and antiestrogenic ligands. In the first series of studies, Dr. Greene used the technique of RNA differential display to identify transcripts in breast cancer cells that are regulated by estrogens and antiestrogens. This tech— nique identified a number of transcripts with expression altered by ER ligands. Sequence and northern blot analysis of one clone that was decreasingly regulated by estradiol revealed its identity as monocyte chemoattractant protein-1 (MCP-l). The basal ex- pression of MCP-l is low in MCF-7 breast cancer cells, but it is stimulated by TNF-a, and estradiol blocks induction of the MCP-l message by the cytokine. TNF-(x is known to regulate MCP-l expression via the NF- KB pathway. The MCP-l gene is known to contain an NF-KB regulatory element, and reporters containing this element are induced by TNF-a in MCF-7 cells. Despite the fact that report- ers containing the MCP-l promoter do not contain any se— quences resembling the classic ERE, estradiol blocks induc— tion of such reporters following transfection into breast cancer cells. Extracts from estrogen-treated MCF-7 cells also decrease the binding of NF-KB to its regulatory element in gel shift stud- ies, and the ER and NF-KB can be co-immunoprecipitated. Col- lectively, these studies suggest that the ER blocks TNF-a in- duction of MCP-l by directly or indirectly decreasing the binding of NF-KB to its regulatory site in the 5’-regulat0ry re- gion of the MCP—l gene. This mechanism is illustrated sche— matically in Fig. 4. TNF-a is known to act via a membrane receptor to stimulate a protein kinase cascade that leads to the phosphorylation of IKB (Fig. 4). Before phosphorylation, this protein forms a dimeric complex with NF—KB in the cytoplasm to prevent its movement to the nucleus. On phosphorylation, the IKB inhibitor dissociates from the complex and is degraded by an ubiquitin-mediated pathway. This process allows the NF-KB to translocate to the nucleus, where it binds to NF-KB sites in target genes and ac- tivates their transcription. The ER appears to decrease transcrip- tion by preventing the binding of NF- KB to its regulatory site in the 5’-flanking region of the gene, most likely by a direct inter- action of the two proteins, as illustrated in Fig. 4. This finding emphasizes that estrogens and antiestrogens can regulate expres— sion of target genes that do not contain hormone response ele- ments, and such regulation is thus likely to occur via protein— protein interactions. Another recent similar example of regulation by protein—protein interactions occurs via binding of the ER to AP-l components (50,51). A second series of studies was aimed at identifying cellular factors that modulate the activity of the ER. To identify such factors, Dr. Greene and his colleagues utilized the ligand- binding domain B (52) of the ER to “capture” proteins from breast cancer cells that interact with the receptor (53). In this approach, a fusion protein between glutathione S-transferase and the ligand-binding domain of the ER is used as an affinity matrix to bind proteins in cell extracts that bind to this domain of the receptor in the presence or absence of estrogens and/or anties- trogens. At present, this approach has already identified at least five to six proteins from MCF-7 cell extracts that bind to the ER. One such “modulator” protein, which has been identified, is a kinase that binds to the ER in the presence of estrogenic ligands and dissociates from the receptor when it is liganded with antiestrogens, such as hydroxy-tamoxifen. This action en- abled the kinase to be purified by first binding proteins in cell extracts to glutathione S-transferase-ER in the presence of es- tradiol, followed by washing to remove unwanted proteins, and then eluting in the presence of antiestrogens. Similar purification (3) E2 (Inactive) NF-KB MCP—l Gene '> Active Fig. 4. Proposed model for regulation of monocyte chemoattractant protein—1 (MCP-l) in human breast cancer cells. Tumor necrosis factor-a binds to a membrane receptor and initiates a kinase cascade (step 1) that phosphorylates the [KB protein shown as a filled rectangle (step 2). Before phosphorylation, IKB forms a dimer with the transcription factor NF-KB (filled oval) in the cytoplasm to prevent its entry to the nucleus. Following phosphorylation, IKB is degraded, Journal of the National Cancer Institute Monographs No. 27, 1999 and the free NF-KB can translocate to the nucleus (step 3). In the absence of estrogens, NF—KB is “active” as a transcription factor and drives transcription of the MCP—l gene via a regulatory element in the 5’-f1anking region of the gene. When occupied by estradiol, the estrogen receptor (open oval) forms a nuclear complex with NF—KB so that it is inactive as a transcription factor. 141 steps were performed by displacing the ER-bound kinase with a peptide with sequences similar to the motifs in coactivators that bind to steroid receptors. This kinase phosphorylates the ER on a serine residue, although the functional consequences of phos- phorylation at this site are unknown at present. This series of experiments also emphasizes that protein—protein interactions are increasingly being recognized as playing potential roles in estrogen action. In a third major series of studies, Dr. Greene and a number of colleagues examined the effects of ligand binding on the struc- ture of the ER. In this work, they solved the crystal structures of the human ER ligand—binding domain (amino acids 301—553) complexed with either estradiol or the mixed antagonist raloxi- fene (54). This study provided definitive evidence that different ligands produce distinct structural alterations in the receptor. The crystal structures revealed that both ligands bind to the same site within the core of the ligand-binding domain of the receptor, but the two ligands induce major conformational differences in the positioning of the most c-terminal a helix in the receptor (helix-12). This evidence is of major significance because helix— 12 is located in the transactivation domain of the ER and appears to be a major site for contact with coactivators and corepressors. Hence, the different structures observed following binding of the two ligands are expected to interact quite differently with these accessory proteins, which drive the transcriptional responses to steroid hormones. Structure and Function of ER-B A novel form of the ER, termed ER-B, was originally cloned from the rat prostate (55) and has also been identified in the mouse (56) and in humans (57). The rat ER-B complimentary DNA encodes a protein of 485 amino acids with a predicted molecular weight of 54 200 that is highly homologous to the ER-or, particularly in the DNA-binding domain (97% amino acid homology) and the c-terminal ligand-binding domain (59% ho- mology). The amino acid homologies between the three ER-Bs identified to date and the human ER-or are illustrated in Table 2 (58), and it is clearly seen that there is nearly perfect homology in the DNA—binding domains of the or and B receptors, substan— tial homology in the ligand-binding domains, but far less ho- mology in the N-terminal regions. The genes for the two recep— tors are on separate chromosomes in the human—ER—B on chromosome 14 and ER-or on chromosome 6—removing all doubts that the two receptors are totally distinct species. The major difference in the structures of the two receptors is in the N-terminal region, which is considerably shortened in the B-re- ceptor (58). In addition, another form of the ER-B is present in the rat that contains an in-frame insertion of 54 nucleotides Table 2. Homologies among various ERs, given as percentage of amino acid identity to the human ER-B* Domain Receptor NH-term DBD Hinge LBD F Overall Human B 100.0 100.0 100.0 100.0 100.0 100.0 Human or 17.5 97.0 30.0 59.0 17.9 47.0 Mouse B 80.6 98.5 84.4 91.9 78.6 88.0 Rat B 79.6 98.5 85.6 93.4 78.6 89.0 *ER = estrogen receptor; NH-term = amino terminal; DBD = DNA- binding domain; hinge = Hinge region; LBD = ligand-binding domain; F = F domain. 142 coding for an insertion of 18 amino acids with the ligand- binding domain. A major difference in the genes for the two ERs is that the or-gene is much larger (approximately threefold) than the B-gene, which has led to the speculation that the latter might be preferentially expressed at certain times in development when shorter genes are more rapidly transcribed, but this speculation remains to be established. Despite considerable differences in sequence in the ligand- binding domain, both ER-or and ER-B bind the endogenous hor- mone l7B-estradiol with about the same affinity. Of interest, however, the B-receptor seems to bind some androgens (e.g., 5-androstenedione) with reasonable affinity, leading to the speculation that this receptor might be activated by androgenic steroids in some situations. In contrast to binding of estradiol, the two receptors show differences in the binding of phytoes- trogens, such as genistein and coumestrol, with ER—B having substantially better affinity for these compounds than its Ot coun— terpart (59). It is also clear that both receptors can stimulate transcription from the consensus ERE and that phytoestrogens, as well as estradiol, can stimulate the transcriptional activity of both re- ceptors ( 59). In addition, ER—or and ER-B are able to form het- erodimers that have transcriptional activity when assayed with the traditional ERE. It now appears that both receptors may also stimulate transcription by other non-ERE mechanisms, such as protein—protein interactions with AP-l components. Of interest, classic ER-or agonists, such as estradiol and DES, function as antagonists in situations in which ER-B stimulates transcription via such AP—l-dependent mechanisms, whereas classic antago- nists, such as tamoxifen, act as agonists (51). In the prostate, ER-B appears to be under androgenic regu- lation, because its levels decrease with castration and can be restored by testosterone administration. Its expression in the adult prostate is highest in the epithelium and very low in the stroma. This pattern is developmentally regulated, however, be— cause ER-B is present at high levels in both the epithelial and mesenchymal layers of the tissue at birth but is then lost from the stroma in the adult. ER-B does not appear to be regulated by ER-or, because levels of the B receptor are similar in wild-type and ERorKO mice. Although estrogens do not seem to directly regulate ER-B expression, neonatal estrogen treatment appears to decrease the expression of this receptor in certain regions of the adult prostate (60). Dr. Gustafsson and his colleagues have also investigated the expression of ER-B in models of vascular injury because estro- gens appear to offer protection against atherosclerotic disease. By using a model of aortic lesions in mice, it is established that estrogens promote healing and that this effect occurs equally in ERorKO mice and wild-type animals (61). This finding suggests that ER—B may have an important role in the vascular response to estrogens. Of interest, ER—B expression (but not ER—or) is dramatically increased in both the endothelial cells and smooth muscle cells following vascular injury, again suggesting a po- tential protective role for estrogens acting via ER-B. In the female reproductive system, ER-B may play a promi- nent role in the ovary. During follicular development, the gran- ulosa cells express high amounts of this receptor, and its level seems to associate with mitotic activity, whereas little receptor is seen in the thecal cells. During the second half of the cycle, ER-B levels then decline. The B-receptor is also widespread through the urogenital tract, leading to speculation that it may Journal of the National Cancer Institute Monographs No. 27, 1999 mediate estrogen action on tissues such as the bladder. In this regard, there are many anecdotal reports that estrogen replace- ment has beneficial effects on urogenital atrophy and micturition in postmenopausal women even though the bladder does not appear to contain significant levels of ER-a, thus suggesting that the newly discovered ER—B may mediate these actions. Another major site of ER-B expression in female animals is the mammary epithelium of pregnant animals. This finding is particularly sig- nificant because these cells have previously been reported not to express ER-a, and it was thus thought that estrogen effects on these epithelial cells were mediated by stromal ERs. This recent finding now suggests that ER-B may directly mediate hormonal actions on the mammary epithelium. Because of increasing interest in the possible actions of es- trogens on cognitive function and Alzheimer’s disease, Dr. Gus- tafsson and his colleagues have compared the expression pat- terns of ER-or and ER-B in the central nervous system of developing and mature rodents. Expression of both receptors is widespread in the central nervous system, but differences are seen in the relative expression in different brain regions (62), suggesting that the two receptors may mediate different func- tions in the brain. One speculation is that the a-receptor may play a more prominent role in reproductive behaviors and the B-receptor might play an important role in certain aspects of cognitive function, but these roles remain to be established. Other sites in which the ER-B shows substantial levels of expression include the bone, kidney, lung, adrenal cortex, intes- tinal mucosa, lymph nodes, testis, sperm, thymus, spleen, and peripheral leukocytes. This expression raises the possibility that ER-B-selective agonists and antagonists might be able to pro- duce selective effects in such tissues. These selective effects might offer some distinct advantages (e. g., for hormone replace- ment therapy, when one wishes to minimize the hyperprolifera- tive actions of estrogens on the endometrium and breast). The possibility for producing selective estrogenic actions via this newly discovered receptor has thus prompted the search for such receptor selective agents. SUMMARY AND POTENTIAL IMPLICATIONS FOR HORMONAL CARCINOGENESIS A number of key points emerged from this session that are likely to have particular relevance for understanding the tran- scriptional actions of estrogens and antiestrogens in breast, pros- tate, and other cancers. 1) Studies with transgenic animals overexpressing the ER-OL have clearly shown that the level of expression of this protein can affect the rate of progression of several cancers. In ad— dition, it is now clear that the ER interacts with a large number of other proteins to regulate transcription. These pro- teins include the so-called coactivators and corepressors as well a variety of other regulatory proteins [for recent reviews, see (63—65)]. Thus, in addition to the levels of the classic ER-a and the newly discovered ER-B, the levels and activi- ties of these proteins may affect the etiology of hormone- dependent cancers, their growth responses to estrogenic sub- stances, and their response to hormonal therapies. 2) It is now clear that different estrogens and antiestrogens can have differential effects on the multiple activation functions of ERs and that these activation functions provide the sur- faces that interact with coactivators and corepressors to regu— Journal of the National Cancer Institute Monographs No. 27, 1999 late target gene expression. This finding has radically changed our thinking about the pharmacology of estrogens, and it now appears theoretically feasible to design highly selective estrogens with minimal growth—promoting effects on breast and other tumors (66,67). Conversely, this finding raises the possibility that certain estrogens might play a greater role in breast and prostate cancers than in others. 3) At the cellular level, it is now clear that we must consider hormone and antihormone effects on cell death, as well as cell proliferation. We must also understand how estrogens affect the interaction of cells with their environment (e.g., substratum, vascular system, etc.) as well as mechanisms by which estrogens regulate internal production of factors that regulate cell function at intracellular sites and understand the basis for the genomic instability commonly seen in cancer cells. 4) One area briefly mentioned by Dr. Greene was the possibility that antiestrogens may play a more “active” role in the treat- ment of breast cancer than the simple competitive blockade of estrogen actions at the receptor site. Thus, several reports are available that antiestrogens, acting through the ER or other mechanisms, may induce the synthesis of factors nor- mally suppressed by estrogens, or at least not expressed in the absence of antiestrogens. One area of great potential sig— nificance when considering estrogens as endogenous car- cinogens are the reports that antiestrogens may induce ex- pression of quinone reductase (68,69). This action would certainly be expected to decrease effects of reactive estro- genic metabolites, such as catechol estrogen quinones, or other genotoxic chemical species generated by redox cycling. Finally, it should be emphasized that, although there is in- creasing interest in understanding the role of estrogens as en— dogenous carcinogens (Chapters 3—5) that may have actions in- dependent of the classic ER, it is clear that ER-mediated processes play significant roles in normal and cancer cells. An increased knowledge of both types of processes is thus certain to enhance our understanding of the causes, treatments, and mecha- nisms to prevent many of the most prevalent human cancers. REFERENCES (1) Elledge RM, Fuqua SA. Estrogen and progesterone receptors. In: Harris JR, Lippman ME, Osborne CK, editors. Diseases of the breast. Philadelphia (PA): Lippincott, Williams & Wilkins; 2000. p. 471—88. (2) Mueller-Fahrnow A, Egner U. Ligand—binding domain of estrogen recep- tors. Curr Opin Biotechnol 1999;10:550—6. (3) Warner M, Nilsson S, Gustaffson IA. The estrogen receptor family. Curr Opin Obstet Gynecol 1999;11:249—54. (4) Gustafsson JA. Estrogen receptor B—a new dimension in estrogen mecha— nism of action. J Endocrinol 1999;163:379—83. (5) Hall JM, McDonnell DP. The estrogen receptor B—isoform (ERB) of the human estrogen receptor modulates EROL transcriptional activity and is a key regulator of the cellular response to estrogens and antiestrogens. En- docrinology 1999;140:5566—78. (6) Di Croce L, Okret S, Kersten S, Gustafsson IA, Parker M, Wahli W, et a1. Steroid and nuclear receptors. Villefranche-sur-Mer, France, May 25—27, 1999. EMBOJ 1999;18:6201-10. (7) Webb P, Nguyen P, Valentine C, Lopez GN, Kwok GR, McInemey E, et. al. The estrogen receptor enhances AP—l activity by two distinct mecha— nisms with different requirements for receptor transactivation functions. Mol Endocrinol 1999;13:1672—85. (8) Enmark E, Gustafsson JA. Oestrogen receptors—an overview. J Intern Med 1999;246:133—8. (9) Cowley SM, Parker MG. A comparison of transcriptional activation by ER or and ER [3. J Steroid Biochem Mol Biol 1999;69:165—75. 143 (10) McKenna NJ, Xu J, Nawaz Z, Tsai MJ, O’Malley BW. Nuclear receptor coactivators: multiple enzymes, multiple complexes, multiple functions. J Steroid Biochem Mol Biol 1999;69:3—12. (11) McKenna NJ, Lanz RB, O’Malley BW. Nuclear receptor coregulators: cellular and molecular biology. Endocr Rev 1999;20:321—44. (12) Xu L, Glass CK, Rosenfeld MG. Coactivator and corepressor complexes in nuclear receptor function. Curr Opin Genet Dev l999;9:l40—7. (13) Cosman F, Lindsay R. Selective estrogen receptor modulators: clinical spectrum. Endocr Rev 1999;20:418—34. (14) Mitlak BH, Cohen FJ. Selective estrogen receptor modulators: a look ahead. Drugs 1999;57:653—63. (15) Fink BE, Mortensen DS, Stauffer SR, Aron ZD, Katzenellenbogen JA. Novel structural templates for estrogen—receptor ligands and prospects for combinatorial synthesis of estrogens. Chem Biol 1999;6z205—19. (16) Sun J, Meyers MJ, Fink BE, Rajendran R, Katzenellenbogen JA, Katzenel— lenbogen BS. Novel ligands that function as selective estrogens or anties— trogens for estrogen receptor-o: or estrogen receptor-[3. Endocrinology 1999;140:800—4. (17) Kuiper GG, van den Bemd G], van Leeuwen JP. Estrogen receptor and the SERM concept. J Endocrinol Invest 1999;22:594—603. (18) Couse JF, Korach KS. Estrogen receptor null mice: what have we learned and where will they lead us? Endocr Rev 1999;20:358—417. (19) Curtis SH, Korach KS. Steroid receptor knockout models: phenotypes and responses illustrate interactions between receptor signaling pathways in vivo. Adv Pharmacol 2000;47:357—80. (20) Couse JF, Hewitt SC, Bunch DO, Sar M, Walker VR, Davis B], et al. Postnatal sex reversal of the ovaries in mice lacking estrogen receptors or and [3. Science 1999;286:2328—31. (21) Krege JH, Hodgin JB, Couse JF, Enmark E, Warner M, Mahler JF, et al. Generation and reproductive phenotypes of mice lacking estrogen receptor b. Proc Natl Acad Sci U S A 1998;95:15677—82. (22) Love RR, Mazess RB, Barden HS, Epstein S, Newcomb PA, Jordan VC, et al. Effects of tamoxifen on bone mineral density in postmenopausal women with breast cancer. N Engl J Med 1992;326:852-6. (23) McDonnell DP, Clemm DL, Hermann T, Goldman ME, Pike JW. Analysis of estrogen receptor function in vitro reveals three distinct classes of anti» estrogens. Mol Endocrinol l995;9:659—69. (24) Tzukerrnan MT, Esty A, Santiso—Mere D, Danielian P, Parker MG, Stein RB, et al. Human estrogen receptor transactivational capacity is determined by both cellular and promoter context and mediated by two functionally distinct intramolecular regions. Mol Endocrinol 1994:8121—30. (25 ) Norris JD, Fan D, Stallcup MR, McDonnell DP. Enhancement of estrogen receptor transcriptional activity by the coactivator GRIP-l highlights the role of activation function 2 in determining estrogen receptor pharmacol- ogy. J Biol Chem 1998;273:6679—88. (26) Willson TM, Norris JD, Wagner BL, Asplin I, Baer P, Brown HR, et al. Dissection of the molecular mechanism of action of GW5638, a novel estrogen receptor ligand, provides insights into the role of estrogen receptor in bone. Endocrinology 1997;138:3901—11. (27) Couse JF, Davis VL, Korach KS. Physiological findings from transgenic mouse models with altered levels of estrogen receptor expression. In: Pav» lik EJ, editor. Estrogens, progestins, and their antagonists. Vol 2. Boston (MA): Birkhauser; 1996. p. 69—98. (28) Couse JF, Davis VL, Hanson RB, Jefferson WN, McLachlan JA, Bullock BC, et al. Accelerated onset of uterine tumors in transgenic mice with aberrant expression of the estrogen receptor after neonatal exposure to diethylstilbestrol. Mol Carcinog 1997;19:236—42. (29) Davis VL, Couse JF, Goulding EH, Power SG, Eddy EM, Korach KS. Aberrant reproductive phenotypes evident in transgenic mice expressing the wild—type mouse estrogen receptor. Endocrinology 1994;135:379—86. (30) Lubahn DB, Moyer JS, Golding TS, Couse JF, Korach KS, Smithies O. Alteration of reproductive function but not prenatal sexual development after insertional disruption of the mouse estrogen receptor gene. Proc Natl Acad Sci U S A 1993;90:11162—6. (31) Couse JF, Curtis SW, Washbum TF, Lindzey J, Golding TS, Lubahn DB, et al. Analysis of transcription and estrogen insensitivity in the female mouse after targeted disruption of the estrogen receptor gene. M01 Endo— crinol 1995;921441—54. (32) Tsukamoto AS, Grosschedl R, Guzman RC, Parslow T, Varmus HE. Ex~ pression of the int-1 gene in transgenic mice is associated with mammary 144 gland hyperplasia and adenocarcinomas in male and female mice. Cell 1988;55:619—25. (33) Bocchinfuso WP, Korach KS. Mammary gland development and tumori— genesis in estrogen receptor knockout mice. J Mammary Gland Biol Neo— plasia 1997;22323—34. (34) Kenney NJ, Dickson RB. Growth factor and sex steroid interactions in breast cancer. J Mammary Gland Biol Neoplasia 1996;1z189—98. (35) Nass SJ, Rosfjord E, Dickson RB. Regulation of the cell cycle progression and cell death in breast cancer. Breast J 1997:3215—25. (36) Nass SJ, Li M, Amundadottir LT, Furth PA, Dickson RB. Role for Bcl-XL in the regulation of apoptosis by EGF and TGFBI in c-myc overexpressing mammary epithelial cells. Biochem Biophys Res Comm 1996; 2772248— 56. (37) Rosfjord EC, Dickson RB. Growth factors, apoptosis, and survival of mam- mary epithelial cells. J Mammary Gland Biol Neoplasia 1999; 4:229—37. (38) Nass SJ, Dickson RB. Defining a role for c—Myc in breast tumorigenesis. Breast Cancer Res Treat 1997;44:1—22. (39) Deming SL, Nass SJ, Dickson RB, Trock, BJ. c—myc amplification in breast cancer: a meta analysis of its frequency and associations with risk factors. Proc Am Assoc Cancer Res 1999;40:205. (40) Amundadottir LT, Johnson MD, Merlino GT, Smith GH, Dickson RB. Synergistic interaction of transforming growth factor 0L and c—myc in mouse mammary and salivary gland tumorigenesis. Cell Growth Differ l995;6: 737—48. (41) Amundadottir LT, Nass SJ, Berchem G], Johnson MD, Dickson RB. C0— operation of TGFOL and c-Myc in mouse mammary tumorigenesis: coordi— nated stimulation of growth and suppression of apoptosis. Oncogene 1996', 13:757—65. (42) Wang JK, Johnson MD, Rosfjord EC, Jamerson MH, Dickson RB. EGF- dependent survival signaling pathways in c—myc-overexpressing mouse mammary tumor cell lines: roles of Erkl/Erk2 and PI3K pathways. Proc Am Assoc Cancer Res 1999;40:164. (43) Rosfjord EC, Dickson RB. Adhesion to collagen IV via [3, integrins in- hibits apoptosis of c—myc-overexpressing breast cancer cells through a PI—3 kinase dependent mechanism. let Annual San Antonio Breast Cancer Symposium, San Antonio, TX, 1998. (44) Nass SJ Dickson RB. Epidermal growth factor-dependent cell cycle pro- gression is altered in mammary epithelial cells which overexpress c-myc. Clin Cancer Res 1998;4:1813—22. (45) Liyanage M, Coleman A, du Manoir S, Veldman T, McCormack S, Dick- son RB, et al. Multicolor spectral karyotyping on mouse chromosomes. Nat Genet 1996;14:312—5. (46) Benaud C, Dickson RB. Role of c—myc in bypassing an anchorage— dependent checkpoint in human mammary epithelial cells. Proceedings of the American Society of Cell Biology, 1998: San Francisco. (47) McCormack SJ, Weaver Z, Deming S, Natarajan G, Torri J, Johnson MD, et al. Myc/p53interactions in transgenic mouse mammary development, tumorigenesis and chromosomal instability. Oncogene 1998;16:2755—66. (48 ) Weaver Z, McCormack S, Liyanage M, du Manoir S, Coleman A, Dickson R, et al. Chromosomal aberrations in the mammary tumors of c-myc and Wnt—l transgenic mice detected by multicolor spectral karyotyping (SKY) and CGH. The Mouse Mammary Carcinogenesis Meeting, 1997; The J ack- son Laboratory, ME. (49) Landel CC, Potthoff SJ, Nardulli AM, Kushner PJ, Greene GL. Estrogen receptor accessory proteins augment receptor-DNA interaction and DNA bending. J Steroid Biochem Mol Biol 1997;63:59—73. (50) Webb P, Lopez GN, Uht RM, Kushner PJ. Tamoxifen activation of the estrogen receptor/AP-l pathway: potential origin for the cell—specific es— trogen-like effects of antiestrogens. Mol Endocrinol 1995;92443—56. (5]) Paech K, Webb P, Kuiper GG, Nilsson S, Gustafsson J, Kushner PJ, et al. Differential ligand activation of estrogen receptors ER): and ERB at AP-l sites. Science 1997;277:1508—10. (52) Seielstad DA, Carlson KE, Kushner PJ, Greene GL, Katzenellenbogen JA. Analysis of the structural core of the human estrogen receptor ligand bind— ing domain by selective proteolysis/mass spectrometric analysis. Biochem- istry 1995;34:12605—15. (53) Landel CC, Kushner PJ, Greene GL. Estrogen receptor accessory proteins: effects on receptor-DNA interactions. Environ Health Perspect l995;103(suppl 7):23—8. (54) Brzozowski AM, Pike AC, Dauter Z, Hubbard RE, Bonn T, Engstrom 0, Journal of the National Cancer Institute Monographs No. 27, 1999 et al. Molecular basis of agonism and antagonism in the oestrogen receptor. Nature 1997;389:753—8. (55) Kuiper GG, Enmark E, Pelto-Huikko M, Nilsson S, Gustafsson JA. Clon- ing of a novel estrogen receptor expressed in rat prostate and ovary. Proc Natl Acad Sci U S A 1996;93:5925—30. (56) Tremblay GB, Tremblay A, Copeland NG, Gilbert DJ, Jenkins NA, Labrie F, et a1. Cloning, chromosomal localization, and functional analysis of the murine estrogen receptor [3. Mol Endocrinol l997;11:353~65. (57) Mosselman S, Polman J, Dijkema R. ERB: identification and characteriza— tion of a novel human estrogen receptor. FEBS Lett 1996;392:49—53. (58) Enmark E, Pelto—Huikko M, Grandien K, Lagercrantz S, Fried G, Norden— skjold M, et al. Human estrogen receptor beta—gene structure, chromosomal localization, and expression pattern. J Clin Endocrinol Metab 1997;82: 4258—65. (59) Kuiper GG, Lemmen JG, Carlsson B, Corton JC, Safe SH, van der Saag PT, et al. Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor [3. Endocrinology 1998;139:4252—63. (60) Prins GS, Marmer M, Woodham C, Chang W, Kuiper G, Gustafsson JA, et a1. Estrogen receptor—B messenger ribonucleic acid ontogeny in the prostate of normal and neonatally estrogenized rats. Endocrinology 1998;139: 874—83. (61) Lindner V, Kim SK, Karas RH, Kuiper GG, Gustafsson JA, Mendelson ME. Increased expression of estrogen receptor-b mRNA in male blood vessels after vascular injury. Circ Res 1998;83:224—9. (62) Osterlund M, Kuiper GG, Gustafsson JA, Hurd YL. Differential distribu— tion and regulation of estrogen receptor-0t and —B mRNA within the female rat brain. Brain Res M01 Brain Res 1998;54:175~80. Journal of the National Cancer Institute Monographs N0. 27, 1999 (63) Shibata H, Spencer TE, Onate SA, Jenster G, Tsai SY, Tsai MJ, et al. Role of co-activators and co—repressors in the mechanism of steroid/thyroid re- ceptor action. Recent Prog Horm Res 1997;52:141—64. (64) Glass CK, Rose DW, Rosenfeld MG. Nuclear receptor coactivators. Curr Opin Cell Biol 1997;92222-32. (65) White R, Parker MG. Molecular mechanisms of steroid hormone action. Endocrine Related Cancer 1998;521—14. (66) Katzenellenbogen JA, O‘Malley BW, Katzenellenbogen BS. Tripartite ste- roid hormone receptor pharmacology: interaction with multiple effector sites as a basis for the cell— and promoter—specific action of these hormones. M01 Endocrinol 1996; 10: 1 19—3 1. (67) McDonnell DP, Norris JD. Analysis of the molecular pharmacology of estrogen receptor agonists and antagonists provides insights into the mechanisms of action of estrogen in bone. Osteoporosis Int 1997;7(supp1 1):SZ9—34. (68) Montano MM, Katzenellenbogen BS. The quinone reductase gene: a unique estrogen receptor-regulated gene that is activated by antiestrogens. Proc Natl Acad Sci U S A 1997;94:2581—6. (69) Montano MM, Jaiswal AK, Katzenellenbogen BS. Transcriptional regula— tion of the human quinone reductase gene by antiestrogen—liganded estro— gen receptor-(x and estrogen receptor-B. J Biol Chem 1998;273:25443—9. NOTE Supported by Public Health Service grants CA72460 (National Cancer Insti- tute) and HD08615 (National Institute of Child Health and Human Develop- ment), National Institutes of Health, Department of Health and Human Services. 145 Chapter 9: Factors Critical to the Design and Execution of Epidemiologic Studies and Description of an Innovative Technology to Follow the Progression From Normal to Cancer Tissue Montserrat Garcia-Closas, Susan E. Hankinson, Shuk-mei Ho, Donald C. Malins, Nayak L. Polissar, Stefan N. Schaefer, Yingzhong Su, Mark A. Vinson The results obtained from experimental studies of estrogen carcinogenesis need validation in epidemiologic studies. Such studies present additional challenges, however, because variations in human populations are much greater than those in experimental systems and in animal models. Because epidemiologic studies are often used to evaluate modest dif- ferences in risk factors, it is essential to minimize sources of errors and to maximize sensitivity, reproducibility, and specificity. In the first part of this chapter, critical factors in designing and executing epidemiologic studies, as well as the influence of sample collection, processing, and storage on data reliability, are discussed. One of the most important requirements is attaining sufficient statistical power to assess small genetic effects and to evaluate interactions between genetic and environmental factors. The second part of this chapter describes innovative technology, namely, Fourier transform-infrared (FT-IR) spectra of DNA that reveal ma- jor structural differences at various stages of the progression from normal to cancer tissue. The structural differences be- come evident from wavenumber-by-wavenumber statistical comparisons of the mean FT-IR spectra of DNA from nor- mal to cancer tissues. This analysis has allowed distinguish- ing benign tissues from cancer and metastatic tissues in hu- man breast, prostate, and ovarian cancers. This analysis, which requires less than 1 pg of DNA, is predicted to be used for detecting early cancer-related changes at the level of DNA, rather than at the cellular level. [J Natl Cancer Inst Monogr 2000;27:147—56] In the study of estrogen carcinogenesis, it has become appar- ent that results obtained from experimental studies need valida- tion in epidemiology/population studies. However, because variations in a human population are much greater than those existing in experimental systems and in animal models, popula- tion studies present additional challenges. In addition, because epidemiologic studies are frequently used to evaluate modest differences in risk factors and, therefore, in their design, it is essential to minimize sources of errors and technical variations and to choose methods with maximum sensitivity, reproducibil- ity, and specificity. In these regards, this chapter focuses on two important topics: One deals with technical issues in study design and in statistical power requirements, and the other focuses on the development of a new technology to measure a surrogate cancer risk marker. Several methodologic challenges and technical hurdles in de— signing and executing epidemiologic/population studies are dis— cussed in this chapter. Important issues that are discussed in- clude reproducibility of laboratory assays, limitations imposed Journal of the National Cancer Institute Monographs No. 27, 1999 by the small amount of plasma/serum collected, and the validity of using a single sample per subject. The chapter also discusses in detail the influences of sample collection, processing, and storage methods on data reliability. Finally, the importance of attaining adequate statistical power by reaching the required sample sizes is highlighted. Several important lessons are enumerated. 1) Collection pro- tocols need to minimize variations in factors that are not of etiologic interest by standardizing case and control subjects on these factors. 2) Sample collection, processing, and storage pro- cedures must be subjected to stringent scrutiny to ensure that variations in these steps will not mask the modest differences expected to exist in the risk factors of interest. 3) Study design must take into consideration the limitations linked to within- person variation over time as well as the single sample per subject collection method and, therefore, whenever possible, re— peated sampling should be considered. 4) Comparison of data collected from different laboratories may be difficult because large variations exist in different study populations and in labo- ratory methods. Introduction of a standardization or validation program should be considered for multisite analyses. 5) To attain statistical power in case—control studies, larger sample sizes are needed for studies that are assessing small genetic effects. Fur- thermore, if the goal of the study is to evaluate interaction among factors, sample sizes need to be increased accordingly. The second half of this chapter focuses on breakthrough tech— nology referred to as the Fourier transform-infrared/statistics model, which has been successfully adapted for analyses of DNA changes in cancer and precancerous tissues (1—6). Infrared spectra generated by applying infrared beams to sample DNA produced a large number of spectra. Fourier transform spectral data analyses, coupled with statistical comparisons, yield a few Afliliation of authors: M. Garcia-Closas, Environmental Epidemiology Branch, National Cancer Institute, Bethesda, MD; 5. E. Hankinson, Department of Epidemiology, Harvard School of Public Health and Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA; S. Ho, University of Massachusetts Medical School, De— partment of Surgery, Division of Urology, University Campus, Worcester; D. C. Malins, Molecular Epidemiology Program, Pacific Northwest Research Institute, Seattle, WA; N. L. Polissar, The Mountain—Whisper Light Statistical Consulting, Seattle, and Department of Biostatistics, University of Washington, Seattle; S. N. Schaefer, Y. Su, M. A. Vinson, Molecular Epidemiology Program, Pacific Northwest Research Institute, Seattle. Correspondence to: Shuk-mei Ho, Ph.D., University of Massachusetts Medi- cal School, Department of Surgery, Division of Urology, Office of Urologic and Translation Research, 55 Lake Ave., No., Worcester, MA (e-mail: Shuk- Mei.Ho@UMASSMED.EDU). See “Notes” following “References.” © Oxford University Press 147 principal components that were shown to be sufficient for dis- criminating DNA alterations in precancerous and cancer DNA samples. The technology has recently been developed further to produce cancer probability-risk score models for breast and prostate cancers. With the use of a model developed for breast cancer, normal tissue, primary tumors, and metastasizing tumors were correctly discriminated at more than 80% probability. The method also has the added advantage of requiring only a small amount of DNA (25%), and the ratio of between-person variation to labo- ratory error was often less than 2.0. Several other studies have also reported variability in assay reproducibility of both plasma (41,42) and urinary (43) steroid hormones, although results de- pended on the laboratory conducting the assay, the specific hor— mone, and the menopausal status of the woman. A number of factors may have influenced the variable repro~ ducibility observed. First, differences in laboratory methods may be important. For example, in our study, although all laborato- ries used radioimmunoassay (RIA) to measure estradiol, two laboratories used celite column chromatography and one labo- ratory used LH20 Sephadex column chromatography prior to RIA, whereas the fourth laboratory did not use a separation step prior to RIA. The laboratory method could not have been the only source of error, however, because results also varied within a single laboratory (e. g., the CV for estradiol ranged from 8% to 59%). These substantial differences might relate to a change either in the laboratory personnel or in the reagents and equip- ment used in the assays or perhaps varying levels of performance by the same technician or piece of equipment over time. In addition, some laboratories may be set up primarily to assay clinical specimens. The level of error tolerable in a clinical set— ting, in which the distinction between normal and abnormal hormone levels is of primary interest, is substantially greater than that which can be tolerated in epidemiologic research, in which relatively small differences within the spectrum of normal hormone levels are the subject of investigation. Another technical challenge in studies of hormones is that a number of different laboratory methods are used to measure the same hormone, and no standardization or validation programs exist. For example, in several studies in which plasma estradiol was measured in postmenopausal women, mean levels were 9 pg/mL (44), 13 pg/mL (45), and 28 pg/mL (46). To what degree these differences represent different study populations or simply differences in laboratory methods is unclear and complicates any comparison of results between the studies. The comparison of different laboratory methods against a “gold standard” would be helpful in resolving this issue; however, it is unclear which analytic method would be most appropriate as the gold standard. Summary and General Recommendations On the basis of our current knowledge, several recommenda- tions can be made to epidemiologists wanting to use hormone measurements in their research. Close collaboration with labo- ratory experts should be obtained in the planning stages of a study and should continue through its conclusion. Any variation from the standard collection and processing procedures should be evaluated prior to their implementation. Before having any study blood samples analyzed, laboratory performance should be independently evaluated. After this initial assessment, a propor- tion of samples sent to the laboratory with each batch of study samples should be quality—control specimens that are indistin- guishable from the case and control specimens. Matched case— control pairs should be handled identically and together, shipped 152 in the same batch, and assayed in the same analytical run. All assays should be conducted without knowledge of the case/ control status. Identical handling of all case and control speci— mens is critical to validity, as any possible deterioration related to collection, processing, or storage should affect case and con- trol specimens equally and will not appreciably affect measures of association. Finally, collection of repeated blood or urine samples from a subset of study subjects should be considered; this collection will allow both the evaluation of within-person variability over time and the use of measurement error correc- tion techniques in the calculation of relative risks. FOURIER TRANSFORM-INFRARED/STATISTICS MODELS Fourier transform-infrared (FT—IR) spectra of DNA have re- vealed major structural differences at various stages in the pro- gression of morphologically normal estrogen-responsive tissues (ERT) to cancer (1—5). Reactions of the hydroxyl radical (-OH) with the base (1—6,47—50) and deoxyribose (1—5) structures have been implicated as major contributors to these modifications, although other factors, to include hypermethylation (51) and the formation of depurinating adducts (52), may modify DNA spec- tra. In ERT (e.g., the human breast), the 'OH is believed to arise from the metal-catalyzed decomposition of H202, which is pro- duced from redox cycling of catechol estrogen metabolites (48) and certain xenobiotics (e.g., aromatic hydrocarbons) (53). The structural differences are evident from wavenumber-by- wavenumber statistical comparisons of the mean FT-IR spectra of DNA (extracted with pheonol) (I) from normal and trans— formed tissues (e.g., normal prostate versus prostate cancer) (5). Principal component analysis (PCA) (4) allows most of the in- formation in each spectrum to be represented by a few principal components (PCS), the first three usually accounting for more than 80% of the total variance. Each PC score is a weighted sum of spectral absorbances. Plots can be constructed on the basis of the first two or three PCs. In these plots, a point represents a single spectrum, and groups (clusters) of points represent the DNA from a particular tissue type (e.g., prostate cancer). In the carcinogenic transformation of one tissue type to another (e.g., normal —> cancer), the location of the cluster and its diversity in PC space are important measures of DNA change (49). When spectral differences exist between the DNA of tissue groups in a disease progression (e.g., normal tissue —> cancer), discriminant analysis can be used to establish cancer prediction models, such as those reported for breast (1,4) and prostate (4,5) cancers. Prototype prediction models, based on multivariate analysis of infrared spectral data, have been developed, and they have an ability to potentially differentiate between tissue groups that were not satisfactorily differentiated by simpler statistical mod- els. These models can potentially distinguish nonmetastatic pri- mary tumors from those with disseminated metastases. Ex- amples of FT—IR/statistics models for predicting cancer-related changes in DNA prior to evidence for cellular transformations are presented, together with discussion of their clinical and etio- logic implications. Significant differences were found between the mean absor- bances of DNA from the morphologically normal ovary (On) and ovarian adenocarcinoma (AC) over most of the spectral region (Fig. 3, A) (54). The P values are presented for each wavenum- ber (Fig. 3, B). Statistically significant differences (from about 1650 cm—1 to 1680 cm], 1200 cm—1 to 1260 cm“, and 1000 cm” to 1150 cm“) are evident in spectral areas assigned to Journal of the National Cancer Institute Monographs No. 27, 1999 BE 20: 2- we E0 ‘53 1‘ Z< . 0 . . . . . . . . . . 1700 1600 1500 1400 1300 1200 1100 1000 900 800 (D 3 005-»- (U > 0.017 o. 0.001- 1700 1600 1500 1400 1300 1200 1100 1000 900 800 Wavenumber (cm") 10 1.0 C. ,_ D. Q) 5- g 0.8‘ DO (U A‘ O m 0- A (£86? 3 0.6- 0 z “- -5 . A E 0.4- .0 2 AAC 5 AAC -15 . . 0.0 . 1 -15 -10 -5 0 5 10 -5 o 5 10 PC 2 Risk Score Fig. 3. Spectral comparisons of ovarian DNA (53). A) Grand mean DNA spectra of morphologically normal ovarian tissue (0“; n = 13) and primary ovarian adenocarcinoma (AC; n = 6); B) P values for spectral comparison in A (unequal variance t test) (53); C) principal component (PC) plot comparing spectra of ovarian DNA from morphologically normal tissues (On) and primary adenocar— cinoma (AC) (53); and D) plot of the probability of ovarian cancer with the risk score for the On and AC. The null hypothesis that the PC scores do not dis- criminate between the groups is rejected with P<.001. Reprinted with permission © 1998 from the National Academy of Sciences, U.S.A. vibrations of the nucleotide bases, the PO; group and deoxy— ribose, respectively (54). PCA of the spectral data provided two major PCs that were plotted against each other (Fig. 3, C). The plot revealed that the On formed a tight, ordered group of points, whereas the ACs were highly diverse and relatively disordered. The relationship between the probability of ovarian cancer and the risk score derived by discriminant analysis is shown in Fig 3, D. The ovarian cancer group is located primarily at the top of the sigmoid-like curve, and the noncancer group is located at the bottom. The predicted probability scores rise rapidly over a nar- row range, which reflects a high degree of discrimination be- tween the groups. The disorder reflected in the AC and the metastasized primary ovarian adenocarcinomas (ACm) contrasts with the order in the On and distant ovarian metastases to the colon (ACdm), as apparent from the mean spectral comparisons (Fig. 4, A) and the PC plot in which the points of each group substantially overlap (Fig. 4, B) (54). Despite the inability to discriminate between the two ordered DNA systems with the use of spectral comparisons and PCA, comparisons of standard de- viations of absorbances at each wavenumber over the entire spectral range revealed increased spectral diversity in the ACdm in regions assigned to base vibrations but not in those relating to the furanose ring. This finding is consistent with the presence of increased base mutations in the DNA of the distant metastases ( 54 ). Comparisons of the mean spectra of DNA from morphologi- cally normal breast tissues obtained from breast reduction sur— gery (reduction mammoplasty tissues, RMT) and invasive ductal carcinoma (IDC) tissues revealed characteristic differences in spectral regions assigned to the base and deoxyribose structures (1—4). A three-dimensional plot of the points from PCA is given in Fig. 5, A. The points representing the RMT are clustered primarily in the upper-left region of the plot. The IDC, compris— ing primary tumors with and without evidence for axillary me- tastases, are broadly dispersed, thus indicating considerable structural diversity. Discriminant analysis of the spectral data provided a relationship between the probability of cancer and the risk score (Fig. 5, B), having a sensitivity of 86% and a speci- ficity of 81%. The DNA of “benign” (microscopically normal) tissues from near the breast tumors of 11 women (not included CD ’5’» a a m “5 E E a O z 2 . 1700 1600 1500 1400 1st 1200 1100 1000 900 860 Wavenumber (cm'l) 10 B 5 - I O . 0 (V) o. O ‘b D. _5_ . . ° On 10 . ACdm -15 - - . . -15 -10 -5 O 5 10 PC 2 Fig. 4. Spectral comparisons of ovarian DNA ( 53 ). A) Mean spectra of DNA from the morphologically normal ovary (0“; n = 13) compared with mean spectra of DNA from ovarian adenocarcinoma metastases to the colon (ACdm; n = 7); B) principal components plot comparing the mean DNA spectra from each group shown in A. Reprinted with permission © 1998, National Academy of Sciences, USA. Journal of the National Cancer Institute Monographs No. 27, 1999 153 A. C. 15 I 10 A A I N 5 0' v ‘ A ELL) A A A‘ 0 a o ‘ A ‘l . '7‘ A A A ' A ' 0 Normal prostate '5 ' A BPH v Prostate cancer I Outlier -10 , . . -10 -5 0 5 10 15 20 PC 1 B. 1.0 a 0.8 - o c (U 2 0.6- o 3‘ E 0.4 (U .o 2 0- 0.2 - 0 Normal Breast U Breast Cancer 00- I . . I I -10 -5 0 5 10 15 20 Risk Score D- 1.0 - 3 Points 5 0.8 - o : G! 2 0.6- o 3* 5 0.4- m .o E a. 0.2 - _ 6 Pornts 0 Normal prostate -10 -5 0 5 10 Risk Score Fig. 5. Spectral comparisons of breast DNA (4). A) Three-dimensional plot of principal component (PC) scores of DNA spectra from normal breast (n = 21) and breast cancer (invasive ductal carcinoma [IDC]; n = 37) tissues showing distinct clustering of each group (1,3); B) plot of the probability of cancer with the risk score for the normal breast and breast cancer. The null hypothesis that the PC scores do not discriminate between the groups is rejected with P<.0001 (I); C) two-dimensional plot of PC scores of DNA spectra from normal prostate in the predictive model) was analyzed ( 4 ). When the scores were used in the breast cancer probability-risk score model (Fig. 5, B), 10 of 11 had a predicted cancer probability of more than 75%. This finding is consistent with data showing that tissue near a breast tumor has a high risk for forming a second cancer lesion (55). In studies of the human prostate, the mean spectral differ— ences between the DNA of normal tissue and the DNA of pros- tatic adenocarcinoma were substantial. The PC plot revealed pronounced discrimination between DNA spectra of normal and cancer tissues (4,5). A similarly effective separation was ob- tained between the clusters of DNA points representing normal tissue, prostatic cancer, and benign prostatic hyperplasia (BPH) (Fig. 5, C). The discriminant analysis models that predict disease probability (normal prostate tissue versus prostate cancer; nor- mal prostate tissue versus BPH) had sensitivities and specifici- ties of 100% for both comparisons. These models are based on more PC scores than the two—dimensional PC plots. The rela- tionship between the normal prostate DNA and the DNA of prostatic adenocarcinoma, expressed in terms of cancer prob- abilities, is shown in Fig. 5, D. Prototype statistical models, based on FT—IR spectroscopy, are being tested in our laboratory. These models hold promise for distinguishing the DNA from primary tumors and metasta- sizing primary tumors (those that have given rise to dissemi- 154 (n = 8), benign prostatic hyperplasia (BPH; n = 18) and prostate cancer (adenocarcinoma; n = 8) in which the clustering is distinct (5); D) plot of the probability of cancer versus the risk score for normal prostate and prostate cancer. The null hypothesis that the PC scores do not discriminate between the groups is rejected with P = .009 (5). Reprinted with permission by Nature Medicine. Portions A and B originally appear in Nat Med 1997;32927—30. Fig. 6. Three-dimensional ellipsoids based on a multivariate model for principal component (PC) scores of DNA spectra of invasive ductal carcinoma (IDC) and metastasized IDC (those that give rise to disseminated metastases; IDCm). The ellipsoids contain an expected 90% of the population of each group (e.g., IDC samples). (See text for details.) Journal of the National Cancer Institute Monographs No. 27, 1999 nated metastases). The FT-IR/statistics models based on simple linear logistic regression, such as those shown in Fig. 5, B, did not effectively differentiate these groups. By use of models based on multivariate normal distributions of the first three PCs, a three-dimensional projection (Fig. 6) was constructed to con- tain a designated percentage (i.e., 90%) of the population of a group. In a model with 90% probability, such as that shown in Fig. 6, a randomly selected IDCm spectrum would likely fall inside the appropriate three-dimensional figure (i.e., only an ex- pected 10% of DNA spectra in the population of IDCm spectra would fall outside the model). By use of this DNA model, nor— mal breast tissue (RMT), primary breast tumors (IDC), and me- tastasizing primary breast tumors (IDCm) were correctly classi- fied as follows: 89% (16 of 18), 97% (31 of 32), and 82% (18 of 22), respectively. The discrimination between the IDC and the IDCm is a potentially important basis for identifying metastasis in primary tumors, prior to evidence for malignant cells at dis- tant sites. The prototype model (Fig. 6), which is presently based on a limited number of samples, can be applied to other systems having larger databases. The FT-IR/statistics models have the ability to identify subtle changes in DNA in relation to the progression of normal tissues to diseased states. We are unaware of other techniques with the power to accomplish such a high degree of discrimination be- tween DNA of natural systems. It is now possible to analyze less than 1.0 ug of DNA with the use of FT-IR spectral techniques recently developed in our laboratory. This will eventually allow the FT -IR/statistics technology to be applied to less than 1.0 mg of tissue, thus broadening the application to small biologic samples (e.g., fine-needle biopsy tissues). Future uses of the technology would be expected to encompass diverse areas of cancer research and clinical practice, as previously described (4). For example, with the use of the FT-IR/statistics technology, the potential exists for detecting early cancer-related changes at the level of DNA, rather than at the cellular level, thereby af— fording a distinct advantage in patient treatment. REFERENCES (I) Malins DC, Polissar NL, Nishikida K, Holmes EH, Gardner HS, Gunsel— man SJ. The etiology and prediction of breast cancer: Fourier transform— infrared spectroscopy reveals progressive alterations in breast DNA leading to a cancer-like phenotype in a high proportion of normal women. Cancer 1995;75:503—17. Malins DC, Polissar NL, Gunselman SJ. Tumor progression to the meta— static state involves structural modifications in DNA markedly different from those associated with primary tumor formation. Proc Natl Acad Sci U S A 1996;93:14047—52. (3) Malins DC, Polissar NL, Gunselman SJ. Progression of human breast can— cers to the metastatic state is linked to hydroxyl radical-induced DNA damage. Proc Natl Acad Sci U S A 1996;93:2557—63. (4) Malins DC, Polissar NL, Su Y, Gardner HS, Gunselman SJ. A new struc- tural analysis of DNA using statistical models of infrared spectra. Nat Med 1997;3z927—30. (5 ) Malins DC, Polissar NL, Gunselman SJ . Models of DNA structure achieve almost perfect discrimination between normal prostate, benign prostatic hyperplasia (BPH), and adenocarcinoma and have a high potential for predicting BPH and prostate cancer. Proc Natl Acad Sci U S A 1997;94: 259—64. Malins DC, Holmes EH, Polissar NL, Gunselman SJ. The etiology of breast cancer. Characteristic alterations in hydroxyl radical-induced DNA base lesions during oncogenesis with potential for evaluating incidence risk. Cancer 1993;71:3036—43. ( 7) Feigelson HS, Henderson BE. Estrogens and breast cancer. Carcinogenesis 1996;17:2279—84. (2 \7 (6 \7 Journal of the National Cancer Institute Monographs No. 27, 1999 (8) Khoury MJ, Beaty TH. Applications of the case—control method in genetic epidemiology. Epidemiol Rev 1994;16:134—50. (9) Hwang SJ, Beaty T, Liang KY, Coresh J, Khoury MJ. Minimum sample size estimation to detect gene—environment interaction in case—control de— signs. Am J Epidemiol 1994;140:1029—37. (10) Goldstein AM, Falk RT, Korczak JF, Lubin JH. Detecting gene—environ— ment interactions using a case-control design. Genet Epidemiol 1997;14: 1085—9. (11) Foppa I, Spiegelman D. Power and sample size calculations for case-con- trol studies of gene—environment interactions with a polytomous exposure variable. Am J Epidemiol 1997;146:596—604. (12) Garcia-Closas M, Lubin JH. Power and sample size calculations in case— control studies of gene—environment interactions: comments on different approaches. Am J Epidemiol 1999;149:689—92. (13) Rothman N, Garcia-Closas M, Stewart WT, Lubin JH. The impact of misclassification in case—control studies of gene—environment interactions. In: Vineis P, Malats N, Lang M, d’Errico N, Caporaso N, Cuzick J, et al., editors. Metabolic polymorphisms and susceptibility to cancer. Lyon (France): IARC; 1999. p. 89—96. (14) Garcia-Closas M, Rothman N, Lubin J. Misclassification in case—control studies of gene—environment interactions: assessment of bias and sample size. Cancer Epidemiol Biomarkers Prev 1999;8z1043—50. (15) Garcia-Closas M, Thompson WD, Robins JM. Differential misclassifica— tion and the assessment of gene—environment interactions in case—control studies. Am J Epidemiol 1998;147:426—33. (16) Khoury MJ, Flanders WD. Nontraditional epidemiologic approaches in the analysis of gene—environment interaction: case—control studies with no controls! Am J Epidemiol 1996;144:207—13. (17) Whittemore AS, Nelson LM. Study design considerations in genetic epi- demiology: theoretical and practical considerations. J Natl Cancer Inst Monogr 1999;26:61—9. (I8) Yen SS, J affe RB, editors. Reproductive endocrinology: physiology, patho- physiology, and clinical management. Philadelphia (PA): WB Saunders; 1986. (I9) Lenton EA, Landgren BM, Sexton L, Harper R. Normal variation in the length of the follicular phase of the menstrual cycle: effect of chronological age. BrJ Obstet Gynaecol 1984;91:681—4. (20) Hankinson SE, London SJ, Chute CG, Barbieri RL, Jones LA, Kaplan LA, et al. Effect of transport conditions on the stability of biochemical markers in blood. Clin Chem 1989;35:2313—6. (21) Hankinson SE, Willett WC, Colditz GA, Hunter DJ, Michaud DS, Deroo B, et al. Circulating concentrations of insulin-like growth factor-I and risk of breast cancer. Lancet 1998;351:1393—6. ( 22 ) Su SC, Garbers S, Rieper TD, Toniolo P. Temperature variations in upright mechanical freezers. Cancer Epidemiol Biomarkers Prev 1996;5z139—40. (23) Cauley JA, Gutai JP, Kuller LH, Dai WS. Usefulness of sex steroid hor— mone levels in predicting coronary artery disease in men. Am J Cardiol 1987;60:771—7. (24) Koenig KL, Toniolo P, Bruning PF, Bonfrer JM, Shore RE, Pastemack BD. Reliability of serum prolactin measurements in women. Cancer Epidemiol Biomarkers Prev 1993;21411—4. (25) Toniolo P, Koenig KL, Pastemack BS, Banerjee S, Rosenberg C, Shore RE, et a1. Reliability of measurements of total, protein-bound, and unbound estradiol in serum. Cancer Epidemiol Biomarkers Prev 1994;3z47—50. (26) Bolelli G, Muti P, Micheli A, Sciajno R, Franceschetti F, Krogh V, et al. Validin for epidemiological studies of long-term cryoconservation of ste— roid and protein hormones in serum and plasma. Cancer Epidemiol Bio- markers Prev 1995;42509—13. (27) Thomas HV, Key TJ, Allen DS, Moore JW, Dowsett M, Fentiman IS, et al. A prospective study of endogenous serum hormone concentrations and breast cancer risk in premenopausal women on the island of Guernsey. Br J Cancer 1997;75:1075—9. (28) Orentreich N, Brind JL, Rizer RL, Vogelman JH. Age changes and sex differences in serum dehydroepiandrosterone sulfate concentrations throughout adulthood. J Clin Endocrinol Metab 1984;59:551—5. (29) Siiteri PK, Simberg N, Murai J. Estrogens and breast cancer. Ann N Y Acad Sci 1986;464:100—5. (30) Langley MS, Hammond GL, Bardsley A, Sellwood RA, Anderson DC. Serum steroid binding proteins and the bioavailability of estradiol in rela- tion to breast diseases. J Natl Cancer Inst 1985;75:823—9. (3]) Cauley JA, Gutai JP, Kuller LH, Powell JG. Reliability and interrelations 155 among serum sex hormones in postmenopausal women. Am J Epidemiol 1991;133:50—7. (32) Hankinson SE, Manson JE, Spiegelman D, Willett WC, Longcope C, Speizer FE. Reproducibility of plasma hormone levels in postmenopausal women over a 2— to 3—year period. Cancer Epidemiol Biomarkers Prev 1995;42649—54. (33) Goodman-Gruen D, Barrett-Connor E. Epidemiology of insulin—like growth factor-I in elderly men and women. The Rancho Bernardo Study. Am J Epidemiol 1997;145:180—6. (34) Micheli A, Muti P, Pisani P, Secreto G, Recchione C, Totis A, et al. Repeated serum and urinary androgen measurements in premenopausal and postmenopausal women. J Clin Epidemiol 1991;44:1055—61. (35) Muti P, Trevisan M, Micheli A, Krogh V, Bolelli G, Sciajno R, et a1. Reliability of serum hormones in premenopausal and postmenopausal women over a one-year period. Cancer Epidemiol Biomarkers Prev 1996; 52917—22 (36) Pasagian—Macaulay A, Meilahn EN, Bradlow HL, Sepkovic DW, Buhari AM, Simkin—Silvennan L, et a1. Urinary markers of estrogen metabolism 2— and 160L-hydroxylation in premenopausal women. Steroids 1996;61:461—7. (37) Michaud DS, Manson JE, Spiegelman D, Barbieri RL, Sepkovic DW, Bradlow HL, et al. Reproducibility of plasma and urinary sex hormone levels in premenopausal women over a one—year period. Cancer Epidemiol Biomarkers Prev 1999;8z 1059—64. (38) Willett WC. Nutritional epidemiology. New York (NY): Oxford University Press; 1990. (39) Rosner B, Spiegelman D, Willett WC. Correction of logistic regression relative risk estimates and confidence intervals for random within-person measurement error. Am J Epidemiol 1992;136:1400—13. (40) Hankinson SE, Manson JE, London SJ, Willett WC, Speizer FE. Labora- tory reproducibility of endogenous hormone levels in postmenopausal women. Cancer Epidemiol Biomarkers Prev 1994:3251—6. (41) Potischman N, Falk RT, Laiming VA, Siiteri PK, Hoover RN. Reproduc- ibility of laboratory assays for steroid hormones and sex hormone-binding globulin. Cancer Res 1994;54:5363—7. (42) Gail MH, Fears TR, Hoover RN, Chandler DW, Donaldson JL, Hyer MB, et a1. Reproducibility studies and interlaboratory concordance for assays of serum hormone levels: estrone, estradiol, estrone sulfate and progesterone. Cancer Epidemiol Biomarkers Prev 1996;52835—44. (43) Ziegler RG, Rossi SC, Fears TR, Bradlow HL, Adlercreutz H, Sepkovic D, et a1. Quantifying estrogen metabolism: an evaluation of the reproducibility and validity of enzyme immunoassays for 2—hydroxyestrone and 16a— hydroxyestrone in urine. Environ Health Perspect 1997;105(suppl 3): 607—14. (44) Hankinson SE, Willett WC, Manson JE, Colditz GA, Hunter DJ , Spiegel- 156 man D, et al. Plasma sex steroid hormone levels and risk of breast cancer in postmenopausal women. J Natl Cancer Inst 1998;90:1292—9. (45) Dorgan JF, Longcope C, Stephenson HE Jr, Falk RT, Miller R, Franz C, et al. Relation of prediagnostic serum estrogen and androgen levels to breast cancer risk. Cancer Epidemiol Biomarkers Prev 1996;52533—9. (46) Toniolo PG, Levitz M, Zeleniuch—Jacquotte A, Banerjee S, Koenig KL, Shore RE, et al. A prospective study of endogenous estrogens and breast cancer in postmenopausal women. J Natl Cancer Inst 1995;87:190—7. (47) Von Sonntag C, Hagen U, Schon-Bopp A, Schulte-Frohlinde D. Radiation- induced strand breaks in DNA: chemical and enzymatic analysis of end groups and mechanistic aspects. Adv Radiat Biol 1981;9:109—42. (48) Liehr JG. Horrnone—associated cancer: mechanistic similarities between human breast cancer and estrogen—induced kidney carcinogenesis in ham- sters. Environ Health Perspect 1997;105(suppl 3):565—9. (49) Yager JG, Liehr JG. Molecular mechanisms of estrogen carcinogenesis. Annu Rev Pharrnacol Toxicol 1996;36:203—32. (50) Liehr JG. Dual role of oestrogens as hormones and pro-carcinogens: tu- mour initiation by metabolic activation of oestrogens. Eur J Cancer Prev 1997;6z3—10. (51) Weitzman SA, Turk PW, Milkowski DH, Kozlowski K. Free radical ad- ducts induce alterations in DNA cytosine methylation. Proc Natl Acad Sci U S A 1994;91:1261—4. (52) Cavalieri EL, Stack DE, Devanesan PD, Todorovic R, Dwivedy I, Higgin- botham S, et al. Molecular origin of cancer: catechol estrogen—3,4-quinones as endogenous tumor initiators. Proc Natl Acad Sci U S A 1997;94: 10937—42. (53) Frenkel K, Wei L, Wei H. 7,l2-dimethylbenz[a]anthracene induces oxida- tive DNA modification in vivo. Free Radic Biol Med 1995;19:373—250. (54) Malins DC, Polissar NL, Schaefer S, Su Y, Vinson M. A unified theory of carcinogenesis based on order—disorder transitions in DNA structure as studied in the human ovary and breast. Proc Natl Acad Sci U S A 1998; 95:7637—42. (55) Henderson CI. Risk factors for breast cancer development. Cancer 1993; 71:2129s-40s. NOTES Supported by Public Health Service grants CA49449 and CA67262 from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services (S. E. Hankinson); and by US. Army Medical Research and Materiel Command Contract DAMD17-95—1-5062 (D. C. Malins). S. E. Hankinson thanks Stacey Missmer for her input into this manuscript and Sandra Melanson for excellent technical assistance. D. C. Malins thanks the National Cancer Institute Cooperative Human Tissue Network for tissues and pathology data. Journal of the National Cancer Institute Monographs No. 27, 1999 Chapter 10: Hope for Prevention—Perspective of the Cancer Advocate Elizabeth A. Hart Breast and prostate cancer survivors and advocates partici- pated as panelists with scientists in an interactive panel dis- cussion following 2 days of scientific presentations on “Es- trogens as Endogenous Carcinogens in the Breast and Prostate.” Advocates raised several issues of concern and questions related to the research presented. Concerns in- cluded the following: 1) a global fear of developing either breast or prostate cancer and recurrence from these tumors, 2) a specific fear that estrogen replacement therapy could enhance the development of new breast cancers and stimu- late recurrence in breast cancer survivors, and 3) a concern that researchers examining minority communities should have sensitivity to the specific culture under study and an understanding of specific research issues that are relevant in those communities. The questions raised included the follow- ing: 1) What are the implications of resistance to antiestro- gen therapies and what is the appropriate sequencing of hormone therapy for longer-term benefit? 2) Can one iden- tify women and men at risk for cancer who do not have the usual risk factors? 3) Where does the development of blood or urine tests to screen for cancer currently stand? 4) Can research findings be translated into effective therapies more rapidly? 5) Can the status of this translational process be communicated to the public in a meaningful way by break- ing down language barriers? 6) What means are available to develop more creative ways to fund pilot studies that do not require preliminary data and to create new funding mecha- nisms to respond to the needs of scientists, particularly those that work collaboratively from multiple institutions and multidisciplines? 7) How can the need for increased empha- sis on and visibility for prostate cancer be communicated? Following the interactive dialogue, scientists and advocates suggested more collaborative research with sustained fund- ing avenues, continued dialogue and collaboration between scientists and advocates, and more collaborative research groups like the Cancer Cube. [J Natl Cancer Inst Monogr 2000;27:157—9] COMPOSITION OF PANEL A multidisciplinary panel, chosen to create a dialogue be— tween scientists and patient advocates, discussed concepts aris- ing from 2 days of intense scientific discussion at a meeting held on March 16—17, 1998, entitled “Estrogens as Endogenous Car- cinogens in the Breast and Prostate.” This landmark meeting was convened by a collaborative group called “The Cancer Cube,” which is interested in demonstrating the causes of breast and prostate cancer. Panelists included the following: 0 a young breast cancer survivor diagnosed at age 38 years, ° 3 breast cancer survivor with metastatic disease, ° a prostate cancer survivor with a brother and father (deceased) diagnosed with the disease, 0 a prostate cancer advocate charged with the development of Journal of the National Cancer Institute Monographs No. 27, 2000 prostate cancer support groups nationwide and advocacy on behalf of those with prostate cancer, - a high-risk individual with multiple family members either deceased or surviving breast, prostate, kidney, uterine, throat, and lung cancer. ' a scientist expert in chemical carcinogenesis, - a scientist expert in clinical/translational sciences, and - a scientist expert in hematology, oncology, and endocrinology. In addition, all scientists (both presenting and in the audience) and advocates participated in an interactive dialogue following brief presentations by panel members. Advocates on the panel listened to the research presented for 2 days and then articulated issues important to the advocacy community and to the public at large relating to the presentations. QUESTIONS AND CONCERNS The issue of cancer prevention is the most critical concern to breast and prostate cancer survivors, their families, advocates, and the public at large. The rapid advances in cancer research in recent years raise expectations that an answer may well be forth- coming in the not—too-distant future. For some, the answer will come too late. Within the advocacy community and the public, there is a tremendous sense of urgency to advance research from the laboratory to the clinical setting as quickly as possible. Fear was an underlying theme throughout the discussion: fear of can- cer in general and fear of recurrence. A specific fear related to the possibility that estrogen replacement therapy could increase the risk of developing new breast cancers and the rate of recur- rence in breast cancer survivors ( I ). IMMEDIATE BENEFITS or CURRENT RESEARCH Recent development of “designer antiestrogens” and use of surrogate drugs to give the benefits of estrogen without using estrogen itself were considered to be important practical ad- vances. These approaches could immediately benefit patients. Designer estrogens that act positively to reduce bone loss, sub- sequent osteoporosis, and bone fractures, and yet do not ad— versely affect the breast, are now available (I). The potential of these agents to prevent cardiovascular disease is currently under Study. Following in this same vein are compounds that are being developed and tested as new antiestrogens. The potential is for longer-term benefit to women with metastatic breast cancer whose hormone-dependent tumors have become resistant to the standard antiestrogens (I). Blocking estrogen production with Correspondence to: Elizabeth A. Hart, R.N., B.A., Hart International, 9051 Oak Path Lane, Dallas, TX 75243 (e—mail: hart.elizabeth@worldnet.att.net). See “Notes” following “References.” © Oxford University Press 157 aromatase inhibitors, such as anastrazole (Arimadex) and letrO- zol (Femara), continues to provide benefit to patients whose tumors are resistant to standard antiestrogens. “Pure” antiestro- gens, which do not have the potential to exert estrogen—like effects on certain tissues, are under study. These agents appear to be promising for the treatment of women whose breast tumors are resistant to tamoxifen. It has been suggested that some com— bination of these therapies, tailored to the individual’s particular parameters, can extend the benefits of hormonal therapy. Quality of life is an important issue for women surviving breast cancer. A number of treatments are currently available to alleviate estrogen—deficiency symptoms experienced by these women and to serve as surrogates for estrogen. As examples, the bisphosphonates act to prevent osteoporosis, the statin drugs lower cholesterol and ultimately prevent heart disease, low-dose vaginal estrogens provide relief from urogenital atrophy, and antidepressants deal with depression exacerbated by estrogen deficiency in susceptible individuals (I). For prostate cancer survivors who experience medical or sur- gical castration and have vasomotor instability, many have ben- efitted from the administration of Clonidine or Megace. Again, the bisphosphonates are useful in treating osteoporosis in set- tings where androgens and estrogens are deficient (2). INTERMEDIATE BENEFITS OF CURRENT RESEARCH Research presented offering intermediate benefits (research in process and not yet available for immediate application) to breast and prostate cancer survivors involves the continuing de— velopment of new drugs acting as hormonal antagonists. Some of these work specifically on the B—estrogen receptor and might have beneficial estrogen effects without having detrimental ef~ fects on the breast. For the prostate, researchers are beginning to look at the use of aromatase inhibitors in men with advanced prostate cancer. This approach is based on the hypothesis that there are mutations of the androgen receptor in advanced pros- tate cancer that make the receptors promiscuous in the sense that they are stimulated to a greater extent with estrogen than with androgen (1). LONG-TERM BENEFITS OF RESEARCH For the long term, the hope is to prevent breast and prostate cancers. Greater understanding of the metabolic activation of estrogens in the body may suggest potential new prevention strategies. Scientists at the meeting suggest that certain metabo- lites of estrogen, both exogenous and endogenous, can generate mutations that could lead to cancer (Chapters 3 and 4). The theory suggests that estrogen receptor—mediated processes would allow these mutations to be propagated. Estrogens acting through receptors could induce cellular proliferation and in- crease the replication of mutated genes. Together, the genotoxic and cell proliferative effects of estrogen would enhance the pro- cess of cellular transformation and, eventually, cause cancer (see Symposium Overview). Metabolic activation of estrogens involves the formation of catechol estrogen metabolites (products of estrogen metabolism, Chapter 5), which, when oxidized in a specific pathway, bind to DNA. These DNA—estrogen complexes cause depurination of DNA (adenine and guanine bases that fall out of DNA) and other DNA damage that leads to tumor initiation (Chapter 4). There is mounting evidence that the pathway leading to the formation of 4-hydroxy estrogens (carcinogenic in animals) is the real culprit, 158 particularly when the enzymes (catechol—0—methyltransferases) that normally neutralize these products of estrogen metabolism are not present or are present at very low levels and, therefore, are not effective protectors (3—6). Of interest, the 2—hydroxylated estrogens, the major products of estrogen oxidation in mamma- lian species, form stable DNA adducts and are not carcinogenic (3—5). If these concepts are borne out, both breast and prostate, as well as other cancers, would share the same initiation process. The implications for prevention are immediate: inhibit metabolic activation of estrogens (particularly to 4-hydroxy estrogens) and enhance their metabolic protection (3,4). RISK FACTORS Another issue of significance to advocates was the ability to identify women early on who do not have the usual risk factors. It was considered important as well to utilize blood or urine tests to screen for cancer. There are currently studies in process look- ing at biomarkers of DNA damage prior to the development of breast cancer, one of which is a blood assay indicating the pres- ence of high anti-HmdU (5—hydroxymethyl-2’-deoxyuridine: an oxidized thymidine) autoantibody titers in healthy women with a family history of breast cancer (7). This would permit the screening of individuals prior to the clinical manifestation of cancer and provide the possibility of prevention. HORMONE RESISTANCE Resistance to antiestrogen therapy and the appropriate se- quencing of hormone therapy for longer-term benefit was dis- cussed at length as a major issue. Clearly, there is benefit to complete blockage of estrogen in individuals relapsing on tamoxifen (Nalvodex) in some settings, yet, in other settings, it is inappropriate to begin with complete estrogen blockade and then expect an antiestrogen to work effectively. So, the most effective sequence for maximum benefit over time seems to be the treatment with adjuvant antiestrogens that are strong antago- nists but weak agonists, followed later on by a pure antiestrogen (1 ). TRANSLATION In addition to the specific research issues raised, advocates focused on translating research findings to the public in a mean— ingful way. How could one break down the language barrier between scientists, advocates, and the public at large? Advocates have become increasingly involved in the research process in the last 10 years, having successfully led the push for increased funding for cancer research, having participated in peer review Of basic, clinical, and translational research proposals [(8); Andejeski Y, Sharp-Breslau E, Hart E, Lythcott N, Alexander L, Rich 1, et al.: manuscript in preparation for publication], and are increasingly included on policy, review, and decision—making bodies. They read the scientific literature. Many have large con— stituencies in need of accurate research information that is readily available in layman’s terms and can be easily dissemi- nated to the public. Scientists were urged to begin to break through the language barrier with the idea of making research findings readily accessible to the lay individual in terms that they could understand. Advocates want to be involved and to help in moving research forward on the fastest track possible, including assistance with funding. Journal of the National Cancer Institute Monographs No. 27, 2000 FUNDING or RESEARCH Funding of research was a major point of discussion. Advo- cates felt that more creative ways of funding pilot studies should be forthcoming. Funding of those studies should not necessarily require preliminary data. At the moment, limited ways of fund— ing exist for pilot studies that can subsequently lead to applica— tions for more traditional funding. Collaborative efforts such as the Cancer Cube (see Introductory Remarks) have an equally difficult time finding appropriate funding mechanisms, although efforts are being made by the National Cancer Institute (Bethesda, MD) to address this issue. Members of the Cancer Cube, which include an advocate, work in multiple research centers. Collaboration among centers provides a unique way to enhance research because it allows wide sharing of specific ex- pertise and cutting-edge technology. Highly technical resources are shared, and specific objectives are chosen by the group to advance research more efficiently and effectively. Funding of research on prostate cancer remains definitively lower than that of breast cancer or acquired immunodeficiency syndrome (AIDS). In the 1997/1998 Labor HHS Appropriation, signed in November 1997, prostate cancer received approxi~ mately $89.5 million compared with $348.6 million in breast cancer, with AIDS at $226.4 million ( 9). Prostate cancer needs more visibility, as do a number of other cancers. However, the advocates and scientists both felt strongly that one disease should not be pitted against another. The need is to figure out how to raise sufficient monies to fund the necessary basic, clini- cal, and translational research which, in specific instances, has applicability in a number of diseases, not just breast or prostate cancer. It is important as well to have trained researchers from and in minority communities, who have sensitivity to the culture and understanding of specific research issues that are relevant in those communities. The US. Department of Defense is currently conducting a consensus conference strategy to develop a blue— print for including training of minority researchers at the pre- doctoral through postdoctoral level. CONCLUSION The interactive dialogue between scientists and advocates was bold and enlightening, with articulation from both groups as to needs, cooperative efforts, and future directions. Advocates suggested more collaborative “Cubes” and creative funding mechanisms, while scientists suggested more collaborative re- search with sustained funding avenues and continued dialogue and collaboration with advocates. It was expressed and abun- dantly clear that both advocates and scientists share an equal Journal of the National Cancer Institute Monographs No. 27, 2000 passion for finding a cure and, even more important, preventing the initiation of breast and prostate cancers, as well as other cancers. REFERENCES (I) The Hormone Foundation, The Susan G. Komen Breast Cancer Foundation, Canadian Breast Cancer Research Initiative, National Cancer Institute of Canada, Endocrine Society, and the University of Virginia Cancer Center, Woman’s Place. Consensus statement. Treatment of estrogen deficiency symptoms in women surviving breast cancer. J Clin Endocrinol Metab 1998; 83:1993—2000. (2) Santen RJ, Borwhat M, Gleason S, Gore MJ, editors. Menopause-treatment options for women surviving breast cancer or concerned about estrogen replacement therapy. The Hormone Foundation, June 1998. (3 ) Liehr JG. Dual role of estrogens as hormones and pro—carcinogens: tumor initiation by metabolic activation of estrogens. Eur J Cancer Prev 1997;6z 3—10. (4 ) Cavalieri EL, Stack DE, Devanesan PD, Todorovic R, Dwivedy I, Higgin- botham S, et a]. Molecular origin of cancer: catechol estrogens-3,4-quinones as endogenous tumor initiators. Proc Natl Acad Sci U S A 1997;94: 10937—42. (5) Stack DE, Byun J, Gross ML, Rogan EG, Cavalieri EL. Molecular charac- teristics of catechol estrogen quinones in reactions with desoxyribonuncleo- sides. Chem Res Toxicol 1996;9:851—9. (6) Service RF. New role for estrogen in cancer? Science 1998;279:1631—3. (7) Frenkel K, Karkoszka J, Glassman T, Dubin N, Toniolo P, Taioli E, et a1. Serum autoantibodies recognizing 5-hydroxymethyl—2’-deoxyuridine, an oxidized DNA base, as biomarkers of cancer risk in women. Cancer Epide- miol Biomarkers Prev 1998;7z49—57. (8) Waller M, Batt S. Advocacy groups for breast cancer patients. CMAJ 1995; 152:829—33. (9) Labor, Health and Human Services Appropriation 105—78, November 1997. NOTES The Panel Members include the following: chair— Elizabeth A. Hart, Presi- dent & CEO, Hart International, Dallas, TX; moderator—David G. Longfellow, Ph.D., Chief, Chemical & Physical Carcinogenesis Branch, Division of Cancer Biology, National Cancer Institute (NCI), Bethesda, MD; panel members— Winston Dyer, Community Coordinator of Minority Clinical Trials, CaP CURE, New York, NY; Carol Hochberg, Board Member and Committee Chairman of the Advocacy and Public Policy Committee of SHARE, New York, NY; Edison Liu, M.D., Ph.D., Director of Clinical Sciences, NCI; M. Brooke Moran, Direc- tor of Patient Advocacy and Government Affairs for the American Foundation for Urologic Disease, Baltimore, MD; Diana Rowden, Chairman of the Board of the Susan G. Komen Breast Cancer Foundation, Dallas; and Richard Santen, M.D., Professor, Division of Hematology, Oncology and Endocrinology, Uni— versity of Virginia Health Science Center, Charlottesville. I gratefully acknowledge all members of the Cancer Cube for their warm inclusion of an advocate in the formation and continuing work of the group and specifically Joachim Liehr, Ercole Cavalieri, Eleanor Rogan, David Longfellow, and Richard Santen for their counsel in preparation of this manuscript. 159 . ,. :wfiawa , _., ka 3+ ' * «Exam mm: mm \ V 1,. . , y. W: K 4. 3.3.. U. C. BERKELEY LIBRARIES CDLHHLSQHH