TECHNOLOGIES FOR DETECTING HERITABLE MUTATIONS IN HUMAN BEINGS 733¢27zq PMBL Recommended Citation: U.S. Congress, Office of Technology Assessment, Technologies for Detecting Heritable Mutations in Human Beings, OTA-H-298 (Washington, DC: US. Government Printing Office, September 1986). Library of Congress Catalog Card Number 86-600523 For sale by the Superintendent of Documents US. Government Printing Office, Washington, DC 20402 (3615:" ,1 .381 ‘ ,« I / ‘7 ‘6 L9 f") « FMfL Ensuring the health of future generations of children is of obvious importance to American society. Heritable mutations, permanent changes in the genetic material that can be passed on to succeeding generations, are the cause of a large but currently un- quantifiable share of embryonic and fetal loss, disease, disability, and early death in the United States today. The methods now available to study heritable mutations, however, offer relatively little information about the kinds of mutations that can oc- cur, their frequency, or their causes. Recent advances in molecular genetics have opened the door to new and innovative technologies that may offer a great deal more informa- tion about DNA. It may soon be possible to characterize mutations precisely, to meas- ure their frequency, and perhaps also to associate particular mutations with exposures to specific mutagenic influences. While some of the new technologies are still on the drawing board, they are developing quickly and several of them may become available for wide-scale use in the next 5 to 10 years. ,A Foreword The Senate Committee on Veterans’ Affairs, the House Committee on Science and Technology, and the House Committee on Energy and Commerce requested that OTA assess the available information about current and proposed means for detecting heritable mutations and on the likelihood and potential impact of such technological advances. These committees have wrestled with the problems of determining whether past exposures to potential mutagens have affected the health of Americans, and of framing reason- able public health laws, given current knowledge and technologies. This report sum- marizes OTA’s findings as they relate to these issues. An advisory panel, chaired by Arno G. Motulsky, provided guidance and assis- tance during the assessment. The OTA Health Program Advisory Committee, OTA staff, and scientific and policy experts from the private sector, academia, and the Fed- eral Government provided information during the assessment and reviewed drafts of the report. We thank all who assisted us. As with all OTA reports, the content of the assessment is the sole responsibility of OTA and does not necessarily constitute the con- sensus or endorsement of the advisory panel or the Technology Assessment Board. Key OTA staff involved in the assessment were Michael Cough, Julie Ostrowsky, and Hellen Gelband. afigwok JOHN H. GIBBONS Director iii Advisory Panel for Technologies for Detecting Heritable Mutations in Human Beings Arno G. Motulsky, Panel Chair Center for Inherited Diseases University of Washington School of Medicine Richard J. Albertini Department of Medicine University of Vermont College of Medicine Michael S. Baram School of Medicine and Public Health Boston University Charles R. Cantor Department of Human Genetics and Development Columbia University College of Physicians and Surgeons Dale Hattis Center for Policy Alternatives Massachusetts Institute of Technology Ernest B. Hook Bureau of Maternal and Child Health New York State Department of Health Alfred G. Knudson, Ir. Institute for Cancer Research Fox Chase Cancer Center Nan M. Laird Department of Biological Statistics Harvard University School of Public Health Mortimer L. Mendelsohn Lawrence Livermore National Laboratory Jeffrey H. Miller Department of Biology University of California at Los Angeles James V. Neel Department of Human Genetics University of Michigan School of Medicine Norton Nelson Department of Environmental Medicine New York University School of Medicine Mark L. Pearson E.I. du Pont de Nemours & Co. Richard K. Riegelman Department of Medicine George Washington University School of Medicine Liane B. Russell Oak Ridge National Laboratory Richard B. Setlow Brookhaven National Laboratory William J. Schull The University of Texas Health Science Center William G. Thilly Department of Nutrition and Food Science Massachusetts Institute of Technology Richard M. Myers, Special Consultant University of California at San Francisco NOTE: OTA appreciates and is grateful for the valuable assistance and thoughtful critiques provided by the advisory panel members. The panel does not, however, necessarily approve, disapprove, or endorse this report. OTA assumes full responsibility for the report and the accuracy of its contents. OTA Project Staff—Technologies for Detecting Heritable Mutations in Human Beings Roger C. Herdman, Assistant Director, OTA Health and Life Sciences Division Clyde 1. Behney, Health Program Manager Michael Cough, Project Director Iulia T. Ostrowsky, Analyst Hellen Gelband, Analyst Other Contributing Staff Cheryl M. Corsaro, Analyst1 Virginia Cwalina, Administrative Assistant Carol Ann Guntow, Secretary/ Word Processor Specialist Diann G. Hohenthaner, Word Processor/P. C. Specialist Eric Passaglia, Senior Distribution Specialist Contractors Elbert Branscomb, Lawrence Livermore National Laboratory Neal Cariello, Massachusetts Institute of Technology Leonard Lerman, Genetics Institute Harvey Mohrenweiser, University of Michigan Medical School Richard M. Myers, University of California at San Francisco Janice A. Nicklas, University of Vermont School of Medicine Maynard Olson, Washington University Medical School Cassandra Smith, Columbia University College of Physicians and Surgeons L.H.T. Van der Ploeg, Columbia University College of Physicians and Surgeons Diane K. Wagener, Washington, DC 1On detail from the National Institutes of Health, April-June 1985‘ Vi Contents Chapter Page 1. Summary and Options ..................................................... 3 2. Genetic Inheritance and Mutations .......................................... 23 3. Kinds and Rates of Human Heritable Mutations .............................. 37 4. New Technologies for Detecting Heritable Mutations .......................... 55 5. New Methods for Measuring Somatic Mutations .............................. 73 6. Laboratory Determination of Heritable Mutation Rates ........................ 81 7. Extrapolation ............................................................. 91 8. Mutation Epidemiology .................................................... 103 9. Mutagens: Regulatory Considerations ........................................ 113 Appendixes Page A. Federal Spending for Mutation Research ..................................... 121 B. Acknowledgments and Health Program Advisory Committee .................. 123 References .................................................................. 129 Index ...................................................................... 139 7 Glossary of Acronyms and Terms Glossary of Acronyms A —Adenine AEC —Atomic Energy Commission ATSDR —Agency for Toxic Substances and Dis- ease Registry (Centers for Disease Control) BEIR —Committee on the Biological Effects of Ionizing Radiations (NRC) C —Cytosine CERCLA —Comprehensive Environmental Response, Compensation, and Liability Act ("Superfund”) DNA —Deoxyribonucleic Acid DOE —Department of Energy EBV —Epstein-Barr Virus EDB —Ethylene dibromide EPA —Environmental Protection Agency ENU —Ethylnitrosourea EtO —Ethylene oxide FDA —Food and Drug Administration (DHHS) FRC —Federal Radiation Council G —Guanine HbA —Hemoglobin A HbM —Hemoglobin M hprt —Hypoxanthineguanine phosphoribosyl transferase (gene) HPRT —Hypoxanthineguanine phosphoribosyl transferase (enzyme) HTT —Heritable Translocation Test ICPEMC —International Commission for Protection Against Environmental Mutagens and Carginogens ICRP —International Commission on Radiolog- ical Protection mRNA —Messenger RNA MSHA —Mine Safety and Health Administration (Department of Labor) NCRP —National Council on Radiation Protec- tion and Measurements NIH —National Institutes of Health NRC —Nuclear Regulatory Commission NRC —National Research Council (National Academy of Sciences) NTP —National Toxicology Program (DHHS) OSHA —Occupational Safety and Health Admin- istration (Department of Labor) OTA —Office of Technology Assessment (US. Congress) PBL —Peripheral Blood Lymphocyte PFGE —Pulsed Field Gel Electrophoresis PKU —Phenylketonuria R —Roentgen RFLP —Restriction Fragment Length Poly- morphism RNA ——Ribonucleic Acid SCE —Sister-chromatid Exchange SLT —Specific Locus Test T —Thymine TGr —Thioguanine resistant TSCA —Toxic Substances Control Act UNSCEAR—United Nations Scientific Committee on the Effects of Atomic Radiation 2DDGGE —Two Dimensional Denaturing Gel Elec- trophoresis ZDPAGE —Two Dimensional Polyacrylamide Gel Electrophoresis 6TG —6Thioguanine 6TGr -—6Thioguanine resistant Glossary of Terms Achondroplasia: A disease marked by a defect in the formation of cartilage at the ends of long bones (fe- mur, humerus) that leads to a type of dwarfism. There are a number of hereditary forms, the most common of which is autosomal dominant. Allele: An alternative form of a gene, or a group of functionally related genes, located at the corre- sponding site on a homologous chromosome. Each allele is inherited separately from each parent. Alleles can be dominant, recessive, or co-dominant for a particular trait. Alpha thalassemia: A genetic defect caused by an alteration in a portion of the gene coding for the alpha globin molecule. The result is an insufficient number of alpha globin molecules and a deficiency of adult hemoglobin. Amino acid: One of a group of 20 molecules that bind together in various sequences to form all protein molecules. The specific sequence of amino acids determines the structure and function of a protein. Amniocentesis: A procedure that involves withdraw- ing a sample (usually 2 to 8 milliliters) of the amniotic fluid surrounding the fetus in utero. This fluid con- tains cells shed by the developing fetus. The cells can be grown in cell culture and analyzed either bio- chemically or cytogenetically to detect a variety of genetic abnormalities in the fetus, including genetic diseases such as Down syndrome and Tay Sachs disease. Autoradiography: A technique for identifying radio- actively—labeled molecules or fragments of mole- cules, useful for analyzing DNA for the presence of mutations. vii Autosome: A chromosome not involved in sex—deter- mination. In a complete set of human chromo- somes, there are 44 autosomes (22 pairs of homologous chromosomes) and a pair of sex-determining chro- mosomes. Base pair (of nucleic acids): A pair of hydrogen-bonded nitrogenous bases (one purine and one pyrimidine) that join the two strands of the DNA double helix. Adenine pairs with thymine, and cytosine pairs with guanine. Beta globin: A constituent of hemoglobin, which is a molecule consisting of four globin subunits and a heme group (two alpha and two beta globins). Carcinogen: A chemical or physical agent that causes cancer. Cell: The smallest membrane-bound protoplasmic body, consisting of a nucleus and its surrounding cytoplasm, capable of independent reproduction. Cell culture: Growth in the laboratory of cells isolated from multicellular organisms. Each culture is usually of one cell type (e.g., lymphocytes, fibroblasts, etc.). Cell line: A sample of cells, having undergone the proc- ess of adaptation to artificial laboratory cultivation, that is now capable of sustaining continuous, long- term growth in culture. Chromosome: A threadlike structure that carries genetic information arranged in a linear sequence. It consists of a complex of nucleic acids and proteins. Chromosome abnormalities: A group of pathological conditions associated with abnormalities in the number or structure (e.g., insertions, deletions, rearrangements) of chromosomes. Chromosome banding: The chemical process of stain- ing chromosomes to identify each pair of homo- logous chromosomes. Staining by various techniques produces patterns of light and dark bands (visible under the microscope) that are characteristic of each chromosome pair. Chromosome banding is particu— larly useful in detecting structural chromosome ab- normalities such as inversions and deletions. Codon: A sequence of 3 adjacent nucleotides in mRNA that specifies 1 of the 20 amino acids or the initia- tion or termination of the peptide chain. The lin- ear sequence of codons determines the order in which amino acids are added to a polypeptide chain during translation. Complementary DNA (cDNA): Single-stranded DNA that is synthesized from a messenger RNA template. It is often used as a probe to help locate a specific gene in an organism. Congenital: Refers to a condition that is present at birth. Congenital abnormality: Any abnormality, genetic or nongenetic, that is present at birth. Cytoplasm: The contents of a cell exclusive of the nu- cleus. It consists of an aqueous solution and the Viii ‘ organelles suspended in it and is the site of most chemical activities of the cell, including protein synthesis. Denaturation: The separation of double-stranded DNA into its complementary, separate strands by treatment with chemicals or heat. The narrow range of temperature or chemical concentration at which DNA denaturation occurs is characteristic of the nucleotide sequence of the particular molecule. Diploid: The chromosome state in which each homologous chromosome is present in pairs. All human somatic cells are diploid (i.e., they have 46 chromosomes), whereas reproductive cells, with 23 chromosomes, are haploid. DNA: Deoxyribonucleic acid. The nucleic acid in chro- mosomes that contains the genetic information. The molecule is double-stranded, with an external ”back- bone" formed by a chain of alternating phosphate and sugar (deoxyribose) units and an internal ladder- like structure formed by nucleotide base-pairs held together by hydrogen bonds. The nucleotide base pairs consist of the bases adenine (A), cytosine (C), guanine (G), and thymine (T), whose structures are such that A can hydrogen bond with T, and C with G. The sequence of each individual strand can be deduced by knowing that of its partner. This com- plementarity is the key to the information-trans- mitting capabilities of DNA. (See also nucIeotide and gene.) DNA probe: A segment of complementary DNA that is used to detect the presence of a particular nucleo- tide base sequence. DNA sequence: The order of nucleotide bases in DNA. Dominant: A term used to refer to a genetic trait that is expressed in an individual who is heterozygous for a particular gene. Compare recessive. Down syndrome: A genetic abnormality characterized by mental retardation, congenital heart defects, immune system abnormalities, various morphologi- cal abnormalities and a reduced life expectancy. Down syndrome is caused by either an extra copy of chromosome 21 (called trisomy 21) or by two copies of chromosome 21 and another chromosome 21 translocated to a different chromosome (usually to chromosome 14). The latter results in ”trans- location Down syndrome.” Trisomy 21 has been shown to increase in frequency with advanced maternal age. Electrophoresis: A technique used to separate mole- cules (such as DNA fragments or proteins) from a mixture of similar molecules. By passing an electric current through a medium containing the mixture, each type of molecule travels through the medium at a rate that corresponds to its electric charge and size. Separation is based on differences in net elec- trical charge and in size or arrangement of the molecule. Enzyme: A protein that catalyzes a chemical reaction without being permanently altered or consumed by the reaction, so that it can be used repeatedly. Enzyme deficiency variants: Abnormal proteins char— acterized by reduction or elimination of enzymatic activity. These changes may be caused by the ab- sence of a gene product, or the presence of proteins that are non-functional or abnormally unstable. Epidemiologic studies: Studies concerned with the re- lationships of various factors determining the fre— quency and distribution of diseases in a human pop- ulation. Escherichia coli (E. coli): A species of rod-shaped, gram negative bacteria that inhabit the normal intestinal tract of vertebrates. Many nonpathogenic strains of E. coli are hosts in recombinant DNA tech- nologies. Eukaryote: Cells or organisms with membrane—bound, structurally discrete nuclei and well—developed cell organelles. Eukaryotes include all organisms except viruses, bacteria, and blue—green algae. Compare prokaryote. Exons: DNA sequences that determine the sequence of amino acids in proteins. Exons are separated on the DNA by introns, or intervening sequences, that are transcribed and later removed, or spliced out, during the production of mature messenger RNA. Gametes: Mature male or female reproductive cells (germ cells: sperm or ova) with a haploid chromo- some content (23 chromosomes in humans). Gene: A linear sequence of nucleotides in DNA that is needed to synthesize a protein and/ or regulate cell functions. A mutation in one or more of the nucleotides in a gene may lead to abnormalities in the structure of the gene product or in the amount of gene product synthesized. Genetic code: The sequence of nucleotide triplets along the DNA that determines the sequence of amino acids in protein synthesis. This code is common to nearly all living organisms. Genome: A term used to refer to all the genetic mate- rial carried by a single gamete. Genotype: The genetic constitution of an organism, as distinguished from its physical appearance (its phenotype). For example, an individual may have a heterozygous genotype for eye color consisting of an allele for brown eyes (which is dominant) and an allele for blue eyes (which is recessive) or a homozygous genotype, with two alleles (both dom- inant) for brown eyes. In either case, the phenotype is the same: brown eyes. Germinal mutations: Mutations in the DNA of repro— ductive cells—egg or sperm. Germinal mutations can be transmitted to the offspring only if one of those particular germ cells is involved in fertili- zation. Germ line: The tissue or cell line that produces gametes and is used for reproductive purposes, as opposed to tissue or cell line from somatic cells. Also known as “germinal tissue." Haploid: Half of the full set of genetic material, or one chromosome of each homologous pair. Gametes have a haploid set of DNA. Fertilization of an ovum by a sperm produces a diploid number of chromo- somes in the zygote. Hemoglobin: The oxygen—carrying molecule found in red blood cells. Adult hemoglobin is a protein composed of a single heme group bound to two alpha globin chains and two beta globin chains. Hemoglobinopathies: A collection of hereditary dis— orders of hemoglobin structure and/ or function. Examples are thalassemia and sickle cell anemia. Hemophilia: A genetic disorder distinguished by a deficiency of one or more coagulation factors—e.g., Factor VIII (hemophilia A) or Factor IX (hemophilia B). The underlying mutations are located on the X chromosome. Hemophilia occurs most often in males who have only one X chromosome, and is transmitted to offspring by asymptomatic females, who have two X chromosomes, one of which carries the hemophilia mutation. Heritable mutation: A mutation that is passed from parent to offspring and was present in a germ cell of one of the parents. Heteroduplex: A double-stranded nucleic acid mole- cule composed of individual strands of different origin, e.g., a parent’s DNA strand hybridized to a child’s DNA strand, or RNA hybridized to DNA. Heterozygote: An individual who has two different alleles of any one particular gene, e.g., an individ- ual who has one copy of the gene for thalassemia at the locus for beta globin is heterozygous for this trait. Homoduplex: A double-stranded nucleic acid mole- cule composed of two strands of the same origin, e.g., genomic DNA isolated from human cells (DNA hybridized with DNA from the same indi- vidual). Homozygote: An individual with the same allele at both genes responsible for a particular trait. Homozygous: Having identical alleles at a given locus. Compare heterozygous. Hybridization: The process of joining two single- strands of RNA or DNA together so that they become a double-stranded molecule. For hybridiza- tion to occur, the two strands must be nearly or perfectly complementary in the sequence of the nucleotide base pairs. Induced mutation: A change in the structure of DNA or the number of chromosomes, caused by exposure of the DNA to a mutagenic agent. In vitro: Literally, “in glass,” pertaining to a biologi- cal process or reaction taking place in an artificial environment, usually a laboratory. Sometimes used to include the growth of cells from multicellular organisms under cell culture conditions. ix In vivo: Literally, "in the living,” pertaining to a bio- logical process or reaction taking place in a living cell or organism. Karyotype: A photomicrograph of an individual’s chromosomes arranged in a standard format, showing the number, size, and shape of each chro- mosome. Kilobase: A unit of measurement for the length of nu- cleic acids along a chromosome. One kilobase (abbreviated kb) consists of 1,000 nucleotide bases. Genes can be several kilobases in length. Lambda: A bacterial virus that infects E. coli, produc- ing many copies of itself within the bacterium. Lesch-Nyhan syndrome: An X-linked recessive dis- order characterized by compulsive self-mutilation and other mental and behavioral abnormalities. It is caused by a defect in the gene that produces a particular enzyme (hypoxanthine-guanine phos- phoribosyl transferase) important in metabolism. The causal relationship to the behavioral aspects of the disorder is not yet understood. Locus: The position of a gene or of a group of functionally related genes on a chromosome. Locus Test: A measurement of genes examined in a study of mutant proteins, based on the number of proteins examined, on the number of gene loci rep- resented by each such protein, and on the number of individual samples obtained. Lymphocytes: Specialized white blood cells involved in the body's immune response. B-lymphocytes originate in the bone marrow and when stimulated by an antigen produce circulating antibodies (hu- moral immunity). T-lymphocytes are produced in the bone marrow and mature in the thymus gland and engage in a type of defense that does not depend directly on antibody attack (cell-mediated im- munity). Marfan’s syndrome: A genetic disorder characterized by abnormally long fingers and toes and by abnor- malities of the eye lenses and heart. (Abraham Lincoln is thought by some to have suffered from this disease). Also called arachnodactyly (“spider fingeredness"). Meiosis: The process of cell division in reproductive cells that reduces the number of chromosomes to half of the full set so that each parent contributes half of the genetic material to each offspring. When the egg and sperm fuse at conception, the full set of chromosomes (a total of 46 chromosomes in humans) is restored in the offspring. Mendelian: A term used to refer to a trait that follows Mendel's laws of inheritance and is controlled by a single gene, and which therefore shows a simple pattern of inheritance (dominant or recessive). So named because traits of this sort were first recog- nized by Gregor Mendel, the Austrian monk whose early research laid the basis for modern genetics. Mendel’s laws include the Law of Segregation, which describes how each pair of alleles separates into different gametes and the Law of Independent Assortment, which describes how different alleles are assorted independently of the other alleles in gametes and how the subsequent pairing of male and female gametes occurs randomly. Messenger RNA (mRNA): Ribonucleic acid that serves as the template for protein synthesis. It is produced by transcribing a DNA sequence into a complemen- tary RNA sequence. Messenger RNA molecules carry the instructions for assembling proteins on the ribosomes. Mitosis: The process of division involving DNA rep- lication, and resulting in two daughter cells with the same number of chromosomes and cytoplasmic material as the parent cell. Multifactorial traits: Traits that are not determined solely by a single gene. Multifactorial, or polygenic, traits (e.g., height) have variable phenotypic effects that depend for their expression on the interaction of many genes and environmental influences. Mutagen: A chemical or physical agent that causes changes in the structure of DNA. Mutation rate: The number of mutations per unit of DNA (e.g., per gene, per nucleotide, per genome, etc.) occurring per unit of time (usually per gen- eration). Mutations: Changes in the composition of DNA, gen- erally divided according to size into “gene muta- tions” (changes within a single gene, such as nucleo- tide substitutions) and "chromosome mutations” (affecting larger portions of the chromosome, or the loss or addition of an entire chromosome). A “her- itable mutation” is a mutation that is passed from parent to offspring and therefore was present in the germ cell of one of the parents. See also induced mutation and spontaneous mutation. Nucleic acids: Macromolecules composed of sequences of nucleotides that carry genetic information. There are two kinds of nucleic acids: DNA, containing the sugar deoxyribose and RNA, containing the sugar ribose. Nucleotide: A subunit of DNA or RNA, consisting of a nitrogenous base (adenine, guanine, thymine, cytosine, or uridine), a phosphate molecule, and a sugar molecule (deoxyribose in DNA or ribose in RNA). The linkage of thousands of these subunits forms the DNA or RNA molecule. Nucleus: The intracellular structure of eukaryotes that contains DNA. Oligonucleotide: A polymer made up of a few (gener- ally fewer than 10 or 20) nucleotides; a short se- quence of DNA or RNA. Peptide: A compound consisting of two or more amino acids linked together, formed through a chemical process that produces one molecule of water for each joining of one amino acid to another. Peptides are the building blocks of proteins. Phenotype: The appearance of an individual or the observable properties of an organism that result from the interaction of genes and the environment. Phenylketonuria (PKU): An autosomal recessive ge- netic disorder of amino acid metabolism, caused by the inability to metabolize phenylalanine to tyro- sine. The resulting accumulation of phenylalanine and derived products causes mental retardation, which can be avoided by dietary restriction of phenylalanine beginning soon after birth. Polymorphism: A single gene trait (e.g., a red blood cell surface antigen) that exists in two or more alter- native forms (such as types A, B, AB, and 0 blood). Genetic variants are considered polymorphisms if their frequency exceeds 1 percent each, but are con- sidered “rare mutations” if they are found in less than 1 percent of the population. Polypeptide: A sequence of amino acids joined in a chain. Probe: A molecule that has been tagged or labeled in some way with a tracer substance (a radioactive isotope or specific dye-absorbing compound) that is used to locate or identify a specific gene or gene product. For example, a radioactive mRNA probe for a DNA gene, or a monoclonal antibody probe for a specific protein. See also DNA probes. Prokaryote: A cell or organism lacking membrane- bound, structurally discrete nuclei. Compare eukaryote. Protein: A molecule composed of hundreds of linked amino acids in a specific sequence, which is, in turn, determined by the sequence of nucleotides in DNA. Proteins are required for the structure, function, and regulation of cells, tissues, and organs in the body. Recessive: A genetic trait that is manifested pheno- typically only when both alleles for the trait are present at a locus. E.g., sickle cell anemia is mani- fest only when both copies of the gene for beta globin contain the beta5 mutation, whereas when individuals have only one copy of the beta5 gene, they do not develop the disease, but have sickle- cell trait, a generally benign condition. X-linked traits act as if they were recessive in females and dominant in males. Compare dominant. Recombinant DNA (rDNA) technology: Techniques involving the incorporation of DNA fragments, generated with the use of restriction enzymes, into a suitable host organism’s DNA (a vector). The host is then grown in culture to produce clones with multiple copies of the incorporated DNA fragment. The clones containing this particular DNA fragment can then be selected and harvested. Replication: The synthesis of new DNA from existing DNA. Restriction enzyme: An enzyme that has the ability to recognize a specific nucleotide sequence in a nucleic acid (ranging from 4 to 12 base pairs in length) and cut or cleave the nucleic acid at that sequence. They are termed “restriction” enzymes because, occurring naturally in bacteria, they recognize foreign nucleic acid (e.g., the DNA of a bacterial virus as it begins to infect and destroy its host) and destroy it, thus restricting the ability of the virus to prey upon certain potential host strains. Over 400 different re- striction enzymes are known, recognizing a great variety of different nucleotide base sequences. Use of restriction enzymes has made possible the cutting and splicing together of nucleic acid within and between different organisms and species. Restriction Fragment Length Polymorphism (RFLP): Fragments of DNA that may vary in length from individual to individual, depending on the presence of polymorphisms or rare mutations. Ribosome: A cellular organelle that is the site of protein synthesis, the process of reading the instructions in mRNA and using it as the guide to constructing the specified protein. RNA (ribonucleic acid): The nucleic acid found mainly in the nucleus and ribosomes and is involved in the control of cellular activities. There are several classes of RNA that serve different purposes, includ- ing messenger RNA (mRNA), transfer RNA (tRNA), and ribosomal RNA (rRNA). The functions of vari- ous kinds of RNA include "translating” selected gene sequences coded in the cell's DNA, and trans- ferring that information outside of the nucleus to structures that then synthesize the proteins indicated by the RNA vehicle. RNaseA: An enzyme that cleaves RNA where certain sequences of nucleotide bases occur. RNaseA has been used to cut fragments of RNA/ DNA hetero- duplexes where particular mismatches in the hetero- duplexes occur. Sentinel phenotypes: A group of autosomal dominant or X-linked conditions'that can occur sporadically. They are thought to result from new germinal mutations in the parents' reproductive cells. Sex chromosomes: The X and the Y chromosomes, 2 of the 46 chromosomes in human cells, that deter- mine the sex of the individual. Females have two X chromosomes, while males have one X and one Y chromosome. Sickle-cell disease: A genetic disorder of hemoglobin function caused by the presence of an abnormal beta globin chain. Patients with sickle-cell disease have red blood cells that tend to deform into a sickle-like shape. The specific defect is caused by an abnormal gene, resulting in the replacement of the usual amino acid, glutamic acid, with valine, in the sixth amino acid position in the beta-hemo— globin molecule. This increases its propensity to crystallize, thus rupturing the red blood cell and causing the cells to lodge in small blood vessels. Also called "sickle-cell anemia.” See sickle—ceII trait. Sickle-cell trait: The generally benign condition shown xi by individuals carrying the variant betaS gene as well as the normal betaa gene. Such individuals are heterozygous for the sickle-cell gene, and are healthy (i.e., are usually asymptomatic), but two hetero- zygous parents have a 25 percent risk with each pregnancy of having a child with homozygous sickle-cell disease. Single gene disorder: A genetic disease caused by a sin- gle gene and showing a simple pattern of inherit- ance (e.g., dominant or recessive, autosomal or X- linked). Also called ”Mendelian disorder." Somatic: A term used to refer to body tissues, as op- posed to reproductive (germinal) tissues. Somatic cell: Any cell in the body except reproduc- tive cells or their precursors. Spontaneous mutation: In the absence of any known causative agent, a change in the structure of DNA or in the number of chromosomes. Also called a “background” mutation. Also see mutations. Stem cells: Undifferentiated cells in the bone marrow that have the ability to replicate and to differenti- ate into blood cells. Tay-Sachs disease: An autosomal recessive genetic dis- ease resulting in developmental retardation, paral- ysis, dementia and blindness, usually fatal in early childhood. The defective gene codes for hexosamini- dase A, an enzyme that is involved in certain chem- ical pathways in the brain. Symptoms are caused by an accumulation of cerebral gangliosides, fatty acid and sugar molecules found in the brain and nervous tissue. The gene is found in highest fre- quency among Jews of Eastern European origin. Teratogen: A physical or Chemical agent (e.g., thalido- mide, radiation, alcohol, etc.) that can cause con- genital abnormalities as a result of exposure in utero. xii Tetramer: A complex molecule consisting of four ma- jor portions joined together (e.g., hemoglobin, in which two alpha chains and two beta chains are joined to a central heme group.) Thalassemia: A group of autosomal recessive genetic disorders characterized by abnormalities in synthe— sis of the globin polypeptides of hemoglobin. The two most common forms are alpha-thalassemia and beta-thalassemia, disorders of the alpha- and beta— globin polypeptides, respectively, which cause im— balances in the production of the globins and lead to an overall deficiency of adult hemoglobin. The thalassemias are most common in people of Medi- terranean, Middle Eastern, and Asian descent. T-lymphocytes: See lymphocytes. Transcription: The synthesis of messenger RNA (mRNA) on a DNA template. The resulting RNA sequence is complementary to the DNA sequence. Transfer RNA (tRNA): Specialized RNA molecules that function to bring specific amino acids to ribo- somes that translate messenger RNA (mRNA) into proteins. Translation: The process in which the genetic code contained in the nucleotide base sequence of mes- senger RNA directs the synthesis of a specific or- der of amino acids to produce a protein. Trisomy: The presence of an extra chromosome, re- sulting in three homologous chromosomes instead of two, e.g., Down syndrome can result from Trisomy 21, or the presence of an extra chromo- some number 21 in each body cell. tRNA: See transfer RNA. X-linked mutation: A mutation that occurs in a region of the X-chromosome. Zygote: A fertilized egg that results from the fusion of sperm and egg. Chapter1 Summary and Options Contents Page Introduction ............................................................ 3 Request for the Assessment ............................................. 4 Scope of the Report .................................................... 4 Background ............................................................. 5 Kinds and Effects of Mutations in Human Beings .......................... 5 Current Methods for Studying Mutations ................................... 6 Animal Studies ........................................................ 6 Studies in Human Beings ............................................... 7 New Technologies for Detecting Human Mutations .......................... 10 Detection and Measurement of Heritable Mutations ........................ 11 Detection and Measurement of Somatic Cell Mutations .................... 13 Use of New Technologies in Research and Public Policy ...................... 14 Feasibility and Validity Testing .......................................... 14 Epidemiologic Activities ................................................ 14 Extrapolation ......................................................... 14 Regulation ............................................................ 15 Federal Spending for Mutation Research .................................. 15 Options ................................................................ 16 Development of New Technologies ...................................... 16 Feasibility and Validity Testing .......................................... 17 Field Studies .......................................................... 18 Integration of Animal and Human Studies .................. -.' ............. 19 FIGURE Figure No. Page 1. Organizational Hierarchy of DNA, the Carrier of Genetic Information in Human Cells ....................................................... 6 Chapter 1 Summary and Options INTRODUCTION Mutations, lasting changes in the genetic in- formation carried in the deoxyribonucleic acid (DNA) of cells, can cause severe diseases and dis— abilities, none of which is curable and relatively few of which can be treated effectively. Such genetic diseases represent a significant fraction of chronic disease and mortality in infancy and child- hood; they generally impose heavy burdens ex- pressed in premature mortality, morbidity, infer- tility, and physical and mental handicap. Some of the most common of the 3,000 or more differ— ent disorders known to result from mutations in- clude Down syndrome, Duchenne muscular dys— trophy, and hemophilia. In addition, mutations have been associated with increased susceptibili— ties to certain chronic diseases, including some forms of diabetes, heart disease, and cancer. Most mutations that are expressed as genetic disease al- ready exist in the population and are carried from generation to generation. A smaller proportion of mutations arises anew, ”sporadically," in each generation, and the specific causes of these mu- tations are unknown. The public and the government have expressed concern about the possibility that environmental exposures are contributing to or increasing the fre- quency of mutations. Mutations are among the chronic health effects singled out in the Toxic Sub- stances Control Act of 1976 (TSCA) and in the Comprehensive Environmental Response, Com- pensation, and Liability Act of 1980 (CERCLA or “Superfund”). In those laws, Congress specified public protection from exposures that can cause mutations. The other major environmental stat- utes contain language broad enough to include protection from mutagens. Unfortunately, little is currently known about the kinds and rates of mutations that occur in human beings. Available methods to study such mutations are inadequate to provide sufficient in- formation for evaluating mutagenic risks. Much of our knowledge of genetic risks to hu- man health from exposures to environmental agents has been derived from the study of the ef- fects of mutagens on experimental animals. These experiments are useful in manipulating various aspects of the mutagenic process, for example, to examine how mutagens act on DNA and to study effects of varying doses and rates of exposure to mutagenic agents administered either singly or in combination. Experimentation with animals is es- sential for assessing potential hazards of new chemical and physical agents before human pop- ulations have been exposed to them. However, at present, technical problems in detecting and measuring mutations limit animal experiments as they limit human studies, so the results from ani- mal experiments, using current methods, detect only a small proportion of the kinds and num- bers of mutations that can occur. Recent advances in molecular biology have led to the development of new technologies for exam- ining DNA that may provide insight into the kinds and rates of mutations that occur in human be- ings. This report assesses these developments and discusses their potential for predicting risks of mu- tation from particular exposures. At present, these new technologies propose rea- sonable and verifiable ways of detecting herita- ble mutations in human DNA and proteins, but none is efficient enough to be used on a large scale. However, there is considerable optimism in the scientific community that these new technologies can provide, for the first time, the means to ob- tain basic knowledge about the primary causes of mutation and the means to assess the kinds and rates of mutations that occur in human beings. Data derived from studies in human beings, along with verifiable methods to extrapolate from corresponding animal data, will permit a more in— formed assessment of the medical and biological 4 - Technologies for Detecting Heritable Mutations in Human Beings consequences of mutagenic exposures. At present, without such comparative data, it is difficult to know whether general extrapolations from ani- mal data would lead to underestimates or over- estimates of the genetic risks for humans. Con— tinuing to rely on inadequate data may incur both human and financial costs, since conclusions drawn from this information contribute to deci- sions about acceptable levels of exposure and the level of society's resources that are devoted to pro- viding protection from such exposures. A combination of factors—concern that envi- ronmental exposures may be contributing to hu— man mutations, questions about the fundamen- tal nature of mutations, and increasing knowledge of the structure and function of DNA—increase the likelihood that new technologies will be de- veloped and field tested. However, studies using these technologies may be expensive and will probably require the collaboration of a large num— ber of scientists; their continued development, pi- lot testing, and large-scale application may require sufficient interest and financial support outside the scientific community. With such support, and with continued development of the techniques, some of these techniques could be ready for large- scale use in the next 5 to 10 years. Congressional interest in supporting basic re- search on human mutations and in the continued development of these technologies is necessary if the regulatory agencies are eventually to have the tools to evaluate risks from most occupational or environmental exposures. The current lack of in- formation on kinds and rates of human mutations is largely due to the inadequacy of present meth- ods to study heritable mutations. Efforts to com- ply with the agencies' mandates to protect peo- ple from mutagens may be impeded unless basic knowledge of causes, kinds, and rates of human mutations is obtained. Request for the Assessment This assessment was requested by the Senate Committee on Veterans' Affairs, the House Com- mittee on Science and Technology, and the House Committee on Energy and Commerce. Interest in the assessment was also expressed by the Senate Committee on Public Works and the Environ- ment, the Senate Committee on Labor and Hu- man Resources, and the House Committee on Vet- erans' Affairs. These committees have wrestled with problems of determining whether past exposures to poten- tial mutagens have affected the health of veterans and civilians and of framing reasonable public health laws that can be implemented, given cur- rent knowledge and technologies. OTA was asked to assess the available information about current means for detecting mutations as they relate to these issues and on the likelihood and potential impact of technological developments. Scope of the Report This chapter summarizes current knowledge about the kinds and rates of human mutations and the methods that have been used to detect herita- ble mutations in human beings and in experi- mental animals. New technologies assessed in this report for detecting and measuring human herita- ble mutations are briefly described. Methods for measuring human somatic mutations are discussed as tools for evaluating the risks of heritable mu— tations. The final section of this chapter presents options for congressional action. Chapter 2 provides background information about human genetics and DNA, and discusses the types of mutations that can occur and their potential health effects. Chapter 3 reviews the literature on current methods for studying muta- tions and summarizes current knowledge about the frequency of heritable mutations in human populations. The new technologies for examining human DNA for heritable mutations are described in chapter 4, followed by descriptions of new so- matic mutation tests in chapter 5. Chapter 6 summarizes data from experimental animals on spontaneous and induced mutations, and discusses the possible use of such data for identifying human mutagens and determining their potency. Chapter 7 focuses on the problems of extrapolating from the results of animal exper- iments to human risks. Chapter 8 discusses epidemiologic considera- tions in the application of the new technologies, Ch. 1—Summary and Options 0 5 such as validation of the new methods and selec- tion of at-risk populations to study. Chapter 9 dis- cusses Federal involvement in protecting against genetic risks and the regulatory mechanisms avail- able to control exposures to mutagenic agents. BACKGROUND Kinds and Effects of Mutations in Human Beings Mutations can occur ”spontaneously,” that is, in the apparent absence of any unusual stimuli, or they can be “induced” by particular agents. It is likely that many or most “spontaneous” muta— tions are caused by external forces, possibly in- cluding ionizing radiation, ultraviolet radiation, viruses, and certain chemicals, but the appropri— ate links have not been made. Some mutagens present around us may also be necessary for sus- taining life, for example, oxygen, components of our food, and some of the body's own metabo- lites. Experiments in animals have shown that many substances present in agricultural, indus- trial, and pharmaceutical chemicals in use today are mutagenic in some test systems. Which of these cause mutations in human beings is still a matter of speculation. Precise causes for essen- tially all mutations that have been identified in human beings are unknown. At present, more than 3,000 different genetic diseases have been identified, including disorders resulting from mutations in DNA, and disorders resulting from the interaction of genetic and envi- ronmental components. Approximately 10 in 1,000 liveborn infants are born with a single gene disorder and an additional 6 in 1,0001iveborn in— fants are born with a major chromosome abnor- mality. It is estimated that approximately 80 per- cent of the single gene disorders are the direct result of mutations that occurred in germ cells of distant ancestors and were passed along to suc- ceeding generations. The remaining 20 percent of these cases (0.2 percent of all livebirths) and the majority of chromosome abnormalities are be- lieved to be due to sporadic mutations in the re- productive cells of one of the parents of the in- fant. An additional 10 in 1,000 liveborn infants A summary of current Federal expenditures in mutation research and potential costs of studies to detect mutations using the new technologies is presented in appendix A. manifest a serious genetic disease sometime after birth, and a far higher proportion of newborns will show indirect effects of one or several parental or ancestral mutations in later life as, for exam- ple, in increased susceptibilities to some forms of heart disease, diabetes, or cancer. Mutations are changes in the composition of the genetic material, DNA (see fig. 1), and are gen- erally divided according to size into gene muta- tions and chromosome mutations. Gene muta- tions refer to changes within a single gene, for example, substitutions of single component nu- cleotides, or small losses or additions of genetic material in expressed or nonexpressed regions of the gene. Chromosome mutations affect larger portions of the chromosome (e.g., structural rear- rangements of genetic material in the chromo- somes) or result in the loss or addition of an en— tire chromosome. Since DNA directs the synthesis and regulation of molecules in the body, either group of mutations can influence a wide range of biological and physiological functions, including reproduction, longevity, intelligence, and physi- cal development. Individual differences in suscep- tibility to disease may result from the effects of one gene, several genes, or combinations of genes and environmental factors. Depending on the nature and location of the mutations and on the function of the genes in which they occur, mutations may, in theory, be beneficial, neutral, or harmful to the individual.1 The kinds and effects of known mutational events range from single nucleotide substitutions (the smallest unit of change in DNA) with no clini- cally observable effects, to single nucleotide sub- 1Current theory maintains that most newly arising mutations in regions of DNA that directly determine the structure and regula- tion of proteins are more likely to be detrimental than beneficial. Little is known of effects of mutations in other regions of DNA. 6 0 Technologies for Detecting Heritable Mutations in Human Beings Figure 1.—0rganizational Hierarchy of DNA, the Carrier of Genetic Information in Human Cells Nucleotides Adenine, guanine, cytosine, and thymine, the basic building blocks of DNA. Codons Nucleotides arrang- ed in a triplet code, each corresponding to an amino acid (components of pro- teins) or to a regulatory signal. AAA CGC GAC CGA Genes Functional units of DNA needed to syn- thesize proteins or regulate cell func- tion. -ACGAAAATCCGCGCTTCAGATACCTTA— Chromosome Thousands of genes arranged in a linear sequence, consist- ing of a complex of DNA and proteins. " ~ Genome 3' §\! ,9 ‘§ The complete set oi fl\ genetic Informatlon, “ u 3‘ $6 each human repro- ,5 4 ' " f ductive cell contains .. X 9“ 23 chromosomes, ' / ¥¥ and all other cells in the body contain a A A \q‘fl full setof46 7‘ I ‘ chromosomes. SOURCE: Oilice of Technology Assessment. stitutions resulting in severe diseases; from major structural and numerical chromosome abnormal- ities (the largest observed unit of change in DNA) leading to various abnormalities and impairments, to those resulting in embryonic, fetal, or neona- tal death. A child's entire genetic endowment comes from the DNA of two single reproductive cells (or ga— metes), one egg and one sperm, from his parents.2 A mutation occurring in the DNA of either of these germ cells, a germinal mutation, is passed on to the child, who is born with a ”new" herita- bIe mutation. Mutations in germ cells that are not involved in fertilization are not passed on to the offspring. If a mutation arises in the DNA of the parents’ nonreproductive cells (collectively termed somatic cells), such somatic mutations are not transferred to the reproductive cells. Somatic mu- tations may, however, affect the parents’ health, and indirectly, their ability to bear a healthy child. 1Each of the gametes (the male spermatozoan and the female ovum) is the product of a series of developmental stages of repro- ductive (or germ) cells. CURRENT METHODS FOR STUDYING MUTATIONS Current empirical methods to study mutations in human beings focus on physiological and bio- chemical effects of mutations because, until re— cently, it was not possible to examine changes in DNA directly. Each of the current methods de- tects only a limited portion of the spectrum of mutational changes. These methods have been used to derive estimates of baseline frequencies of some kinds of human mutations. Animal Studies Much of our knowledge about how substances interact with DNA and how they may cause mu- tations is derived from studies with experimental animals. In addition, some estimates of human mutation rates have been derived, by extrapola- tion, from animal studies. Beginning soon after World War II and still continuing, spontaneous and induced gene mutation rates have been stud- ied in laboratory mice. The rate of spontaneous heritable gene mutations, as detected in these ex- periments, is roughly two to eight mutations per 1 million genes per generation of mice. Radiation and approximately two-thirds of about 20 chem- icals3 tested so far increase the frequency of de- 3On the basis of animal experiments using the specific locus test and the heritable translocation test, in which they were found to be mutagenic in all germ cell stages, these chemicals are strongly suspected to be mutagenic in humans (154). They included both com— mon environmental agents and chemicals available in the labora- tory that are not normally found in the environment. Ch. 1—Summary and Options 0 7 tectable heritable gene mutations in these experi- mental mice. Generally, the chemical substances that induce heritable mutations in mature and maturing germ cells also induce somatic muta— tions, and vice versa. However, substances that induce somatic mutations do not necessarily in- duce heritable mutations in immature germ cells (stem cells). Unlike mature germ cells or somatic cells, these germinal stem cells may have efficient systems for repair of mutational or premutational damage. There is, in fact, indirect evidence from dose-response and dose-rate experiments that ger- minal stem cells have good repair systems. It may be useful to quantify relationships be- tween somatic and germinal cells with regard to mutagenic potency of different chemicals, and to study mutations in equivalent sets of genes in both types of cells. Animal experimentation is useful for determining the feasibility of using human so- matic mutation rates for predicting the risk of hu- man heritable mutations. This work in animals may demonstrate whether it is possible to extrap- olate from the occurrence of somatic to that of heritable mutations at all, and may help deter- mine whether it is possible to generalize from ani— mal data to human beings. In addition, animal studies on heritable mutations are useful not only for determining whether a given agent is muta- genic but also for more general explorations of the factors that may influence the occurrence of mutations. Studies in Human Beings Spontaneous heritable mutations in human be— ings have been studied by examining: 1) the inci- dence of certain genetic diseases ("sentinel pheno- types”), 2) gross changes in chromosome structure or number, and 3) changes in the structure or function of blood proteins. Epidemiologic studies of specific populations, in particular, the survivors of atomic bombs in Japan, provide some infor- mation about induced heritable mutations in hu- man beings. Sentinel Phenotypes The classic method for identifying human herit- able mutations is the empirical observation of in- fants and children for the presence of certain rare genetic diseases known as sentinel phenotypes. Ex— amples include achondroplasia (dwarfism), aniridia, and some childhood cancers, such as retinoblas- toma and Wilms' tumor. By recording the occur- rence of sentinel phenotypes as a proportion of the total number of livebirths in a defined popu- lation over time, the frequency of each disorder (and of its corresponding mutation) can be esti- mated (165). The characteristic of sentinel phenotypes that is most useful for mutation studies is that these conditions are ”sporadic” in most or all cases; they almost always result from a new germinal muta— tion in one of the parents of the afflicted individ- ual.4 Each different sentinel phenotype is thought to result from a different, single, mutant gene, al- though precise genetic information to confirm the single gene hypothesis is lacking in most cases. Despite the distinctive characteristics of sentinel phenotypes, the relevance of existing data on the frequency of the various sentinel phenotypes to the study of kinds and rates of human mutations is limited by a lack of knowledge of the genetic bases of the phenotypes, and by the small frac- tion of DNA that accounts for these phenotypes. Of the several thousand known genetic diseases, only 40 are thought to satisfy the criteria for in- clusion as ”sentinel phenotypes." Roughly 40 genes are involved in the 40 sentinel phenotypes, among a total of an estimated 50,000 to 100,000 expressed genes in an individual’s DNA. Sentinel phenotypes are severely debilitating conditions that require accurate diagnosis and long-term medical care. However, practical dif- ficulties arise in gathering and maintaining data on the incidence of sentinel phenotypes for the purpose of tracking mutation rates. Infants with sentinel phenotypes are rare, numbering approx- imately 1 in 10,000 to 1 in 10 million liveborn infants, depending on the particular disease. Con- sequently, a huge number of infants must be ob- served in order to find even a few infants with sentinel phenotypes. Diagnosis of individual phe— notypes is complicated by the genetic heterogeneity of these disorders, so that highly trained special- ists in various pediatric subdisciplines, which are ‘Various tests are done to exclude nonsporadic cases, which could result from X-linked recessive inheritance, mistaken parentage, and the occurrence of other genetic or nongenetic conditions that mimic the appearance of sentinel phenotypes. 8 - Technologies for Detecting Heritable Mutations in Human Beings few in number, would be needed to make these observations. Millions of consecutive newborn in- fants would have to be monitored thoroughly and accurately for many years, and registries, much larger than those currently in use, would be needed to collect and maintain the necessary data. Chromosome Abnormalities Another method for identifying a certain class of human heritable mutations is the examination of chromosomes under a light microscope for the presence of chromosome abnormalities (“cyto— genetic analysis”). Normal human DNA is orga- nized into 46 chromosomes (44 autosomes and 2 sex chromosomes), distinguishable by size, pro- portional shape, and staining pattern, or ”band- ing." Chromosome abnormalities are defined as either numerical (extra or missing whole chromo- some[s]) or structural (deletions, insertions, trans- locations, inversions, etc., of sections of chromo- somes). In total, chromosome abnormalities are estimated to occur in at least 5 percent of all hu— man conceptions. The majority of such concep- tuses are spontaneously aborted, but the few that survive comprise approximately 0.6 percent of all liveborn infants. The incidence of Down syn- drome is 1 in 650 at birth, making it the most com- mon chromosomal disorder in newborns. With the most advanced chromosome staining methods currently available, approximately 1,000 bands are distinguishable in one set of human chromosomes, although far fewer bands are pro- duced with routine methods. Mutations in DNA sometimes cause a change in the banding pattern, particularly if such mutations involve large sec- tions of a chromosome. With routine banding methods, there may be several hundred genes present in each visible band, and with higher reso- lution banding methods, a single band may con- tain about 100 genes. However, smaller muta- tions, from single nucleotide changes within genes on up to some deletions and insertions of entire genes, generally are not visible by any banding method. Measurement of Mutant Proteins Most of the available information on rates of spontaneous human mutations has been derived from studies of sentinel phenotypes and chromo- some abnormalities, but data on different kinds of mutations in humans is now emerging from studies of mutant proteins. In general, mutant proteins are more precise indicators of genetic damage than are clinical and cytogenetic obser— vations. Certain mutations in genes that determine the structure of proteins alter the chemical char- acteristics of the proteins, causing them to behave differently in separation and purification proce— dures. These differences suggest that a mutation has occurred because proteins are constructed according to blueprints in DNA, and changes in DNA can lead to the production of altered pro- teins. If the protein under study is an enzyme, a mutation within the gene that codes for it can alter, diminish, or eliminate the enzyme's bio— chemical activity. Operationally, mutant proteins are identified by taking samples of blood from each member of a “triad," including both parents and the child. Proteins are extracted from blood components, and the proteins are separated by electrophore- sis. Putative mutations are identified when a pro- tein from a child behaves differently from the cor- responding protein from both parents. One-Dimensional Separation of Proteins.—The technique most commonly used to study mutant proteins is electrophoresis, a method of separat- ing proteins on the basis of their electrical charges. The term ”electrophoretic variant,” or “electro- morph,” is used to describe a protein that behaves differently in electrophoresis from the correspond— ing protein found in the parents. Although one-dimensional electrophoretic anal- ysis of proteins is well established and can be improved by including functional assays for ad- ditional enzymes, it is limited to detecting: 1) mu- tations that do not eliminate the functional abil- ity of the proteins, 2) nucleotide substitutions only in coding regions of genes for the proteins exam- ined, and 3) only those nucleotide substitutions that alter the electrical charge on these proteins; such substitutions are thought to account for about one-third of all nucleotide substitutions in coding regions, which, in turn, account for a frac- tion of all the kinds of mutations that can occur. Electrophoresis does not detect many other types of mutations, including small duplications, re- arrangements, or mutations that result in the Ch. 1—Summary and Options v 9 absence of gene products; these mutations are thought to constitute the majority of the muta- tions induced by certain mutagens, including ra- diation. Moreover, electrophoresis does not de- tect mutations that occur anywhere outside the coding regions of a certain set of genes, includ— ing mutations in other coding genes and in non- expressed regions of the DNA. Data from several studies using one-dimen- sional electrophoresis have been used to estimate the rate at which mutations produce electropho- retic variants, and from this estimate, to infer the total rate of amino acid substitutions in proteins, and the corresponding mutation rate per nucleo- tide in human DNA. Two-Dimensional Separation of Proteins.—An extension of one-dimensional electrophoresis in— volves separation of proteins in a second dimen- sion. With two-dimensional electrophoresis, about 300 proteins from each person can be separated and examined, compared with about 100 proteins per sample that can be separated in one-dimen— sional electrophoresis. Further improvements may be possible with the use of computer algorithms to assist in interpreting the complex two-dimen— sional gels. This technique is currently feasible, and although it detects the same types of muta- tions as one-dimensional electrophoresis, it can examine more proteins per sample. Epidemiologic Studies An extensive body of data from experimental animals demonstrates that exposure to radiation and to certain chemicals can induce mutations in mammalian germ cells. In humans, exposure to ionizing radiation is known to cause somatic mu- tations, and it is suspected to enhance the prob- ability of heritable mutations. To date, however, the availabIe methods have provided no direct evi- dence for the induction (by Chemicals or by radi- ation) of mutations in human germ ceIIs. The single largest population studied for in- duced mutations is the group of survivors of the atomic bombs detonated in Hiroshima and Naga- saki in 1945. Many survivors of the bombs re- ceived doses of radiation that could have caused germinal mutations; in experiments with muta- tion induction by radiation in mammals, similar kinds and doses of radiation were sufficient to cause observable mutations in offspring. There- fore, it was assumed that germinal mutations could have been induced in people exposed to the radiation from the blasts. Medical examinations of the survivors soon af- ter the blasts revealed the immediate effects of whole body irradiation: loss of hair; reduction in bone marrow activity; and reduction in circulat- ing white blood cells, associated with a reduction in the body’s resistance to infection. Among those who recovered from the immediate effects of the radiation, there was a significant excess of can- cer deaths later in life. Certain types of leukemia were the first cancers to appear in excess, but con- tinuing followup has revealed later increases in other cancers, such as multiple myeloma and cancers of the breast, thyroid, colon, esophagus, stomach, lung, ovaries, and possibly of the spi- nal cord and nerves (24,55). Exposure of pregnant women to radiation was found to be associated with an increased incidence in their liveborn infants of small head circumfer- ence, mental retardation, and an increased inci- dence of childhood cancers. The critical time for fetal brain damage from radiation exposure was identified as the period of 8 to 15 weeks of gesta- tion (99). Analysis of the chromosomes prepared from peripheral blood lymphocytes of survivors ex- posed to the radiation has indicated an excess of chromosome aberrations (7). Certain types of chromosome aberrations (mainly balanced struc- tural rearrangements, such as reciprocal translo- cations and inversions) have been found to per— sist in circulating lymphocytes long after exposure to radiation, whereas other types of chromosome aberrations (e.g., unbalanced rearrangements) in lymphocytes declined in number soon after ex- posure. Overall, the frequency of chromosomally aberrant cells in the survivors' blood was found to be proportional to the estimated dose of radi- ation received at the time of the bombing. It has not yet known whether these somatic mutations are correlated with specific cancers or other dis- eases in the survivors. Survivors’ children who were conceived after the acute radiation exposure were examined to 10 0 Technologies for Detecting Heritable Mutations in Human Beings study mutagenic effects on the parents’ reproduc— tive cells. Using various methods available from 1945 to the present, survivors' offspring were stud- ied for “untoward pregnancy outcomes,"5 for cer- tain chromosome abnormalities, or most recently, for abnormal blood proteins. The offspring of par- ents exposed to atomic radiation were compared with the offspring of parents who were beyond the zone of radiation (greater than 2,500 meters from the hypocenter at the time of the bombings). Observation and analysis of some 70,000 off— spring has revealed no statistically significant ex- cess in the incidence of stillbirths, congenital mal- formations, neonatal deaths, or chromosome abnormalities. These findings suggest that the fre- quency of radiation-induced germinal mutations that led to certain gross abnormalities in newborn infants was not high enough to be detectable in a population of that size and genetic heter— ogeneity. However, they do not rule out the pos- sibility of other manifestations of genetic damage in these children, or of latent expressions of such damage, since the methods used to study this pop— ulation could examine only a small subset of DNA and only a limited number of genetic endpoints. Analysis of the children’s blood proteins for electrophoretic variants was later done to detect recessive mutations, that is, mutations that are not expressed as disease (unless present in both 5These were defined as major congenital defects and/or stillbirths and/or death in the survivors' offspring during the first postnatal week. These abnormalities can be caused by exposure to radiation, as well as to other environmental agents, and by socioeconomic factors. copies of a particular gene). This analysis, begun in 1976, found few mutations in either the exposed or control groups, making interpretation problem- atic. While the results indicate no significant ex- cess of mutant proteins in the children of exposed parents, they do not exclude the possibility that an excess exists undetected. Unfortunately, elec- trophoresis is inefficient at detecting deletions, one of the most likely types of radiation-induced mu— tations. Overall, these findings do not rule out the possibility of genetic damage to the offspring of survivors of the atomic bombs, but they put upper limits of the frequency of occurrence of cer- tain types of mutations that the current methods are able to detect. Taking these findings at face value, and cog- nizant of an enormous body of data on the genetic effects of radiation on experimental animals, the investigators suggest that the dose of radiation necessary to double the human mutation rate (the ”doubling dose”) was between 139 and 258 rem,° but they caution that there is a possibility of large error attached to that estimate since genes other than the ones sampled may demonstrate differ- ent sensitivity to radiation, and some types of mu- tations may be repaired more efficiently than others. Their estimate of the doubling dose, if cor— rect, however, indicates that man could be con— siderably less sensitive to radiation than labora— tory mice. 6A rem (Roentgen—Equivalent—Man) is a measure of absorbed ra— diation dose. For comparison, a chest X-ray exposes an individual to about 0.1 rem. NEW TECHNOLOGIES FOR DETECTING HUMAN MUTATIONS Recent advances in molecular biology have led to the development of techniques that allow di- rect examination of DNA for evidence of muta— tions. These methods can examine large regions of human DNA without requiring detailed knowl— edge of the genes contained in those regions and without preparing genetic probes for particular sequences. Unlike current methods—observing sentinel phenotypes, chromosome abnormalities, and electrophoretic variants—that are limited to detecting a small fraction of all kinds and num- bers of mutations, the new techniques have the potential for detecting a wide, unselected spectrum of mutations across the DNA. These techniques are examples of the state-of—the-art in molecular genetics and they are now promising to provide the basis for better approaches to studying muta- genesis. Ch. I—Summary and Options 0 11 Applying recent developments in molecular bi- ology to the problem of detecting sporadic mu— tations, the new techniques described in this report propose reasonable and verifiable ways of exam- ining human DNA for alterations in sequence and structure. These developments include the abil- ity to clone specific genes, to cut up DNA into predictable fragments, to hybridize complemen- tary DNA strands, to detect less-than-perfect hybridizations due to single base pair changes, and to separate large fragments of human DNA. Some of these methods detect similar types of mutations and some complement each other by detecting dif- ferent types. Some of these may merit further de- velopment and, eventually, pilot testing. Several new technologies, representing different approaches, are discussed below. None of these new techniques has been applied to large human populations or to experimental animals and it is not known how well they will perform. At present, none of these techniques is approaching the efficiency needed for examining the kinds and rates of mutations in a population or for determining whether mutation rates are in- creasing. With technical improvements in effi- ciency, some of the techniques, or derivatives of them, could be available in the next 5 to 10 years for large—scale use. Since these technologies pro- vide new information about DNA, the health im- plications of any newfound mutations may not be immediately known. Additional research and methods would be needed to examine biological and physiological implications of the identified mutations for the populations studied and for their descendants. Detection and Measurement of Heritable Mutations Restriction Fragment Length Polymorphisms Restriction endonucleases are enzymes, isolated from bacteria, that can be used experimentally to cleave isolated DNA molecules into fragments at specific sites in the DNA sequence that they rec- ognize. If any of these sites have been altered by mutation, the resulting pattern of fragment sizes would also be altered. Restriction enzymes can be used to detect mutations that either: 1) create a new restriction site, 2) eliminate an old one, or 3) change the distance between existing restriction sites. (These mutations may include single nucleo- tide substitutions, or multiple nucleotide deletions or insertions.) To use this method to detect mutations, DNA would be treated with a set of restriction en— donucleases and the resulting DNA fragments sep- arated by electrophoresis and examined for differ- ences between those present in the child’s DNA and those in either parent’s DNA. Restriction site analysis does not allow examination of every nu- cleotide. However, the use of a set of combined restriction enzymes increases the number of re- striction sites identified, allowing a larger portion of the DNA to be examined, including both ex- pressed and nonexpressed regions. Genomic Sequencing The most straightforward approach to looking for mutations is by determining the sequence of every nucleotide in a child’s genome and then comparing this with the DNA sequences of the child's parents. To determine its sequence, human DNA is cut with restriction endonucleases into fragments, and then each fragment is analyzed for its sequence of nucleotides. Genomic sequencing would detect mutations regardless of where they occur—in regions that code for specific proteins and regulatory functions as well as in regions without known functions—and is therefore poten- tially very informative. While it is technically pos- sible at present, sequencing is currently feasible only for very limited sections of DNA, such as the length of DNA comprising only a few genes. Because of current technical inefficiencies, it would be an enormous task, involving many labora— tories, a large number of scientists, and at least several decades to sequence even one entire ge- nome, the complete set of DNA in an individual’s germ cell. At present, it is not feasible to use genomic sequencing to examine several peoples’ genomes for mutations, although sequencing can be used in conjunction with other techniques to examine small sections of DNA. One-Dimensional Denaturing Gradient Gel Electrophoresis A modification of the standard electrophoretic gel procedure, denaturing gel electrophoresis al- 12 - Technologies for Detecting Heritable Mutations in Human Beings lows DNA fragments to be separated not only on the basis of size, but also on the basis of sequence of nucleotides. Double-stranded DNA separates (“denatures”) into its constituent strands when it is heated or when it is exposed to denaturing chemicals. A gradient of increasing strength of such chemicals can be produced in an electro- phoretic gel so that DNA samples will travel in the direction of the electric current, separate by size, and begin to dissociate as they reach their particular critical concentration of denaturing chemical. Every unique strand of DNA dissociates at a different concentration of denaturant. In fact, a sequence difference of only one nucleotide be- tween two otherwise identical strands of double— stranded DNA is enough to cause the strands to dissociate at different concentrations of denatur- ant chemical, and to stop traveling at different locations in the gel. Using this technique, the par- ents' and child's DNA would be cut into fragments with restriction enzymes, dissociated into single stranded DNA, and reannealed with radioactively labeled probe DNA. The resulting heteroduplex fragments would then be separated on the basis of their DNA composition in a denaturing gra- dient gel. Mismatches between the sequences of probe and child's DNA and not between probe and parents' DNA would appear as different banding patterns on the gel. Again, a compari- son between the banding pattern of parents’ and child's DNA analyzed in this way may identify a wide range of mutations in all DNA regions. Two-Dimensional Denaturing Gradient Gel Electrophoresis Another approach to detecting mutations is a technique whereby sizing and denaturing gels are used to differentiate among DNA sequences com- mon to parents' and child’s DNA, polymorphisms in either parent’s DNA which are transmitted to the child, and any new mutations in the child’s DNA. Like the two-dimensional polyacrylamide gel procedure for protein separation described earlier, this approach compares locations of spots on a gel (in this case, DNA spots) for evidence of new mutations. In this method, various com- binations of parents' and child’s DNA are pro- duced and compared on the basis of the dena- turant concentration at which they dissociate. Parents’ and child’s DNAs are compared to each other, rather than to relatively small probes. This approach, which would allow detection of muta- tions in the complete genome, would detect differ- ences (”mismatches”) between a child’s DNA and his or her parents’ DNA. Such mismatches would represent various types of mutations in the nucleo- tide sequence in expressed and nonexpressed DNA regions. DNA-RNA Heteroduplex Analysis This technique hinges on the production of DNA bound to complementary strands of ribo- nucleic acid (RNA) or ”DNA-RNA heteroduplexes," and on the use of specific enzymes, such as “RNaseA,” that cleave the DNA at particular se- quences where the DNA and RNA are not per— fectly bound at every nucleotide. This is similar to the use of restriction enzymes, which cleave DNA at particular normal sequences, except that RNaseA cleaves the RNA strand in RNA/ DNA hybrid molecules where there are mismatched nucleotides, indicating mutations. The resulting fragments are then separated electrophoretically to detect differences between parents’ and child’s DNA. The efficiency of this approach depends on the number of differentwmismatches that can be recognized and cleaved. This method would de- tect nucleotide substitutions over a large portion of the DNA. Subtractive Hybridization Detecting mutations would be much easier if it were possible to ignore the majority of DNA sequences that are the same in parents and child and, instead, focus only on the few sequences that may be different. Subtractive hybridization is an idea for selecting and characterizing sequences in a child’s DNA that are different from either par- ent’s DNA. First, the double—stranded DNA of both par- ents is cut into fragments with restriction enzymes, dissociated into single-stranded DNA, and then mixed together with a set of single-stranded refer- ence DNA sequences. These reference sequences represent all possible sequences of 18 nucleotides (analogous to a dictionary of 18-letter words using Ch. 1—Summary and Options . 13 only 4 different letters.) Each reference sequence binds to its complementary sequence in the paren— tal DNA and can be removed from the mixture. Any reference sequences left over, not bound to parental DNA, represent sequences not found in the parents’ DNA. If the child’s DNA binds with any of these left-over sequences, such hybrids would indicate that the child’s DNA contains different sequences from those in the parents’ DNA. These hybrids could be separated and ana- lyzed for mutations. This approach is the least well developed of all the ones discussed in this report, and its feasibil- ity is unknown. If it does prove feasible, how- ever, this approach would identify short sequences containing mutations in any part of the DNA, al- lowing further detailed study (e.g., by DNA se- quencing) of the kinds of mutations that may occur. Pulsed Field Gel Electrophoresis If human DNA were short and simple, it could be cut up with restriction enzymes and separated electrophoretically into discrete bands, each rep- resenting a particular segment of the total DNA. However, human DNA is so long that when re— striction—digested DNA is electrophoretically sep- arated, the resulting fragments of the whole set of DNA form a continuous smear of bands. Even if the DNA could be cut into 100 or 200 fragments, the pieces would be too big to pass individually through the pores of a standard electrophoretic gel. A new technique, pulsed field gel electropho- resis is being developed to separate large frag— ments of human DNA and to examine such frag- ments for evidence of mutations. The procedure may detect submicroscopic chromosome muta- tions, including rearrangements, deletions, breaks, and transpositions. At the present time, however, the method cannot handle whole human chromo— somes, though it works well with fragments of human chromosomes and with smaller whole chromosomes from lower organisms. This tech— nique may be useful in detecting chromosome mu- tations that are intermediate in size between ma- jor rearrangements (observable by cytogenetic methods) and single base pair changes, potentially a large proportion of all possible mutations. Detection and Measurement of Somatic Cell Mutations The methods for detecting heritable mutations rely on comparing the DNA of parents with the DNA of their children to infer the kinds and rates of mutations that previously occurred in parents' reproductive cells. While this information is val— uable in learning about heritable mutations, it may come months or years after the mutations have actually occurred, and this temporal sepa- ration of events makes it difficult to draw asso- ciations between mutations and their causes. Tests to detect somatic mutations may be useful in sig- naling the probability of heritable mutations. Such tests may be useful in relating exposures to spe- cific mutagens with particular genetic events in the cell, and they may help to identify individ- uals and populations at high risk for mutations.7 It is thought that the mutation process is fun- damentally similar in germinal and somatic cells. If this is true, then it may be possible to predict the risk of germinal and heritable mutations on the basis of measurements of somatic mutations, which are inferred from the frequency of mutant cells. Several investigators are currently working on methods to relate the frequency of mutant cells to the number and kinds of underlying mutations. Several new techniques for detecting and meas- uring somatic mutations are described in this report. Mutant somatic cells may occur during growth and development and may appear as rarely as one in a million normal cells. Detection of somatic mutants requires methods for scanning through a million or so nonmutant cells to find a single variant cell. Two general approaches are used: 1) screening, which uses high-speed machin- ery to look at the total population of cells and either count or sort out the variants; and 2) selec- tion, in which a population of cells is cultured in the laboratory under conditions that permit the growth of variant cells and that restrict the growth of the majority of cells that are nonvariants. The 7Even without knowing the exact relationship between rates and kinds of somatic and heritable mutations, it is reasonable to pre- dict that people with high somatic mutation rates might be at higher risk for heritable mutations. either because of a particular environ— mental exposure or a genetic susceptibility to mutations. 14 o Technologies for Detecting Heritable Mutations in Human Beings new DNA technologies could be used to charac- terize the mutations in any such somatic variants found. Different mutagens have been shown to pro- duce distinguishable types of mutations (“muta- tional spectra") in human cells grown in culture. Determining mutational spectra may be useful in associating specific mutational changes in somatic cells with particular mutagenic agents to which individuals may be exposed. Such information could be useful in understanding the causes of mutational changes as well as in monitoring at- risk populations. Data on kinds and rates of somatic mutations may provide a monitor for exposure to mutagens and carcinogens, and are relevant to the study of carcinogenesis and of aging. However, measure- ments of somatic mutation rates per se have lit— tle direct applicability to intergenerational (or transmitted) effects, without corresponding infor- mation on heritable mutations. USE OF NEW TECHNOLOGIES IN RESEARCH AND PUBLIC POLICY Feasibility and Validity Testing The new technologies now range from ideas on paper to being in various stages of laboratory de- velopment, but none is yet ready for use in the field. A critical step before a technology is used in an investigation of a population thought to be at risk for mutations is that the technology be tested for validity and feasibility. To assure that a technology is a “valid" method, that is, that it detects the types of mutations that it theoretically is capable of detecting, and to characterize the de- gree to which it gives the ”right” answer, it will be necessary to test a technology in a series of vali- dation studies. A first step might be to test the technologies against pieces of DNA with known mutations, and at a later stage, in offspring of ani— mals exposed to known mutagens. Feasibility test- ing will be required to make sure the technology can be efficiently scaled up for analyzing large numbers of samples. Epidemiologic Activities If the value of new technologies for detecting mutations is to be realized, it will be as tools for determining rates and patterns of mutations in epi- demiologic studies of human beings. Once a tech- nology has successfully passed through validation and feasibility tests, it will become a candidate for use in three major types of epidemiologic activities: surveillance, monitoring, and ad hoc studies. Surveillance is a routine activity whose aim, in the context of this report, would be to measure the “baseline" rate of mutations in a defined pop- ulation over the course of time and to facilitate rapid recognition of changes in those rates. Mon- itoring consists of observations over time in a pop- ulation thought to be at increased risk of, in this case, heritable mutations, because of exposure to a known or suspected mutagen, for the purpose of helping the specific at—risk population in what— ever way is possible. People living around haz- ardous waste disposal sites have been monitored for endpoints other than mutations, e.g., cancer and birth defects, and they would be likely can- didates for mutation monitoring when technol— " ogies become available to do so efficiently. Ad hoc studies of a variety of designs are carried out i to test hypotheses about suspected causes of mu- tations. Ideally, the results of ad hoc studies can be generalized to populations other than those specifically studied. Extrapolation Making predictions from observations of cause and effect in one system to probable effects in another is one form of extrapolation. The proc- ess involves a set of assumptions in moving from one system to another. The practical importance of extrapolation for mutagenicity is, ideally, to be able to predict mutagenic effects on human be- ings from the response in laboratory animals or lower test systems. The ability to extrapolate to Ch. 1—Summary and Options 0 15 human responses addresses one of the major goals of public health protection, the ability to iden- tify substances harmful to human beings before anyone is exposed, thereby providing a rational basis for controlling exposure. Extrapolation can be qualitative or quantita— tive. Qualitative, also called biologic, extrapola- tion involves predicting the direction of a result, for example, if a chemical causes mutations in a laboratory test, can we also expect mutations in human beings? Quantitative extrapolation in- volves translating a quantitative result in an ani- mal test into a quantitative estimate of mutagenic risk in humans. Going a step further in extrapo- lation, can an estimate of mutagenic damage be translated into a measure of genetic disease? A number of theoretical models for extrapolat— ing mutagenic effects have been proposed, based on various parallel relationships. For instance, it might be true that if the relationship between so— matic and heritable effects in animals were known after exposure to a specific mutagen, and if one could measure a somatic effect in human beings who had been exposed to the same substance, a heritable effect in human beings could be pre- dicted, assuming the relationship between somatic and heritable effects is parallel in animals and hu- man beings. Because of the paucity of data, par— ticularly from human studies, it has been impos- sible to validate such an extrapolation model. The new technologies should allow a major increase in the database which, in turn, should allow re- searchers to more fully explore relationships among various types of test results. Regulation Congress has mandated public protection from mutagens in certain environmental health laws (e.g., TSCA and CERCLA), and other laws pro- vide mandates broad enough to empower agen- cies to take action against mutagens in virtually any appropriate situation. Except for radiation, however, very little regulatory evaluation has taken place on the subject of heritable mutations. This is directly related to the lack of sensitive methods to detect heritable mutations in human beings, and the related difficulty in extrapolating from results in nonhuman test systems to prob- able human responses. The Environmental Protection Agency (EPA) has recently issued ”Guidelines for Mutagenicity Risk Assessment." EPA's approach is relatively simple and pragmatic. It requires only the types of information that can be acquired with current technologies, but allows for information from new technologies, as they become available. The guidelines require evidence of: 1) mutagenic activ— ity from any of a variety of test systems, and 2) chemical interactions of the mutagen in the mam- malian gonad. Using a “weight-of—evidence" de- termination, the evidence is classified as ”suffi- cient,” ”suggestive,” or ”limited" for predicting mutagenic effects in human beings. Although chemicals have not generally been regulated as mutagens, it is probable that ex— posures to mutagens have been reduced by regu- lations for carcinogenicity. Strong evidence sup— ports the idea that a first stage in many cancers involves mutation in a somatic cell, and one of the most widely used screening tests for poten- tial carcinogens, the Ames test, is actually a test of mutagenicity. The extent to which people are protected against heritable mutations if their can- cer risk from a specific agent is minimized is at present unknown. The new technologies should greatly improve our ability to make that judgment. Federal Spending for Mutation Research OTA queried Federal research and regulatory agencies about their support of research directed at understanding human mutations. For fiscal year 1985, they reported about $14.3 million spent on development or applications of methods for de- tecting and/ or counting human somatic or herita- ble mutations. An estimated $207 million was spent in the broader category of related genetic research. 16 - Technologies for Detecting Heritable Mutations in Human Beings OPTIONS Research related to the new technologies de- scribed in this report has the dual aims of increas- ing the knowledge base in human genetics about the causes and effects of mutations, and produc- ing information that could be used to estimate mutagenic risks for the purpose of protecting pub- lic health. The pace and direction of research toward developing these methods and the qual— ity and efficiency of preliminary testing of meth- ods could be influenced by congressional and ex- ecutive branch actions and priorities. Integration of research in different test systems and progress in developing extrapolation models also can be influenced by actions now and in the near future. Continued progress in the development and ap- plication of new technologies will depend not only on support of the individuals and laboratories directly involved in this research, but also on work in other areas. Although not directly ad- dressed in this assessment, support of research in medical, human, mammalian, and molecular genetics will be essential to a full understanding of the mutation process. The options that follow are grouped in four sec- tions: 1) options to influence the development of new techologies, 2) options to address various aspects of feasibility and validity testing, 3) op- tions concerning the use of new technologies in field studies, and 4) options to encourage coordi- nation of research and validation of extrapolation models. Development of New Technologies The Department of Energy (DOE), several agencies of the Department of Health and Human Services, and EPA currently provide funding to researchers in independent and government lab- oratories for various research and development activities pertaining to human mutation research and the development of new technologies. Each agency, appropriately, proceeds down a slightly different path. OTA estimates that a total of about $14.3 million was spent on human mutation re- search in fiscal year 1985. Progress in this research could be speeded up by increased funding, though it is difficult to quantify an expected gain. It is clear also that less money spent on human muta- tion research will slow progress in laboratories al- ready engaged in this research, and could deter other scientists from pursuing research in this field. Option 1: Congress could assure that funding levels for human mutation research and closely related studies do not decline without the re— sponsible agencies assessing the impact of fund— ing cu ts on research progress. This requirement could be expressed in appropriation, authori- zation, or oversight activities. Of the several funding agencies, DOE has taken the lead in funding much of the research on new methods for mutation detection described in this report, and researchers at some of DOE’s National Laboratories are among the leaders in the re- search. DOE also is the agency responsible for funding U.S. participation in the Radiation Effects Research Foundation, the joint United States- Japan body that continues to study the health of Japanese atomic bomb survivors. Option 2: A lead agency for research related to detecting and characterizing mutations could be designated. The lead agency would be re- sponsible for tracking the development of new technologies, facilitating the interchange of in- formation among scientists developing the tech- nologies and those in related fields, and en- couraging and facilitating coordinated studies involving different subdisciplines. The lead agency also would keep Congress informed about activities in this area. DOE may be the logical choice for a lead agency. The types of activities that a lead agency might engage in are described below. These activities are important whether or not a lead agency is desig- nated, and Congress should consider directly en— couraging them if there is no lead agency. Tracking the Development of New Technologies All the new technologies require improvements in efficiency before they become useful tools for studying human beings. As research proceeds, some techniques will develop more quickly than Ch. 1—Summary and Options 0 17 others, some will be dropped, and some may change in character, altering the kinds of muta- tions they can detect. The lead agency would be responsible for keeping track of these devel- opments. It would be useful if the "tracking” responsi- bility could lead to actions on the part of the lead agency that would promote the rapid and efficient development of the technologies. At some point, the technologies will be ready for feasibility test- ing and eventually, field testing. It would help re- searchers to know the stages of development of various methods, and it could help agencies make decisions about funding studies using certain methods. In addition, the lead agency could as- sist researchers by anticipating needs that will be common to all research programs and by encour- aging efficient use of resources. Facilitating Information Exchange Among Researchers In 1984, DOE organized and funded a meeting that brought together for the first time many of the researchers involved in laboratory-based mu- tation research. This meeting is acknowledged among those who attended as a milestone for in- formation exchange and the generation of new ideas concerning detection of heritable mutations. In fact, the ideas for some of the new methods described in this report were born at that meet- ing. There is a continuing need for this type of information exchange. Keeping Congress Informed Congress has already directed regulatory agen- cies to reduce exposures to environmental agents that may cause mutations. As part of its oversight of both regulatory and research activities, Con- gress could benefit from up-to-date information about the development of various technologies. The lead agency could report in a specified man- ner, e.g., a brief annual report describing the level of current research, its goals, results of completed or ongoing work, and expected near—term devel- opments. This information could also be the ba- sis for informing the public about mutation re— search. This activity will continue to be valuable to Congress through later phases of development and application of new technologies. Feasibility and Validity Testing After technologies pass through a phase of de— velopment to improve their efficiency and to work out technical details, a period of feasibility and validity testing will be necessary before a tech- nology can rationally be used as a tool in a large- scale study of heritable mutations in human be- ings. There are some ways to make validity test- ing an efficient process. As an example, a “DNA library," a collection of known DNA sequences, particularly sequences carrying known mutations, could be established and maintained by one lab— oratory, which could make DNA sequences avail- able to all researchers developing mutation detec- tion technologies. This material could be used to determine whether a particular technology detects those mutations that it is designed to detect, anal— ogous to testing chemical procedures and equip— ment against known chemical “standards." At a more advanced stage, new technologies might also be tested in animal experiments with a selection of mutagens known to cause specific types of mu- tations. If a lead agency is designated, some of these options would logically be among that agency's responsibilities. If there is no lead agency, these functions could still be encouraged by Con- gress through oversight activities. Banking Biological Samples Biological samples, especially blood samples, are often collected during the course of medical examinations for people thought to be exposed to environmental or occupational agents. Imple- mentation of a plan to bank those samples would facilitate human mutation studies and other re- lated research. Examining stored samples is not nearly so dis— ruptive as collecting samples from a population. The very act of specifically collecting samples for a study of mutations would raise expectations that definitive information about risks would be forth- coming. Examining stored samples would avoid that human cost and could, at the same time, pro- vide a realistic test of new technologies before they are applied to people who are anxious about the effects of environmental exposures on their genes. Although stored samples offer many advan- tages, preparing human samples for storage and 18 - Technologies for Detecting Heritable Mutations in Human Beings maintaining the stored samples is a significant task. Currently the high cost of storing samples inhibits establishment of sample banks. There has been little research directly aimed at improved methods for storing biological samples. Option 3: Congress could encourage banking of biological material obtained from at—risk indi- viduals and their spouses and offspring, with the objective of studying somatic and herita- ble mutations from these stored samples before technologies are used directly in a study of an exposed population. A centralized samples bank available to all researchers would be the most efficient means of establishing a reposi— tory of sufficient size and scope for validity testing. Option 4.- Congress could encourage the appro- priate agencies (through the lead agency, if one is designated) to investigate the potential for im- proving technologies for storing biological sam- ples. Where possible, funding should be en- couraged for such improvements. This option is relevant to a broad spectrum of human health research, and collaboration among disciplines should be encouraged for determining the spe- cific storage needs of different lines of research. Field Studies Sometime in the next 5 to 10 years, it is likely that one or more of the technologies for study- ing human mutations will be brought to the point of readiness for epidemiologic studies of human beings thought to be at risk for mutations. Such a study could be initiated by an independent sci- entist or a group of scientists involved in the de- velopment of a new technology who have iden— tified an at-risk population; a new technology or technologies could be used in an environmental health investigation triggered by the discovery of a potentially at-risk population; or Congress could mandate a study of mutations in a specific popu- lation. Depending on the way a study is initiated, different agencies will be drawn in, and different funding mechanisms will come into play. Should a study be initiated by one or a group of investigators who submit a grant proposal for funding, the current system for review of research awards is probably appropriate for making sci— entific judgments about whether a method is ready for large-scale testing. From that point on, how— ever, there are significant differences between a proposal to use a new technology in a human pop- ulation and most other proposals. The major dif- ferences are size and money. Application of any promising technology will require that state-of- the-art methods in molecular biology be expanded from small-scale laboratory use to large-scale ex- amination of blood samples collected from hun- dreds or thousands of people. A study such as that described above would re— quire a significant commitment of funds over a period of years, and could account for a large per- centage of a granting agency's funds. The amount of money necessary can only be estimated, but it could amount to tens of millions of dollars, not a large amount in government spending, but large in comparison to most biomedical research grants, which usually range from less than $100,000 to about $500,000 annually. The necessary commit— ment of funds from any single agency, consider- ing current spending for this type of research, is likely to be impossible, no matter how worthwhile the proposed study. Option 5.- Congress could consider providing spe- cific add-on funding to finance an in vestigator- initiated realistic test of a promising method to study human heritable mutations. A study mandated by Congress or an agency— initiated study using new technologies will not necessarily undergo the rigorous review and crit- icism that a proposal submitted to a granting agency would. Congress already has some experi— ence in mandating epidemiologic studies. Con- cerns about cancer and birth defects convinced Congress to mandate studies of military veterans who were exposed to Agent Orange or who par- ticipated in atomic bomb tests. In addition, Con- gress has vested the Public Health Service with the authority to carry out a wide range of health investigations of exposures from toxic waste dumps through ”Superfund.” The role of the man- dated ”Agency for Toxic Substances and Disease Registry” (ATSDR) is to “effectuate and imple- ment the health related authorities" of Superfund. ATSDR, located within the Centers for Disease Control, is a logical place for new technologies to be used as exposed populations are identified through other provisions of Superfund. Ch. 1—Summary and Options - 19 An unusual array of health problems or an ap- parent excess of disease among people living around toxic waste sites could trigger a study by ATSDR. Such problems could also prompt local residents to press Congress for studies to deter- mine whether, among other things, the disorders had been caused by mutations, and whether these mutations could have resulted from chemicals in the waste site. Some scientific societies, such as the American Society of Human Genetics, could be asked to participate in a review of the feasibility of detect- ing possible health effects from environmental ex— posures. If the study is determined to be feasible, Congress and the public can be reassured that the study’s findings are likely to be useful in decision- making. Alternatively, a decision that no study is currently feasible would underline the impor- tance of developing and testing new methods. Re- viewing the feasibility of a study may be perceived as delaying an investigation unnecessarily. How- ever, performing a study that has little or no likeli- hood of answering questions about exposure and mutations has marginal value at best and would not serve the people who requested it. Option 6.- If Congress mandates a study of herita- ble mutations in an at—risk population using the new technologies, the mandate should include a feasibility assessment by a panel of experts before the study begins. Option 7: Congress could require agencies to plan for a rigorous outside review by a panel of ex- perts of any agency-initiated study using a new technology, before such a study can proceed. Integration of Animal and Human Studies Much of our current ability to estimate the ef— fects of various external agents on human beings is derived from animal studies. In the future, ani- mals will continue to be used to test for mutage- nicity and, ideally, to predict effects in human be— ings. Right now, the available methods and body of data provide an inadequate basis for making predictions from animal results to human effects. The new technologies described in this report for detecting heritable and somatic mutations can be applied in both animal systems and in human be- ings. This information and information from cur- rently available animal systems could be used in an integrated fashion to study relationships be- tween somatic, gerrninal, and heritable mutations, and to pursue the development of extrapolation models for predicting effects in human beings. The kind of integrated research necessary will not occur spontaneously if the current pattern holds. There are few researchers engaged in studies that integrate comparable information from differ- ent systems for the purpose of elucidating rela- tionships among such systems. Improvements in extrapolation from animal to human data and from somatic to heritable mutations await efforts to encourage and coordinate the appropriate re- search. Single laboratories or centers are unlikely to be able to perform all the different tests neces- sary to develop and test extrapolation models, making coordination among laboratories essential. The National Toxicology Program (NTP) is the center of Federal mutation research using experi- mental animals, funding research grants and con- tracts nationally and internationally. The National Center for Toxicological Research, a laboratory of the Food and Drug Administration that is part of NTP, has facilities and experience to carry out large-scale animal tests. EPA also has a genetic toxicology program, and it is actively pursuing development and application of methods for studying human somatic mutations. DOE’s Na- tional Laboratories have directed or carried out the majority of large-scale animal studies of mu- tation rates and mechanisms, and this experience, as well as the advanced technology that these lab- oratories have developed for sorting and study- ing chromosomes and cells, will be useful in im- proving methods for extrapolation. These three agencies could encourage the development of methods for linking information from animal studies to somatic and heritable mutation risks in humans. Option 8: Congress could encourage studies of so- matic, germinal, and heritable mutations in ex- perimental animals using both currently avail- able and new technologies. Further, research funding agencies should encourage animal studies directed at identifying the mechanisms of mutagenesis and elucidating relationships be- tween mutagenic potencies in animals and in human beings. Chapter2 Genetic Inheritance and Mutations Contents Page Function and Organization of DNA ........................................ 23 Protein Synthesis ...................................................... 25 Inheritance of Genetic Traits ............................................ 27 Kinds of Mutations ...................................................... 29 Gene Mutations ....................................................... 29 Chromosome Mutations ................................................ 29 Health Effects of Mutations ............................................... 30 Early Mortality and Morbidity .......................................... 31 Diseases During Adult Life ............................................. 31 Effects of Somatic Mutations ............................................ 31 Persistence of New Mutations in the Population ............................. 32 FIGURES Figure No. Page 2. The Structure of DNA ................................................ 24 3. Replication of DNA ................................................... 25 4. Diagrammatic Representation of Protein Synthesis According to the Genetic Instructions in DNA ........................................... 26 The Genetic Code .................................................... 27 5. 6. Chromosomal Translocation ........................................... 30 Chapter 2 Genetic Inheritance and Mutations ”Like begets like" is an expression that describes repeated observations in families. Except in rare instances, most children are normal and gener— ally resemble their biological parents, brothers, and sisters. In these cases, the genetic informa- tion has been passed intact from the parents to their children. This report is about the instances when some- thing goes wrong, when a genetic mistake is made; a mutation in the DNA of the parents' reproduc- tive cells, both of whom are physiologically nor- mal, is passed to a child who may or may not ap- pear normal, depending on the nature and severity of the mutation. All types of mutations, from the benign to the severe, are considered in this report. As an introduction to human genetics for non- specialists, this chapter summarizes some basic in- formation about normal functions of DNA, the various kinds of mutations that can occur, and health effects of mutations. More complete infor- mation can be found in a general reference such as Vogel and Motulsky (165). FUNCTION AND ORGANIZATION OF DNA Deoxyribonucleic acid, DNA, is the carrier and transmitter of genetic information. Its functions require that it is relatively stable and that it is able to produce identical copies of itself. Certain agents, such as some forms of radiation, viruses, and certain chemicals, alter DNA’s ability to . maintain these characteristics. Mutations, changes . in the composition of DNA, can occur as a resultx of these agents acting on DNA or on the systems in place to repair DNA damage. Human DNA can be thought of as an immense encyclopedia of genetic information; each person's DNA contains a unique compilation of roughly 3 billion nucleotides in each of its two chains, the exact sequence shared by no other person except an identical twin. The genetic information en- coded in DNA is contained in the nucleus of cells, and is packaged in units, or chromosomes, that consist of long twisted double strands of DNA sur- rounded by a complex of proteins. Each chromo- some contains thousands of genes, the functional units of DNA, which are ”read," or transcribed, by the cell so that genetic information can be used to make proteins or to regulate cell functions. It is thought that only a small proportion, approx— imately 1 to 10 percent, of human DNA is trans— lated into specific proteins. Functions of the non— translated majority of DNA are largely unknown. There are 23 pairs of chromosomes in all nucleated human cells except human reproductive cells (germ cells). The latter contain one of each pair of chromosomes, since two germ cells (one egg and one sperm) fuse at conception and re-create a full set of DNA in the offspring. DNA is composed of two chains of nucleotides bound together in a double helical structure (fig. 2). The backbones of the two chains, formed by sugar (deoxyn'bose) and phosphate molecules, are held to each other by hydrogen-bonded nitroge- nous bases: adenine, thymine, guanine, and cyto- sine, abbreviated A, T, G, and C, respectively. Each of the chains' units, or nucleotides, consist of a sugar, phosphate, and nitrogenous base. The linear sequence of nucleotides repeated in vari— ous combinations thousands of times along the DNA determines the particular genetic instruc- tions for the production of proteins and for the regulation of cell functions. A change in even a single nucleotide, e.g., sub— stituting G for T, among the 3 billion nucleotides in the DNA sequence, constitutes one type of mu- tation. Depending on where it occurs, such a mu- tation could be sufficient to cause significant pathological changes and dysfunction, or it could cause no impairment at all. In thalassemia, for example, a genetic disorder of hemoglobin syn— thesis, approximately 40 different “spelling er- 23 24 - Technologies for Detecting Heritable Mutations in Human Beings Figure 2.—The Structure of DNA Guanine Adenine 2a. The pairing of the four nitrogenous bases of DNA: Adenine(A) pairs with Thymine (T) 0\ Guanine (G) pairs with Cytosine (C) "0 sugar-phosphate backbone 2b. The four bases form the four letters in the alphabet of the genetic code. The sequence of the bases along the sugar—phosphate backbone encodes the genetic in formation. A«--T A---T A---T l T---A G--- Base pairs C---G. A-«T Sugar-phosphate T... backbone I---A "-G A---T A schematic diagram of the DNA double helix. A three-dimensional representation of the DNA double helix. 2c_ The DNA molecule is a double helix composed of two chains. The sugar-phosphate backbones twist around the out- side, with the paired bases on the inside serving to hold the chains together. SOURCE: Office of Technology Assessment. Ch. 2—Genetic Inheritance and Mutations - 25 rors," or (single or multiple) nucleotide changes, have been identified in a particular gene that speci- fies the production of globin, an essential constit- uent of hemoglobin. In complex diseases such as heart disease, susceptibility is thought to be in- fluenced by the interaction of environmental fac- tors such as diet, and genetic factors such as mu- tations in one or several genes. During DNA replication, each chain is used as a template to synthesize copies of the original DNA (fig. 3). New mutations are transmissible to daughter cells. To replicate, the two chains of the DNA double helix separate between nucleo- tides, and each is copied by a series of enzymes that insert a complementary nitrogenous base op- posite each base in the original strand, creating two identical copies of the original one. As a re- sult of specific pairing between nitrogenous bases, each chain builds its complementary strand using the single chain as a guide.1 Protein Synthesis The normal functioning of the human body— including the digestion of food; the production of muscle, hair, bone, and skin; and the function- ing of the brain and nervous system—is depen— dent on a well-ordered series of chemical re- actions. Proteins, especially the enzymes, promote the chemical reactions on which these processes are based. Each enzyme has a specific function; it recognizes a chemical, attaches to it, reacts with it, alters it, and leaves, ready to promote the same reaction again with another chemical. To synthesize a protein, the genetic informa- tion contained in one or more genes along the DNA is transcribed to small pieces of ribonucleic acid (RNA), which are faithful replicas of one strand of DNA (see fig. 4). Each strand of RNA moves out of the nucleus to the cytoplasm of the cell, where it serves as a template for protein synthesis. Proteins are composed of hundreds of linked subunits called amino acids. At the ribo— somes, RNA is used to gather and link amino acids in a specific sequence and length according to the design specified in the DNA to form the different proteins. 1The molecular structure of the nitrogenous bases in DNA requires that A pairs only with T (or with U in RNA) and G pairs only with C. Figure 3.—Replication of DNA Old Old Old New New Old When DNA replicates, the original strands unwind and serve as templates for the building of new complementary strands. The daughter molecules are exact copies of the parent, with each having one of the parent strands. SOURCE: Office of Technology Assessment. Genetic information in DNA is organized in a triplet code, or sequence of three nucleotides. Each triplet specifies a complementary codon in “mes- senger RNA” (mRNA) which, in turn, specifies a particular amino acid. The sequence of such triplets in a gene ultimately determines the amino acid sequence of the corresponding protein. Codons exist for each of the 20 amino acids that make up the myriad of different proteins in the body. Additional codons exist to signal “start” and 26 0 Technologies for Detecting Heritable Mutations in Human Beings Figure 4.—Diagrammatic Representation of Protein Synthesis According to the Genetic Instructions in DNA 1. TRANSCRIPTION 5, l 3/ RNA UUAAGAGCGUGUGGGAAGC l Remove introns, ds 2. RNA PROCESSING process en W: U U U G A G G U G A A G Transfer to Nucleus cytoplasm Nuclear membrane Cytoplasm , 3, Mature 5 messenger UUUGAGGUGAAG RNA (mRNM 3. TRANSLATION mFlNA 5/ l 3, transfer RNA 4 URNA) Valine Ribosome Protein polypeptide chain composed of linked amino acids SOURCE: Adapted from AAE.H. Emery, An introduction to Recombinant DNA (Chichester: John Wiley & Sons, 1984). Ch. 2—Genetic Inheritance and Mutations - 27 Figure 5.—The Genetic Code Codon Amino Acid Codon ' Amino Acid Codon Amino Acid Codon Amino Acid Each codon or trl let of nucleotides UUU Phenylalanine UCU Serlne UAU Tyrosine UGU Cysteine In RNA codes'for anpamlno acid (AA). UUC Phenylalanine UCC Serine UAC Tyrosine UGC Cysteine Twenty’dlfferent amino acids are pro- UUA “WW ”CA s°"“° ”M "W ”GA 3‘” duced from a total of 64 different FINA UUG Leucine UCG Serine UAG stop UGG Tryptophan codons, but some amino acids are llied m re hen o n e. ., cuu Leucine ccu Proiine CAU Histidine ceu Arginine gfigfiymafiige I; 83°C.,,eg°b§°gfiu (mg CUC Leucine CCC Proilne CAC Histidine CGC Arglnlne by UUC). In addition one codon (AUG) CUA Leucine CCA Prelim ' CAA Glutamine CGA Arginine specifies the “start", of a protein and 3 Gus Leuclne CCG Proilne GAG Glutamine CGG Arginlne COdOl’IS (UAA, UAG, and UGA) Snglfy termination t a r tein. M all ns in AUU isoleucine ACU Threonlne AAU Aeparaglne AGU Serine the nucleotide sezgence can: choange AUG isoleuclne ACC Threonlne AAC Aeparaglne AGC Serlne the resultlng protein structure it the AUA leoleuclne ACA Threonine AAA Lysine AGA Arginine mutation alters the amino acid speci. AUG Methionine ACG Threonine AAG Lysine AGG Arginlne fled by a triplet codon or If It alters the ("a") reading frame by deleting or adding 3 cl ti . GUU Vaiine GCU Alanine GAU Aepartic acid GGU Glycine "U 90 de GUC Veline ccc Alanine GAG Aspartlc acid 666 Glycine, U - uracil (thymine) A - adenine GUA Valine GCA Alanine GAA Glutamlc acid GGA Glycine C - cytosine G - guanine GUG Vaiine GCG Alanine GAG Glutamic acid GGG Glycine SOURCE: Ollice of Technology Assessment and National Institute of General Medical Sciences. “stop” in the construction of a polypeptide (pro- tein) chain. The triplet code provides more infor- mation, however, than is needed for 20 amino acids; all except 2 of the amino acids are speci- fied by more than one codon (see fig. 5). Inheritance of Genetic Traits A child’s entire genetic endowment comes from the DNA in a single sperm and egg that fuse at conception. The fertilized egg then divides, be- coming a multicellular embryo, and the cells differentiate into specialized tissues and organs during embryonic development. In terms of genetic material, there are two general types of cells: somatic cells and germinal cells. Those that have a full set of DNA at some time in their de- velopment and do not participate in the transmis- sion of genetic material to future generations are somatic cells, which include all cells in the body except the reproductive cells. Germ cells include the gametes (egg and sperm, each of which con- tains half of the total set of DNA) and the cell types from which the gametes arise. A normal, full set of human DNA in somatic cells consists of 46 chromosomes, or 23 pairs of homologous chromosomes: 22 pairs of ”auto- somes" and 1 pair of sex—determining chromo- somes. Females have 22 pairs of autosomes and two X chromosomes, while males have 22 pairs of autosomes and an X and a Y chromosome. Re— productive cells, with half the normal set of DNA, are produced by reduction division, or meiosis, from sex-cell progenitor cells in the female ovary or male testis. These progenitor cells are set aside early in fetal development and are destined to be the only source of germ cells. Each ovum or sperm has 23 single, unpaired chromosomes: ova have 22 autosomes and one X chromosome; sperm have 22 autosomes and one X or one Y chro- mosome. Most genes are present in two copies, one on each member of a homologous pair of chromo- somes. For instance, if gene A is found on chro- mosome 1 it will be present at the same location or locus on both copies of chromosome 1. The exact nature of the DNA sequence of gene A may differ between the two chromosomes, and the word allele, classified as either normal or mutant, is used to refer to different forms of the same gene. Many different mutant alleles are possible, each one defined by a particular change in the DNA sequence. A mutation that is not expressed when a normal gene is present at the same locus on the sister chromosome is called recessive and a mu— tation that is expressed even in the presence of a normal gene is called dominant. 28 - Technologies for Detecting Her/table Mutations in Human Beings A Normal Karyotype of 46 Human Chromosomes ii it it t 3 4 5 o'- O‘- O..." >”———-'*~» ‘1 ~.-. .0... \ nut-fl .3 ‘~ ’ ‘— .'~ III..- 6 I QC I) .II. 12 . ' o V ' ll it it if 13 14 IS 1g 17 18 D E ii if .2. n I , I9 20 2| 22 F G X Y Photo credlt: Gail Stetten, Johns Hopkins Hospital Karotypes typically show all the chromosomes of a single cell, and are used mainly to identify abnormalities in the number or structure of chromosomes. Chromosomes can be isolated from any nucleated cell type in the body, but are most often isolated from white blood cells derived from a sample of venous blood or from amniotic fluid cells obtained at amniocentesis. The cells are stimulated to grow and divide in the laboratory, yielding a large number of cells which contain a sufficient amount of DNA to isolate and examine. The cells are then blocked in their dividing phase by the addition of a chemical, such as colcemid, to their growth media. After harvesting the cells, they are treated with a hypotonic solution to cause them to swell, and then with a chemical fixative to maintain this fragile, swollen state. When small amounts of the cell suspension are carefully dropped onto microscope slides, the cells break open, spreading out their chromosomes. To visualize the chromosomes and to distinguish the different homologous pairs, the chromosomes are stained to produce a characteristic banding pattern for each of the 23 pairs of homologous chromosomes. The chromosomes can then be photographed under the microscope, cutout from the final print, and numbered and arranged in order of decreasing size as shown in this karyotype. Ch. 2—Genetic Inheritance and Mutations 0 29 KINDS OF MUTATIONS Mutations are changes in the composition of DNA. They may or may not be manifested in out- ward appearances. Some mutations, depending on where they occur in the DNA, have no appar- ent effects at all, and some cause changes that are detectable, but are without obvious impairment. Still other mutations lead to profound effects on human health and behavior: stillbirths, neonatal and early childhood deaths, various diseases, severe physical impairments, and mental defi- ciencies. If mutations occur in genes that determine the structure or regulation of essential proteins, the results are often detrimental to health. Since most DNA in human cells is not expressed or used in any known way, mutations in these regions may or may not be clinically apparent, although it is possible that subtle physiological variations may occur. Interactions among several such mutations may result in unusual responses to environmental stimuli or they may influence susceptibility to various chronic diseases. Mutations can be generally divided according to size into two groups: gene mutations and chro- mosome mutations. Chromosome mutations in- clude numerical and structural abnormalities. Gene Mutations Gene mutations occur within or across a sin- gle gene, resulting from the substitution of one nucleotide for another or from the rearrangements (e.g., deletions) within the gene, or they may duplicate or delete the entire gene. The mutation responsible for sickle cell anemia, for example, is a point mutation: a single nucleotide in the gene coding for globin, a constituent of hemoglobin, is substituted for another nucleotide, resulting in the substitution of one amino acid (glutamic acid) for another (valine) at a certain position in the globin chain. One of the mutations responsible for a form of alpha-thalassemia, another form of anemia, involves the deletion of an entire alpha- globin gene, resulting in the absence of alpha— globin synthesis. Since proteins are produced from the instruc— tions in genes, a mutation in a gene that codes for a specific protein may affect the structure, reg- ulation, or synthesis of the protein. A mutation that results in a different codon may or may not change the resulting protein, since different codons can code for the same amino acid. Alter- natively, such a mutation may result in the sub- stitution of one amino acid for another, and pos- sibly change the charge distribution and structure of the protein. Another mutation in the same gene may result in gross reduction or complete loss of activity of the resulting protein. Detailed analy- sis of the mutations underlying thalassemia dis- ease has shown that mutations occurring in the bordering areas, not within the translated parts of the gene, can also impair gene function (171). Chromosome Mutations Major changes affecting more than one gene are called chromosome mutations. These involve the loss, addition, or displacement of major parts of a chromosome or chromosomes. They are often large enough to be visible under the light micro- scope as a change in the shape, size, or staining pattern (”banding”) of a chromosome. Major chromosome abnormalities are often lethal or re- sult in conditions associated with reduced lifespan and infertility. Structural Abnormalities Structural mutations in the chromosomes change the arrangement of sets of genes along the chromosome. Sections of one chromosome, in- cluding many genes, may break and reattach to another chromosome (e.g., in a balanced trans- location, shown in fig. 6) or another location on the same chromosome. Alternatively, sections of chromosomes can be deleted, inserted, duplicated, inverted. Numerical Abnormalities Numerical chromosomal mutations increase or decrease the number of whole chromosomes, 30 - Technologies for Detecting Heritab/e Mutations in Human Beings Figure 6.—Chromosomal Translocation Translocation chromosome 8q — Normal chromosome 8 Translocation Normal chromosome chromosome 14 14q+ 0,. 9 9. ’ Sites of breakage SOURCE: Office of Technology Assessment, HEALTH EFFECTS OF MUTATIONS In Western countries, genetic diseases are more visible now than in the earlier part of this cen- tury. Advances in medical care, public health measures, and living conditions have contributed to a gradual decline in the contribution of envi- ronmental factors to certain types of diseases, particularly, infectious diseases and nutritional deficiencies (32). Genetic disorders as a group now represent a significant fraction of chronic diseases and mortality in infancy and childhood, although each disorder is individually rare. Such disorders generally impose heavy burdens, expressed in mortality, morbidity, infertility, and physical and mental handicap. Approximately 90 percent of them become clinically apparent before puberty. The incidence of genetic disease is not known precisely, but it is estimated that genetic disorders are manifested at birth in approximately 1 per- cent of liveborn infants, and are thought to ac- count for about 7 percent of stillbirths and neonatal deaths and for about 8.5 percent of child- hood deaths. Nearly 10 percent of admissions to pediatric hospitals in North America are reported for genetic causes. The majority of these genetic diseases lack effective treatment. A small, but in- creasing number are avoidable through prenatal diagnosis and selective abortion (21,32,40). without changing the structure of the individual chromosomes. Errors in the production of the germ cells can lead to abnormal numbers of chro- mosomes in the offspring. For example, Down syndrome can be caused by the presence of an ex- tra chromosome (Trisomy 21); three copies of chromosome 21 are present while all other chro- mosomes are present in normal pairs, giving a to- tal of 47 chromosomes in each cell. A similar, but opposite, error in production of the germ cells re- sults in the lack of one chromosome. Turner syn— drome (Monosomy X), in which affected females have 44 autosomes and only one sex chromosome, an X chromosome, results in abnormal develop- ment of the ovaries and in sterility. The effects of these mutations may result not from alteration in the nature of the gene products but from the abnormal amount of gene products present. In most of these cases, the particular genetic dis- order is passed along in families, or the asymp- tomatic carrier state for the disease is passed along for many generations, before it may be expressed in a child with the disorder. In some of these cases, however, the disorder may suddenly appear in a child when no relatives have had the disorder. Not uncommonly, such an event can be explained by mistaken parentage or by one of the parents show- ing mild or almost unnoticeable signs of the dis- order, so that this parent has a disease-causing gene that can be passed on to his or her children. However, the sudden, unexpected appearance of a new condition in a family can also result from a new mutation in the reproductive cells of the parents. Examples of conditions that result from new mutations in a large percentage of cases in- clude Duchenne muscular dystrophy, osteogenesis imperfecta (a condition resulting in brittle bones), and achondroplasia (dwarfism). The exact causes of new heritable mutations are unknown. It is thought that the risk of some new heritable mutations increases with increasing age of the father. Certain rare disorders due to mu- tations occur at as much as four times the aver- age rate when fathers are 40 years of age or older at the time of conception of their children. Such Ch. Z—Genetic Inheritance and Mutations - 31 disorders include achondroplasia, Apert’s syn- drome (a disorder with skull malformations and fusion of bones in the hands), and Marfan’s syn- drome (a complex syndrome including increased height). Advanced maternal age is a risk factor for some newly arising chromosome abnormal- ities, including Down syndrome. Very small deletions of a particular chromo- some have been found in children having certain cancers such as Wilm’s tumor (chromosome num— ber 11) and retinoblastoma (chromosome num— ber 13). Some chromosome mutations are com- patible with normal life, and others may be associated with increased morbidity and repro- ductive problems. For example, when a section of a chromosome breaks off and reattaches to another chromosome, both copies of all of the genes may still be present, but they are located abnormally. If no genetic material has been lost in the move, this type of rearrangement is called a balanced translocation. However, a gene that was previously ”turned off" may, in its new loca— tion, be suddenly “turned on”; this explanation has been invoked to explain how cancer genes can be activated. Furthermore, in the production of germ cells, an abnormal complement of DNA could result, so that a parent with a balanced translocation may have offspring with unbalanced chromosomes, an outcome that usually has dele- terious consequences for the child. Early Mortality and Morbidity Numerical and structural chromosome abnor- malities have been detected in as many as 50 to 60 percent of spontaneous abortions, 5 to 6 per- cent of perinatal deaths, and 0.6 percent of live- born infants (43). Fetuses that survive to term with such abnormalities are apparently only a small percentage of the number of fetuses with chro- mosome abnormalities that have been conceived. Abnormalities such as trisomy 13 or trisomy 18 are among the more common disorders found in the liveborn group, whereas many numerical and structural abnormalties, most of which are not found in surviving infants, have been found in spontaneous abortuses. Some of the numerical chromosome abnormal- ities, such as trisomy 13 or trisomy 18, can be compatible with a full-term pregnancy but may cause severe handicap in infancy and childhood and may be associated with a reduced lifespan. The health effects of structural chromosome aber- rations (involving only parts of chromosomes) de- pend on the size and location of the aberration. However, features shared by many of these con- ditions include low birthweight; mental retarda- tion; and abnormalities of the face, hands, feet, legs, arms, and internal organs, often neccessitat- ing medical care throughout an individual's life. Diseases During Adult Lite Some genetic disorders, although present in an individual's DNA since his conception, do not manifest themselves until adult life. Huntington disease, a degenerative neurological disorder char- acterized by irregular movements of the limbs and facial muscles, mental deterioration, and death usually within 20 years of its first physiological appearance, is an example of a genetic disorder with adult onset. Many disorders that develop during adulthood and that tend to Cluster in families do not have classical patterns of genetic inheritance. Several genes may be involved in the development of these ”multifactorial diseases," and environmental fac- tors may also influence the manifestation and severity of these disorders. Susceptibilities to con- ditions such as diabetes mellitus, hypertension, ischemic heart disease, peptic ulcer, schizophre- nia, bipolar affective (manic depressive) disorder, and cancer are thought to have genetic as well as environmental components. Effects of Somatic Mutations Mutations that occur in somatic cells may af- fect the individual, but they are not passed on to future generations. Somatic mutations may play a role in the development of some malignant tumors by removing normal inhibition to cell growth and regulation, producing cells with a selective growth advantage. A wide range of ex— periments in animals and observations in humans indicate that somatic mutation may be one step in the development of certain types of cancer, in- 32 - Technologies for Detecting Heritable Mutations in Human Beings eluding leukemia and cancers of the breast and thyroid. Mutations that cause deficiencies in the normal DNA repair system, such as in the genetic disease xeroderma pigmentosum, can also lead to cancer following exposure to a mutagenic agent. PERSISTENCE OF NEW MUTATIONS IN THE POPULATION The “spontaneous mutation rate" represents the sum of ”natural" error arising during the course of life and reproduction, and external influences, both natural and manmade. The mutation rates that have prevailed throughout recent human his- tory may be in a state of relative equilibrium, al— though accurate information is lacking. Suppose, though, that mutagens are suddenly introduced heavily into the environment, and that there is an increase in mutation rates. It is impossible to accurately predict the number of people born with genetic diseases in future generations that will re- sult from that increase, but some predictions can be made based on what is known about the ef— fects of different types of mutations on subsequent generations. Two cases are considered below: 1) the effect of a doubling of the spontaneous mu— tation rate in one generation, and 2) the effect of a permanent doubling of the spontaneous muta- tion rate. Assume that a population was exposed to a “pulse" of a mutagenic agent, such as radiation from an atomic bomb or from chemical agents in a toxic spill, and that the exposure caused a doubling of the heritable mutation rate. Overall, the number of extra cases of disease associated with mutation in the first generation after the pulse (the generation in which the effect would be greatest), will be comparatively small. One esti— mate (from a 1977 United Nations Scientific Com- mittee on the Effects of Atomic Radiation report) is that 10.5 out of every 100 liveborn infants has some kind of genetic disorder (88). Almost all of those cases are from mutations already present in the gene pool, and not from new mutations. The National Research Council (NRC) estimates that one mutational pulse would add an extra 0.66 cases, so instead of 10.5 per 100, 11.16 cases would occur. The effects on subsequent genera- tions would be smaller, but even after hundreds of generations, there might be some small incre— ment of the mutational load as a result of the pulse. To understand this phenomenon, it is use- ful to consider different kinds of mutations sep- arately. Chromosome mutations would be the shortest lived in the population, because in most cases they result in sterility, so they would not be passed on to the next generation. They would probably be eliminated completely after about 1.25 gener- ations. Dominant and X—linked (sex-linked) mutations are the next shortest lived. Many of these, because they often cause such severe disease, interfere with a normal lifespan and reproduction. After about four or five generations, on average, they would no longer be transmitted to future generations. New recessive mutations will probably have the greatest chance of being maintained in the popu- lation. Virtually none would be eliminated in the first generation, because a single individual will have only one copy of the gene, and two would be needed to cause overt disease. Recessive mu- tations may take generations to surface as over disease. There is much less certainty about the ulti- mate fate of recessive mutations than about dom- inant, X-linked, or chromosome mutations. Many “common” or multifactorial diseases are partially influenced by genetic traits carried in families or arising anew. These include congeni- tal malformations, and such common diseases as heart disease, diabetes mellitus, bipolar affective disorder, immunological disorders, and cancer. Since it is difficult to estimate the proportion of disease that is currently influenced by mutations, it is also difficult to estimate how long such mu- tations induced from a pulse of a mutagenic agent would persist in the population. However, the NRC report (88) estimates that these mutations would persist for an average of 10 generations. These estimates suggest a surprisingly small in- crease over background rates of genetic disease Ch. 2—Genetic Inheritance and Mutations - 33 in the first few generations after intense exposure to mutagenic agents, which would decline slowly over many generations. The NRC report (88) sug- gests that about half of the total effect would oc- cur in the first six generations. The most severe mutations would be more easily observable as dominant genetic diseases. These would be elim- inated within one or two generations since they interfere with survival and fertility. Mutations with a less severe initial impact would persist in the population for a longer period of time. A permanent doubling of the current mutation rate would, however, after perhaps hundreds of generations, lead to a new equilibrium, with the incidence of genetic disease at about twice its cur- rent level. One of the biggest unknowns about predicting future effects of mutations is the im- pact of an accumulation of recessive mutations in the population. There is also little information to bring to bear on the question of the impact of an increase in mutations to the prevalence of com— mon diseases. Chapter3 Kinds and Rates of Human Heritable Mutations Contents Page Introduction ............................................................ 37 Surveillance of Sentinel Phenotypes ........................................ 38 Mutation Rates Estimated From Phenotypic Data .......................... 39 Limitations of the Phenotype Approach .................................. 4O Cytogenetic Analysis of Chromosome Abnormalities ......................... 40 Mutation Rates Derived From Chromosome Abnormalities ................. 42 Limitations of Cytogenetic Analysis ...................................... 42 Detection of Blood Protein Variants ....................................... 44 One-Dimensional Electrophoresis ........................................ 44 Quantitative and Kinetic Studies of Enzyme Deficiency Variants ............. 45 Molecular Analysis of Hemoglobin M and Unstable Hemoglobins ........... 46 Studies of Populations Exposed to Atomic Radiation ......................... 48 Early Epidemiologic Studies ............................................. 49 Cytogenetic Studies .................................................... 49 Biochemical Analyses .................................................. 49 Mutation Studies in Populations Exposed to Radiation in the Marshall Islands ............................................... 50 Conclusions ............................................................. 50 TABLES Table No. Page 1. Candidate Sentinel Phenotypes ......................................... 39 2. Prevalence of Chromosome Abnormalities at Birth ........................ 43 3. Rates of Mutation Leading to HbM and Unstable Hb Disorders ............ 48 4. Results in Studies To Detect Heritable Mutations by One—Dimensional Electrophoresis in Human Populations Not Known To Be at High Risk for Germinal Mutations ............................................... 51 S. Spontaneous Mutation Rates of Genetic Abnormalities in Human Populations ................................................... 51 Kinds Chapter 3 and Rates of Human Heritable Mutations INTRODUCTION The genetic code specified in DNA directs the transcription of RNA, which is translated into polypeptide chains; these chains, in turn, are as- sembled into proteins, which are the body's tools for the regulation of physiological, biochemical, and behavioral functioning. Mutations, transmis— sible alterations in DNA, can change the mes- sages, causing alterations in RNA molecules and proteins. Some of these changes can lead to sub- tle abnormalities, diseases, and disabilities. Environmental factors and specific interactions between environmental and genetic factors are thought to account for the majority of diseases or susceptibilities to disease. Mutations alone are thought to account for only a fraction of the cur— rent burden of disease. However, there is a group of disorders for which new mutations are the pri- mary causes. Among these “genetic" diseases is a group of rare clinical conditions, ”sentinel phe— notypes,” that occur sporadically in each gener- ation; children with sentinel phenotypes are usu- ally born to parents who do not have the same disease, indicating a new mutation that arose in a germ cell of one of the parents. There are several other indicators of new mu- tations in human beings that do not rely on de— tecting diseases. The most widely used methods have been cytogenetic analysis of chromosomes and biochemical analysis of proteins. The structure and number of chromosomes vis— ible by light microscopy has been used to detect a certain class of heritable mutations. Gross chro- mosome abnormalities, often associated with high fetal and neonatal death rates, are identifiable with increasing precision by cytogenetic analy- sis. A large portion of these chromosome abnor- malities arise anew each generation. Studies of mutationally altered proteins have been done on a limited basis. The proteins most accessible for analysis are those circulating in blood; other body proteins are not routinely ob- tained for analysis. The method available for detecting abnormal proteins is electrophoresis. Heritable mutations have been detected by iden- tifying electrophoretic variants of blood proteins. In addition, a combination of approaches has been used to estimate the mutation rate per nucleotide in genes coding for globin polypeptides, constit- uents of hemoglobin. These approaches to iden- tifying mutations are discussed in this chapter. The term ”mutation rates” for particular types of mutations in specific regions of DNA is used frequently in this report. Mutation rates for differ— ent kinds of mutations, or corresponding to differ- ent overt effects of mutations, can be expressed as mutations per locus, per gene, per nucleotide, and per gamete. All of these indicate a specific type of mutation occurring per generation, reflect- ing mutations arising anew from one generation to the next. This chapter summarizes current knowledge about such mutation rates as they have been measured thus far; it is not yet known whether these rates apply also to other popula— tions, to other gene regions, or to other types of mutations than the ones examined. Rather than looking directly at germinal mu- tations—mutations in egg and sperm DNA—the methods described in this report focus on herita- ble mutations in liveborn offspring, a subset of all germinal mutations.1 A larger subset of ger- minal mutations includes new mutations observed in offspring before birth. In recent years, cyto— genetic techniques have been used to identify chromosome abnormalities in spontaneous abor- tuses. These investigations have shown that the frequency of new mutations in spontaneous abor- 1At present it is not feasible to examine egg or sperm DNA for genetic damage, although some information on major chromosome abnormalities can be obtained by cytogenetic examination of sperm. 37 38 - Technologies for Detecting Heritable Mutations in Human Beings tuses is much higher than in liveborn infants; the observed incidence of new mutations at birth only partially represents the “true” incidence of such heritable mutations in all conceptuses, and may give an incomplete picture of the potential risk from environmental exposures to mutagenic agents. Nevertheless, the most lasting medical and social consequences of new heritable mutations are derived from those mutations that are com- patible with life at least through infancy and childhood. SURVEILLANCE OF SENTINEL PHENOTYPES The United Nations Scientific Committee on the Effects of Atomic Radiation (144) and the Com— mittee on the Biological Effects of Ionizing Radi— ations (87) estimate that 1 percent of liveborn in- fants carry a gene for an autosomal dominant disease; in 20 percent of these cases (0.2 percent of livebirths) their disease is due to a new, or ”spo- radic,” mutation that arose in the reproductive cells of one of their parents. Observing and re- cording the birth of infants with such sporadic conditions is the classical approach to identify- ing heritable mutations that lead to serious dis— ease (36).2 A subset of the group of autosomal dominant and X-linked conditions, indicator conditions known as “sentinel phenotypes," may be particu- larly useful in identifying heritable mutations. Sentinel phenotypes are clinical disorders that oc- cur sporadically, probably as the result of a sin— gle mutant gene. They are manifested at birth or within the first months of life, and usually require long-term, multidisciplinary medical care (83). In addition, sentinel phenotypes are associated with low fertility, so their appearance suggests muta— tions that have been passed not from affected par- ents to offspring, but instead, have been passed from unaffected parents to offspring as a result of a newly arising mutation in a germ cell of one of the parents of that offspring. Affected infants would have this mutation in the DNA of all their somatic cells and 50 percent of their germ cells. 1Rates of a wide variety of new heritable mutations leading to severe diseases have also been estimated by an ”indirect method.” This method is a theoretical approach, based on the loss of disease- causing genes from the population (due to reduced reproductive ca- pabilities of individuals with these genes) and on the sporadic fre- quency of the genes in the population. The mutation rate necessary to maintain such genes in the population, despite their recurrent loss, is inferred from these observations. The types of populations and conditions suitable for this method are limited, and its precision is unknown (91,165). Mulvihill and Czeizel (83) compiled a list, shown in table 1, of 41 genetic disorders (36 dom- inant and 5 X-linked) that satisfy the above cri— teria. Each of these phenotypes is individually rare, some of them occurring in only one in a mil— lion liveborn infants. Since children with sentinel phenotypes have serious health problems associated with their con- ditions, they may first come to the attention of the family practitioner or pediatrician, and then be referred to subspecialists for a complete diag- nosis. Clinical expertise in dysmorphology,3 med— ical genetics, pediatric ophthalmology, and on- cology may be needed to diagnose accurately the small numbers of children with sentinel pheno— types present among the vast majority of un- affected children and among children with pheno- typically similar, but nongenetic conditions (called ”phenocopies") (71,83,165). Reliable incidence data are available only for 13 dominant and 5 X-linked sentinel phenotypes (166); only those that are manifested in infancy, rather than in childhood or adolescence,4 are sys- tematically recorded. Diagnostic records of sen- tinel phenotypes may be available in population— based registries of childhood cancers, birth de- fects, and genetic diseases, with initial entries based on birth certificates and later entries added as the conditions are diagnosed during infancy. After investigators rule out the possibility of a dis- crepancy between stated and biological parent- age by genetic testing (166), sporadic cases of a sentinel phenotype are accepted as evidence for 3Physicians who specialize in diagnosing syndromes character- ized by unusual physical form and stucture, for example, congeni- tal malformations. ‘Such disorders include achondroplasia, acrocephalosyndactyly, osteogenesis imperfecta, retinoblastoma, and Wilms’ tumor (66). Ch. 3—Kinds and Rates of Human Heritable Mutations - 39 Table 1.—Candidate Sentinel Phenotypes‘ Inheritancea Phenotypes Identlflable at blrth: Achondroplasia .......................... AD Cataract, bilateral, isolated ................ AD Ptosis, congenital, hereditary .............. AD Osteogenesis Imperfecta type i ............ AD Oral-faciaI-digital (Gorlin-Psaume) syndrome type I ................................. XD lncontinentia pigmenti, BIoch-Sulzberger syndrome ............................. XD Split hand and foot, bilateral atypical ....... AD Aniridia, isolated ......................... AD Crouzon craniofacial dysostosis ........... AD Holt-Oram (heart-hand) syndrome .......... AD Van der Woude syndrome (cleft lip and/or palate with mucous cysts of lower lip)... . AD Contractural arachnodactyly ............... AD Acrocephalosyndactyly type i, Alpert’s syndrome ............................. AD Moebius syndrome, congenital facial diplegia ............................... AD Nail-patella syndrome .................... AD Oculodentodigital dysplasia (ODD syndrome) ....................... AD Polysyndactyly, postaxial ................. AD Treacher Collins syndrome, mandibulofacial dysostosis ............................ AD Cleidocranial dysplasia ................... AD Thanatophoric dwartism .................. AD EEC (ectrodactyly, ectodermal dysplasia, cleft lip and palate) syndrome ........... AD Whistling face (Freeman-Sheldon) syndrome ............................. AD Acrocephalosyndactyly type V, Pfeiffer syndrome ............................. AD Spondyloepiphyseal dysplasia congenita. . . . AD Phenotypes not Identifiable at blrth: Amelogenesis imperfecta ................. AD Exostosis, multiple ....................... AD Marian syndrome ........................ AD Myotonic dystrophy ...................... AD Neurofibromatosis ....................... AD Polycystic renal disease .................. AD Polyposis coli and Gardner syndrome AD Retinoblastoma, hereditary ................ AD Tuberous sclerosis ....................... AD von HippeI-Lindau syndrome .............. AD Waardenburg syndrome ................... AD Wiedemann-Beckwith (EMG) syndrome ..... AD Wilms’ tumor, hereditary .................. AD Muscular dystrophy, Duchenne type ........ XR Hemophilia A ............................ XR Hemophilia B ............................. XR aAD refers to autosomal dominant inherltance, XD refers to X-Iinked dominant inheritance, and XFi refers to X-Ilnked recessive inheritance. SOURCE: J.J. Mulvlhlll and A. Czeizel, “Perspectives in Mutations Epidemiology 6: A 1983 View of Sentinel Phenotypes," Mutat. Res. 123:345-361, 1983. a new germinal mutation originating in a germ cell of one of the parents (65). In general, recording the incidence of sentinel phenotypes is not useful for monitoring a small population exposed to a suspected mutagen. Sur- veillance of sentinel phenotypes in large popula- tions is potentially more useful for estimating the background mutation rates leading to dominant genetic disorders. Given the rarity of the individ- ual sentinel phenotypes, the most reliable data for estimating mutation rates would be derived from the largest target populations with, for example, international cooperation among study centers (62,66). Mutation Rates Estimated From Phenotypic Data Data are available on the incidence of various sentinel phenotypes from different time periods in particular regions worldwide, though represen- tation is incomplete. Crow and Denniston (22) summarized the most current data; these data have not been significantly updated since the 19405 and 19505. The frequencies for any given phenotype are generally consistent between studies and the frequencies of occurrence of the different phenotypes vary over a thousandfold range, from 1 in 10,000 to 1 in 10 million depend- ing on the particular disorder. The arithmetic mean of the rates for the different disorders is ap- proximately 2 mutations per 100,000 genes per generation5 (17,89,166). This average rate is often cited as the “classical” rate of heritable mutations. 5Mutation rates corresponding to sentinel phenotypes are expressed as the frequency of mutations per gene. It assumes one gene per dis— order and one disorder per person and it is designated “per genera- tion” because the mutations are detected in offspring of individuals whose germ cells have incurred mutation. In general, the mutation rate is defined as the incidence of a sporadic case of some indicator (e.g., a sentinel phenotype or a protein variant) divided by two times the total number of cases examined (e.g., newborn infants or pro— tein determinations). The two in the denominator is to express the mutation rate, by convention, as the rate “per fertilized germ cell," instead of ”per conceptus,” which is made from two germ cells (162). 40 - Technologies for Detecting Heritable Mutations in Human Beings Limitations of the Phenotype Approach Several sources of error may affect estimates of heritable mutation rates based on the incidence of sentinel phenotypes. The most important one results from the choice of phenotypes studied: phenotypes that occur in the population most often are the ones most likely to be recorded. Phenotypes that are so severe as to cause prena— tal death, or so mild as to be indistinguishable from the normal variety of phenotypic traits, are difficult or impossible to include in these popu- lation surveys. It is not known how large a bias this selection occasions. In addition, the mutation rates corresponding to clinically recognizable phenotypes may not be representative of muta- tion rates for other traits. Other problems contributing to an over- or un- derestimated mutation rate may include: 1) in- complete ascertainment among certain popula— tions of liveborn infants; 2) mistaken identification of parentage, a significant issue since the num- ber of cases of mispaternity may be high enough to exceed the number of true sporadic cases of sen- tinel phenotypes; and 3) the occurrence of pheno- copies (phenotypically similar nongenetic causes of the same condition) and genocopies (recessive forms of the same phenotype inherited from un- affected carrier parents). In addition, particular phenotypes do not necessarily correspond to a unique, single gene (163); often, mutations in any one of several genes can lead to the same pheno— type. This is not likely to influence the rate esti- mate significantly unless a large number of differ- ent mutant genes are involved, in which case the rate of mutation of the particular phenotype would be the sum of the rates of mutation at each of the alternative genes. Surveillance for sentinel phenotypes is an im- portant part of assessing the clinical impact of new mutations. The number of sentinel phenotypes available to study is limited, however, so that sur- veillance of these phenotypes illuminates only a fraction, albeit an important one, of this overall impact. Individually, the sentinel phenotypes are quite rare. Keeping track of the frequencies of the differ- ent phenotypes over periods of time would require large-scale surveillance of huge numbers of infants over many years. A large team of medical spe- cialists and a well-organized database would be needed to obtain complete and accurate ascertain- ment (65). Unfortunately, the practical difficul- ties in screening huge populations of newborns, diagnosing these rare disorders, and organizing and maintaining the data, make it difficult, though not impossible, to generate reliable data. The first such effort is currently being organized by the European Collaborative Study (62). CYTOGENETIC ANALYSIS OF CHROMOSOME ABNORMALITIES The adverse effects of chromosome abnormal- ities on human health are well established and provide sufficient medical reasons for screening infants for them. Data on the frequency of chro- mosome abnormalities is also useful in assessing the impact of detectable new chromosome muta- tions on the current generation. Chromosome abnormalities are defined as either numerical (extra or missing whole chromo- somes) or structural (deletions, insertions, trans- locations, inversions, etc., of sections of chromo— somes). In liveborn infants, the presence of an extra or missing sex Chromosome is often associ- ated with physical, behavioral, and intellectual impairment. The presence of an extra autosome is even more detrimental: it is usually associated with severe mental and physical retardation and often with premature death (39,119). There is no doubt that fetuses with recognized numerical chromosome aberrations are at a much higher risk for aborting spontaneously than fe- tuses without such defects. Such chromosome ab- normalities contribute to a large proportion of spontaneous abortions and stillbirths, and in live— born survivors may also contribute to repeated early abortions and fertility problems in adult- Ch. 3—Kinds and Rates of Human Heritable Mutations - 41 hood (169). The significance to human health of structural rearrangements of the chromosomes is less well defined, but such defects have been asso- ciated with mental retardation, physical malfor- mations, and a variety of malignant diseases (43). In addition, carriers of balanced translocations may produce offspring with unbalanced comple- ments of DNA; these offspring may die in utero and may account for a high spontaneous abor- tion rate among such carrier parents. It is believed that at least 5 percent of all rec- ognized human conceptions have a chromosome abnormality, although this is a crude estimate be- cause it is not corrected for factors which could influence the occurrence of such abnormalities (such as maternal age) or for factors which could affect the rate of embryonic or fetal death in some of the disorders (43). Numerical and structural chromosome abnormalities have been detected in 50 to 60 percent of recognized spontaneous abor- tions (data summarized in ref. 12), 5 to 6 percent of perinatal deaths (ibid.), and 0.6 percent of live- born infants (42). Chromosome abnormalities are useful for meas- uring heritable mutation rates since they are usu- ally identified as new mutations, either because the defect causes virtual infertility or sterility (as in the case of numerical chromosome abnormal- ities) and therefore could not have been inherited " ‘ from a parent with the same defect, or because the parents of an individual with a chromosome abnormality (e.g., a structural abnormality) have been shown not to carry the same mutation. Nu- merical chromosome abnormalities and some struc- tural abnormalities are detectable with cytogenetic methods using current chromosome staining and banding techniques. In general, little is known about the genetic mechanisms of either numerical or structural rear- rangements. They appear to result from complex processes during gametogenesis in which various defects may lead to the same or different outcomes (18,119). Consequently, frequencies of the vari— ous cytogenetic abnormalities may correspond to a combined rate of several different mutational events or to a single mutational event. Individuals with chromosome abnormalities can be identified on the basis of karyotypes (see ch. 2). However, since the vast majority of con- ceptuses with chromosome abnormalities such as trisomies and unbalanced structural rearrange- ments are spontaneously aborted, many before a pregnancy is recognized, the frequency of these mutations at birth is known to represent only a small fraction of the presumed frequency of these mutations at conception. Data on the frequency of chromosome abnor- malities in fetuses at the time of amniocentesis (generally done at 16 to 18 weeks gestation) pro- vides information not available in newborn screening. In addition, recent progress in cyto- genetic analysis of chorionic villi cells makes it possible to identify chromosome abnormalities in fetuses at 8 to 10 weeks gestation (116). This tech- nique provides even earlier information on pos- sible chromosome abnormalities, before fetuses with certain abnormalities (particularly trisomies and unbalanced structural rearrangements) would have aborted spontaneously. The analysis of fe- tal cells obtained at amniocentesis has provided valuable data on the frequency of chromosome abnormalities in the second trimester of pregnancy (12,45,159.) However, estimates of mutation rates from those measurements may be inflated unless certain biases are corrected. Women who elect amniocentesis—generally older mothers, couples who have had previous reproductive problems, and/ or couples who may have been exposed to mutagenic agents (101) —are at increased risk for chromosome abnormalities in their offspring. Cytogenetic surveys of consecutive livebom in- fants have provided the most extensive data on chromosome mutation rates, although it is known that these data underestimate the ”true" rate of new chromosome mutations. It should also be noted that the prevalence of sentinel phenotypes and biochemical variants in liveborn infants may also underestimate their “true" rate at conception, but little is known about selection against these mutations in the developing fetus (e.g., molecu- lar repair of the defect, or spontaneous abortion of the fetus), in contrast to the documented evi- dence for selection against Chromosome mutations in the fetus (135.) However, the frequency at birth is probably the best indicator of events with the highest cost in terms of the social and economic burden of these diseases. 42 - Technologies for Detecting Heritable Mutations in Human Beings Mutation Rates Derived From Chromosome Abnormalities At least 10 major research efforts worldwide have attempted to determine the prevalence of chromosome abnormalities in liveborn infants. It is possible to estimate the rate of new mutations underlying these defects, but difficult to compare with rates of sentinel phenotypes, since chromo- some mutations, by definition, are multigenic changes and sentinel phenotypes correspond to single gene mutations. The combined data from studies surveying a total of 67,014 newborn infants for various periods between 1974 and 1980 in the United States, Canada, Scotland, Denmark, the USSR, and Japan (120) are shown in table 2. Two of the studies used newer banding techniques that iden- tify more abnormalities than conventional stain- ing methods used in the remaining eight studies. For conditions that are associated with virtual infertility, complete sterility, or prereproductive death, it is often assumed that each case is the re- sult of a new mutation (unless there is evidence for gonadal mosaicism"). Such conditions include the autosomal trisomies (e.g., Down syndrome) and the sex chromosome aneuploidies (e.g., Turner’s syndrome). For these defects, the apparent mutation rate observed in liveborn infants is calculated as the number of individuals identified with a given ab- normality divided by the total number of indi- viduals examined. This mutation rate is expressed as the number of mutations per generation. For conditions which could be the result of either new mutations or inherited mutations, such as the vari— ous balanced and unbalanced structural rearrange- ments, it is necessary to karyotype the parents of infants found with these chromosome abnormal— ities, and to exclude inherited mutations from cal- culations of the de novo mutation rate. For these °Mosaicism refers to the presence of two genetically different cell populations in the same individual. This can result from a muta- tion during one of the earliest divisions in the zygote, a few days after conception, that led to the proliferation of a large number of cells that were derived from the original cell bearing the mutation. conditions, the observed mutation rate is deter- mined using the following formula: Proportion who are likely Number of patients to have de novo mutations7 Mutation _ with abnormality rate Total number examined The data show that the frequencies of chromo- some abnormalities in liveborn infants range from 2 in 100,000 newborns per generation for inver- sions to 121 in 100,000 newborns per generation for Trisomy 21 (Down syndrome), the most com— mon newly arising numerical chromosome abnor- mality found at birth. Overall, the numerical abnormalities are much more frequent than the structural abnormalities. However, the rates for balanced structural rearrangements may be low in these data, since only two of the surveys used more sensitive banding methods which are need- ed to identify these more subtle aberrations. Limitations of Cytogenetic Analysis Screening for chromosome mutations in new- born infants gives only a partial picture of their incidence. The prevalence of a particular genetic mutation at birth can be thought of as the result of the incidence of the mutation at conception in- teracting with the probability of survival of con- ceptuses with the mutation. Fetal survival, in turn, depends on individual fetal and maternal attri- butes, as well as on the interaction between the two (136). Some mutations in the fetus may af- fect fetal survival, whereas other mutations may have no effect on survival. Major chromosome aberrations, in particular, numerical abnormal- ities and most unbalanced structural rearrange- ments, are often incompatible with fetal survival and are associated with an increased risk of prena- tal loss (45). Balanced chromosome rearrangements, how— ever, may not lead to excessive prenatal loss. For this reason, the balanced structural rearrange— ments may be a useful indicator of new germinal mutations observed in liveborn offspring. Since some of these infants identified as having a bal- 7Such proportions are based on empirical data for each type of chromosome aberration. In this case, data from Jacobs, 1981 were used and are shown in table 2 in parentheses in the third column. Ch. 3—Kinds and Rates of Human Heritabie Mutations o 43 Table 2.—Prevalence of Chromosome Abnormalities at Birth Number New chromosome Total of new abnormalities per 100,000 Chromosome abnormality population mutantsa newborns per generation Numerical anomalies: Autosomal trisomies:b Trisomy 13 ....................................... 67,014 3 5 Trisomy 18 ....................................... 67,014 8 12 Trisomy 21 ....................................... 67,014 81 121 Male sex chromosome anomalies: 47,XYY .......................................... 43,048 males 43 100 47,XXY .......................................... 43,048 males 42 98 Female sex chromosome anomalies: 45,X ............................................. 23,966 females 2 8 47,XXX .......................................... 23,966 females 24 100 Balanced structural rearrangements: Robertsonian translocations:C D/Dd ............................................ 67,014 48(2/29)=3.3 5 D/G ............................................. 67,014 14(2/11)=2.5 4 Reciprocal translocations and insertions ............... 67,014 60(13/43)=18 27 Inversions ......................................... 67,014 12(1/8)=1.5 2 Unbalanced structural rearrangements: Translocations, inversions, and deletions .............. 67,014 37(7/16)=16 24 aSee discussion in text for proportion of new mutants among all those identified with structural rearrangements. Trisomies refer to conditions in which there are 47 chromosomes (instead of the normal 46 chromosomes or 23 pairs of homologous chromosomes, indicated as 46,XX for females or 46,XY for males) because of the presence of an extra copy of one chromosome; three copies of this particular chromosome would be present instead of the normal two. Autosomal trisomies indicate an extra autosome, which includes any chromosome except one of the sex chromosomes. cRearrangements in which the long arms of two chromosomes fuse, resulting in one "hybrid” chromosome and the loss of the short arms of each chromosome. dD and G refer to groups of chromosome numbers 13-15 and 21-22, respectively. SOURCE: K. Sankaranarayanan, Genetic Effects of ionizing Radiation in Muiticeiluiar Eukaryotes and the Assessment of Genetic Radiation Hazards in Man (Amster- dam: Elsevier Biomedical Press, 1982), pp. 161.164. anced chromosomal mutation may have inherited the relatively subtle abnormality from one of their parents, it is necessary to determine whether the mutation observed in the child is actually a new germinal mutation or not. Although it is not understood how bands ob- servable under the microscope correspond to the structure and composition of DNA, mutations in DNA can cause a visible change in the banding pattern, particularly if such mutations involve a large section of a chromosome. Increased resolu- tion and sensitivity of cytogenetic techniques, spe- cifically the ”high-resolution” banding methods (72), makes it possible to detect smaller and more subtle aberrations than was previously possible. Currently, routine cytogenetic methods can pro— duce 200 to 300 bands (containing several hun- dred genes in each band) in a complete set of DNA, and the high-resolution methods can dis- tinguish as many as 1,000 bands (containing about 100 or fewer genes per band) (59). With more smaller bands distinguishable, fewer genes are present per band, and a change in one gene has a greater likelihood of showing up as a change in the banding pattern. At present, data available for large populations rely on only the routine staining and banding methods. Like the observations of sentinel pheno- types in newborn infants, cytogenetic methods focus on a subset of mutations, albeit those with serious clinical consequences. Even though chro- mosome abnormalities are not as rare as sentinel phenotypes, these two groups of abnormalities to- gether probably account for only a small fraction of heritable mutations. Mutations detectable by the methods discussed in the remainder of this chapter and in the fol- lowing chapter do not necessarily have adverse clinical consequences. Their detection broadens the spectrum of identifiable mutations by includ- ing those that may not have an immediate adverse effect on the individual, and also offers more pre- cise information about the nature of the genetic changes. 44 0 Technologies for Detecting Her/table Mutations in Human Beings DETECTION OF BLOOD PROTEIN VARIANTS In a single-gene, dominant genetic disease, if the causative mutant gene occupies one allele of the pair of genes and the normal form of the gene occupies the other allele, the one mutant gene is dominant over the normal gene and is sufficient to cause overt disease. Dominant diseases are said to be expressed in individuals who have a single dose, or are heterozygous, for the disease-causing gene. Autosomal recessive diseases, however, are clinically apparent only when the mutant gene is present at both alleles, or in the homozygous state; a single dose, the heterozygous state, can, for in- stance, destroy the activity of an enzyme or pro- duce a defective gene product, but the correspond- ing normal gene supplies a sufficient amount of normal gene product to prevent overt clinical symptoms. It is difficult to infer the frequency of recessive genes by population studies of pheno— types since individuals with one copy of the reces- sive mutant gene are phenotypically indistinguish- able from individuals with only normal genes for the particular protein. Biochemical means of iden- tifying of heterozygotes, who are phenotypically normal, is one way of identifying new mutations that are expressed as recessive traits. Biochemical studies on asymptomatic individ- uals with protein variants (indicative of single gene, recessive mutations) can be useful in express- ing the cumulative, ”invisible" effect of an in- creased mutation rate. Any protein in the body could be studied, but since peripheral blood is eas- ily accessible, proteins from blood are most com- monly used for these studies. Three major tech- niques for the detection and characterization of variant proteins have been used in studies of hu- man populations: 1. electrophoresis of blood proteins, 2. quantitative and kinetic studies of enzyme deficiency variants from erythrocytes, and 3. molecular analysis of Hemoglobin M (HbM) and unstable hemoglobins. One-Dimensional Electrophoresis Electrophoresis separates proteins according to net molecular charge. A mixture of proteins is ap- plied to a starch or acrylamide gel and exposed to an electric current. Each protein moves accord- ing to its own electric charge and separates from proteins with different charges. The subsequent addition to the gel of stains that bind to the pro- teins makes it possible to visualize the location of the proteins. Electrophoresis can be used to de- tect variant proteins in individuals and in their parents. The identification in an offspring of a var- iant protein that is absent from either parent is evidence for a new mutation. As in the other ap- proaches, it is necessary to exclude the possibil- ity of a discrepancy between stated and biologi- cal parentage so that only new, not inherited, mutations are considered. Electrophoretic variants are detectable because a mutation results in a change in the net charge of a protein. On a theoretical basis it is expected that approximately one-third of amino acid sub- stitutions produce a Change in molecular charge of a protein, and there is also evidence that elec- trophoresis can detect changes in a molecule’s con- figuration due to a nucleotide substitution. All in all, it is estimated that electrophoresis probably detects about 50 percent of nucleotide substitu- tions in expressed regions of coding genes (96). The protein variants'detected by electrophoresis are usually not associated with clinically recog— nizable problems. However, this technique has an important advantage for mutation research over the phenotypic approach; it measures changes at a level closer to the level of the mutational events and therefore may provide more accurate esti— mates of mutation rates. It also reflects more sub- tle genetic changes, broadening the spectrum of detectable genetic endpoints. Mutation Rates Derived From Electrophoresis Several large-scale studies to detect protein var- iants in human populations have been reported (excluding studies in Japanese atomic bomb sur- vivors, which are described below). Neel, et al. (94), and Mohrenweiser (76) devised a large-scale pilot program to study mutation rates using pla- cental cord blood obtained from newborn infants in Ann Arbor, Michigan (approximately 3,500 samples). They found no new heritable mutations Ch. 3—Kinds and Rates of Human Heritable Mutations - 45 in 36 different proteins in a total of 218,376 10- cus tests in Caucasian and 18,900 such tests in black infants.8 Harris and colleagues (38) summa- rized data from studies in the United Kingdom in which 43 different loci coding for blood proteins were analyzed, and no new mutations were found in 133,478 locus tests.9 Altland and colleagues (5) examined blood from filter paper submitted origi- nally for screening for phenylketonuria in approx- imately 25,000 newborn infants in West Germany. They identified one putative new mutation among five different hemoglobin proteins in approxi- mately 225,000 locus tests. Feasibility of Electrophoresis The availability of human blood for analysis of variant proteins makes electrophoresis feasi— ble for studies of large populations. In addition, it permits comparisons of the frequency of occur- rence of variant proteins between humans and other species, since comparable proteins are usu- ally associated with homologous gene sequences (93). Perhaps the most important advantage of this approach is that it permits the study of a de- fined set of genes expressing functional proteins and offers more precise information on the num- ber of genes involved in the biochemical obser— vations made and on the effect of mutation on some proteins (90). However, electrophoretic analysis can provide information only on mutations in genes that are transcribed into RNA and subsequently translated into proteins, and only for mutations that alter the protein’s net molecular charge or configura- tion. It cannot provide information on mutations in nontranscribed regions of the DNA, which comprise the larger portion of the genome (165). Since the baseline frequency of new protein var- iants is still poorly defined, it is difficult to esti- mate the size and scope of a population study nec- essary to detect an increase over a period of time or to detect differences in rates between two populations. One-dimensional electrophoresis is ”The number of locus tests takes into account the number of pro- teins examined, the number of gene loci represented by each such protein, and the number of individual samples obtained. 9The number of individuals tested varies depending on the en- zyme examined, but for some of the enzyme tests, more than 10,000 individuals were tested. limited by the number of proteins that can be examined for variants, approximately the same number of different phenotypes included in the list of sentinel phenotypes. However, one-dimen— sional electrophoresis in a large population is a more straightforward approach to studying new heritable mutations than is population surveil- lance for sentinel phenotypes. Vogel and Altland (164) estimated that analy- sis of two populations of 10 million individuals each would be needed to have a 95-percent chance of detecting a 10-percent increase in spontaneous mutation rates. This calculation is based on a theoretical rate of 1 mutation per 1 million genes per generation and 50 loci tested per individual in the study. The population size could be reduced if the proteins that were examined mutated more frequently or if detection of only large increases in the mutation rate (e. g., 50 percent) were accept- able, or if many more proteins could be exam- ined from each individual tested. Quantitative and Kinetic Studies of Enzyme Deficiency Variants Mutations that result in variant proteins with greatly reduced or no biological activity, when present in the heterozygous state, are not readily detectable in electrophoretic studies. Such "en- zyme deficiency variants," or “nulls,"10 can be identified through automated biochemical tests that measure enzyme activity in red blood cells. These studies increase the detectable range of mutational events. Loss of function of a protein may result from: 0 the absence of a gene product, 0 the presence of a protein which is non— functional catalytically, or 0 abnormally unstable enzymes (75). The types of mutations causing these abnormal— ities include chromosome rearrangement or loss, deletions, frameshift mutations, and nucleotide 10These mutations are characterized by the presence of an enzy- matically inactive protein, or by an absence of a protein. These can be caused by gene deletions, nucleotide substitutions in critical loca- tions in the gene, mutations in the ”start” and ”stop" codons for the protein, or even by nucleotide substitutions at splicing junctions for the introns. 46 . Technologies for Detecting Her/table Mutations in Human Beings substitutions in coding or noncoding regions of DNA. The range of genetic events underlying the production of enzyme deficiency variants is thought to be larger than for electrophoretic mo- bility variants, since mutations in noncoding re- gions (e.g., intervening sequences, flanking re- gions, etc.) as well as in coding regions could be detected (74). Four studies have provided data on the fre— quency of enzyme deficiency variants. The fre- quencies of such variants have been measured in: 1) placental cord blood samples from approxi- mately 2,500 individuals in Ann Arbor, Michi- gan (74,75); 2) blood samples obtained from par- ticipants of the studies in Japan coordinated by the Radiation Effects Research Foundation (122); 3) blood samples obtained from 3,000 hospitalized individuals in West Germany (30); and 4) blood samples obtained from Amerindians of South and Central America (78). No new mutations were found in any of these studies. However, rare but not new enzyme deficiency mutations that have been inherited over many generations have been measured at an overall frequency of 2 variants per 1,000 determinations in a total of approxi- mately 110,000 locus tests. If new mutations producing enzyme deficiency variants occur, as predicted, no more than twice as often as new mu- tations leading to electrophoretic variants, the lack of new mutations in these 110,000 locus tests is not inconsistent with electrophoretic data. Molecular Analysis of Hemoglobin M and Unstable Hemoglobins Stamatoyannopoulos and Nute (131) calculated mutation rates for certain nucleotides in the globin genes, which, when mutated, produce autosomal dominant disorders expressed as chronic hemo- lytic anemia or methemoglobinemic cyanosis. Their data provided the first direct measurements of heritable mutation rates for particular nucleo— tides in human DNA. Although these data are limited to the specific characteristics of the globin genes, they suggest the kind of information which may be possible to obtain for other genes as the genetic basis for various single gene disorders is discovered. Even if the mutation rates derived from these studies of the globin genes are not rep- resentative of rates of mutation at other loci, the hemoglobin system described below provides a model for combining epidemiologic, clinical, and molecular information to estimate heritable mu- tation rates in human populations. Structure and Function of Hemoglobin Hemoglobin molecules, which exist in a num- ber of slightly different forms, are the universal carriers of oxygen in the blood from the lungs to all cells of the human body. Different mixtures of hemoglobins are present at all stages of devel- opment from embryonic to adult, although a par- ticular type of hemoglobin tends to predominate at any one stage. In addition, hundreds of abnor- mal hemoglobins have been identified, ranging in clinical severity from benign through serious hematologic disorders. Human hemoglobin is a protein consisting of four separate globin poly- peptide chains and four iron-containing heme groups. In adults, the predominant form of he- moglobin is hemoglobin A (HbA), composed of two alpha-globin chains and two beta-globin chains each joined to a heme group. Hemoglobin's primary function of binding, car- rying, and releasing oxygen is closely regulated by several factors, including the maintenance of a precise three-dimensional conformation of the hemoglobin molecule, the composition and bal— ance of globin chains, and the distribution of ionic charges in and around the molecule. Mutations which result in amino acid substitutions at par- ticular sites in the globin polypeptide chain can markedly alter these specific functional proper- ties of the hemoglobin molecule. Two types of abnormal hemoglobin, HbM and unstable hemoglobin, usually occur sporadically as a result of a new mutation and are expressed as dominant mutations.11 Stamatoyannopoulos and Nute (131) described the genetic changes lead- ing to these particular abnormal hemoglobins and their incidence in human populations. From this “A heterozygous individual, having one mutant globin gene and one normal allele, would have clinically recognizable symptoms of the disease. This is in contrast to other, more common hemoglobin diseases, such as sickle cell anemia and thalassemia, where the het- erozygote is asymptomatic and the disease occurs only when an in- dividual is homozygous for the variant gene. Ch. 3—Kinds and Rates of Human Heritab/e Mutations 0 47 information, they were able to calculate the rate of mutation leading to these abnormal hemo— globins. Unstable Hemoglobins Various mutations in the globin genes can lead to an altered sequence of amino acids in the globin chains and to a major change in the three-dimen- sional structure of the hemoglobin molecule. Often these changes create a less stable hemoglo— bin molecule than its normal counterpart. “Un— stable hemoglobin” can result in the hemoglobin precipitating within the red cell, leading to mem- brane damage and premature destruction of the red cells by the liver and spleen (165). The degree of clinical severity of the unstable hemoglobin depends in part on the site at which the amino acid replacement occurs and on the properties of the amino acid that is substituted (37). However, manifestations of the disorder vary from mild hemoglobin instability which may not be clinically significant, to severe hemoglo- bin instability, which causes chronic hemolytic anemia. Unstable hemoglobin is diagnosed by isolating and characterizing the abnormal hemoglobin, and by determining its molecular stability. Since the amino acid sequence of normal hemoglobin is known, and the sequences for some abnormal hemoglobins have been determined, the genetic mutation underlying these unstable globins have been identified. Hemoglobin M—The Methemoglobinemias A blue, cyanotic appearance is the chief clini- cal manifestation of methemoglobinemia, caused by HbM. Cyanosis occurs whenever a significant proportion of the hemoglobin in the circulating red blood cells is not carrying oxygen, and in this case it is because of an inability of the hemoglo- bin to bind and carry oxygen (37). Five different mutations in the globin gene which lead to the production of HbM have been identified. All of these are expressed as dominant phenotypes. The common characteristic of these mutations is that each results in substitution of an amino acid in the globin peptide where the heme groups are attached. (These mutations are specific substitutions of histidine for tyrosine on the alpha-globin chain [at positions 58 and 87] or on the beta-globin chain [at positions 63 and 92], or by substitution of valine for glutamic acid on the beta-globin chain [at position 67].) By alter- ing the site of attachment of heme on the globin chains, these mutations change the three—dimen- sional organization of the entire hemoglobin mol- ecule and interfere with its ability to combine with oxygen (100). Measurement of Mutation Rates in the Globin Genes Using published data on the number of cases of hemoglobin disorders from 10 countries (134), Stamatoyannopoulos and Nute (131) determined the number of children with HbM and unstable hemoglobin among an estimated number of live- births in the defined populations. They recorded a total of 55 cases of de novo hemoglobin mu- tants, each of which derived from substitutions of single nucleotides in the coding portions of the globin genes: 40 children with unstable hemoglo- bin (all beta chain mutations) and 15 with HbM (10 beta chain and 5 alpha chain).12 The requi- site paternity data, also from published informa- tion on these mutations, had been collected on 19 of the 55 cases which were used to calculate mutation rates (131). Their data are shown in table 3. The mutation rates derived from these data are expressed in mutational events per alpha- or beta- gene nucleotide per generation, and range from 5.9)(10‘9 to 19 ><10‘9 for alpha- and beta—gene nucleotides, respectively. The rate for all beta— gene variants together is 7.4X10‘9 per nucleo- tide per generation. These estimates correspond to mutation rates per globin chain gene per gen- eration of 2.6X1O‘6 to 8.3X10‘6. The rate per gene is obtained by multiplying the rate per nu— 1ZThe number of alpha chain mutations detected was less than expected, given the existence of two alpha-globin genes per beta- globin gene in an individual. A possible reason for this may be that alpha-chain defects would produce physiological abnormalities at an earlier stage of fetal development than would beta-chain abnor- malities, since the alpha gene is activated much earlier, and may be causing a greater loss of fetuses early in gestation. 48 - Technologies for Detecting Heritable Mutations in Human Beings Table 3.—Rates oi Mutation Leading to HbM and Unstable Hb Disorders Mutation rate/ nucleotide/ Type of disorder generationa Unstable Hbs . .. 5.9)(10‘9 2.6x10‘6/438 nucleotides AlphaM variants . 10.0>< -9 4.2x10‘6/423 nucleotides BetaM variants .. 18.9)(10‘9 8.3xlo-6/438 nucleotides All beta-chain variants ...... 7.4x10‘9 3.2X1O‘6/438 nucleotides aThe mutation rate per nucleotide is based on the rate of appearance of the par- ticular disorder (number of mutants observed divided by 2 multiplied by the num ber oi births in the generation(s) included). This is the average mutation rate tor the particular disorder expressed per globin gene affected per generation. To calculate the mutation rate per nucleotide per generation, Stamatoyannopou- toe and Nute took the frequency of each mutant divided by 2 (for two beta genes in a genome) multiplied by 3 (since a given nucleotide can be substituted by any one of three other nucleotides), then divided that product by the number of different substituted nucleotides actually observed in their data. ”The estimated mutation rate per alpha-globin and beta-globin gene was calcu- lated as the mutation rate per nucleotide multiplied by the number of nucleoA tides constituting the coding portion of the corresponding gene (423 in the alpha-globin gene and 438 in the beta-globin gene). SOURCE: G. Stamatoyannopoulos and PE. Nute, “De Novo Mutations Produc- ing Unstable Hbs or HbsM. ll: Direct Estimates of Minimum Nucleo- tide Mutation Flatas in Man,“ Hum. Genet. 60:181-88, 1982. Mutation rate/ globin geneb cleotide by the number of nucleotides in the cod- ing portion of the gene. The rates calculated from these data are likely to be minimum rates corresponding to these dis— orders. The diagnosis of these hemoglobin dis- orders requires structural and functional analy- ses of the aberrant proteins, and not all cases would be identified initially by their phenotypic abnormalities; some degree of underascertainment is inevitable in these data. This is particularly true in the data for the unstable hemoglobins, where clinical manifestations of the genotype range from mild to severe. The more severe cases are likely to be identified and reported in the literature. Overall, the order of magnitude of the estimates for gene mutation rates, 1 mutation in 1 million genes per generation, is consistent with that of the estimates of the spontaneous heritable mutation rate made by Neel and colleagues in the Japanese populations that comprised the control group in their ongoing study of the genetic effects of the atomic bombs (see below). This suggests that even though the rates derived from these data are based on mutations observed in only alpha- and beta— globin genes, they are not very different from mu- tation rates in other expressed genes. STUDIES OF POPULATIONS EXPOSED TO ATOMIC RADIATION The first demonstration that ionizing radiation could induce genetic mutation is usually credited to Muller (81) for his work with Drosophila, and since then, many scientists have examined the na— ture and frequency of induced mutations in vari- ous species of animals, including whole mammals (120). There is an extensive body of experimental data that suggests that exposure to radiation and to certain chemicals can induce mutations in mam- malian germ cells. In humans, exposure to ioniz- ing radiation is known to cause somatic mutation, and is suspected to enhance the probability of ger— minal mutation (165). To date, however, the available methods have provided no documented evidence for the induction (by chemicals or by ra- diation) of mutations in human germ cells. The single largest population at risk for induced heritable mutations is the group of survivors of the atomic bombs dropped on Hiroshima and Nagasaki in 1945. Survivors of these bombs are estimated to have received doses of radiation con- sidered to be ”biologically effective"; in experi- ments with mutation induction by acute exposure to radiation in mammals, similar kinds and doses of radiation were sufficient to cause phenotypi- cally apparent mutations in offspring (and pre- sumably other nonexpressed mutations as well). On the basis of this kind of experimental data, the assumption was made that germinal mutations could have been induced in people exposed to the radiation. Efforts to detect genetic consequences of the atomic bombs detonated in Hiroshima and Nagasaki have been in progress continuously since 1946, sponsored by the Atomic Bomb Casualty Commission and by its successor in 1977, the Ra— diation Effects Research Foundation. These studies have attempted to determine whether germinal mutations are expressed as heritable mutations in the survivors' offspring, particularly as ”untoward Ch. 3—Kinds and Rates of Human Heritable Mutations - 49 pregnancy outcomes,” as single gene disorders, as chromosome abnormalities, or most recently, as abnormal blood proteins. Early Epidemiologic Studies The earliest clinical studies, as well as the con- tinuing mortality surveillance study, sought to identify a variety of health problems in the off- spring of exposed atomic bomb survivors. These studies consisted of epidemiologic surveys of 70,082 newborn infants born in Hiroshima and Nagasaki between 1948 and 1953, including about 38,000 infants for which at least one parent had been proximally exposed to the radiation (defined as being within 2,000 meters of the hypocenter, while distally exposed was defined as being greater than 2,500 meters from the hypocenter or not exposed at all). Stillbirths, infant mortality, birth- weight, congenital abnormalities, childhood mor- tality, and sex ratio were examined. An “un- toward pregnancy outcome” was defined as a stillbirth, a major congenital defect, death dur- ing the first postnatal week of life, or a combina- tion of these events. The investigators found no statistically significant, radiation-related increase for any “untoward pregnancy outcomes" (123). One reason for this finding may be that these health problems each have some genetic basis, but that all are influenced by a variety of nongenetic factors as well, some of which may have obscured an effect of genetic mutation. In addition, a con- tinuing study (the F1 Mortality study) of mortal— ity among children born between 1946 and 1980 has not demonstrated a relationship between ra— diation exposure of the parents and mortality in the offspring (56). Cytogenetic Studies Since 1967, Awa and colleagues have screened the children of survivors of the atomic bombs for cytogenetic abnormalities to determine whether exposure of the parents’ gametes to the radiation caused a higher frequency of chromosome abnor- malities in the offspring. After first reporting a higher frequency of sex-chromosome abnormal- ities among 2,885 children of exposed parents compared with 1,090 children of the control group of parents, Awa (6) concluded that a mutagenic effect of the radiation was not demonstrated in these data. The total frequency of chromosomally abnormal children of exposed parents (0.62 per- cent) was higher than in the control group (0.28 percent), although the difference was not statis- tically significant, and was similar to the fre— quency of such abnormalities observed in several studies of unselected, consecutive newborn pop- ulations (120). Results of their more recent analysis of the data showed that 0.52 percent of the offspring of ex— posed parents were determined to have an abnor— mal chromosome constitution, compared to 0.49 percent of the offspring of the control group of parents, which suggests that there is no signifi- cant difference between the two groups of off- spring in frequency of autosomal balanced and unbalanced rearrangements and of sex-chromo- some numerical abnormalities (8). This study did not provide information on chromosome abnor- malities which are associated with high mortal— ity rates, such as most of the unbalanced struc— tural rearrangements and numerical chromosome abnormalities involving the autosomes. Biochemical Analyses Beginning in 1976, one-dimensional electropho- resis was used to look for mutations affecting pro- teins found in red blood cells and blood plasma. Rare variants of these proteins, differing by a sin- gle constituent amino acid, are taken to represent mutations which occurred in the germ cells of one of the parents if the variant is detected in a child and is absent from both parents. Classification as a “probable mutation” is made after inquiry into possible errors of diagnosis and of stated parent- age (124). Neel and colleagues (93,96,121) analyzed 28 en- zyme and serum proteins from blood samples ob- tained from children who were identified in the F1 Mortality study and who were born between 1959 and 1975 to survivors of the atomic bombs. Over 10,000 children were examined in each group of exposed and unexposed parents. These children were also participants in the cytogenetic studies, and the biochemical analyses made use of the same blood samples that were used for the 50 - Technologies for Detecting Heritable Mutations in Human Beings chromosome analyses. To date, two probable mu— tations have been identified among offspring of proximally exposed parents (estimated to have re— ceived an average conjoint gonadal exposure of approximately 600mSv [60 rem”]) in 419,666 10- cus tests, while 3 new mutations have been found among 539,170 locus tests of unexposed individ- uals. Although the number of new mutants is small, a crude mutation rate can be calculated from these data. The rate of mutation in the ex- posed group is approximately 5 mutations in 1 million genes, and the rate in the unexposed group is 6 mutations in 1 million genes per generation. Such rates have wide margins of error and are not statistically different (96). In 1979, the tests of electrophoretic variants were expanded to include assays for reduced activ- 13The doses of radiation that the survivors are estimated to have received are currently being re—evaluated in light of new informa- tion concerning the quantity and quality of radiation released by the two bombs (125). The revised estimate of exposure will prob— ably be lower than the current estimate of 60 rem. CONCLUSIONS This chapter has described several currently fea— sible approaches for determining rates of herita- ble mutations. Each of the methods has its par- ticular problems. The genetic defects underlying the sentinel phenotypes are largely unknown. These phenotypes may result from mutations within genes at one or several genes, and may be complicated by the presence of phenotypically similar but nongenetic traits, all serving to bias estimates of their frequency upwards and to cre— ate a wide margin of error. Even with a collec- tion of sentinel phenotypes that can be reliably identified in newborn infants, population surveys of a sufficient size are difficult to perform. To detect a small or moderate increase of sentinel phenotypes, consecutive newborns of entire na- tional populations would have to be monitored thoroughly for many years, perhaps as part of an ongoing surveillance program for birth defects and childhood cancers. ity variants of certain red cell enzymes.” The pres- ence of such variants suggests a deletion in the DNA, producing ”null" mutations which would not be identified by one-dimensional electropho- resis. No such variants were found among 11,852 locus tests of 10 erythrocyte enzymes in children of exposed and unexposed parents (122). Mutation Studies in Populations Exposed to Radiation in the Marshall Islands Neel and his colleagues (93) reported results of electrophoretic studies of blood proteins from off- spring of parents exposed to radioactive fallout from the Bravo thermonuclear test explosion at Bikini in the Marshall Islands in 1954. No vari- ant proteins indicating new mutations were found in 1,897 locus tests of 25 proteins measured in chil— dren of unexposed parents, and in 1,835 such tests in children of exposed parents. “An enzyme level three standard deviations below the mean and/or less than or equal to 66 percent of normal was defined as a deficiency variant. The routine staining and banding of chromo— somes is the most straightforward of the current methods, but they allow only major chromosomal changes to be identified. More advanced band- ing techniques, which are not performed rou- tinely, allow smaller chromosome abnormalities to be detected in some cases. Among the current methods described here, the detection of electrophoretic protein variants comes closest to mirroring precise changes in DNA. However, one-dimensional electrophore- sis is limited by the relatively small number of pro- teins it can examine and by the fact that it does not readily detect the absence of gene products, caused by deletions or by specific nucleotide sub- stitutions. Detection of enzyme deficiency vari— ants, although still in early stages of development, could complement electrophoresis by detecting loss-of—activity variants. In addition, large pop- Ch. 3—Kinds and Rates of Human Heritable Mutations 0 51 ulation surveys are required to generate statisti- cally significant results, and many more paternity tests are likely to be needed as a result of the bio- chemical approach compared to the phenotypic and cytogenetic methods. (Since the protein var- iants are not associated with reduced fertility, as are most of the sentinel phenotypes and chromo- some abnormalities, only a fraction of the pro- tein variants will be new mutations.) An interesting example of an approach for de- tecting mutations in a single gene or in a group of genes is provided by the analysis of mutations underlying unstable hemoglobins and HbM. Analyses of other accessible and well-studied genes may help to generate better estimates of mu- tation rates per nucleotide and per gene on the human genome, although it is not known how representative any one gene is of the remainder of the DNA. Moreover, characterizing the mu— tations at the molecular level provides important information, especially where specific mutagens are suspected to have caused the mutations. The different approaches have led to various estimates of heritable mutation rates in human be- ings (see table 4). Sentinel phenotypes, represent— ing the occurrence of a single dominant mutant gene, are detected in about 2 individuals, on aver- ' age, per 100,000 examined per generation. Indi- vidual types of chromosome abnormalities are up to 50 times more frequent than the average fre- quency of sentinel phenotypes. Analyses of the mutant hemoglobins studied by Stamatoyan— nopoulos and Nute (131) give estimated spontane— ous mutation rates of about three mutations per 1 million genes per generation. The rate measured in one—dimensional electrophoresis of blood pro- teins is also about three mutations per 1 million genes per generation, a figure based on the ex- amination of over 1 million locus tests involving 30 to 40 different proteins. The different approaches currently available for detecting heritable mutations (listed in table 5) in human beings generate data that are generally not comparable. Each set of data reflects the nature of the approach itself, not necessarily the nature of the different genes examined or their spontane- ous mutation rates. It is particularly difficult to compare directly the results of studies of muta— Table 4.—Results of Studies To Detect Heritable Mutations by One-Dimensional Electrophoresis in Human Populations Not Known To Be at High Risk for Germinal Mutations Population Locus tests Mutations United Statesa: Caucasian ...... 218,376 0 Black .......... 18,900 0 United Kingdomb .............. 133,478 0 West Germanyc ............... 225,000 1 Central and South Americad . . . . 118,475 0 Japan“l ....................... 539,170 3 Marshall Islands6 .............. 1,897 0 Total ....................... 1,255,296 7 Overall “spontaneous" mutation rate = 4 mutations/1,255,296 locus tests is equiv— alent to approximately 3 mutations/1 million genes per generation. SOURCES: aJ.V. Neel, H,W. Mohrenweiser, and M.H. Meisler, ”Rate of Spontaneous Muta- tion at Human Loci Encoding Protein Structure," Proc. Natl. Acad. Sci. USA 77(10):6037-6041, 1980; H.W. Mohrenweiser, 1986, unpublished data, cited in Neel, et al., 1986. bH. Harris, D.A. Hopkinson, and EB. Robson, "The Incidence of Rare Alleles De- termining Electrophoretic Variants: Data on 43 Enzyme Loci in Man," Ann. Hum. Genet. Lond. 37:237-253, 1974. OK. Altland, M. Kaempfer, M. Forssbohm, and W. Warner, ”Monitoring for Chang- ing Mutation Rates Using Blood Samples Submitted for PKU Screening," Hu- man Genetics, Part A: The Unfolding Genome (New York: Alan R. Liss, 1982), pp. 277-287. ‘ dJ.V. Neel, C. Satoh, K. Gorikl, et al., "The Rate With Which Spontaneous Muta- tion Alters the Electrophoretic Mobility of Polypeptides,” Proc. Nat. Acad. Sci. , USA 83:389-393, 1986; J.V. Neel, unpublished data. ‘ eJ.V. Neel, H. Mohrenwelser, S, Hanash, et al., “Biochemical Approaches To Mon- itoring Human Populations for Germinal Mutation Rates: I. Electrophoresis," Utilization of Mammalian Specific Locus Studies in Hazard Evaluation and EsA timation of Genetic Risk, F.J. deSerres and W. Sheridan (eds) (New York: Ple- num Press, 1983), pp. 71-93. Table 5.—Spontaneous Mutation Rates of Genetic Abnormalities in Human Populations Mutationsl Genetic anomaly generation References Sentinel phenotypes 2 mutations (166) per 100,000 individuals Chromosome 2 to 121 mutations Table 3-2 abnormalities per 100,000 individuals Electrophoretic 3 mutations per protein variants million genes Analysis of unstable 3 mutations per hemoglobins million globin genes SOURCE: Office of Technology Assessment. Table 3-4; (96) Table 3-3; (131) tion rates based on clinical syndromes with bio- chemical studies. Measurements from each of the current methods can be used to generate a sepa- rate mutation rate and, for the time being, those rates can be viewed as baselines for the frequency of various kinds of human mutations and as refer- ence points for some of the most severe effects of heritable mutations. 52 - Technologies for Detecting Heritable Mutations in Human Beings The different approaches currently available re— flect the various forms that mutations and their consequences take. Each of the current approaches to studying mutations is generally limited to par- ticular kinds of mutations or to a particular set of outcomes. The cytogenetic and biochemical ap- proaches discussed are limited to detecting a small subset of mutations that occur in human DNA. Identification of sentinel phenotypes and chromo- some abnormalities limits the observations to a particular set of outcomes, mostly in the severe end of the spectrum. The available data are based on mutations observed in a small fraction of the genes, about 40 to 50, out of a total of about 50,000 to 100,000 in the human genome. Aspects of the current methods that limit their scope also make them inefficient. The proportion of all mutations that these methods can identify is relatively small and the limited number of suit— able genetic endpoints (acceptable sentinel pheno- types, chromosome abnormalities, and protein variants) makes it difficult to sample more than a few mutational events per individual. As a re- sult, huge populations of ”exposed" and “un- exposed" people would be needed to find out whether the exposure resulted in excess heritable mutations (71). Not surprisingly, perhaps, there is no documented evidence to date for a meaning- ful difference in rates for any type of mutation between two populations (or in the same popu- lation at different times). Sentinel phenotypes and cytogenetic analyses provide a highly selective approach to some of the worst possible consequences of new muta- tions. However, they do not permit more than a partial view of the broader consequences of mu- tations in human beings. Detailed molecular, physiological, and clinical information on the kinds and rates of new mutations will eventually be needed to assess the immediate or potential im- pact of new mutations. Chapter4 New Technologies for Detecting Heritable Mutations Contents Page Introduction ............................................................ 55 Short Descriptions of the New Technologies ................................ 56 Two-Dimensional Polyacrylamide Gel Electrophoresis ..................... 56 DNA Sequencing ...................................................... 57 Restriction Fragment Length Polymorphisms .............................. 58 One-Dimensional Denaturing Gradient Gel Electrophoresis ................. 59 Two—Dimensional Denaturing Gradient Gel Electrophoresis ................. 61 Ribonuclease Cleavage of Mismatches in RNA/ DNA Heteroduplexes ........ 64 Subtractive Hybridization .............................................. 66 Pulsed Field Gel Electrophoresis ......................................... 68 Conclusions ............................................................ 68 TABLE Table No. Page 6.Current and New Methods for Detecting Human Heritable Mutations ....... 55 FIGURES Figure No. Page 7. Production of Restriction Fragment Length Polymorphisms Using Restriction Enzymes and Separation of Different DNA Fragment Sizes by Agarose Gel Electrohoresis .................................... 59 8. Restriction Fragment Length Polymorphism ............................. 60 9. One-Dimensional Denaturing Gradient Gel Electrophoresis ................ 62 10. Two-Dimensional Denaturing Gradient Gel Electrophoresis ............... 63 11. Ribonuclease Cleavage of Mismatches in RNA/ DNA Heteroduplexes ....... 65 12. Subtractive Hybridization ............................................. 67 Chapter 4 New Technologies for Detecting Heritable Mutations INTRODUCTION New developments in molecular biology have opened the way for direct examination of large portions of human DNA to detect kinds and rates of mutations. Until recently, it has been possible only to infer the occurrence of genetic damage from indirect, limited, and often imprecise obser— vations (see ch. 3). In principle, these new tech- nologies would allow samples of DNA from peo— ple exposed to potential mutagens to be examined for precise genetic damage. Little is known about the nature and frequency in humans of the majority of mutational events, since previous information was available only on DNA sequences that could be selected on the ba— sis of gene expression. Until recently, human mu- tation was studied in fine detail only in particu— lar genes (e.g., hprt and globin genes) that could be fished out of genomic DNA by using some gene-specific probe, or in genes whose products (mRNA, proteins, etc.) could be isolated and examined. In this chapter, new methods are described for examining proteins and for examining large re- gions of genomic DNA (see table 6), regardless of their function or location in the genome and without devising selection methods for specific known genes. These new technologies are cur- rently under development or being considered by the scientific community. Most of them were ex- plored as potentially feasible technologies at a workshop cosponsored by the International Com- mission for Protection Against Environmental Mutagens and Carcinogens (ICPEMC) and the U.S. Department of Energy, in December 1984. Table 6.—Current and New Methods for Detecting Human Heritable Mutations Indicators of mutation Method of detection Current methods: Sporadic genetic disease Numerical or structural Cytogenetic analysis chromosome abnormalities Variant blood proteins Sentinel phenotypes 1D-electrophoresis Quantitative enzyme analysis 2D-electrophoresis New methods: Altered DNA sequences DNA sequencing Restriction fragment length polymorphisms 1D-denaturing gel electrophoresis 2D-denaturing gel electrophoresis Ribonuclease cleavage Subtractive hybridization Pulsed field gel electrophoresis SOURCE: Office of Technology Assessment. Several different approaches are used in these new technologies. Two-dimensional electropho- resis of proteins follows from the techniques of one-dimensional electrophoresis and incorporates new methods of analyzing protein spots on the gel. In several of the DNA methods, restriction enzymes are used to fragment genomic DNA be— fore sizing and sequence analyses are performed. Other technologies hinge on the production of hy- brid DNA molecules, or heteroduplexes, and on gradient denaturing gels that are used to identify heteroduplexes with base pair changes. Another method devises a way of separating larger frag- ments of DNA than were previously possible to separate. These new methods are discussed below. 55 56 - Technologies for Detecting Heritable Mutations in Human Beings SHORT DESCRIPTIONS OF THE NEW TECHNOLOGIES Two-Dimensional Polyacrylamide Gel Electrophoresis A development following from the experience with one-dimensional electrophoresis for protein variants (see ch. 3) is two—dimensional polyacryl- amide gel electrophoresis, “2-D PAGE," the only technique described in this chapter that is cur- rently in use in studies of human beings. With this technique, proteins are separated first by iso- electric focusing on the basis of their molecular charge, and then by electrophoresis on the basis of their molecular weight. From 500 to 1,000 pro- teins may be visible from a single sample on one gel, and about 100 of these may be sufficiently clear to detect rare variants in a child that are not present in the parents (92). To search for new mutations in proteins using 2—D PAGE, blood samples are taken from child/ mother/ father trios. The program that has had most experience with this technique for mutation research, the University of Michigan Medical School, collects cord blood from newborns and parental blood samples at the time of the child’s birth. Each blood sample is fractionated into six parts: platelets, plasma, nonpolymorphonuclear cells, polymorphonuclear cells, erythrocyte (red blood cell) membranes, and erythrocyte cytosol (the contents of the cell) (95). Each fraction is treated and analyzed separately. Separation in the first dimension is by the charge of the protein molecules, qualitatively sim- ilar to the separation using one-dimensional elec- trophoresis. Each sample is run on a long, thin (1 to 2 mm in diameter) cylinder of polyacryl- amide in a hollow glass tube using a technique called isoelectric focusing. The separation time can be speeded up or slowed down depending on the voltage applied. After the first separation, the gel, which looks like a clear noodle, is "extruded" from the tube. In the second separation, the ”noodle” is laid across the top of a gel about 7 inches square, that is held between two glass or plastic plates. A cur- rent is passed through the gel, and the protein molecules migrate down through the gel at a speed that is proportional to their size. A final step renders the proteins visible by autoradiography (for proteins that can be isotopically labeled) or by stains such as Coomassie Brilliant Blue and silver- or nickel-based stains. The trained eye or a computer program detects patterns of spots on the gel, though most of the stained areas are too crowded or indistinct for drawing conclusions about specific proteins. Only spots in relatively clear areas and of a sufficient intensity are chosen for scoring. The analysis en- tails comparing the parent/child trios for differ— ences. No two gels are run in precisely the same way, so it is often difficult to superimpose a set of three gels over one another and to look for differences. While protein spots bear the same relationships to each other on different gels, they routinely appear at slightly different coordinates. For each type of sample mentioned above, a standard constellation of protein spots is scored for genetic variation. A mutant protein can ap- pear on the gel as a spot for which there is no cor— responding spot on parental gels. The gels can be interpreted visually, which is the traditional method, and currently the most accurate, but computer—assisted interpretation is desirable. Two steps in analysis lend themselves to computeriza- tion. First, and now routinely done, the protein spots can be ”read” by computer for location and intensity. The second step is computerized com— parison of different gels. Two approaches for comparison are being developed in different lab— oratories, each of which will probably be useful on a ”production scale" in the near future (92). Eventually, computers will be relied on to screen gels, and to eliminate from further consideration the overwhelming majority of parent/child trios with no variants. In comparison to one-dimensional protein sep— aration, a relatively small number of parent/child trios have been analyzed with 2-D PAGE, and no heritable mutations have been found. However, the results of these analyses have revealed a wide array of largely unknown proteins, and distinct differences in the spectrum of proteins from differ- ent cell types. The range of electrophoretic vari- Ch. 4—New Technologies for Detecting Heritable Mutations - 57 ants of a number of these proteins have been char- acterized in plasma (104) and in erythrocyte lysate (105). In plasma, the variation is substantially higher than has been reported in other blood fractions. The advantage of 2-D PAGE is that it can be done now without a major developmental break- through and without a great initial cost, though the running expenses are significant. It may be some years before any of the DNA analytic meth- ods catch up to protein analysis in practical ap- plication. The future of protein analysis for de- tecting mutations after DNA methods become routine is unclear, however. There are technical limitations of the 2-D PAGE technique. First, the technique itself is technically demanding. Second, the number of proteins that can be scored on each gel is limited by the degree of separation possible on plates that are of a prac— tical size. As with most of the other techniques, only a certain spectrum of mutations is detectable with 2-D PAGE. Chromosome rearrangements in gen- eral are not revealed through electrophoretic var- iants. This technique has its greatest strength in detecting point mutations and larger deletions in the DNA that change the size or charge of a pro- tein molecule. Point mutations that do not cause such changes, and small insertions or deletions, are less likely to be detectable with this method. Currently, null mutations cannot be detected relia- bly, and this is probably the greatest disadvantage of 2-D PAGE. Overall, 2-D PAGE should detect an estimated 25 to 35 percent of all spontaneous gene mutations. Several aspects of 2—D PAGE of blood proteins, in addition to computerized com- parisons of gels, are amenable to improvement, and at least one approach is being developed to improve detection of null mutations (92). The identity of most of the proteins that can be visualized on two-dimensional gels is un- known. Although it has not yet been done, the technology exists to recover proteins from gels, purify them, and then determine their amino acid sequences (92). Once amino acid sequences are worked out, the nucleotide sequences that code for them can be deduced and nucleic acid probes can be assembled to correspond to the protein— coding nucleotide sequences. The probes could theoretically be used to locate the corresponding DNA within the human genome, thereby deter- mining exactly where in the genome the coding sequence lies. DNA Sequencing Conceptually, the most thorough and straight- forward analysis of the genome would consist of lining up the chromosomes and determining the nucleotide sequence from end to end on each one. Once the sequence was known, it could be stored in computer memory. The advantage of determining the complete genomic sequence would be that any nucleotide sequence that was identified in any laboratory could immediately be mapped to its chromosomal location. As nucleotide sequences for various parts of the genome were obtained in different lab- oratories, both common polymorphisms1 and rare mutations could be identified by making compar- isons between the total genomic sequence and par- tial sequences. Furthermore, the method would detect mutations and polymorphisms no matter where they occur—in introns, exons, and repeti- tive sequences. Walter Gilbert, Fred Sanger, and their col- leagues developed the theoretical basis for the techniques that makes it possible to consider se- quencing the genome (67,118) and in 1984, Ge- orge Church and Walter Gilbert described a method for sequencing any part of the genome (20), although the method could be applied to se— quencing the entire genome. The genomic DNA would be cut into pieces small enough to be han- dled in the laboratory, and then chemical meth- ods would be used to determine the nucleotide se- quence within each fragment. The methods are well understood, are being applied in dozens of laboratories around the world, and are a stand- ard exercise in the education of molecular biol— ogy graduate students (33). Through February 1Polymorphisms are "common" mutations that are present in the population at a relatively high frequency. Several alternative ver— sions of the same basic gene sequence, all equally functional, can co-exist. New mutations are defined as occurring in less than 1 per- cent of the population, excluding the more common polymorphic variants. 58 - Technologies for Detecting Her/table Mutations in Human Beings 1985, the methods have been used to examine dis- crete and defined segments of DNA from organ- isms with large genomes and complete genomes from smaller organisms, sequencing a total of 4,045,305 base pairs (150)—roughly 1/800 of the human genome. Practically, we are far away from being able to sequence the entire genome, not because the techniques are unavailable but because they are expensive and slow when applied to a task of the magnitude of sequencing 3 billion nucleotides. A meeting during the summer of 1985 considered this enormous task and estimated the cost at about $1 per nucleotide, or $3 billion overall. The ef- fort can also be estimated in terms of technician— years, and that comes to about 200 technician- years to sequence 100 million nucleotides, which would be expected to yield about one new mu- tation. Although it is conceptually satisfying, the enor- mous resources necessary to sequence the entire genome make it unlikely that the task will be un- dertaken soon. However, genomic sequencing technology is still progressing, and further auto- mation of the process may increase its efficiency. Continued encouragement of those efforts and provision of sufficient resources to develop tech- niques may make it possible to sequence the en- tire human genome in a reasonable time. When it is done, it will probably be undertaken as part of an effort to provide more general information about human genetics, rather than to identify hu- man mutations. If the genome is sequenced, for whatever reason, it will be a great boon to muta— tion studies by providing a reference for all tech- niques that involve sequencing or the use of re— striction enzymes. Restriction Fragment Length Polymorphisms A method that offers a less exhaustive survey than genomic sequencing, but which is potentially more efficient, involves the use of restriction en- zymes and gel electrophoresis to detect nucleo- tide substitutions or small deletions in the DNA (98). Genomic DNA is isolated from white blood cells and treated with restriction enzymes, chem- icals that recognize specific sequences in the DNA and “cut” the DNA wherever such sequences oc— cur (see fig. 7). (The number of fragments pro- duced is determined by the frequency of occur- rence of the particular enzyme recognition site in the DNA sequence.) Restriction site analysis does not, in practice, examine every nucleotide. How- ever, the use of a set of combined restriction en- zymes increases the number of restriction sites identified, allowing examination of a larger por- tion of the DNA, including both expressed and nonexpressed regions. After the DNA is fragmented (see fig. 8), there are two alternative ways of proceeding, one using a gene cloning method, and the other using a non- cloning, direct method. Cloning is the process of ”growing” human DNA in other organisms; it reduces the amount of DNA needed—in most cases corresponding to the amount of blood that must be drawn—but it increases the number of laboratory manipulations. Additionally, it introduces some uncertainty about the possibility that DNA changes will oc- cur during cloning. This can be checked, but re- quires diligence. In the gene cloning method, the human DNA fragments are incorporated into the genetic ma- terial of a virus called “lambda.” Lambda repro- duces in the bacterium E.c011', and large quanti- ties of viral DNA containing the human segments are produced. Following isolation of the cloned human DNA, each clone is treated with a set of restriction enzymes that cut it into smaller frag- ments. These collections of fragments are then separated by gel electrophoresis, producing dis- tinct bands visible by gel staining methods. A visual comparison of the location of the bands de- rived from parental DNA and from children’s DNA shows whether a mutation has occurred in a site for a restriction enzyme. The presence of a band in the child's DNA that is not present in the parents’ DNA, or vice versa, indicates a nucleotide substitution or deletion restriction en- zyme recognition site and suggests that a herita- ble mutation has occurred. Ch. 4—New Technologies for Detecting Heritable Mutations ' 59 Figure 7.—Production of Restriction Fragment Length Polymorphisms Using Restriction Enzymes and Separation of Different DNA Fragment Sizes by Agarose Gel Electrophoresis <—c————> 4—b-> 4———d-—-——> <—a-> Agarose sizing gel electrophoresis ( ~ ) Larger fragments DNA l Ir V W 4 fragments result: c, b. d, and a (I <———c———> ¢ ‘¥——l Direction current DNA It If N l 1 Base change leads to loss of restriction enzyme site and creates 3 fragments c, e, and a 4v»— T = restriction enzyme site of cleavage SOURCE: Olfice of Technology Assessment. An alternative method bypasses the lambda cloning procedure, and analyzes DNA fragments directly. Fragments from the entire human ge- nome are spread out on an electrophoretic gel, forming smears of indistinguishable bands. In or- der to visualize the location of particular segments of DNA and to evaluate their position relative to the equivalent band in the reference sample, the naturally double-stranded DNA is dissociated into single-stranded DNA while still in the gel and al- lowed to reassociate with small pieces of specific genes which are added to the gel and which are radioactive (probes). Those original DNA se- quences that pair with the probes are now tagged. The DNA is then transferred to a flexible plastic membrane conserving the spatial arrangement of the DNA (a procedure known as Southern blot- ting) and autoradiographed; bands are detected where the tagged sequences are present. As with the cloning method above, the appearance of a Smaller fragments (+) DNA l DNA ll Positions of different fragment sizes after treatment with restriction enzyme new band or the disappearance of an old band suggests a nucleotide substitution or deletion in a recognition site of one of the restriction enzymes. One-Dimensional Denaturing Gradient Gel Electrophoresis A modification of the standard electrophoretic gel procedure, proposed by Fischer and Lerman (34), allows DNA to be separated not only on the basis of size, but also on the basis of sequence of nucleotides even if differences do not occur at restriction enzyme recognition sites. Double- stranded DNA dissociates into single-stranded DNA when it is heated or when it is exposed to denaturing chemicals (e.g., formamide or urea). A gradient of increasing strength of such chemi- cals can be produced in a gel so that DNA sam- ples will travel in the direction of the electric cur- 60 - Technologies for Detecting Heritable Mutations in Human Beings Figure 8.—Restriction Fragment Length Polymorphism 6) Mother‘s DNA Father‘s DNA Child's DNA ’ ‘ . ’il . ’I l [I \ Treat wrth [I i Treat wrth / \ Treat with /’ \ restriction 1' I i restriction ’1’ ‘| $ restriction / l| enzymes 1' ;\z enzymes N enzymes I, lB L—/\ J ~\ X \ \l V N N / \_,\ k V Clone into lambda Fragments ~ — — — ’ .Of -" \ ’\ 1—/ Viral , F’ A; DNA fi " Separate fragments by size by agarose gel electrophoresis (‘) M F 0 Electric _ - current ¢ _ (+) ‘ /\ DIRECT (NONCLONING) METHOD Separate fragments in an electrophoretic gel (-) Associate and reanneal with “PcDNA probes M F C Visualize bands using autoradiography Isolate geonomic DNA from white blood cells. Using restriction enzymes, cut DNA into double-stranded fragments of various lengths. Gene Cloning Method: Clone each fragment into the bacterial virus lambda, infect into Eco/i, grown in petri dishes, and allow the virus to replicate within the bacteria. Isolate large quantities of the viral DNA, which now contains segments of human DNA, and cleave it with restriction enzymes. a Separate DNA fragments using agarose gel electrophoresis. Visualize bands corresponding to different sizes of fragments using fluorescent staining. Fragments found in the child’s DNA samples and not in the parents’ DNA may con- tain heritable mutations. These bands can be removed and the DNA analyzed for specific mutations. Direct (noncloning) Method: Apply samples of DNA fragments to an electrophoretic gel, producing smears of indistinguishable bands. Dissociate DNA fragments into single strands and incubate with radioactive, single-stranded, 32P-labeled DNA probes for specific human genes. The probes hybridize with complementary sequences in the DNA and label their position in the gel. Visualize the position of bands using autoradiography. Changes in the position of bands in the child’s DNA compared to the parents‘ DNA suggest possible new mutations. Bands can be isolated and DNA analyzed for sequence differences. SOURCE: Office of Technology Assessment Ch. 4—New Technologies for Detecting Heritable Mutations - 61 rent, separating by size, and will also begin to dissociate as they reach their particular critical concentration of denaturing chemical. Dissocia- tion causes the molecule to split into constituent parts, or unravel, and get stuck in the pores of the gel (84). Every unique strand of DNA dissociates at a unique concentration of denaturant. In fact, the difference of only one nucleotide between two otherwise identical strands of double-stranded DNA of 250 nucleotides in length is enough to cause the strands to dissociate at different con- centrations of denaturant chemical, and to stop travelling at different locations in the gel. Again, a comparison between the banding pattern of par- ents’ and child's DNA analyzed in this way may identify a wide range of mutations in all regions of the DNA. In this method, total genomic DNA is isolated from individuals and is heated to form single- stranded DNA (see fig. 9). It is then mixed with radioactive, single-stranded DNA probes that cor- respond to a small portion of the genome. Fol- lowing incubation so that double-stranded hybrid molecules (“heteroduplexes”) will form, they are treated with restriction enzymes and electrophoresed in a denaturing gradient gel. The presence of a single mismatch between a nucleotide in the “nor- mal” probe and the corresponding segment from a person bearing a mutation makes the heter- oduplex slightly more sensitive to denaturing chemicals so that such a molecule will stop trav- eling in the gel before the point at which a per- fectly paired heteroduplex will stop. The gel is dried and exposed to X-ray film, and the result- ing autoradiogram is examined visually for differ- ences in banding patterns between parents’ and child’s DNA. Two-Dimensional Denaturing Gradient Gel Electrophoresis Leonard Lerman has proposed a technique whereby a two-dimensional separation of parents’ and child's DNA is used to differentiate among DNA sequences common to all three members, polymorphisms in either parent which are trans- mitted to the child, and new mutations in the child’s DNA (57). Like the two—dimensional poly- acrylamide gel procedure for protein separation described above, two-dimensional denaturing gradient gel electrophoresis (ZDGGE) compares locations of spots on a gel (in this case, DNA spots) representing various types of mutations in the nucleotide sequence in expressed and nonex- pressed regions of the DNA. This method sepa— rates DNA fragments on the basis of differences in base composition (or sequence), after separat- ing them on the basis of differences in length. 2DGGE detects differences between DNA hetero— duplexes by responses of their structure to grad- ual changes in denaturant concentration in the gel. Genomic DNA from the parents in one sample and from the parents and children in another sample would first be treated with restriction en- zymes and separated by size on separate agarose gels in a single lane (see fig. 10). The gel strips are then laid across the top of two denaturant gra- dient gels and the DNA is electrophoresed through the increasing gradient of denaturant. Each gel is then cut into horizontal slices, each of which con- tains pieces of the original homoduplex DNA of all sizes, but including only those pieces that dis- sociated at the same denaturing concentration ("isomelting” groups). These gel slices containing homoduplexes of parents’ DNA or of parents’ and child's DNA are physically removed (or collected on removable membranes) heated, and allowed to reassociate as heteroduplex molecules (see fig. 10). In order to differentiate between common areas of the DNA among parents and children, another gradient denaturing gel is used to denature the mixtures of double-stranded, isomelting DNA of the parents (M1/ M; + F1/ F2) and another of the parents' DNA mixed with the child's (Ml/M2 + F1/ F; + Cl/Cz). Various double stranded combi- nations of parents' and child’s DNA are produced (e.g., Mz/Cl, Fl/Cz, etc.). The two sets of mix- tures are compared in the final two gels. Sequences that pair up perfectly represent com- mon sequences to all three (level “i” in fig. 10). At the bottom, a narrow region, dense with hetero- duplex fragments, represents the largest fraction of the sample—those that are perfectly matched between base pairs and therefore are located at the same denaturant concentration as the origi- nal homoduplexes contained in the narrow slice 62 0 Technologies for Detecting Heritable Mutations in Human Beings Figure 9.—One-Dimensional Denaturing Gradient Gel Electrophoresis Mother's DNA Father’s DNA Child’s DNA Ill i“ I ‘ I ‘ / \ Treat with ,’ \\ $ Treat with / 'I‘ i Treat with l/ \\ restriction / \\ restriction / I restriction I t enzymes ,’ ‘. enzymes / ‘. enzymes 1’ V N / \v ”\J / .% % ’\/ w ”L N . / \_/ ® Dissociate and Dissociate and Dissociate and i reanneal with l reanneal with i reanneal with ”P-DNA probe ”P-DNA probe "P-DNA probe v M /\—r N N N N N Forms heteroduplexes w /\_/ 1 Forms Forms heteroduplexes heteroduplexes W w @ w Heteroduplex % w witha mismatch l w W Separate fragments in gradient denaturing gel Probe Father Mother Child (+) Increasing concentration of denaturant @lsoiate genomic DNA from white blood cells. Using restriction enzymes, cut DNA into (double-stranded) fragments of various lengths. Dissociate double-stranded DNA fragments into single strands, and reanneal in the presence of radioactive 3’P-labeled, single-stranded DNA probes. Heteroduplexes form between probe and sample DNA even if the base sequences are not perfectly complementary; if mutations are present, some of these heteroduplexes will contain mismatches. @Separate heteroduplex fragments in a denaturing gradient gel. Fragments with mismatches denature in a lower analyzed for new mutations. concentration of denaturant than fragments that are perfectly complementary. Visualize the position of the fragments with autoradiography. Fragments containing the child’s DNA that denature sooner than the parents‘ fragments can be SOURCE: Office of Technology Assessment. Ch. 4—New Technologies for Detecting Heritable Mutations - 63 FIgure 10.—Two-Dimenslonal Denaturing Gradient Gel Electrophoresis CD Mother’s DNA Fathers DNA Child 3 DNA Treat with II I". Treat with Treat with I ‘, restriction , “ restriction 1.\) restriction ,’ Q enzymes by enzymes \J ;A\e_n’zymes I v V (9 (M./M:/ (“Rafa (MI/M2 + F./F2) (MI/M2 + FI/Fz + CJCz) Parents‘ DNA Parents’ & child’s DNA @ (—) I —) Larger fragments Separate fragments Larger fragments in an agarose sizing gel Smaller fragments electrophoresis Smaller fragments (+) (+) ¢ Denaturing & ® I I I II I I l gradient " IIIIIIII ll . gels . Increasmg Increasmg concentration concentration of denaturant Remove , , . of denaturant slice ‘ ' ‘ Homoduplexes Homoduplexes © Irvin-A .- -2--. 3:1.“ 7""- :~- m4 .‘I‘,'.-I-;,.--A -. "I l'\"’v 1-3 .,_...'..-I.“~;‘. _3_ siQh --,--..«'-"-. _,_ l Dissociate & reanneal * Dissociate & reanneal ta. .. .o‘-;;_. . .9...“ ,1 -__-. ... -; .g--, A _-;. .~:- 1 u‘,'-‘“~-.._‘-.-;~y- .-—.- .-. “-1.; Mt ..- 5-323"; . 7...-J UT] Heteroduplexes . Heteroduplexes DenaturIng gradient gels Increasing concentration of denaturant Increasing .. .. ‘ ”' concentration <———II ofdenaturant mat-ram“. m. rf’i Heteroduplexes from Heteroduplexes from parents’ DNA parents’ 8. child’s DNA Isolate genomic DNA from white blood cells. Using restriction enzymes, cut DNA into double-stranded fragments of various lengths. Mix mother's and father‘s DNA fragments (M,/M2 + F,/F,) in one set, and mix parents‘ and child's DNA fragments (M,/M2 + F,/F2 + C,IC,) in another set. Separate fragments by length in an agarose sizing gel. Cutout alengthwise strip of the gel, containing a range of fragment sizes, and lay the strip across the top of a denaturing gradient gel. @Separate fragments along an increasing concentration of denaturant. Cut out a slice of the gel at a narrow interval of the temperature gradient, or collect a similar interval on a removable membrane While maintaining spatial arrangement of fragments in the narrow slice, dissociate the double-stranded fragments into single strands, and reanneal them to form heteroduplexes (e.g., M,/C,, F,/C,, F,/M,, etc.) of various combinations of strands of the original homoduplexes (M,/M,, F,IF,, and C,/C,). Lay the thin strip of heteroduplex fragments arranged by size across the top of a new gradient gel, and separate in the direction of an increasing concentration of denaturants. Comparing the two final gels, the DNA spots are grouped in three distinct regions (see text for explanation). Analyze spots in the highest region of the gel for DNA fragments with possible new mutations. SOURCE: Office of Technology Assessment. 64 0 Technologies for Detecting Heritable Mutations in Human Beings removed from the gel or collected in the mem- brane. Polymorphisms that are present in one parent or the other show up as spots that denature earlier than the perfectly matched DNAs (level “ii"). In the middle region of the gel, a diffuse distribu- tion of fragments represents heteroduplexes with single mismatches between the two strands, caus- ing them to dissociate earlier in the gel (at a lower denaturant concentration, occurring above the perfectly matched heteroduplexes). This region would contain inherited polymorphisms as well as new mutations, but the region would be too dense with DNA spots to distinguish individual spots. Molecules that contain more than one polymor— phism per fragment denature even earlier (level ”iii") since multiple mismatches in a fragment have an additive effect on the fragments stability in the denaturing gradient. The highest area in the gel, where heteroduplexes dissociate in the lowest denaturant concentration, is where fragments with more than one mismatch would be located. These mismatches could be new mutations or inherited polymorphisms. Since the number of such frag- ments in the sample is likely to be small, this re- gion would show discrete, nonoverlapping spots; DNA spots in this third level of the gel can be removed from the gel and analyzed to identify possible new heritable mutations. By comparing the two final gels, new spots should be apparent to a trained eye. These would represent single base pair changes or very small deletions and re- arrangements in all regions of the DNA. In theory, it should be possible to cut many strips of isomelting regions out of the first gel and derive an analysis of each melting region by a ser- ies of second gels. By comparing the child’s DNA to his or her parents’ DNA, this approach allows for large portions of the genome to be examined simultaneously. Ribonuclease Cleavage of Mismatches in RNA/DNA Heteroduplexes A technically simpler approach to detecting mu- tations in cloned and genomic DNA has been pro— posed by Richard Myers (85,86). This technique uses an enzyme, ribonuclease A (RNaseA), that cleaves double-stranded RNA/ DNA heterodu- plexes where a specific mismatch of base pairs oc— curs. RNaseA cleaves the RNA/ DNA molecule where a cytosine (C) in RNA occurs opposite to adenine (A) in DNA. (C normally pairs with G [guanine] and A normally pairs with T [thymi- dine].) The idea behind this method is similar to that of restriction enzymes, each of which cleaves DNA at specific normal sequences, except that RNaseA cleaves RNA/DNA hybrid molecules where mismatches occur. The efficiency of this approach depends on the number of different mis- matches that can be recognized and cleaved. The greater the number of enzymes that can be found to cleave different mismatches, the more types of mutations can be detected. This method should detect nucleotide substitutions over a large por- tion of the DNA. With this method, genomic DNA is first iso- lated and mixed with radioactively labeled RNA probes (see fig. 11). The mixture is heated so that the strands dissociate and then reassociate ran- domly with complementary strands. When RNA molecules bind with DNA molecules of homolo— gous sequences, the result is one type of "heter- oduplex.” “Homologous” means that the sequence is close enough betweenthe two molecules that they will form a stable double-stranded molecule. The presence of a single base pair mismatch does not prevent double-stranded heteroduplexes from forming, however. The heteroduplexes are treated with RNaseA and separated according to size on a standard agarose gel. Perfectly paired molecules will be un- affected by RNaseA and will form bands on the gel that correspond to their original size. How- ever, imperfectly paired molecules will be cut by the enzyme at the site of the C :A mismatch, and two separate fragments will result, showing up as two separate bands on the gel. For analysis, parents' and child’s DNA are treated and elec— trophoresed separately, and the gels compared for different patterns by autoradiography to reveal the occurrence of any possible mismatches. Cur- rently this method is applicable only for (RNA)C: A(DNA) mismatches, which represent 1 of 12 pos- sible types of mismatches between these heter- oduplexes. Mother's DNA Ch. 4—New Technologies for Detecting Heritable Mutations - 65 Father‘s DNA Figure 11.—Ribonuclease Cleavage of Mismatches in RNAIDNA Heteroduplexes Child‘s DNA I / ‘ I I I . I I , I ‘ [I '. Treat wrth I \ $ Treat wrth ,’ ‘. Treat with l ‘. restriction ,’ ‘. restriction I \‘ restriction ® I,’ I enzymes ,’ ‘ enzymes l,’ l‘ enzymes I ‘ N 1' ‘ ”\J I % l N I f\/ I "\9 & \_/, Vt & V, L Dissociate l Dissociate L Dissociate ’\- N V [\d /—\/ N N V \d y N N N \d ssRNA Reanneal with ssRNA Fleanneal with ssRNA Reanneal with probes ’ l ssFtNA probes probes hl ssRNA probes probes "l ssRNA probes Probe 3% w DNA W W m: 5 2; WW (Fragment with my V~mism W, W fl~m,,,. CM: :§ "\_/ atCh) Treat with Treat with Treat with L RNase A ¢ RNase A £ RNase A W W V 5:353: 5 W W ’—\_/ w M, W W MAW W Separate fragments by gel electrophoresis. Look for new fragments in child's DNA not present in parents’ DNA. @ (—) Probe Mother Father Child w . “W \ , (Fragment wrth r—— mismatch cut into 2 pieces) Electric current i (+) lsolate genomic DNA from white blood cells. Using restriction enzymes, cut DNA into double-stranded fragments of various lengths. Dissociate double-stranded DNA fragments to single strands, and reanneal in the presence of single-stranded RNA (ssFlNA) probes, forming RNA/DNA heteroduplexes. Treat the heteroduplexes with RNaseA, an enzyme that cuts RNA/DNA heteroduplexes where certain mismatches occur (e.g., where cytosine in RNA is mismatched with adenine in DNA). @Separate heteroduplex fragments by size using gel electrophoresis. Visualize the position of fragments in the gel by autoradiography. Fragments found in the heteroduplexes containing the child‘s DNA, but not found in those containing parents’ DNA, indicate possible new mutations. SOURCE: Office of Technology Assessment. 66 - Technologies for Detecting Her/table Mutations in Human Beings Subtractive Hybridization Detecting mutations would be much easier if it were possible to ignore the millions of nucleo- tide sequences that are identical in both parents and child and, instead, focus only on the relatively few sequences that are different. Currently, the best estimate for the frequency of human muta- tions is about 1 in 100 million base pairs, or about 30 per genome, so it is necessary to examine 3 bil- lion base pairs from each of the parents and a child and find those 30 that differ. George Church had conceived of a method to find sequences in a child's DNA that are not present in either parent’s DNA (25). While there is no experience with Church’s pro— posed method, and therefore, no information about its feasibility, it is a promising idea. The basis of Church's method is to view the human genome as a group of unique nucleotide se- quences. In this context, “unique sequence" means a sequence that may occur only once in the en- tire genome. To explain what a unique sequence is, consider the shortest and longest sequence in the genome, a single nucleotide and the entire ge- nome. A single nucleotide is not unique; A, T, G, and C occur repeatedly throughout the ge- nome. At the other extreme, the entire genome is unique. Somewhere in between those extremes is a length of nucleotides that is long enough to be unique without being too long to handle ex- perimentally. It turns out that a sequence of 18 nucleotides is long enough to be unique; on statistical grounds, any sequence of that length or greater that occurs once in the genome is not expected to occur a second time. Therefore, if every pos- sible sequence of 18 nucleotides, or ”18-mer," is synthesized, this collection of 18—mers will include sequences that together represent every possible sequence of 18 nucleotides of DNA (there are 4” or 70 billion possibilities). Some proportion of these are actually present in human sequences, some are not. It is feasible to make the entire col— lection of 70 billion different 18-mers, and a sin- gle synthesis provides sufficient 18-mers for many experiments. To look for mutations, DNA from parents and child is isolated and cut with restriction enzymes into relatively small lengths of 40 to 200 base pairs and dissociated into single strands (see fig. 12). This parental DNA alone is mixed with the col- lection of 18—mers under conditions that permit the formation of perfect hybrids: each nucleotide sequence of an 18-mer hybridizes exactly with each corresponding genomic sequence. The con— ditions are also such that an 18-mer that inter- acts with a genomic sequence complementary to just 17 of the 18 nucleotides will not form a sta- ble hybrid. When the hybridization reaction is complete, the 18-mers that did not find a perfect match with any parental DNA sequences are left behind as single-stranded molecules, unchanged by their participation in the hybridization re- action. Because they are single stranded, they can be separated from all the other 18-mers that are hybridized to parental DNA. This unbound frac- tion of the mixture is retained, while the fragments of parental DNA that bind to the 18-mers are dis— carded. This pool of single-stranded 18-mers which did not bind to any parental DNA sequence is then mixed with similarly prepared DNA from the child, and hybridized under the same conditions. This time, however, the molecules of interest are the sequences that bind to the child’s DNA (and not previously to the parents' DNA). Any 18-mer that hybridizes perfectly with a sequence in the child’s DNA will identify a sequence that is present in the child’s DNA and absent from the parents’ DNA. Such sequences may contain new muta- tions present in the child’s DNA. It should then be possible to isolate and characterize this hybrid formed between the 18-mer and child’s DNA, so that the mutant sequence can be determined. If mutations occur at a frequency of 1 per 100 mil- lion nucleotides, and if this technique works per- fectly, it could detect 30 to 40 nucleotide sequences in the child's DNA that are not present in the par- ents’ DNA. This approach is the least well developed of all the ones discussed in this report, and its feasibil- ity is unknown. If it does prove feasible, this ap- proach could identify short sequences containing mutations in any part of the DNA, allowing fur- ther detailed study (e.g., by DNA sequencing) of the kinds of mutations that are occurring. Ch. 4—New Technologies for Detecting Heritabie Mutations 0 67 Figure 12.—Subtractlve Hybridization 6) Mother‘s DNA Father’s DNA Child's DNA Cleave DNA with restriction enzymes I I I Cleave DNA with ,’ I I I .' :3 I. I! ; i restriction enzymes II: _ :4 LN' V E:' " éw : 7- % M ——\_, t; N_/ “V 94/4 % t ’2 § % \\_$——; Parents’ DNA fragments Child’s DNA fragments ® 18mer library Dissociate and reanneal with 18mer sequences. Forms bound heteroduplexes and leaves some unbound fragments. W “~'“"’“" ~. _ Separate ~ 18mers Separate and discard bound fragments Unbound 18mers @ Bound fragments M W (heteroduplexes) Dissociate child’s DNA fragments “:25 “AL": & reanneal in presence of unbound 18mers “- 21/ Clone and characterize heteroduplexes for new Discard :_\\—/_/ mutations @ unbound fragments M N W \ / M Bound fragments are 18mers that hybridized with @ W Unbound 18mers w” and M child‘s DNA, after not child's DNA k4 binding to any fragment of parental DNA ~' in homoduplexes Isolate genomic DNA from parents' and child‘s white blood cells; mix parents' DNA samples together, while keeping ®child‘s DNA separate. Cleave DNA into fragments with restriction enzymes. @Dissociate parents‘ DNA fragments into single strands, and reanneal with a set of unique single-stranded 18mer sequences. 18mer sequences complementary to sequences in the parents’ DNA will form double-stranded, or bound, fragments. The remainder, the unbound fraction, will remain single stranded. @Separate the bound fraction from the unbound fraction. The bound fraction is discarded. Unbound 18mers for which no parental complement exists are allowed to hybridize with the child’s DNA fragments under the same conditions of dissociation and reannealing. Any fragments of the child’s DNA that hybridize with unbound fraction of 18mers represent sequences present in the child’s DNA that are not present in either parents’ DNA. @ Isolate these heteroduplexes and analyze them for specific new mutations, SOURCE: Office of Technology Assessment. 68 - Technologies for Detecting Her/table Mutations in Human Beings Pulsed Field Gel Electrophoresis If human DNA were short and simple, it could be cut up with restriction enzymes and separated electrophoretically into discrete bands, each rep- resenting a particular segment of the total DNA. However, human DNA is so long that electropho- retically separated fragments make a continuous smear of bands. If the DNA were cut in only a few places to produce only 100 or 200 bands, the pieces would be too big to pass individually through the pores of a standard electrophoretic gel. A new technique, pulsed field gel electropho— resis (PFGE) is being developed to allow separa— tion of fragments of human DNA larger than can be separated with other electrophoretic techniques and to examine such fragments for evidence of mutations. In theory, the procedure will detect submicroscopic chromosome mutations, includ- ing rearrangements, deletions, breaks, and trans- positions. At present, the method cannot handle intact human chromosomes although it works very well with smaller chromosomes from lower organisms and with small fragments of human chromosomes (157,158). The standard process for isolating genomic DNA randomly breaks the long molecules into pieces. In the PFGE procedure, genomic DNA is treated so that random breaks are avoided. Whole cells are suspended in liquid agarose and after the agarose solidifies into a gel, enzymes are added to degrade proteins and RNA. This apparently minimizes random damage to the high molecu- lar weight DNA. DNA is cut into exact and pre- dictable fragments by addition of a “rare cutter,” a restriction enzyme (or group of enzymes) that cuts DNA consisting of 6 billion nucleotides, into CONCLUSIONS In a short period of time, enormous advances have been made in the ability to “read" the genetic material and to understand what it means. The techniques described in this chapter are examples of state-of-the-art molecular genetics. Some of the components of these techniques were originally developed for other purposes in genetic research, while others were designed specifically for detect- only 3,000 pieces averaging 2 million nucleotides each (about 1/60 of an average chromosome) (126). Ten different fragments are each labeled with a radioactive probe, so that each lane in the gel will have 10 visible bands, the most that can be analyzed per lane. The fragments are then sepa- rated on an agarose gel in which the electrical field is applied alternately (”pulsed”) in perpendicular directions for 24 to 72 hours. The pulse time is adjusted, according to the particular size of the fragments, to maximize separation of the frag- ments. The optimal pulse time is one in which the fragments are constantly untangling and reorient- ing themselves. Autoradiography is the final step in the procedure. Each of the 10 fragments bound to a labeled probe appears as a visible band on the gel on exposure to X-rays; chromosomal mu- tations would appear as a shift in the position of the fragment containing the mutation. Recent evidence suggests that fairly large pieces of human chromosomes can be separated, al- though most of the experience with the technique is limited to certain lower organisms such as yeasts and unicellular parasites. Results of experiments with DNA derived from these organisms, which have much smaller genomes than humans, sug- gest that chromosome mutations would be appar- ent if they were at least 5 percent as large as the fragment itself. The procedure would have to be modified for human chromosomes due to their larger size and complexity (127). This technique may be useful in detecting chromosome mutations that are intermediate in size between major rear- rangements (observable by cytogenetic methods) and single base pair changes. ing new heritable mutations. They all are successes in the sense that they propose reasonable and ver- ifiable ways of examining human DNA for alter- ations in the nucleotide sequence. However, none of the DNA-based techniques is approaching the efficiency that would be needed for any of them to be appropriate for field use. As these technol— ogies develop and information is shared among Ch. 4—New Technologies for Detecting Heritable Mutations 0 69 investigators, it may be possible to design com- binations of methods that complement each other and that select for those samples in which there is a greater likelihood of finding new mutations. In general, despite impressive achievements that have been made, the technology remains too un- wieldy for large-scale applications. Olson (98) likens the current situation in mutation research to the status of computer technology in the late 19505. The spectacular improvement in computers came not from scale-up of 19503 technology but from development of the integrated circuit. He suggests that big improvements in technology will provide the appropriate methods; doing more of what we have already done is unlikely to be sufficient. Analyzing complete sets of human DNA from large numbers of people using these methods would, at present, require an enormous commit- ment of resources. In considering the wisdom of devoting major resources to such research, Olson (98) emphasizes that the program's ultimate goal would have to be far more ambitious than the detection of a few bona-fide examples of new mu- tations. A program large enough to have a legiti- mate impact on public policy about environ- mental exposures would have to be much larger than a 5 or 10 person laboratory and be several orders of magnitude more efficient than existing restriction fragment length polymorphism tech— nology. One of the outstanding characteristics of cur- rent research in molecular genetics is the rapid pace of new developments and ideas; assuming this continues or accelerates, we can count on bet- ter methods to replace the proposed ones de— scribed in this report. The better ones should be able to examine at least the major portions of the genome and be practical in studying populations at higher risk for heritable mutations. They should also be able to detect a wide, if not the entire, spectrum of mutational events. Until the next generation of techniques is de- veloped, the best course for the next few years may be to maintain a balance between the sup- port of current methods and the support of basic research which is the source of better ones. Main- taining a balance implies that resource—intensive efforts should not pull unwarranted amounts of resources away from developing new techniques, although no special scientific effort seems to be needed to encourage this highly innovative and rapidly paced area of research. Chapter5 New Methods for Measuring Somatic Mutations Contents Page Introduction ............................................................ 73 Detection of Somatic Mutations in White Blood Cells ....................... 74 Selection for HPRT Mutants in T-Lymphocytes ........................... 74 Mutational Spectra From Human T-Lymphocyte HPRT Mutants ............ 75 The hprt Gene ........................................................ 75 Results of Studies of hprt' T—lymphocytes ............................... 75 Detection of Somatic Mutations in Red Blood Cells ......................... 76 Glycophorin Somatic Cell Mutation Assay ............................... 76 Hemoglobin Somatic Cell Assay ........................................ 77 Conclusions ............................................................ 78 TABLE Table No. Page 7. Frequencies of hprt‘ T-Lymphocytes in Human Beings .................... 76 Chapter 5 New Methods for Measuring Somatic Mutations INTRODUCTION All the techniques for learning about heritable mutations described in chapter 4 rely on compar- ing DNA or proteins of parents with DNA or pro- teins of their children to infer the kinds and rates of mutations that occurred in parental germ cells. While that information is of great value in learn- ing about heritable mutations, it comes well af— ter the mutations have actually occurred. Herita- ble mutation tests are not useful for determining rapidly that an individual or a population has been exposed to a mutagen, so that people who have been exposed can be appropriately moni- tored and others, who have not been exposed, can be warned and spared exposure to that same mutagen. For this reason, it is desirable to develop tests for somatic cell mutations that might allow inferences to be made about possible heritable ef- fects of some mutagenic exposure. Instead of pro- viding an exhaustive catalog of somatic cell mu— tation assays, this chapter provides examples of developing technologies that may soon be avail- able to measure somatic mutations in human pop- ulations. It is not known whether the rates and kinds of somatic mutations in blood or any other non- germinal tissue have any predictable relationship with rates and kinds of germinal mutations; it is only with further development and testing of those methods, together with methods for detect- ing germinal mutations, that such relationships might be elucidated. However, even without knowing the precise relationships between somatic and germinal mutations, somatic tests may pro- vide an early warning that people are being ex- posed to environmental mutagens. Furthermore, identifying and characterizing somatic mutants will provide important guidance for reducing ex- posures to mutagens and carcinogens, and is rele- vant to the study of aging and of carcinogenesis. Exposure to mutagens may increase the risk of developing some kinds of cancer. In addition, in- creases in the frequencies of second cancers occur- ring years after treatment of a first cancer with radiation therapy have been noted (see e.g., ref. 11). In a study of survivors of cervical cancer, the risk of developing a second cancer was found to be highest for organs close to where radiation therapy was delivered, strongly suggesting that the development of second cancers may be related to the mutagenic effects of radiation (11). Measurement of somatic mutants offers the pos- sibility of identifying differences in somatic mu- tation rates over time within individuals, among different individuals, and among different popu- lations. Such studies can be used to examine pos- sible associations between exposures to particu- lar environmental or occupational agents and the frequency of mutations. Ultimately, they may be used to identify individuals or populations who are exposed to mutagens. This information could be useful to the regulatory agencies in identify— ing genotoxic compounds and their potencies. To be useful for studying human populations, somatic tests must be based on reasonably acces- sible cells (e.g., blood cells) and on recognizable markers for mutation (e.g., variant proteins or enzymes). The most easily sampled human tissue is blood, and thus far, all the somatic tests that have been developed for studying mutations have used red and white blood cells. Somatic cell mutation assays focus on two types of cellular changes that result from mutations: 1) alterations in phenotypic properties of cells, such as changes in cells' resistance to drugs added to their growth media;1 and 2) alteration in a gene product, such ‘Studying drug resistance in humans is inherently limited to study— ing single copy genes, such as genes on the X chromosome (includ- ing the hprt gene) or dominant selectable markers. 73 74 - Technologies for Detecting Heritab/e Mutations in Human Beings as a protein. The first type of change has been most extensively studied using the mutation as- say based on 6-thioguanine (6-TG) resistance, and the second type has been studied using two types of proteins found in circulating red blood cells. Instead of examining DNA directly, current and developing somatic tests detect rare mutant cells based on marker phenotypes among large num- bers of mostly identical normal cells of a specific type. Although somatic cell mutation assays de- tect phenotypic events that represent, as closely as possible, actual mutations in DNA, the data are expressed as numbers of mutants, rather than in numbers of mutations. Various analytic tech- niques have been used to estimate the mutation frequency from the observed mutant frequency (97). DETECTION OF SOMATIC MUTATIONS IN WHITE BLOOD CELLS Albertini and his colleagues (3) and Morley and his colleagues (79) have developed assays using human peripheral blood lymphocytes to detect mutations that result in a phenotypic change in these cells. The assays are based on the ability of mutant T-lymphocytes, which are present in the peripheral blood at the time the sample is drawn, to replicate DNA and to grow under conditions that do not allow growth of normal cells. The mu- tant cells are identified by their resistance to 6—TG, a drug that is normally toxic to human T-lymphocytes. Normal cells produce an enzyme, hypoxanthine- guanine phosphoribosyl transferase (HPRT), that metabolizes 6-TC to a toxic chemical, killing the cells. Mutations in the hprt gene (the gene that directs the production of HPRT), can result in cells that do not produce HPRT. The mutation from the normal gene (hprtt) to the mutant gene (hprt‘) makes the cell resistant to 6-TG (6-TGr) because they cannot metabolize 6-TG into its toxic form. This particular selective system has been shown to meet all the necessary criteria for a valid genetic assay: 1) the 6-TGJr mutants are genetic variants that ”breed true”: whether grown with or with- out 6-TG, they remain resistant to 6-TG when challenged with the drug (3,79); 2) exposure of laboratory cultures of T-lymphocytes to known mutagenic agents such as X-rays or ultraviolet ra- diation increases the frequency of 6-TGr mutants, indicating an increase in mutation in these cells (17,160); 3) there is a demonstrable change in the gene product—a deficiency in the enzyme HPRT; and 4) there is a demonstrable change in the DNA sequence of the hprt gene—a direct demonstra- tion that the hprt gene has mutated (4,143). Selection for HPRT Mutants in T-Lymphocytes Nicklas (97) described two methods for iden- tifying mutant T-lymphocytes in human blood samples. The first is the autoradiographic assay. After blood is drawn, the T—lymphocytes are pu- rified and frozen. On thawing, the cells are split into two cultures, both of which are treated with phytohemagglutinin, a chemical that stimulates DNA synthesis. One of the cultures is also treated with 6-TG. The other culture is an untreated control. A label, radioactive thymidine, is added to both cultures, and all cells capable of DNA replication and growth incorporate the labeled thymidine into their DNA. In the the control culture not exposed to 6-TG, essentially all the cells divide and incor- porate thymidine. In the culture exposed to 6-TG, normal cells are killed and therefore do not in- corporate the radioactive thymidine into their DNA. The cells from both cultures are then fixed on microscope slides, covered with an X-ray film or emulsion, and then developed using standard radiographic techniques. Cells that have incorpo— rated radioactive thymidine emit radiation, caus- ing a black, exposed spot on the developed film. The developed films are then viewed under a microscope. This technique is quite sensitive; radioactive cells can be counted at a rate of 2 per 10,000 to 1 per 10 million cells. The procedure can also be automated to make it relatively effi- Ch. 5—New Methods for Measuring Somatic Mutations - 75 cient for identifying rare mutants. Dividing the fraction of radioactive cells in the 6-TG-exposed culture by the fraction in the culture not exposed to the drug produces an estimate of the frequency of 6—TG-resistant mutants. The second method, called the clonal assay, selects for cells that are resistant to 6-TG. White blood cells are dispensed in small test tubes or wells in a small plastic device resembling a tiny muffin tin, each test tube or well containing 6— TG in the growth medium. Only the mutant 6- TG-resistant cells grow and divide, forming visi- ble colonies within 10 to 14 days. These colonies can then be analyzed further by isolating their DNA and analyzing it for mutations. The clonal assay allows characterization of mutations found in these cells, whereas the autoradiographic method permits only determination of their fre- quency. Mutational Spectra From Human T-Lymphocyte HPRT Mutants Cariello (16) described a technique for analysis of the hprt gene that may correlate mutational patterns with particular mutagenic exposures. In this method, all of the mutant hprt‘ T—lympho- cytes in a small blood sample are isolated and grown in culture for 2 weeks to produce enough DNA for analysis. DNA of these clones of differ- ent hprt‘ mutants is then examined by first cut- ting it into fragments with restriction enzymes, adding a group of different DNA probes and sep- arating DNA bound to the probes by gradient gel electrophoresis. DNA with different mutations will bind to different probes. The pattern of mu- tations in the original blood sample appears as the pattern of DNA probe positions on a gel. Different mutagens are known to produce clearly different patterns of mutation in human as well as bacterial cells, so the aim of this technique is to identify the probable causes of mutations in human blood cell samples by using these patterns, or “mutational spectra” (139). This approach in- cludes studies of the pattern of spontaneous mu- tation in human T-lymphocyte samples as well as studies of the mutational spectra produced by potentially mutagenic chemicals to which human beings may be exposed. The goal of this approach is to ”fingerprint” for the presence of a particular mutagen in an individual's life history. The hprt Gene One of the advantages of studying the hprt gene is that much has been learned about it from studies of a particular genetic disease. Heritable mutations in the hprt gene located on the X chromosome lead to Lesch-Nyhan syndrome, a severely debilitat- ing disorder characterized by mental retardation and self-mutilation. The hprt gene has been iso— lated from blood samples taken from patients with Lesch-Nyhan syndrome, and the mutations in the hprt gene have been studied. Comparisons have been made between the types of mutations present in cells of Lesch-Nyhan patients and those present in rare, mutant hprt‘ T—lymphocytes selected from blood samples of people without the disease. DNA from 5 of 28 Lesch-Nyhan patients were found to have major hprt gene alterations— deletions of DNA, amplification of DNA, and de- tectable changes in DNA sequence (172). Studies of mutant T—lymphocytes selected from normal blood samples have shown similar gene alterations in several cases (4,143). These results show that the hprt— T-lymphocyte assay detects spontane— ous mutations in the hprt gene in somatic cells that are qualitatively similar to clinically appar- ent heritable mutations in humans, and also dem- onstrates that the clonal assay in T—lymphocytes detects true mutations, not phenotypic changes that mimic mutations. Results of Studies of hprt' T-lymphocytes A number of studies have measured frequen- cies of T—lymphocytes resistant to 6-TG in nor- mal individuals. Average mutant frequencies for the groups of people tested fall within a fifteen- fold range, between about 1 and 15 mutations per million cells (table 7). However, mutant frequen- cies for different individuals vary widely. For in- stance, Albertini reports a range of frequencies from 0.4 to 42 mutants per million T-lymphocytes from 23 individuals, representing a hundredfold range in somatic mutant cell frequencies among a small number of people (2). This wide range in 76 0 Technologies for Detecting Heritable Mutations in Human Beings Table 7.—Frequencies of hprt‘ T-Lymphocytes in Human Beings Average mutant frequency Range Number Reference 1.1><1o*6 NR3 12 (70) 2.2><10—6 NR NR3 (1) 2.9x1o—6 NR 11 (117) 3.o><1owe NR 14 (79) 3.a><1o—6 — 1 (2) 4.6><1o—‘3 3.9—6.0)(10‘6 2 (137) 6.1 ><10—6 NR 45 (26) 1.2><10*5 — 1 (3) 1.4x1o-5 o.4—42><10—6 23 (2) 1.5><10—5 0.8—2.5)(10‘5 24 (160) 3Number of individuals not reported. SOURCE: Office of Technology Assessment. mutant frequencies may be due to different ages and other characteristics of the T—lymphocyte donors. Two investigators have shown that there is a linear increase of hprt’ mutants in T—lym- phocytes with increasing age (141,160). Donor characteristics, such as smoking behavior, have been shown to increase somatic mutant frequency; genetic variation in sensitivity to mutagens may also exist among individuals. The degree to which the observed frequency of mutants reflects the actual frequency of somatic mutations in this assay is not fully understood. Descendants of the original mutant cells increase the observed frequency, so growth of mutants in vivo could falsely lead to a conclusion of higher mutation rates than is actually the case, unless the data are corrected for this problem. In addition, hprt‘ cells may be shorter-lived than hprt+ cells. Before the technique can be used to estimate pos- sible increases in somatic mutation rates from environmental exposures, the wide range of back- ground rates of these mutants has to be investi- gated, including discrimination between true ge— netic events and any possible phenocopies. While imprecision of the assay may contribute to this wide range of mutant frequencies, such a range is not unexpected for genetically and geographi- cally diverse human populations. DETECTION OF SOMATIC MUTATIONS IN RED BLOOD CELLS Unlike white blood cells, mature red blood cells have no nuclei and contain no DNA. They can- not be grown in the laboratory since they do not have the capacity to replicate. Red blood cells can, however, be used to detect mutations expressed in altered proteins on their surface. Red blood cells are also easily obtained from a blood sample. There are about 1 billion red blood cells per mil- liliter of whole blood. Branscomb (13) described two methods to detect variant red cells, using high-speed automated microscopy and flow cy- tometry to analyze and sort hundreds of red blood cells per second to detect rare red cell mutants. These two methods are described below. Glycophorin Somatic Cell Mutation Assay The glycophorin assay detects gene-loss muta— tions expressed by the absence of cellular proteins corresponding to those genes. Such mutations can be detected where two or more variants of a par- ticular protein normally exist, where the genes for the alternative forms are said to be codominant (both variants are expressed when a gene for each is present), and where different variants are ex- pressed in the same cell. A mutation causing the loss of function of the gene for one of the variant forms results in a cell that expresses only the other form of the protein (9,10,50). One such protein is glycophorin A, a glycopro— tein present on the surface of red blood cells. Two variants of glycophorin A normally exist, called ”M" and ”,"N which differ from each other by 2 out of a total of 131 amino acids. The M and N serotypes have no known biological function, and each allele functions independently of the other. The M and N variants of glycophorin A can be labeled independently using monoclonal antibodies carrying different fluorescing mole- cules. As a result of the labeling, the M variant appears green and the N appears red when viewed under a fluorescent microscope. A flow cytome— ter, a machine through which cells are passed at Ch. 5—New Methods for Measuring Somatic Mutations - 77 very high speed, can sort and count the number of cells that are green, containing only the M var- iant, and the number that are red, containing only the N variant. Individuals whose red cells display both vari— ants of glycophorin are heterozygous for the M and N alleles (i.e., they inherited an M gene from one parent and an N gene from the other). As used currently, this assay detects gene loss mutations at glycophorin genes in individuals who are het- erozygous for the M and N variants. Red blood cells from heterozygotes show both red and green fluorescence; if a mutation inactivates one of the glycophorin A alleles, only the other variant would be present on the cell surface, hence only one color would appear. Single-color cells are rec- ognized and distinguished from the majority of double-colored cells, and are counted as somatic mutant cells. The glycophorin A assay is currently being used at the Lawrence Livermore Laboratory to study mutations in two populations: cancer patients’ blood cells examined before and after treatment with known doses of chemotherapeutic drugs, and Japanese atomic bomb survivors whose dose of radiation can be estimated (69). Hemoglobin Somatic Cell Assay The frequency of particular hemoglobin vari- ants that exist at low frequencies in red blood cells of normal individuals can be measured. Cells con- taining these variants presumably arise in bone marrow stem cells, the precursors of red blood cells. Methods have been developed using mono- clonal antibodies, prepared for each type of var- iant, to label particular mutant hemoglobin mole- cules. This method detects small mutations, such as single nucleotide changes resulting in amino acid substitutions, frame shift mutations, etc., in the gene for beta-globin, a constituent of hemo- globin. Investigators at the University of Washington (132,133) and at Lawrence Livermore National Laboratory (50) have developed an approach to detect hemoglobin mutations in red blood cells. In this technique, an antibody to the mutant he- moglobin is bound to a chemical called a fluoro- chrome, which is then mixed with red cells in sus- pension. The fluorochrome emits fluorescent light of a specific color when light of a particular wave- length is shone on it. The cells are screened under a microscope that allows the investigator to di- rect light of a particular wavelength on the sam- ple. If any of the cells contain the mutant hemo- globin bound to fluorescing antibody, they can be viewed and counted. Stamatoyannopoulos and his colleagues have used this technique on blood samples from 15 in- dividuals using antibodies specific for Hemoglo- bin S (the mutant form of hemoglobin present in sickle cell anemia) and Hemoglobin C. An aver— age of 50 million cells per individual was screened, and an average mutant frequency of 1.1 in 10 mil- lion cells per subject was obtained, with a range of 4 in 100 million to 3 in 10 million. These in- vestigators also screened the red blood cells of 10 individuals who had been exposed to known mutagens—X-rays or mutagenic drugs used to treat cancer. The frequencies of mutant hemo- globins obtained from these samples were within the range found for normal individuals. This technique is labor—intensive because it re- lies on visual screening of the cells: it takes one technician nearly 1 month to screen 100 million red cells. The investigators have since collaborated with Dr. Mendelsohn and his associates at the Lawrence Livermore Laboratory to use their auto- mated fluorescence-activated cell sorter to screen the cells. Although some technical difficulties were encountered with this instrument, results obtained on red cells from six normal individuals (on screening 500 million cells per individual) yielded an average mutant frequency of 1 in 10 million, about the same as the average frequency deter- mined from visually counted cells. With an array of different mutant hemoglobins to use for production of antibodies, a variety of mutations can be detected in hemoglobins found in rare cells present in normal individuals: single nucleotide substitutions, deletions, frame-shift mutations, or virtually any mutation that causes the production of mutant hemoglobins. There are currently two technical problems with this technique. First, a significant number of false positives are generated. Better methods of fixa- 78 - Technologies for Detecting Heritable Mutations in Human Beings tion and antibody staining of the hemoglobins are necessary, along with verification of each vari- ant found. Second, the method is not yet fully automated, but research in several laboratories is progressing toward full automation. Significant progress has already been made; Mendelson (68) CONCLUSIONS The methods to detect and quantify somatic mutations suggest some additional approaches for studying the nature and mechanisms of mutation in human beings. Somatic tests offer the possi- bility of drawing associations between exposures to specific mutagens and particular genetic events in the cell, and they may help to identify individ- uals and populations at high risk for mutations. They may lead to understanding the relationships between the frequency and nature of somatic and germinal mutations, and if those relationships can be identified, somatic tests may contribute signif- icantly to understanding why human heritable mutations occur and may help predict their fu- ture occurrence. At present, however, it is not possible to directly extrapolate from mutations in somatic cells to risks of heritable mutations in offspring. Animal studies can be used to measure quan— titatively the correlations between exposure to mutagens and rates of somatic mutation, as well as to investigate possible associations between so- matic and heritable mutations. Coordinated tests of somatic and germinal mutations in experi- mental animals may help guide such research in humans. Such data, combined with data on hu- man somatic mutation rates, may suggest specific risks of heritable mutations in human populations. In this regard, the tests for somatic mutation described in this report appear promising. None of these methods is currently ready for applica- tion to human population studies, however, even though many of the technical details have been reported that with present capabilities using flow sorting, 88 antibody-labeled red cells can be iden- tified in a mixture that contains a known quan- tity of 100 such labeled red cells and 1 billion un- labeled red cells, a relatively good retrieval rate. worked out. Since the assays do not detect DNA mutations directly, it is particularly important to verify the results from these methods by isolat- ing variant cells and confirming that the altered proteins do indeed reflect primary changes in the genes coding for these proteins. It is also neces- sary, although more difficult, to determine how often these methods miss the mutations they are designed to detect. In general, the most useful somatic cell muta- tion assay may be one that detects a spectrum of mutations (single amino acid substitutions to large deletions and translocations) and that can detect mutations in different cell types. Different tissues may demonstrate different frequencies of somatic mutants, and equally as important, some types of mutant cells may have longer lifespans in the body than others. Consideration of such charac- teristics could be useful in choosing the most appropriate assay for particular circumstances. Short-lived mutants could be useful in studying the effects of short-term exposures to potential mutagens, while longer lived mutants could be better indicators of mutagenic effects of long-term, chronic exposure to environmental mutagens. The genetic analysis of white blood cell DNA offers the possibility of identifying genetic ”fin- gerprints,” or particular patterns of mutations in- duced by specific mutagenic agents. Such patterns are an essential first step toward understanding the molecular mechanisms of genetic change in human DNA. Chapter6 Laboratory Determination of Heritable Mutation Rates Contents Page Introduction ............................................................ 81 Rodent Tests for Heritable Mutations ...................................... 81 The Dominant Lethal Test ............................................. 81 The Heritable Translocation Test ....................................... 82 Specific Locus Tests ................................................... 82 Other Studies in Animals .............................................. 83 Factors That Influence Induced Mutation Rates in Mice ...................... 84 Effects in Males .............................................. : ........ 84 Effects in Females ..................................................... 85 Estimating Effects in Human Beings From Animal Data ...................... 86 TABLE Table No. Page 8. Spontaneous Mutation Rates in Mice .................................... 83 FIGURE Figure No. Page 13. Spermatogonial Mutation Rates in Mice After Low—Dose Rate and High-Dose Rate Radiation ............................................ 85 Chapter 6 Laboratory Determination of Heritable Mutation Rates INTRODUCTION The genetic information of all organisms is en- coded in DNA (with the exception of some viruses that depend on RNA), and mutations have been found in all species that have been examined. The biological similarity of DNA structure and func- tion in all species has allowed scientists to study mutation rates and mechanisms of mutagenesis in several species for over half a century. In addi- tion to these efforts in basic research, in recent years the biotechnology industry has provided impetus, including financial incentives, for sophis- ticated inquiry into mutations and mutagenesis in bacteria and yeast. The industry’s interest is in using mutagenesis to benefit from some of the particularly useful characteristics of certain micro- organisms, such as the ability of some bacteria to digest oil, and the fermentation capabilities of various yeasts. Another reason for studying mutations in lower organisms is to identify agents that cause muta- tions in their DNA, agents that might act simi- larly on human DNA. For instance, concern about human risks from the genetic effects of ionizing radiation led to the funding of extensive mouse studies, which have provided quantitative data on factors that influence the yield of various types of heritable mutations. In the case of chemicals, however, concern about cancer, not about herita— ble mutations, has been the primary motivation for developing short-term laboratory test systems to detect mutations. In the last decade, as more and more evidence incriminates somatic mutation as an early event in the development of cancer, the properties of mutagenicity and carcinogenic- ity have been conjoined, making identification of mutagens an important part of efforts to identify cancer-causing agents. Many of the tests for muta— gens measure a chemical’s capacity to alter the DNA of bacteria or yeast (145). While such a short-term test in a lower organism may distin— guish between agents that can and cannot change DNA, it provides no information for making quantitative estimates of mutagenic potency in hu- mans nor does it show that the chemical can nec- essarily reach animal or human gonads in an ac- tive form to cause heritable mutations (108). Whole animal tests specifically designed to detect germ-cell or heritable mutations provide some in- formation of that type. RODENT TESTS FOR HERITABLE MUTATIONS At least 15 tests and variations on them are available for the detection and measurement of heritable mutations induced in mammalian germ cells. Lower organisms, notably Drosophila, are the subjects of similar tests, but most of the in- formation about mutation rates comes from mice. The three most widely used rodent tests are de— scribed below (58,108). The Dominant Lethal Test The dominant lethal test detects mutations that cause the death of rodent embryos. The usual pro- cedure in this test is to treat male mice with a suspected mutagen and then mate them with un- treated females. If the treatment causes a muta- tion in a male germ cell that is lethal at some stage 81 82 0 Technologies for Detecting Heritable Mutations in Human Beings in the development of the embryo (after fertili- zation), the lethal events are detectable by dissect- ing female uteri a few days after implantation of embryos (implantation occurs about nine days af- ter fertilization). If there are more dead embryos than expected (compared with matings of unex- posed mice) and/ or more implantation sites than expected in the uterus where there is no longer an embryo, the effect is assumed to be the result of a mutation in the male germ cell. Chromosomal studies in mice have shown that almost all “dom- inant lethal" events are associated with major chromosome abnormalities. Human analogs of the postimplantation fetal losses counted in dominant lethal tests are spon- taneous abortions (miscarriages). Fifty to 60 per- cent of spontaneous abortions are associated with numerical and structural chromosome abnormal- ities, the same basic types of abnormalities de- tected by dominant lethal tests (12). These simi— larities between mice and men, and the general pragmatic acceptance of higher animals as predic- tors for human risk, underlie the idea that agents that are positive in a mouse dominant lethal test may also cause spontaneous abortions or chromo- some abnormalities in liveborn infants. The Heritable Translocation Test The heritable translocation test detects breakage and rejoining of fragments from different chro- mosomes. When a translocation involves no loss of genetic information (a “balanced transloca- tion”), often it has no adverse effect on the func- tioning of the carrier individual. The unusual chromosomes produced by the breaking and re— joining are present in all cells, and they can pair with each other and divide and function like nor- mal chromosomes. However, such translocations cause reproductive problems because the germ cells may have incomplete sets of genetic mate- rial and may lead to unbalanced chromosome ab- normalities in offspring, often resulting in death during embryonic or fetal development. Herita- ble translocations are thus detectable in mice be- cause they result in reduced litter size in the sec- ond generation after exposure to the suspect agent, or in sterility in the first generation. In addition, cytogenetic studies can detect abnormal chromo- somes and can be used to study the association between reduced fertility and chromosome abnor- malities. Specific Locus Tests Since 1927, when Herman Miiller first demon- strated that radiation causes mutations in fruit flies (81), there has been concern about possible genetic effects of radiation in human beings. Following World War II, in anticipation of widespread use of nuclear power, the Atomic Energy Commission (AEC) contracted for research on the mutagenic effects of radiation in mice to provide an experi- mental underpinning for estimating effects in human beings. Specific locus tests that detect mu- tations in a few specific genes in mice were de— veloped for that purpose. This work and use of specific locus tests to measure chemically-induced mutations have provided the animal data most useful for estimating human risk. Through research supported by the AEC, Wil- liam L. Russell (109) bred a strain of ”tester” mice that differs from wild—type mice in 7 morpho- logical features—coat colors, patterns of pigmen— tation, eye color, and ear shape—that are reces- sive traits. Each of the features is thought to be controlled by a different genetic locus. Tester mice are homozygous for the recessive alleles of these genes, while wild—type mice are homozygous for the dominant alleles. When the tester strain is mated to a strain of wild-type mice, the offspring should all be heterozygous at the seven test loci, and therefore they should appear wild-type be- cause the the wild-type alleles are associated with the dominant phenotype. However, if a germinal mutation occurs in a gene for one of the seven loci in the wild-type strain, causing the dominant allele not to be expressed, the offspring from that mutation will have a phenotype characteristic of the tester strain. In a specific locus test, the wild-type mouse is exposed to radiation or to a known or suspected chemical mutagen and then mated to the tester strain. In most cases, only male wild-type mice are exposed, and the tester strain mice are female. The reason for this is largely a practical one: one exposed male can sire a large number of progeny Ch. 6-Laboratory Determination of Heritable Mutation Ffates - 83 when mated to many females, but an exposed fe— male can have only one litter at a time. In addi- tion, the information so far available suggests that male germ cells are more sensitive to mutagens than are female germ cells. After mating exposed wild-type and tester mice, the offspring are ex- amined for the easily seen morphological features related to the seven loci. Since the morphologic variants are readily identified, large numbers of offspring can be examined quickly by an investi- gator. The frequency of mutant offspring of ex- posed parents is compared with offspring of un- exposed parents. The importance of the ease of recognizing mutants is underscored by the low fre- quency at which mutations occur spontaneously —as few as 50 mutations per 1 million offspring of unexposed mice. (Because mutations can be de— tected at any of the seven loci, examination of one mouse provides information about one observa- tion at each of the seven loci.) Estimates for the rate of spontaneous mutations in mice based on specific locus test results are sum- marized in table 8. These data are based on exam- ining almost 1 million progeny of exposed mice, or about 7 million loci. Though the reliability of the estimate is limited by the small number of mu- tations seen, the calculated rate of between two and eight mutations per 1 million genes is not in— consistent with the human rate estimated by one- dimensional electrophoresis of blood proteins (see ch. 3). Although the ”average" rate of mutations is a useful touchstone, it is used in the full knowledge that the average includes very different rates at the seven loci. For instance, there is a 35-fold difference in radiation—induced mutation rates be- tween the most and least sensitive of these par- ticular seven loci. The choice of a different set of loci for examination might yield quite different results. Other Studies in Animals There is a family of animal experiments called dominant mutation tests, which use the appear— ance of dominant features, such as skeletal anomalies or cataracts, to detect mutations (108). Many of the endpoints used resemble known hu— man genetic disorders, and the results of these tests have been used to estimate human genetic risk. In addition to the types of studies described above, some of the techniques used to study hu- man beings have been applied to experimental ani- mals. Electrophoresis of blood proteins and quan— titative enzyme assays in human beings, described in chapter 3, have been used to study correspond- ing proteins in mice. The electrophoretic approach has been used in studies of mutagenized mice at the National Insitute of Environmental Health Sci- ences (51,52,53,63) and also at the Institute for Genetics, Neuherberg, Germany (19,19a). A small effort in electrophoretic analysis has also been conducted at the Oak Ridge National Laboratory (107). Similar studies have been reported in Dro- sophila (80,103,140,161). Mutagenized mice have been tested for enzyme deficiency variants using quantitative enzyme assay techniques similar to those used to study human populations (19a,73). Enzyme assays have also been used in mouse and Drosophila experi- ments that involve selective mating, similar to the specific locus test strategy. By selecting parental strains that have electrophoretically distinct al— leles at a locus, enzyme deficiency variants can be detected in offspring by the absence of one activity staining band that, in the absence of a Table 8.—Spontaneous Mutation Rates in Micea Number of Number of Sex mutants progenya Mutations/locusb Reference Female ............................................... 3 or BC 204,639 2.1 or 5.6 X 10" (110) Male ................................................. 39 727,319 7.6 X 10—“ (113) aData about spontaneous mutation rates were collected from animals used as controls in experiments to determine the effects of radiation or chemical agents on males and females separately. There are more progeny from males because more experiments have been done in males. bA rate of 1 x 10‘6 means that, on average, a mutation occurs once in a million loci. 0The choice of three or eight mutants exemplifies a problem in counting mutations, Three litters from female mice had mutant offspring. Two litters had one mutant each; the third had six mutants, all of which may have resulted from a single mutation. The result is three probable mutations and eight mutant offspring. SOURCE: Office of Technology Assessment. L—_ 84 ' Technologies for Detecting Heritable Mutations in Human Beings mutation, would be inherited from one of the par- ents (51,52,103,161). Data from animal experiments using electro— phoretic and quantitative enzyme assays are lim- ited, but at this stage a general, cautious conclu- sion can be drawn. The background mutation rate in the mouse derived from these experimental data is in the same range as the estimate of the spon- taneous rate derived from studies in human be- ings (77), recognizing, however, that these esti— mates are based on small numbers and may be heavily influenced by chance fluctuations. FACTORS THAT INFLUENCE INDUCED MUTATION RATES IN MICE When radiation or a chemical is tested for mu- tagenicity, the experiment does not necessarily produce definitive answers to the questions: ”Is this agent mutagenic?" And ”if so, how mutagenic is it?" The answers differ depending on several factors, two important ones being whether the animal is male or female and the stage of devel- opment of the exposed germ cell. There is ample evidence from radiation studies that male and fe- male germ cells differ in their sensitivity to muta- gens. In males, germ cells appear to be more likely to sustain a mutation the further along the devel- opmental path they are. Overall, female germ cells appear to be more resistant to mutagens than are male germ cells. Effects in Males The male testis, from puberty on, contains a mixture of germ cells at different developmental stages, including spermatogonial stem cells, dif- ferentiating spermatogonia, and intermediate stage cells, the spermatocytes and spermatids, which are maturing into spermatozoa. The mech- anisms of sperm production are qualitatively very similar in mice and men, differing largely in tim- ing: in mice, sperm production from spermatogo- nia to spermatozoa takes 35 days; in men, it takes 74 days (60). When a male is irradiated, germ cells at all stages of development are exposed. By vary- ing the times of mating, the effects of radiation on different stages can be examined. For instance, if male mice are irradiated and mated immedi- ately, any mutants that appear in the progeny must result from mutations that were caused in fully mature sperm. In a mating two weeks after exposure, eggs are fertilized by spermatozoa that were spermatids at the time of irradiation. If seven or more weeks are allowed to pass, sperm in the ejaculate would have been irradiated at the sper- matogonial stem-cell stage. Experimental evidences shows that cells at each stage have varying sensi- tivities to induction of mutations by radiation, the later, postspermatogonial stages being most sen- sitive. While spermatogonial stem cells are more resis- tant to the effects of mutagens than are germ cells in later stages of development, mutations in stem cells are of greater concern. Stem cells persist throughout the lives of males, constantly giving rise to new generations of sperm cells. A muta- tion in stem cell DNA results in every cell that buds off from the stem cell bearing a mutation. The lifespan of later germ-cell stages constitutes only a small fraction of the total reproductive lifespan (especially in humans). Thus, in the case of acute exposure to these later germ-cell stages, only conceptions occurring during a short postex- posure interval are at risk. For these reasons, the effects of radiation and some chemicals have been especially closely studied in spermatogonial stem cells in animals. The most common way for people to be ex- posed to ionizing radiation is through acute ex- posure to medical X-rays, or long—term, low-level exposure to natural background radiation. Cer— tain groups are exposed occupationally (e.g., ura— nium miners, nuclear powerplant workers, some medical personnel); and some may have special low—level environmental exposures from living near nuclear installations, nuclear waste sites, ura- nium mine tailings, or areas of high natural ra- don concentrations. In mice, exposure of males to X—rays or other forms of ionizing radiation at low dose rates causes a measurable increase in the number of mutations, while such exposures do not measura- Ch. 6—Laboratory Determination of Heritable Mutation Rates ' 85 bly increase mutation rates in females. Russell and Kelly (113) counted the numbers of mutants caused by varying doses of radiation in male mice. They find no evidence for a dose so low that it is not mutagenic, but the number of mutations decreases with lower and lower doses (see fig. 13). The data in the figure also show the differing re— sults of exposure at high and low dose rates. For example, the number of mutations caused by a total dose of 300 roentgens (R),1 administered at a dose rate of between 72 and 90 R per minute is about 80 X 10—6 per locus; the same total dose administered at dose rates of 0.08 R per minute or lower causes fewer mutations, about 30 X 10‘6 per locus. These data may be interpreted by saying that low dose rates allow time for re- pair of some of the damage caused by the radia- tion. (For further discussion of dose and dose-rate relationships, see ch. 7.) Effects in Females Female mammals differ from males in the bi— ology of germinal stem cells. Whereas male sper- matogonial stem cells divide constantly during adulthood, females are born with their total com- plement of egg precursor cells. Shortly after birth, all cells destined to become eggs (ova) are present in the ovaries as ”arrested oocytes." Exposure of arrested oocytes to radiation has caused no in- crease in mutation rates in mice (110). During the six weeks before ovulation an arrested oocyte ma- tures, undergoing meiosis and, in a mature state, travels from the ovary to the uterus. Irradiation during this maturation period does increase mu- tation rates in female germ cells. Two different explanations can be advanced for the failure to recover mutations from irradiated arrested oocytes. First, it might be that arrested oocyte DNA is somehow protected from the mu- tagenic effects of radiation, perhaps by having very efficient repair mechanisms that correct radiation-induced mutations. On the other hand, in mice, arrested oocytes are very sensitive to the killing effects of radiation (27). These lethal ef— fects could also account for the absence of muta- genic effects in female mice. It is argued that since 1A measure of radiation. In human terms, a chest X-ray exami- nation produces a gonadal radiation dose of less than 0.0005 R (148). Figure 13.—Spermatogonlal Mutation Rates in Mice Atter Low-Dose Rate and High-Dose Rate Radiation 24 20 - (I) 3 0 2 L— _ B a ‘6 / 1‘9 X 2 '9 12 — 16 ‘5 E S 8- g A E 3 Z 4 _ 0 l l l I 100 300 500 700 900 Dose (Roentgenst Straight lines fitted to experimental results from a number of specific-locus tests involving single doses of radiation. Line A: Radiation delivered at low dose rates (0.8 Roentgens/minute and lower). Line 3: Radiation delivered at high dose rates (72 to 90 Roentgenslminute). The figure shows that a dose of radiation delivered at a faster rate causes more mutations per locus than the same total dose of radiation delivered at a slower rate. (90% confidence limits are shown.) SOURCE: W. L. Russell and E. M. Kelly, “Mutation Frequencies in Male Mice and the Estimation of Genetic Hazards of Radiation in Men," Proceedings of the National Academy of Sciences (U,S.) 79:542-544. 1982. radiation kills oocytes very efficiently, oocytes would either die before they were mutated or they would die with mutations that would never be ex- pressed. The extreme sensitivity to the killing effects of radiation in mouse arrested oocytes is not com- mon to all animals. In the squirrel monkey, high sensitivity is limited to oocytes irradiated during fetal life. Arrested oocytes in other monkey spe- cies are no more sensitive to lethal effects of ra- diation than are other cells. Similarly, human oocytes, at least those that are irradiated after 86 - Technologies for Detecting Her/table Mutations in Human Beings birth, appear resistant to the lethal effects of ra- diation. The differing sensitivities of oocytes of differ- ent species to lethal effects of ionizing radiation complicate efforts to extrapolate from mutation rates in female mice to the expected rates in hu- man females (28). In particular, radiation may re- sult in more mutations in human females than would be predicted from female mice because hu- man arrested oocytes are less likely to be killed by radiation. That might allow a human oocyte to survive a mutagenic dose of radiation and to be fertilized. Using experimental data from female mice, Rus— sell (1977) made several calculations relating ef- fects of radiation in mice to potential effects in human females. In each calculation, he made the “worst case” assumption that all stages of female germ cells were as sensitive to the mutagenic ef— fects of radiation as are the most sensitive stages— maturing and mature oocytes—in mice. Because there are different methods of handling the data, he made four different calculations. Three pro- duced estimates indicating that radiation would not increase the mutation rate in human arrested oocytes. The fourth estimated that human ar- rested oocytes would be somewhat less than half as sensitive (from 17 to 44 percent as sensitive) to radiation as are human spermatogonia. ESTIMATING EFFECTS IN HUMAN BEINGS FROM ANIMAL DATA Most extrapolations from animal data to hu- man mutagenic risks are based on studies in male mice because: 1) there are more data for male mice than for female mice, 2) there appear to be fewer biological differences in germ cell development be- tween males of the two species than there are be- tween females (see above), and 3) male germ cells appear to be more sensitive to radiation than fe- male germ cells and may provide a more sensitive indicator of genetic risk. The dose-response curves in figure 13 are de- rived from experiments on male mice and they provide information for estimating human risks. The absence of a detectable threshold for genetic effects in male mice suggests that humans are at some increased risk for mutations at any level of radiation exposure. This finding increases the im- portance of making quantitative estimates of the effects of radiation; since risk does not go to zero except at zero dose, what is it at levels of human exposure? The "doubling dose,” which is the dose of radiation that induces an additional number of mutations equal to the number that occur spon— taneously (resulting from all mutagenic influences, known and unknown), can be derived from the information in figure 13. The doubling dose differs depending on the radiation dose rate. The dou- bling dose in male mice for high dose rates is about 40 R; for low dose rates, about 110 R. This is of practical importance for human beings, since some exposures are at high dose rates—eg. the populations of Nagasaki and Hiroshima and peo- ple receiving radiation therapy for some types of cancer—but most population exposure to radia— tion is delivered at a low dose rate—eg, people who live near radioactive mines. About three times as many mutations are caused by radiation delivered at a high dose rate as opposed to an equal total radiation dose delivered at a low dose rate (114). In chapter 3, the effects on mutation rates of acute irradiation of the populations at Nagasaki and Hiroshima are discussed. The exact doubling dose for those populations is still being debated because of uncertainties about the total dose and because of insensitivity of the methods used to detect mutations. Using the best information available about those populations, however, hu- mans appear to be less sensitive than mice to high dose rate exposures. Current radiation exposures of United States citizens average about 200 millirem (mrem) per year. On a population basis, this exposure is gen- erally divided about equally between high dose rate medical exposures and low dose rate expo- sures to natural sources of radiation. These ex- posures do not approach the doubling dose for Ch. 6—Laboratory Determination of Heritable Mutation Rates 0 87 mutations. If the average rate of human mutations is now on the order of one mutation per 1 mil- lion genes per generation, we can estimate the ef- fect of a hypothetical increase in human radia— tion exposure to twice the current level. If exposures increased to 400 mrem and if that ex- posure were as mutagenic as the high dose rate delivered to mice, it would increase the average human rate to 1.005 mutations per 1 million genes per generation. Such an increase would be un- detectable by any current method. Animal studies have proven useful for learn- ing about the various relationships of dose, par- ticularly for radiation, and germ—cell and herita— ble mutation rates. They are limited, however, in providing information directly applicable to hu- man beings. First and most important, human be- ings differ from animals in ways that we know about and in more ways that are not understood. Second, the endpoints measured in animal tests do not generally correspond directly to endpoints measurable in human beings under nonexperi- mental conditions. Third, the available animal tests, like the currently available methods for studying human beings, detect mutations arising in specific, selected, loci which may not be rep— resentative of other loci. Evidence from Russell's specific locus test, cited above, shows a 35-fold difference between the least sensitive and most sensitive of the loci tested. A fourth considera- tion, again similar to the limitations of studies in humans, is that each test does not detect more than a fraction of the mutations that can occur, and the exact nature of the mutations that are de- tectable cannot be characterized with these methods. When the new technologies discussed in this re— port are further developed, they presumably will be applied to experimental animals as well as to human beings. Studies that examine correspond- ing genes with the same techniques in humans and animals may provide some of the interspecies comparisons that now are lacking. Chapter 7 Extrapolation Contents Page Introduction ............................................................ 91 The Aims of Extrapolation ............................................. 91 Extrapolation Models .................................................... 93 Attempts at Qualitative Extrapolation ..................................... 94 Attempts at Quantitative Extrapolation .................................... 97 Dose-Response Relationships ........................................... 97 Generalizations About Dose-Response ................................... 99 Limits of Extrapolation .................................................. 99 TABLE Table No. Page 9. Weighting of Test Results for Presumption of Germ—Line Mutagenicity ...... 95 FIGURES Figure No. Page 14. Biological Test Systems for Studying Mutations ......................... 92 15. An Example of a Parallelogram ........................................ 93 Chapter 7 Extrapolation INTRODUCTION Currently, predictions of possible risk of herita- ble mutations in human beings are based on infer- ences, or ”extrapolating” results, of mutagenicity tests in other organisms or in laboratory cell cul- tures. One of the key problems in genetic extrap- olation is that, while there is no shortage of muta- genicity tests using a variety of organisms and cell types, researchers have just begun the painstak- ing work of drawing out relationships among re- sults of different tests, and of eventually validat- ing models for extrapolating to heritable effects in human beings (see fig. 14). This chapter reviews the basic constructs that have been devised for framing genetic extrapolations and then presents some efforts that have been made to carry out spe- cific extrapolations. The Aims of Extrapolation For many years, the emphasis in mutagenicity testing has been on developing individual tests and learning about their properties. Test development has not been targeted exclusively toward devel- oping predictive models for heritable mutations in human beings and in fact the quest for tests that could be used to predict carcinogenicity through somatic mutation has predominated. This being the case, the test systems vary tremendously and include, for example, tests in bacteria, insects, ani— mal somatic and germ cells in culture, whole mam- mals, and human somatic cells. Efforts to relate results from one test system to another or from various tests to human beings, even on a qualita- tive level, have been relatively recent and have not progressed very far to date. Bridges (15) sum- marized this situation: Despite the extremely large number of screen- ing tests that have been performed, remarkably little rational thought has been devoted to the use that should be made of the results of such tests. Mutagenicity tests in lower organisms1 are 1“Lower organisms" refers to bacteria, yeast, Drosophila, etc., and not to whole mammals. essentially screening tests warning of a possible human hazard. They do not give any quantita- tive indication of the level of risk to man, not so much because of differences in the organiza- tion of DNA in man and lower organisms, but because of the complex metabolic capabilities of man which may greatly enhance or diminish the mutagenic effectiveness of an agent. In recent years, scientists who use different ex- perimental approaches to study mutations have begun to discuss methods to correlate results from various test systems. The first consideration is whether results from one sytem are quaIitativer predictive of results in another, i.e., are appro- priate endpoints being considered and do the re- sults agree in whether they are positive or nega- - tive. This is called ”biologic extrapolation.” The second type of extrapolation is quantitative, which deals with the relationship between the quantita- tive response in a test to a quantitative estimate of the likely effect in human beings. Examples of questions posed in extrapolation are the following: If experiments demonstrate that a single exposure to a high dose of a chemical in- duces heritable mutations in mice, what would be the result of a single exposure to a lower realis— tic dose or of a long-term exposure to a lower dose of that chemical in humans? How do results of mutagenesis experiments on rodent somatic cells in a test tube cell culture (e.g., Chinese hamster ovary cells “in vitro") relate to the issue of muta- tion rates in human somatic cells, or mutation rates in human germ cells? What are appropriate dose conversions between various experimental systems and human beings? If a worker in a chemical plant has a tenfold increase in mutation frequencies in his or her somatic cells, what are the long-term health implications? Is that worker at increased risk of having a child with a new mutation? These types of questions illustrate the complex issues that arise in using results from one system to make pre- dictions of risks to human health. 91 92 - Technologies for Detecting Heritable Mutations in Human Beings Figure 14.—Biological Test Systems for Studying Mutations Heritable mutations in human beings Germ cells In vltro H ——__————§ Somstlc cells in vivo :m3c Somatic cells in vitro \ Mammalian \ germ-cell systems \ \ \ \ \ \ \ \ _ \ Mammalian \ somatic-cell \ systems Germ cells In vlvo Germ cells In vitro Somatic cells in vlvo \_——_———- Somatic cells In vitro “(DECI— —m33m3 fimEOl‘ mam—‘Jmowo Higher eukaryotes (e.g., higher plants, insects) Lower eukaryotes (e.g., yeast, fungi) Prokeryotes (mainly bacteria) Test systems for studying mutations, arrayed according to biological similarity to the end- point of interest—heritable mutations in human beings. Some test systems can be used to measure a single endpoint; more than one type of endpoint (e.g., both chromosomal and gene mutations) can be measured in other systems. Extrapolation from any system to human germ cells in vivo involves many, mostly untested, assumptions. SOURCE: Office of Technology Assessment. Ch. 7—Extrapolation - 93 In translating results from one system to another (using similar genetic endpoints), a number of sep- arate extrapolations may have to be made: 1) from species to species; 2) from experimental doses to actual environmental doses; 3) from one cell type to another; 4) from in vitro to in vivo physiologi- cal conditions; and 5) the biggest and most un- certain leap, from estimates of mutation frequen- cies to estimates of genetic disease in humans. The EXTRAPOLATION MODELS Several researchers have begun developing mod- els for extrapolating from one test system to another (14,15,46,49,61,64,106,129,130,138). One of the key features common to all the extrapola- tion models developed is that a result from a sin- gle test system would not be used alone to predict a result in another test system. Instead, results from several related test systems are correlated and used together. Several of the methods described are extensions or rearrangements of the first extrapolation meth- od developed by Sobels (1290)—the parallelogram (130): The underlying principle is to obtain informa- tion on genetic damage that is hard to measure directly, for example mutation in mouse germ cells, by comparison with endpoints that can be determined experimentally, e.g., alkylation per nucleotide in mammalian cells in vitro and in mouse germ cells, and mutation induction in mammalian cells in vitro. Sobels' parallelogram is illustrated in figure 15. It is relatively easy to determine the mutation fre- quency in mouse somatic cells (a quantity called “A”) upon exposure to a particular chemical muta- gen in culture. With certain types of chemicals (alkylating agents), it is also possible to derive a measure of the interaction of the mutagen with the DNA of those cells, which is quantified as ”alkylations per nucleotide" (a quantity called ”A”). Alkylations per nucleotide can also be meas— ured in mouse germ cells after exposure to the same mutagen (B'). The relationship of these different values is then used to calculate the expected mu- tation frequency in mouse germ cells (B) on ex- kind of information that would give the biggest boost to the ability to predict effects in humans with information from other test systems is know- ing exactly what kinds of mutations (e.g., point mutations, chromosomal rearrangements, etc.) each of the tests detects. The new technologies dis- cussed in this report may provide this type of in- formation. Figure 15.—An Example of a Parallelogram A ' B , Alkylations per nucleotide <————-> Alkylations per nucleotide A e\ Measured mutations -----> Estimated mutations (Note that the same test system is represented in the vertical dimension and the same genetic endpoint in the horizontal dimension.) SOURCE: F.H, Sobels, “The Parallelogram: An Indirect Approach forthe Assess- ment of Genetic Rlsks from Chemical Mutagens,” pp. 323-327 in KC. Born at al. (eds), Progress in Mutation Research, vol. 3 (Amsterdam: Elsevler, 1982). posure to the same mutagen. A major, untested assumption is that the ratio of A to A ’ is propor- tional to the ratio of B to B ’, i.e., A/A ’ = B/B ’. If that is true, it is then a matter of simple algebra to predict the mutation frequency in mouse germ cells (B) by solving the equation for B, which is the only unknown quantity. A similar parallelogram can be used to extrap- olate results from mouse germ cells to human germ cells. Mutation frequencies are measured in so- matic cells of both humans and mice. Germ-cell mutation frequencies are measured in the mouse and compared to somatic-cell mutation frequen- cies in the mouse. Assuming that the ratio of germ- cell to somatic-cell mutation frequencies is the same in mice and human beings, germ-cell muta- tion frequencies in human beings can be predicted from the measured human somatic-cell mutation frequencies (64). Streisinger (138) has proposed a more complex extrapolation method using two sequential paral- 94 0 Technologies for Detecting Heritable Mutations in Human Beings lelograms. In the first parallelogram, a measure of chemical interaction with human germ cell DNA is estimated from measurements of the effects of the chemical in human and animal (e.g. , monkey) somatic cells, and a measure of interaction of the same chemical with germ-cell DNA in the same animal. In the second parallelogram, the ratio of a measure of chemical interaction with mouse germ-cell DNA to mouse germ-cell mutation rates (from the specific locus test) is used to predict the human germ-cell mutation rate using the estimated value of chemical interaction with human germ- cell DNA from the first parallelogram. This con- struct embodies several untested assumptions. Bridges (1980) also developed a more complex extrapolation model based on the original paral- lelogram model of Sobels. Based on his approach, Bridges outlined the types of results needed to fill information gaps, ultimately to assess the impact of mutagens at the: 1) molecular; 2) cellular; and 3) whole organism levels in both animals and man. He suggested studies to determine: 1) the presence of an effective dose of mutagen at the molecular level by measuring the concentration of mutagen in the gonads or blood or the extent of reaction with DNA; 2) whether there appears to be a rela— tionship between the presence of the mutagen and a biological response at the cellular level by meas- uring somatic mutation frequencies or chromo- somal changes in lymphocytes; and 3) whether there is an effect at the whole organism level by measuring the frequency of heritable genetic defects, congenital malformations, or fetal loss. The values Bridges specifies are obtainable in animal systems. To obtain such values for man, Bridges suggests that use be made of certain other- wise normal human populations that are exposed to large doses of mutagens. Examples of such pop- ulations are patients treated for diseases, such as cancer, with drugs that are known to be muta- genic, and certain occupational cohorts in which there are known excesses of cancer (1,155). Simul— taneous studies using the same mutagens could be carried out in experimental mice to determine the relative sensitivities of mouse and man to these mutagens. Parallelogram models are attractive for their simplicity and inherent logic. They are appropri- ate starting points for exploring relationships among test results when sufficient data become available to do so. However, the assumptions em- bodied in parallelogram models—consistent, pre- dictable relationships among various cell types, translatable among species—are almost entirely untested. The great differences among species make it unlikely that these parallelogram models will survive validation studies intact. While they may continue to be useful research tools for pos- ing logical questions, they may or may not prove practical for predicting risks of heritable muta- tions in human beings. ATTEMPTS AT QUALITATIVE EXTRAPOLATION L. Russell and colleagues (106) compared the results of mutagenicity tests carried out in a vari- ety of systems other than whole mammals with results from specific locus tests (SLTs) and herita- ble translocation tests (HTTs) in mice (see ch. 6 for descriptions of these two tests). The purpose of the comparison was to find out how well re- sults of the nongerm-cell tests corresponded to the qualitative results (positive or negative) of the two germ—cell tests. The analysis was limited by the relatively small number of chemicals that have been tested in either the SLT or the HLT. About 35 chemicals have been tested in one or both of the germ-cell assays, out of a total of about 2,000 chemicals for which some test results are avail- able from any system.2 The comparison tests were grouped into 18 cat— egories and the categories given relative weights according to their biological relationship to one of the germ-cell tests. The categories and their weights are given in table 9. The lowest weighted category includes tests using prokaryotes, such as bacteria, directly treated by the suspected muta- gen. Higher scores signify moving toward higher zThese test systems are not the focus of this report and are not described here in detail. Descriptions of these tests can be found in (88). Ch. 7—Extrapolation - 95 Table 9.—Weighting of Test Results for Presumptlon of Germ-Line Mutagenicity Exposure Exposure not within within mammalian mammalian body Weight body Weight Germ cells Weight Prokaryotes, all endpoints SAL 1 BFT 2 WP_ HMA Lower eukaryotes, all endpoints YEA YEP 2 ASP NEU Higher eukaryotes, chromosome aberrations PYC 3 DAN 8 DHT Higher eukaryotes, gene mutations PGM 3 SFlL 8 Mammals, genetically nondefined endpoints $01 803 6 SPM 4 802 4 $04 UDH UDT 8 UDP Mammals, chromosome aberrations CYC 5 MNT 7 DLT CYE CY9 15 CYB 10 CYO CY5 Mammals, gene mutations CHO 0Y8 V79 5 MST 10 SPF 15 L51 NOTE: In general. the weights increase from top to bottom and from left to right in the table. From top to bottom, the tests progress from lower to higher organisms and from more general endpoints to endpoints of direct relevance to human heritable mutations. From left to right, the categories progress from in vitro tests in both somatic and germ cells, to in vivo germ-cell tests. Explanation of test symbols: ASP Aspergillus, all tests BFT Body fluld tests, all assays CHO Chinese hamster ovary cells in culture CYB Mammalian cytogenetics, in vivo, animal bone marrow CYC Mammalian cytogenetics, in vitro, all cell types CYE Mammalian cytogenetlcs, in vivo. animal lympho- cytes or leukocytes CYO Mammalian cytogenetics, in vivo, oocyte or early embryo studies CY5 Mammalian cytogenetics, in vivo, human bone marrow 0Y8 Mammalian cytogenetics, in vivo, human lympho- cytes or leukocytes 0Y9 Mammalian cytogenetics, all male germ-cell studies DAN Drosophila aneuploidy studies, all tests DHT Drosophila heritable translocatlon test DLT Dominant-lethal test in rodents HMA Host~mediated assay studies L51 Mouse lymphoma cells in culture, gene muta- tions at TK locus MNT Micronucleus test, all species MST Mouse spot test NEU Neurospora crassa, all tests PGM Plant gene mutations, all tests PYC Plant chromosome studies, all tests SAL Salmonella histldine reversion tests 501 Sister~chromatld exchange, human cells in vitro SCZ Sister-chromatid exchange, animal cells in vitro 80:) Sister-chromatid exchange, animals in vlvo $04 Sister-chromatid exchange, human cells in vivo SPF Sperm abnormalities in F. males SPM Sperm abnormalities in treated animals SFlL Drosophila sex-linked recessive lethal test UDH Unscheduled DNA synthesis, human diploid fibroblasts UDP Unscheduled DNA synthesis, rat primary nepatocytes UDT Unscheduled DNA synthesis, testis In vlvo V79 Chlnese hamster lung (V79) cells In culture, all gene mutation studies WP_E. coIi reverse mutation studies YEA Saccharomyces cerevlsiae, all tests YEP Schlzosaccharomyces pombe, all tests SOURCE: LB, Russell, CS. Aaron, F. de Serres, et al., “Evaluation of Mutagenicity Assays for Purposes of Genetic Risk Assessment," Mutation Research 134:143-157, 1984. mammals, toward germ cells, and toward treat- ment with the chemical in a whole mammal. A single composite score was calculated for each chemical tested, adding together scores from each category in which there were test results. There is only one score per category regardless of the number of tests. Positive results are scored as posi- tive numbers; negative results as negative num- bers, e.g., an in-vitro somatic-cell chromosome aberration test with a positive result yields a score of +5, one with a negative result, a score of —5. Russell and colleagues found that nearly all chemicals that tested positive in either or both the SLT and HTT had high composite scores from other tests. A number of chemicals with negative SLT and HTT results also had high, positive com- posite scores, representing “false positives" in the comparison tests. Similar analyses looked separately at the SLT and HTT and the comparison tests that specifi- cally detect gene mutations or chromosome aber- rations, respectively. The results are similar to those matching the results in all comparison tests against both mammalian germ-cell assays: high scores for most chemicals positive in the germ- cell tests, and a number of false positives. 96 - Technologies for Detecting Heritable Mutations in Human Beings In an additional analysis, the comparison tests are ranked according to how well each predicts the results of the two germ—cell tests. In general, the tests in higher numbered categories in the earlier analyses, i.e., those that are closer biolog- ically to whole mammal germ-cell tests, had bet- ter correlations with the SLT and HTT. For the SLT, the best predictors overall were the mouse spot test, unscheduled DNA synthesis in mouse testis, and the micronucleus test. For the l-ITT, unscheduled DNA synthesis in testis, the domi— nant-lethal test, and one lower ranked test, sister- chromatid exchange in cultured animal cells, were the strongest predictors. While it appears that the results of some of the comparison tests correlate relatively well with the mammalian germ—cell tests, in fact, not one of these correlations reaches conventional statistical sig- nificance, meaning that the tests do not predict reliability better than chance. The lack of signifi- cant results is due, in large part, to the small num- ber of comparisons for many tests, and in part because of the process for selecting chemicals for SLT and HTT. From a practical, public policy standpoint, this is an important finding. The lack of statistically significant results does not mean that these comparisons are without value. The study provides a status report on the quality and quantity of existing data. Since the two whole mammal tests (the SLT and HTT) are relatively expensive and time-consum- ing, they are usually reserved for testing chemi- cals highly suspected of causing heritable muta- tions. The suspicion is based on results of other tests, specifically the comparison tests examined in Russell and colleagues' analysis. It is hardly sur- prising, therefore, that the comparison test results are largely positive for chemicals eventually tested in the mammalian germ-cell assays. Russell and colleagues took the preponderance of positive re- sults into account in their analyses. Many chemicals have been tested in mutage- nicity assays because, for reasons of chemical structure or other properties, there is a high likeli- hood that they will be mutagenic. While these chemicals have proved useful as laboratory tools, they are not necessarily useful for drawing con- clusions about what people are actually exposed to. W. Russell (111), using the same database used by L. Russell and colleagues (106), looked exclu— sively at the 11 “environmental chemicals” (those found in the home or workplace) that have been tested in the SLT and examined the results. All 11 are positive in the Drosophila sex-linked lethal test, the 10 that have been tested are positive in mammalian somatic-cell tests, and there are a va- riety of positive results in other test systems. None of the 11, each of which was tested at very high doses, is positive in the SLT, suggesting no increase in mutations in spermatogonia (the pre-meiotic male germ-cell stage) although several have posi— tive results in tests of later sperm developmental stages. What conclusions can be drawn about the va- lidity of qualitative extrapolation from various mutagenicity tests to a risk of heritable mutations in human beings? The available data give no direct information about mutagenic effects of chemicals on human germ cells in vivo. Application to hu- mans rests on a series of assumptions about the response of human germ cells in relation to re- sponses in other types of cells and in other species. The analysis of the these test results does allow some generalizations about biologic extrapolation and about the nature of the available test data- base. On the first point, there is evidence suggest- ing that chemicals that test positive in some com- parison tests for gene mutations or chromosomal aberrations have a likelihood of being positive in the SLT or HTT, respectively, but are not invari- ably so. To date, chemicals testing negative in the comparison tests have not produced clear posi- tives in whole mammal germ-cell tests, but the database supporting this comparison is limited. Biologically, it seems unlikely that chemicals that are consistently negative in comparison tests would, in fact, be germ-cell mutagens in whole animals. But it is unlikely also that very many chemicals with negative results in one or two tests will have been tested in enough systems to allow an empirical test of that hypothesis. Both L. Russell’s and W. Russell’s analyses de- scribed above suggest that a number of chemicals that test positive in the simpler comparison tests will be not be shown to cause mutations in sper— matogonia] germ cells. At present, however, it is impossible to know based on comparison test re- sults, which chemicals are true human germ-cell mutagens and which are not. Ch. 7—Extrapolation 0 97 A major limitation of the database is that nearly all tests have been in exposed male animals. Two strong animal mutagens—radiation and a chemi- cal agent, ethylnitrosourea—do not appear to cause heritable mutations in the immature germ cells of female mice, but the data for females are not sufficient to draw firm conclusions. The new technologies described in this report, which could be used to provide information on the types of mutations detected by the various tests, may improve our ability to apply the re— sults of simpler tests to predicting human risk. ATTEMPTS AT QUANTITATIVE EXTRAPOLATION Not surprisingly, quantitative extrapolation is even less advanced than is qualitative extrapola- tion. However, there are some data bearing on quantitative relationships that may eventually be useful in predicting effects in human beings. First, there is some information about dose-response relationships. Second, some preliminary attempts are being made to determine relationships among certain corners of the parallelogram models de- scribed earlier in this chapter. One such effort is described below. In an ongoing effort, a group of investigators is using a parallelogram approach to evaluate the effects of gamma radiation, cyclophosphamide (a medical drug) and benzo[a]pyrene (an environ- mental chemical) on mouse and human cells by assaying chromosomal aberrations and sister chro- matid exchanges in mitogen-stimulated peripheral blood lymphocytes (PBLs). While these investi- gators have as yet little data, a paper presenting some preliminary results (170) lays out the rigor- ous procedures necessary for proper extrapola- tion of results from different studies of just one substance to predictions in untested systems. For instance, in the case of radiation, the authors re- late an experimental result in irradiated cultured human PBLs to reports from the literature of the same chromosomal endpoint (dicentrics) in PBLs from patients who have received therapeutic ra- diation. Thus far, these authors have not presented any conclusions about the relationships they are studying. Dose-Response Relationships Extrapolation from high to low dose, and from high to low dose rate, requires knowledge of dose— response relationships. For heritable gene muta- tions, adequate data on dose-response relation- ships is limited to the effects of ionizing radiation and one chemical, ethylnitrosourea (ENU), in male mice. The available information comes from re- sults of SLTs in mice, most of the work being done in a small number of laboratories in different parts of the world. Radiation and ENU have also been assayed in many short-term in vivo and in vitro tests in human somatic cells, so some generaliza- tions may be drawn about the nature of dose- response relationships for these two agents. In both cases, there are independent effects of dose rate and of total dose. This means that a fixed dose may cause a different rate of mutations depending on the intensity of the dose, i.e., a more protracted administration may result in fewer mutations than if the total dose is administered at one time. Radiation Specific-Locus Test. —In a series of experiments from the mid-19505 to the present, a range of ra- diation doses, delivered at a range of dose rates, has been tested in male mice. Results are avail- able for both spermatogonial and postspermato- gonial stages. In spermatids and spermatozoa (postspermatogonial stages), the dose-response relationship is generally linear, and there is no ef— fect of dose rate. This means that approximately the same mutation rate results from a short, high- dose exposure and from a chronic, low-dose ex- posure when the same total dose is given. In spermatogonia, the early, pre-meiotic stage, for equal total exposure, radiation given at high dose rates causes more mutations per unit of dose than radiation at lower dose rates. At a high dose rate of 90 Roentgens per minute (R/ min) or inter- mediate dose rate of 8 R/ min (work cited in 111), 98 0 Technologies for Detecting Heritable Mutations in Human Beings the mutation rate decreases with decreasing dose faster than would be predicted by a linear rela- tionship. At dose rates of 0.8 R/ min or below, however, the rate of mutations per unit dose ap- pears to be constant, and without a threshold. At low dose rates, the dose-response relationship is linear, and above about 0.8 R/ min, the mutation rate per unit of dose rises more quickly than would be predicted by a linear relationship. Above a to- tal dose of between 600 and 1,000 rem, the muta- tion rate begins to decline rapidly. Investigators using the SLT offer an explana- tion for the observed dose-response patterns (111, 114,115). They postulate that the difference in dose-rate response between spermatogonia and postspermatogonial stages is a function of an ac— tive repair mechanism in metabolically active sper- matogonial cells, which does not function at later stages. In the earlier, spermatogonia] stages, a larger percentage of changes can be repaired be- fore the spermatogonia complete meiosis when ex- posure is at a lower dose rate than is the case with an acute, high-dose—rate exposure. In post-sperma- togonial stages when capacity for repair is low, the total radiation dose, irrespective of dose rate, is the determinant of the mutation rate. Heritable Translocation Test.—Generoso and co-workers (135) have investigated heritable trans- locations induced by high-dose rate (96 R/ min) irradiation of spermatogonia] stem cells of mice. They report a linear dose-response relation be- tween 0 and 600 R of total irradiation, and repeti- tion of doses in this range gave additive effects up to 2,000 R. From these data, the expected in- crease in heritable translocations at high-dose-rates is calculated to be about 0.00004 per R. Cytogenetics.—Waters and colleagues (170) de- scribe a dose-response relationship for gamma ra- diation after both in vitro and in vivo irradiation, using as an endpoint a certain type of chromo- somal mutation, dicentrics, in PBLs. The in vivo dose-response, which was derived from reports in the literature, is linear at lower dose rates, and quadratic at higher dose rates, meaning that the increase in the number of dicentrics rises faster than it would if the relation continued to be lin- ear. The quadratic component may be explained by an interaction of mutational events causing some dicentrics, and the interactions being increas- ingly more likely as the dose rate increases. Ethylnitrosourea (ENU) Specific Locus Test—As is the case with radia- tion, both total dose of ENU and dose-rate inde- pendently affect the mutation rate in mouse sper- matogonia. Experimental data indicate that the response at doses below 100 mg/ kg of body weight is ”infralinear,” meaning that as the dose is lowered, the mutation rate drops faster than would be pre— dicted by a linear relationship. At doses between 100 and 400 mg/ kg, the response appears to be linear (41). No threshold was detected over the range of doses tested, but the possibility of a , threshold at a dose lower than 25 mg/ kg (the low- est single dose tested) is not excluded by the data. In an experiment examining mutational re- sponses at different dose rates, the mutation rate was measured in mice that were given 10 weekly doses of 10 mg/ kg of ENU, and compared with the mutation rate for a single dose of 100 mg/ kg of ENU. The mutation rate for the fractionated dose was only about 15 percent of the rate for a single dose (112). Russell (111) notes that, in light of information indicating that ENU reaches germ cells in doses proportional to injected amounts (see next section), these results cannot be explained by differences in metabolic processes. He interprets the infralinear portion of the dose-response curve and lower mu- tation rate that follows dose fractionation to be the result of effective mutational repair systems in spermatogonia, the same reasoning as in the case of radiation. Unscheduled DNA Synthesis in Mouse Sperma- tids.—Carricarte and Sega (cited in 111) found the dose-response of “unscheduled DNA synthesis" in mouse spermatids to be linear over the range from 10 to 100 mg/ kg, the same range over which W. Russell (111) found an infralinear relationship in the SLT. Sega (cited in 111) also measured ”ad- duct formation” after injections of ENU, and found a linear response in the range from 5 to 100 mg/ kg. One conclusion from these observations is that chemical interaction with DNA may not always be directly related to the rate of mutation, and Ch. 7—Extrapolation 0 99 this presents a major difficulty for Sobels' parallelogram, described earlier in this chapter. Sister-Chromatid Exchange and Thioguanine Resistance.—]ones and colleagues (54) investigated the dose-response relationship for two somatic- cell endpoints in whole mice exposed to ENU. The frequency of sister-chromatid exchange (SCE) and the frequency of thioguanine resistant (TG') cells were measured in white blood cells in the spleen. Corresponding endpoints exist in human beings, which makes them potentially useful for extrapo- lating to human responses. The investigators meas- ured both the response over time at several dose levels, and the response to a range of doses at sev- eral points in time after exposure. Linear relationships were discovered between dose and reponse for both SCEs and TGr cells, but the timing of these responses was quite differ— ent. For all dose levels, the highest SCE levels were measured on the first day after exposure, and SCE LIMITS OF EXTRAPOLATION The main limitation to validating extrapolation models is a virtual lack of data to complete the parallelogram models. Until there is enough em- pirical information to determine whether qualita- tive and quantitative generalizations can be made, it is impossible to know how useful extrapolation could be. At present, qualitative extrapolation is at least a tentative guide for identification of muta- genic agents. Even when data are available, many questions bedevil the use of experimental animal data to ex- trapolate risks to humans. In particular, little is known about the comparability of species with respect to activation, detoxification, and tissue dis- tribution of specific chemicals, as well as other interspecies differences in metabolism. In addition, mutational responses of germ cells at different stages of development can differ for different types of mutagens. Many of these gaps in information could be filled by studies that are now technically feasible. levels decreased back to baseline after about 70 days, whereas the TGr response rose linearly over time to a peak after about 80 days for each dose level. Generalizations About Dose-Response While dose—response relationships cannot be generally and simply described some generaliza- tions can be made. It appears that mutations in somatic cells are more predictable from the vari- ous experimental test systems available than are mutations germ-cell. Repair of mutations in germ cells may explain this difference. Both a better understanding of the mechanisms of mutation and repair at the molecular level, and empirical com- parisons of test data on more substances will con- tribute to a better understanding of dose-response relationships. Once again, the new technologies may contribute to this understanding. In using test results from somatic cell systems, there are differences in the sensitivities of various types of somatic cells and germ cells to different mutagens. Even within a single cell type, differ- ent gene loci can respond very differently to a spe- cific mutagen. In making comparisons, for in- stance, between somatic and germ cells, if the same locus (or set of loci) cannot be tested, any differ- ences in outcome cannot necessarily be attributed solely to a difference in cell type. These are a few of the many uncertainties in extrapolations from somatic to germ cells. An organizational problem that affects extrap- olation in a practical way is that collaborations have been rare among researchers working with different systems, and are particularly rare among scientists working in different laboratories. In addi- tion, the test systems are not necessarily stand- ardized among laboratories working with the same systems, so results from different laboratories can— not always be combined or directly compared. 100 0 Technologies for Detecting Heritabie Mutations in Human Beings Even relatively simple aspects of experimental pro- cedures, for instance, the conditions under which cells or organisms are exposed to radiation or chemicals, vary enough among laboratories to ren- der results incomparable to varying degrees. Overall, the development of methods to extrap- olate from different experimental test systems to the risk of heritable mutations in human beings is still in its infancy. While the various parallelo- gram models are appealing because of their poten- tial usefulness, some of the data already available suggest that they may never broadly applicable. Both qualitative and quantitative improvements in the database of experimental results will be nec- essary before the usefulness of extrapolation can be reasonably assessed. Chapter 8 Mutation Epidemiology Contents Page Introduction ............................................................ 103 Validation of Measurement Techniques .................................... 104 Surveillance and Disease Registries ........................................ 104 Monitoring and Exposure Registries ....................................... 106 Ad Hoc Epidemiologic Studies ............................................ 107 Population to Study ..................................................... 108 Radiation-Exposed Groups ............................................. 108 Cancer Patients ....................................................... 108 Other Populations ..................................................... 109 Conclusions ............................................................ 110 Chapter 8 Mutation Epidemiology INTRODUCTION This report is about biochemical and genetic techniques for studying heritable mutations. These new techniques will ultimately be used to study people who are suspected of being at high risk for excess heritable mutations. The types of epidemio- logic activities in which mutation detection tech- niques may eventually be used are: surveillance, monitoring, and ad hoc studies. Before any tech- nique can be used for those purposes, a series of validation studies will be needed, calling for dif- ferent populations that are appropriate for study at different stages of development of the tech- nologies. SurveiIIance is a routine activity whose aim, in the context of this report, would be to measure the ”baseline” rate of mutations in a defined pop- ulation over the course of time, and to facilitate rapid recognition of changes in these rates. Fol- lowing the distinction made by Hook and Cross (44), the term monitoring is reserved for obser— vations over time in populations thought to be at increased risk for heritable mutations because of a known or suspected exposure to a known or suspected mutagen, for the purpose of helping the specific population in whatever way possible. Ad hoc studies of a variety of designs are carried out to test hypotheses about suspected causes of mu- tations. The different aims of surveillance, monitoring, and ad hoc studies require that different criteria be applied for deciding when and whether to carry out one or more of those activities. Surveillance and monitoring are not designed as hypothesis- testing activities, though they may be sources for hypothesis development. The reasons a population is chosen for surveil- lance may be largely opportunistic. It is unlikely that an entirely new system of data collection would be put in place for mutation surveillance. It is more probable that mutation surveillance would be added to an existing program that is established for another purpose, for instance, birth defects surveillance. The population covered must be large enough to generate reliable rates for mutational events that may be relatively rare, but there is no fixed requirement for size. Of the three types of activities, surveillance generally would involve the smallest effort and resource expendi- ture per individual, but because large numbers of people would be routinely subject to the surveil- lance test, the total cost could be large. There is, therefore, a great need to consider the costs and benefits of such a program before embarking on one, and for choosing the detection technique ac- cordingly. The threshold for initiating mutation surveillance would be relatively high. Informa- tion about trends sometimes can be obtained by means other than full-scale surveillance, and some such studies for that purpose have been done. The main reason for instituting a mutation monitoring program is concern about the poten- tial effects of a mutagenic exposure in a specific group of people. If there is enough concern, there undoubtedly will be a greater expenditure of re- sources and effort per person than is the case for surveillance, meaning that more extensive con- tact with the population and testing would be justified. Given today's knowledge, it is hard to conceive of a situation in which there would be a concern only about heritable mutations, so any mutation monitoring effort would most likely be part of a larger program. The obvious concurrent concerns for exposures thought to be mutagenic are cancer and birth defects. Strict criteria based on tests of the statistical power to detect certain levels of effects are not appropriate for monitored populations, since the concern is about that par- ticular group of people. Any finding is of inter- est in that situation. The information may be used incidentally to calculate upper limits of risk which could be generalized to other populations with similar exposures. The most important consider- ation in making a decision to monitor is that there is reasonable evidence suggesting that the popu- lation may be at a substantially increased risk. 103 104 o Technologies for Detecting Heritable Mutations in Human Beings This decision may well be influenced by political pressures to act, but ideally there should be a rec- ognition that the best scientific judgment either does or does not support a monitoring effort. The purpose of ad hoc studies is to test hypoth- eses about exposures and effects. This is the one place where it is imperative that studies be de- signed to achieve a high probability of detecting an effect if it is present. Such studies are valuable not only for the sake of the populations involved in the studies, but for their generalizability to other populations. Results of these studies can form the basis for public health actions, if levels of risk can be established. VALIDATION OF MEASUREMENT TECHNIQUES For the most part, the new technologies de— scribed in this report have not been ”validated.” They are new, and have not been applied to large numbers of people. Surveillance, monitoring, and ad hoc studies all require tools—in this case tech- niques for measuring mutations—with an acceptable degree of validity. That is, the tests must meas- ure what they are designed to measure, within some definable limits. Generally, it is not possible to gather reliable information about a population and concurrently gather validating information about a technique used to measure outcome, un- less another technique, with known validity, and known relationship to the new technique, is also applied in the study. Even though that is techni— cally feasible, it is probably not an efficient way to gather validating data. A first step in the validation process would be to use laboratory-prepared samples with known DNA sequences to confirm that the types of mu- tations that should theoretically be detected with a new technique actually are reliably detected. Be- yond that stage, the need to move to clinical sam- ples can be met using stored blood from individ- uals studied previously for other reasons. These stored samples need not be from parent/child triads initially, but triads will be needed at a later stage. A number of research organizations are stor- ing samples that would be appropriate for studies of mutations using new DNA techniques. The Na— tional Cancer Institute for example, is storing blood samples from cancer patients who have been treated with drugs and radiation, and the Radiation Effects Research Foundation has stored blood from Japanese citizens who were exposed to atomic bombs. In both of these cases, DNA is stored according to an established technique that uses Epstein-Barr virus to transform lymph- ocytes, thereby ”immortalizing" the cells so they can be grown indefinitely. The transformed cells can be frozen for the long term in liquid nitro— gen. Both sample preparation and long-term stor— age costs are substantial. With currently available technologies, it is unlikely that large numbers of samples will be stored. This limits the number and variety of samples available for validation studies, and also for later studies of people exposed to po- tential mutagens. SURVEILLANCE AND DISEASE REGISTRIES The methods and aims of disease surveillance have developed based on experience in infectious disease control. Reporting of vital statistics, in particular births and deaths for calculating birth and death rates, is also a form of surveillance. Sur— veillance of noninfectious diseases is a relatively recent development, with roots in the desire to track the incidence of cancer. Although the first national cancer surveillance system, which cov- ers about 12 percent of the population, was put in place as recently as 1972, New York State in- stituted a reporting system for cancer cases in 1940, and Connecticut followed the next year. There are now dozens of cancer surveillance sys- tems operating around the country and interna- tionally, covering a range of populations from counties to whole countries (145). Information about individual patients in can- cer surveillance systems forms the basis for ”can- cer registries." The routine output of registries consists of cancer rates by sex, age, and race (where applicable) for each cancer site. Registries also are an important source for researchers in- vestigating hypotheses about cancer causation. In this sense, cancer surveillance, with information about individuals recorded on registry forms, is similar to mortality statistics, with information about individuals recorded on death certificates. It should come as no surprise that there are no ”heritable mutation surveillance systems” now in place. There are, in various places around the world, registries of birth defects, which include records of at least some sentinel phenotypes. Be- yond that, as this report shows, there are at present no techniques for detecting mutations that are suitable for use in a large-scale population sur- veillance program. Because developments have been so rapid, however, there may be one or sev- eral good candidate techniques within 5 or 10 years. Surveillance traditionally has involved report- ing, to a central place, information already co]- Iected by some segment of the health care system for reasons directly related to the health of indi- viduals. Even for infectious diseases, only cases that come to the attention of physicians are re- ported. Active ”case-finding" in the population is not a usual feature of surveillance. For chronic disease, the same is true. Cancers diagnosed by physicians are entered into registries. Case-finding programs, such as breast cancer screening pro- grams, are instituted on the basis of their effec— tiveness in identifying cases early in the course of disease, for the benefit of the individual with the disease. Infants are examined for birth defects because of the potential impact on the children and their parents' lives, and not mainly for the purpose of computing the rates of birth defects in a popula- tion. Testing programs for newborns, including biochemical tests for metabolic diseases (not nec- essarily a result of a mutation), also have been Ch. 8—Mutation Epidemiology 0 105 instituted because of their importance to the health of the individual. Nearly all States now require testing newborns for phenylketonuria (PKU), and some require additional tests. For instance, New York requires testing for PKU, sickle-cell anemia, and congenital hypothyroidism, which are moder- ately rare, and also for very rare conditions in- cluding maple syrup urine disease, homocystin- uria, histidinemia, galactosemia, and adenosine deaminase deficiency (102). The tests do not im- pose an added burden on the newborn, since all are carried out using the same blood sample. Cytogenetic techniques have not been used for population-based surveillance of chromosome ab- normalities, but some large hospital-born series of newborns have been tested (102). Most of the recorded cases of chromosome abnormalities found in this way might eventually have been detected later in life because of health and reproductive problems, but some others might otherwise have gone undetected. There is no formula for deciding whether to in- stitute a surveillance program, but there are char- acteristics of the disease, of the population, and of the particular test to be used that contribute to the decision: 1) the seriousness of the disease (if the measured endpoint is known to be associ- ated with a disease); 2) the ability to alter its clin- ical course after diagnosis; 3) the prevalence of the disease in the population; 4) the reliability and validity of the test; 5) the acceptability of the test to the population; 6) the cost of the screening pro- gram; and 7) the cost of not screening (i.e., the cost of treatment and social support). It is worth considering these factors in thinking about screen- ing and surveillance for heritable mutations. The idea of surveillance for heritable mutations represents a departure from the traditional appli- cations of surveillance. It appears to be the case that most heritable mutations are not related to disease over the course of an individual’s lifetime, and no predictions useful to the individual can currently be made about the effect of a heritable mutation in the absence of recognizable disease, beyond those that are associated with sentinel phenotypes and major chromosome abnormal— ities. As mutation detection techniques become more and more sensitive, in fact, a greater per- 106 0 Technologies for Detecting Heritable Mutations in Human Beings centage of the mutations detected may not be re- lated to a known effect on health. Heritable mutation surveillance beyond report— ing sentinel phenotypes will require more than just a reporting of events already detected. It will re- quire imposing a test burden on a population for the sole purpose of collecting information about mutations that may never affect an individual's life. This argues against instituting surveillance. A reason in favor of surveillance is that it is clear that increased mutation rates will be looked for in special populations, those being monitored be- cause of worries that they have been exposed to a mutagen. Surveillance systems can provide a range of estimates of ”baseline” or ”background” rates, even though they may be from different populations. In a more general sense, one of the original aims of surveillance is relevant: to sub- stantiate long-term trends and patterns in health events and to detect changes that may be ad— dressed by public health action. MONITORING AND EXPOSURE REGISTRIES Monitoring is the ”long-range observation of individuals who are at presumptive high risk for adverse outcomes because of specific life events,” (44) in particular, exposures to suspected muta- gens. The event may be catastrophic, such as ex- posure at the time of detonation of an atomic bomb, or a chemical plant explosion. Or the “event” may be long term, such as an occupational or an environmental exposure. There are about two dozen populations around the world cur- rently monitored for long-term health effects, and some of those programs include various studies of heritable mutations. The most extensive population monitoring, in- cluding monitoring for mutations, is of the Japa- nese residents of Hiroshima and Nagasaki, many of whom were exposed to substantial amounts of radiation during World War II when atomic bombs were detonated in those cities. The popu- lation around a chemical plant that exploded near Seveso, Italy in 1976, releasing several pounds of dioxin, is the subject of health monitoring activi- ties, including monitoring for birth defects. The people exposed to methyl isocyanate in Bhopal, India, will undoubtedly be followed for years to come. Because these groups were exposed, and because it is conceivable that something could be done to alleviate health problems if they are de— tected early, or if warning signals are picked up, they are being monitored; the scientific knowl— edge gained as a result is a secondary benefit. The most prominent examples of chronic ex- posures are from occupational activities and toxic chemicals in the environment. The populations exposed to hazards often are not geographically determined, but may be a collection of workers from around the country. Workers exposed to radiation in the nuclear power industry are an ex- ample of this. There are several ”exposure regis- tries” in existence worldwide, though none spe- cifically because of a perceived increased risk of mutations. One such registry has the names of all workers who were exposed to dioxin during the manufacture of various chemicals in this coun- try. There also is an international dioxin regis- try, with names of workers from all around the world. The registry does not, however, have in- formation about the health status of those work- ers. A similar registry for workers exposed to beryllium exists in this country. A report prepared for the Nuclear Regulatory Commission in 1980 recommended that a registry be started for work- ers exposed to low-level ionizing radiation in cer- tain types of workplaces, because of a possible increased cancer risk (29). These registries could be used for monitoring and as a potential popu- lation to include in ad hoc studies. Ch. B—Mutation Epidemiology 0 107 AD HOC EPIDEMIOLOGIC STUDIES Surveillance, monitoring, medical case reports, and laboratory research can all lead to hypothe- ses about possible causes of heritable mutations. An investigator wishing to test a hypothesis must find suitable subjects to study, in contrast to a monitoring activity, where the existence of the ex— posed population is the reason for acting. A study should be undertaken only if there is a good chance of answering the question of interest. Disease and exposure registries are common sources of indi— viduals to study, depending on the question. A cohort design will probably prove the most useful approach for studies of heritable mutations, though case-control studies of sentinel phenotypes may also prove valuable. A cohort study involves identifying a group of individuals, some exposed to the suspected mutagen and some not exposed. The health outcomes, i.e., the presence or absence of mutations in offspring, of the two sub-cohorts are compared. A higher rate of mutations in the exposed group would signify an “association" be— tween the exposure and heritable mutations. Sta- tistical tests are applied to the results to estimate the likelihood of the result occurring if in fact there was no real difference in mutation rates between the two groups. In a case-control design, a group of “cases,” in- dividuals with conditions of interest, e.g., sentinel phenotypes, is compared to a group of individ- uals who do not have the condition of interest, but who are otherwise similar demographically. The cases and controls are compared according to their past histories of exposures or other char- acteristics that might be associated with the mu— tation and an assessment made as to whether their histories differ in important ways. The important question for all studies is not just whether the exposure is ”associated with" muta- tions, but whether it causes them. That is a diffi- cult if nearly impossible judgment to make in most instances, but there are some generally accepted guidelines for evaluating the likelihood of an asso- ciation being causal based on epidemiologic evi- dence. These are: 1. Consistency: The association is observed in studies by different investigators, at differ- ent times and in different populations, and in studies of different designs. 2. Strength: The size of the effect of an exposure is the measure of strength of association. This is usually measured as an estimate of relative risk (a ratio of the rate of mutations in an exposed group to the rate in an unex- posed group). The presence of a dose-re- sponse relationship, that is, the size of the effect changes in a logical way with the level of exposure and in at least some cases, with the dose rate. 3. Specificity: Specificity refers to the degree to which the exposure is associated exclu- sively with the outcome of interest, in this case a mutation, and the degree to which a mutation is associated exclusively with the exposure. The concept of specificity derives from study of infectious disease and is rele- vant to the study of mutations (and chronic diseases generally) only in special cases, for example, a specific mutation that almost never occurs in the general population but appears to be exclusively related to a par- ticular exposure. While a highly specific rela- tionship can provide positive evidence for a causal relationship, a lesser degree of speci- ficity does not necessarily argue strongly against causality. 4. Temporal Relationship: The exposure must occur before the effect. In the case of herita- ble mutations, the picture is more compli- cated. See chapter 6 for a discussion of the timing of exposure for males and females for a plausible effect on germ cells. 5. Coherence: All available information from medical and biological science, and from epi- demiologic observations and studies, fits to- gether in a way that supports the hypothe- sis. The greater the variety of information, and types of study designs, the stronger the finding of coherence. These criteria are quite stringent, and even in the best of cases, often cannot be met, but they are useful as standards. 108 - Technologies for Detecting Her/table Mutations in Human Beings POPULATIONS TO STUDY There are elements in the environment that damage human health under certain conditions of exposure. Biologic, chemical, and physical agents cause acute and chronic diseases in hu- mans. At present, there are no exposures une- quivocally known to cause heritable mutations in human beings. A combination of factors, in— cluding the rather insensitive methods for detect- ing heritable mutations that have been available, and the possibility that human germ cells may not be very susceptible to some mutagens, probably contribute to this situation. As a consequence, in- vestigators looking for the effects of mutagens must do so in people who have been highly ex- posed to agents that are very likely to be muta- genic in germ cells. There are not very many large groups of people fitting that description, a fact that many might find surprising. Radiation-Exposed Groups Radiation causes heritable mutations in labora- tory mice and is the most likely potential germ- cell mutagen to which large numbers of human beings have been exposed, either intentionally or accidentally. The largest population with a known high radiation exposure, the Japanese atomic bomb survivors, continue to be followed for ef- fects on cancer incidence, birth outcomes, and heritable mutations. Heritable mutations have been studied by clinical observations, cytogenetic techniques, one-dimensional electrophoresis of blood proteins, and more recently with the most sensitive technique of two-dimensional gel elec- trophoresis of blood proteins (see ch. 3). A report was prepared in 1980, under contract to the Nuclear Regulatory Commission (NRC), that evaluated opportunities for studying the health effects of low-level ionizing radiation (29). The report is focused on cancer, but the evalua- tion methods apply equally to studying muta- tions. The authors initially identified 100 candi- date populations. About 30 remained after two broad criteria were applied: 1) that there be data identifying exposed individuals, and 2) that there were at least 10,000 people in a single population group or one comprising several similar groups. Those 30 populations were evaluated further, and recommendations made that if additional studies were to be undertaken, three occupational groups and one group with environmental exposure held out the greatest promise of yielding a reliable re— sult. Even the best of these, however, has a rela- tively low power: less than a 50 percent chance of finding an excess of cancer if it exists. In gen- eral, this level of power would be unacceptable in an epidemiologic study. Although political con- siderations might influence a decision to go ahead with a study, they do nothing to increase the power of the method. The power figures for these studies refer to can- cer detection, and the probability of detecting heritable mutations is undoubtedly far lower, making it unlikely at best that anything could be learned about radiation and heritable mutations by studying any of these groups with currently- available methods. The report to the NRC contained one other rec- ommendation, that a registry for radiation work- ers be initiated. The registry would maintain information about radiation doses and some in- formation about other exposures. This recommen- dation has not been acted on. There are exam- ples of radiation-exposure registries, but these are mainly for people acutely exposed accidentally, and not for the more usual long-term chronic ex- posures of workers. Cancer Patients Treatment for many cancers includes chemo- therapy with cytotoxic drugs, some of which are carcinogenic in laboratory animals and mutagenic in vitro, and treatment with high doses of radia- tion. There is a growing body of evidence that cancer patients are at a severalfold increased risk of developing second cancers, and some of these second cancers may be attributable to treatment of the first cancer with drugs and radiation (see, e.g., ref. 149). Cancer is mostly a disease of old age, but certain cancers have their peak incidence in younger people. Hodgkin’s disease, for in- stance, occurs with greatest frequency in young Ch. 8—Mutation Epidemiology - 109 men. Childhood leukemias, some brain cancers, and tumors with strong genetic components, e.g. Wilms’ tumor, retinoblastoma, and neuroblastoma, occur in the first few years of life. As treatment for these early cancers has improved over the last two decades, larger numbers of people are sur- viving, and it is these survivors who are at an in- creased risk of a second cancer, and possibly of heritable mutations. Results from four studies of the offspring of childhood cancer survivors, and nine studies of offspring of adult cancer patients have been pub- lished as of mid—1985. Several other studies are in progress (82). The combined published studies represent more than 700 cancer patients (both male and female) and more than 1,500 pregnancies, about 1,200 of which resulted in live births. Four percent of the liveborn babies had major birth defects, which is similar to the incidence in the general population. Two of the liveborn children had cancer. One had a hereditary bilateral retinoblastoma, as his fa- ther had. The other, the daughter of a brain can— cer survivor, had acute myelocytic leukemia. One child had a condition that could have been the result of a new mutation, the Marfan syndrome, which fits the definition of a sentinel phenotype. Several other children had defects that might have had genetic components, but none of these rep- resented sentinel phenotypes. The largest study of offspring of childhood can- cer survivors, including about 2,300 individuals from five population-based cancer registries, is nearing completion. Preliminary results indicate no increased risk of cancer in offspring compared with a control group, but the analysis is not yet final (82). Another long-term followup study, with more than 3,300 cases enrolled to date, is under way in the United Kingdom. No results are yet available from that study (82). The findings of a large international coopera- tive study of second tumors in children treated for cancers are provocative (142). Overall, 12 per- cent of children who survive at least 2 years af- ter a first cancer develop a second cancer some- time during the 25 years following the first cancer. Most of the patients in the study were treated with high—dose radiation therapy. The risk of second cancers was highest among children with cancers known to be strongly genetically influenced. In that group, there may well be a genetic defect that predisposes to mutations, e.g., a faulty repair mechanism, which could also be related to a higher risk of heritable mutations in that group. Cancer registries are the most numerous regis- tries of any type, and cohorts of treated patients and patients with second tumors are relatively easy to identify, compared with identifying other populations potentially exposed to mutagens. These groups should be considered when studies of heritable mutations using the new technologies become feasible. Other Populations A study of birth outcomes in people who had attempted suicide by self-poisoning in Hungary is an example of opportunistic use of available information (23). A cohort of about 1,300 indi- viduals who took large doses of drugs in suicide attempts has been studied since 1976. Early on, the investigators looked for short-term effects on somatic cells, using cytogenetic and biochemical testing. Long—term followup of birth outcomes ex- amined spontaneous abortions, ectopic pregnan- cies, stillbirths, low birthweight, and congenital anomalies. The study suffered a large loss of fol- lowup of study subjects, but in those evaluated, no important excesses in any of these endpoints were discovered. 110 - Technologies for Detecting Heritable Mutations in Human Beings CONCLUSIONS A very important question is answered by sim— ply observing birth outcomes in people thought to be at high risk, namely whether those individ- uals are at risk of having children with serious diseases and disabilities. The new mutation de— tection technologies discussed in this report may greatly increase the power to identify mutations in studies such as those described above, adding another dimension to knowledge about the rela— tionship between exposure to mutagens, the pres- ence of detectable mutations in DNA, and the ex- istence of observable health effects. Chapter 9 Mutagens: Regulatory Considerations Contents Page Introduction ............................................................ 113 Federal Involvement in Protecting Against Genetic Risk ...................... 113 Radiation Protection ................................................... 113 Agents Other Than Radiation .......................................... 114 Regulatory Issues ....................................................... 115 A Regulatory Definition of a Germ-Cell Mutagen ......................... 115 Quantitative Extrapolation ............................................. 117 Mutagenicity and Carcinogenicity ....................................... 117 New Methods for Measuring Mutation Rates and Their Potential Effects on Regulation ...................................................... 118 Chapter 9 Mutagens: Regulatory Considerations INTRODUCTION It is clear from the information gathered in this report that we are some years distant from being able to muster convincing direct evidence of any but very large increases in the rates of heritable mutations in human beings. The ability to pre- dict from experimental data which agents are likely to increase the mutation rate if human be- ings were exposed has not been put to the test. Even for the most likely mutagens, the ability to make quantitative extrapolations is relatively un- developed. Nonetheless, it is reasonable and pru— dent to accept that the environment may contain human germ-cell mutagens and that, to the ex- tent possible, human beings should be protected from them at levels that might cause heritable mu- tations. Ionizing radiation was recognized as a cause of heritable mutations in fruitflies in the 19205, and the Federal Government has since made efforts to protect workers and the public from excessive radiation exposure. Widespread concern about the mutagenic potential of chemicals is more recent, an issue brought into focus by the environmental movement that took shape in the late 19605. While the potential cancer-causing properties of man— made chemicals have been the driving force be— hind environmental health laws, two more recent laws specifically mention mutation as an endpoint against which the public should be protected. About a dozen other statutes include language broad enough to charge the Federal Government with the responsibility to protect against herita- ble mutations. Evidence from current methods for measuring mutation rates suggests that basing reg- ulation of environmental agents on carcinogenic— ity will likely assure protection against heritable mutations, but new, more sensitive detection tech- nologies, such as those discussed in this report, may necessitate a reexamination of that con- clusion. FEDERAL INVOLVEMENT IN PROTECTING AGAINST GENETIC RISK Radiation Protection In 1928, the newly created International X-Ray and Radium Protection Commission was charged by the Second International Congress on Radiol- ogy with developing recommendations for pro— tection against radiation (156). The following year, the Advisory Committee on X-Ray and Ra— dium Protection was formed to represent the U.S. viewpoint to the international commission. These two bodies were the forerunners of the current International Commission on Radiological Pro— tection (ICRP) and the National Council on Ra— diation Protection and Measurements (NCRP), the latter chartered by the U.S. Congress in 1964. The ICRP and NCRP have, since their first rec- ommendations in the 19305, based their accept- able radiation exposure limits on both heritable and somatic effects. The limits recommended have been lowered over the years, reflecting increased knowledge about radiation effects, and particu- larly about the effects on the population of low levels of radiation. Neither the ICRP nor NCRP recommendations have the force of law, but by and large, they have formed the basis for the radiation protection limits adopted by U.S. regulatory agencies. The first Federal entity officially charged with providing the agencies with guidance for developing radia- tion protection standards was the Federal Radia- tion Council (FRC), established in 1959. In 1960, FRC issued recommendations for both occupa- tional exposure and exposure of members of the public, which drew on ICRP and NCRP work (156). Over the years, the National Academy of Sciences Committee on the Biological Effects of Ionizing Radiation and the United Nations Sci- 113 114 - Technologies for Detecting Her/table Mutations in Human Beings entific Committee on the Effects of Atomic Radi- ation have also been influential in providing anal— yses that undergird exposure limits. The National Environmental Protection Act of 1970 transferred the responsibilities of FRC to the new Environmental Protection Agency (EPA). EPA administers several environmental health statutes under which that Agency is responsible for setting standards for radiation exposure in spe- cific conditions. Under its broader responsibili- ties, EPA has provided guidance for exposure from diagnostic X-rays, which is the regulatory responsibility of the Federal Food and Drug Administration (FDA), and for exposure of ura— nium miners, who are the responsibility of the Mine Safety and Health Administration (MSHA). Agencies other than EPA with responsibility for some aspects of radiation protection include: the Nuclear Regulatory Commission, the Department of Energy, the Department of Defense, FDA, MSHA, the Occupational Safety and Health Administration (OSHA), and the Department of Transportation. The States have responsibilities as well. Each entity, depending on its specific charge, is required to protect workers, the pub- lic, or both in accordance with the guidance pro- vided by EPA. The basic occupational and population ex- posure guidelines have not been revised since 1960. In 1981, EPA proposed new occupational guidelines (153), in line with 1977 ICRP recom- mendations (47), but these have not been made final, and they do not represent a change in total acceptable dose from the earlier guidelines. They do, however, place less of the emphasis on muta- genesis, and relatively more on somatic effects than do the 1960 guidelines. The ICRP includes in its risk estimates only genetic effects occurring in the first two generations after exposure. That probably accounts for roughly half of the total genetic effect, whatever its size. The current occupational exposure limit is 5 rem total body dose per year, with not more than 3 rem total body dose from occupational exposure in any one quarter of the year, and a more detailed breakdown for different groups of organs. The 1977 ICRP recommendations abandon the speci- fications by organ, and use a weighted whole- body dose. Today, almost all radiation exposures of US. workers are well below the regulated limits, though there are exceptions. The quantitative limits set by the ICRP and NCRP and adopted by Federal groups, are accompanied by the “ALARA principle”—that radiation exposures should be ”as low as reasonably achievable.” Exposure to the public are to be limited to be- low 25 millirem (mrem) to the whole body, 75 mrem to the thyroid, and 25 mrem to any other organ. These levels mainly affect the regulation of radionuclides in air and the disposal of radio- active waste. The numbers come from ICRP and NCRP recommendations, and are based on con- sideration of both genetic and somatic effects. Agents Other Than Radiation Congress formally recognized the need to pro- tect against chemical mutagens in the Toxic Sub- stances Control Act of 1976 (TSCA), and again in 1980 in the Comprehensive Environmental Re- sponse, Compensation, and Liability Act (CERCLA, or ”Superfund"). These two laws are administered by EPA, as are other statutes that include broad mandates to protect the public from environ- mental hazards. Other laws designed to protect citizens from external agents under which chem- ical mutagens could be regulated include the Fed- eral Food, Drug, and Cosmetics Act, administered by FDA; the Occupational Safety and Health Act, administered by OSHA; the Consumer Product Safety Act, administered by the Consumer Prod- uct Safety Commission; and the Atomic Energy Act, through which the Nuclear Regulatory Com- mission is empowered to protect certain workers from radiation hazards. Although it is almost certain that chemicals that might cause heritable mutations have been regu- lated, no regulations have been written or stand- ards set for these agents because of that property. In a few cases, the mutagenic potential of chemi— cals has been considered by regulatory agencies, but carcinogenic properties have driven standard- setting. Ch. 9—Mutagens: Regulatory Considerations 0 115 OSHA has included thorough reviews of mu- tagenicity data in notices of regulatory actions for two of the best-publicized chemical hazards of the 19805: ethylene oxide (EtO) and ethylene dibro- mide (EDB). Tests for heritable mutagenicity in Drosophila and experimental mammals have yielded positive results in at least some systems for both of these chemicals. While results of mutagenicity tests are included in the Federal Reg- ister notices for these chemicals, quantitative ex- trapolations for both the final EtO standard (152), and the proposed rulemaking for EDB (151), are based on protecting against carcinogenicity. REGULATORY ISSUES Currently there does not appear to be a scien- tific basis for the specifics of regulatory action against mutagens. The following questions face regulators and the scientific community involved in mutation research: 1. What is an appropriate regulatory definition of a probable human germ-cell mutagen and how does that definition relate to what is known about mutagens from epidemiologic studies and experimental studies? 2. Is it possible to derive quantitative estimates of the risk of heritable mutations in humans from experimental evidence in animals or from somatic—cell mutation tests in human beings? If it is not, what kinds of informa- tion are necessary before such extrapolation is possible? 3. How likely is it that a substance will require a more stringent standard as a mutagen than it will as a carcinogen or for other toxic effects? 4. How will information from the new technol- ogies for detecting heritable mutations that are described in this assessment change our perception of the kinds of ”adverse effects” against which regulation should be directed? The discussion in the remainder of this chap- ter addresses these questions. EPA has recognized germ-cell mutagenicity as a class of adverse effects, particularly in its respon- sibilities under the Federal Insecticide, Fungicide, and Rodenticide Act, under which it, in addition to OSHA, has acted to regulate exposures to EtO and EDB. Unique among the regulatory agencies, EPA's Reproductive Effects Assessment Group (in the Office of Research and Development) has pre— pared guidelines for mutagenicity testing, which are described later in this chapter. A Regulatory Definition of a Germ-Cell Mutagen Strategies for regulating mutagens to protect public health cannot today rely on data from cur- rent studies of heritable mutations in human be- ings. Just as is the case in regulating carcinogens, a regulatory definition must serve as a substitute, particularly for making judgments about the po- tential risks of new substances. Certain lessons can be learned from experience in regulating carcinogens. (For a discussion of the issues surrounding carcinogen regulation, see 145.) The most convincing evidence for carcino- genicity, from well—conducted epidemiologic stud- ies, is that human beings have developed cancer after exposure to a given agent. If an increase in genetic disease could be convincingly shown to be related to a specific agent, there would certainly be no problem in acting against that agent. The spirit of the regulatory laws, however, embody the concept of taking protective action before peo- ple are harmed. In the absence of direct evidence from human beings, regulators must rely on in- direct evidence from a variety of experimental test systems which will never be absquter predictive of effects in human beings. A regulatory defini- tion of a mutagen will be pragmatic and rely on 116 - Technologies for Detecting Heritable Mutations in Human Beings information that it is possible to collect, and on a number of untested assumptions. EPA is the first US regulatory agency to have proposed ”Guidelines for Mutagenicity Risk As- sessment" (154). The guidelines require evidence of: 1) mutagenic activity and 2) chemical inter— actions in the mammalian gonad. The decision to regulate is to be based on a “weight-of—evidence" determination. EPA has proposed no formal method for quantitative extrapolation by which acceptable exposure levels could be set. According to the EPA guidelines, evidence for mutagenic activity may come from tests that de- tect point mutations and structural or numerical chromosome aberrations. Structural aberrations include deficiencies, duplications, inversions, and translocations. In the absence of evidence of heritable mutations in human beings, evidence from a variety of experimental test systems may be invoked. For mutagens that cause point mu- tations, whole animals tests (e.g., the mouse specific-locus test) provide the highest degree of evidence, but these tests are relatively more ex- pensive than short-term tests, and there is a limited capacity for laboratories to perform them. Other tests for point mutations include those in bacteria, eukaryotic micro-organisms, higher plants, insects, and mammalian somatic cells. Structural chromosome aberrations can be de— tected either in somatic or germ cells in different assays. The organisms used include higher plants, insects, fish, birds, and several species of mam— mals. Mutagens that cause numerical changes in chromosomes may be missed by the tests that directly measure DNA damage. Tests specifically directed at detecting changes in chromosome num— ber are not as well developed as are those for point mutations or structural changes in chromosomes. Tests are in various stages of development in fungi, Drosophila, mammalian cells in culture, and intact mammals, including mammalian germ- cells tests. Results from tests that measure endpoints other than mutagenicity directly may also be used in judging the potential mutagenicity of a substance. DNA damage, unscheduled DNA synthesis in mammalian somatic and germ cells, mitotic re- combination and gene conversion in yeast, and sister chromatid exchange in mammalian somatic and germ cells are cited by EPA as tests that pro- vide evidence known to be correlated with mu- tagenicity, though they measure other genetic events. Evidence from various kinds of mutagenicity tests is weighted with regard to the relationship of the test to human germ-cell mutation. Greater weight will be given to results from tests in: 1) germ cells over somatic cells, 2) mammalian cells over submammalian cells, and 3) eukaryotic cells over prokaryotic cells. EPA lists two classes of evidence for chemical interactions in the mammalian gonad: sufficient and suggestive. Sufficient evidence is from studies in whole mammals that demonstrate, for exam- ple, unscheduled DNA synthesis, sister chromatid exchange, or chromosomal aberrations in germ cells. Adverse effects on the gonads or on repro— ductive outcomes after exposure, which are con- sistent with the substance reaching the gonads but which do not indicate direct interaction with DNA, are considered as providing suggestive evidence. The final step in EPA’s mutagenicity risk assess— ment is the weight-of—evidence determination, which classifies the evidence for potential germ- cell mutagenicity as ”sufficient,” ”suggestive," or “limited." In this step, results of tests plus any in- formation about effects in human beings is evalu— ated. Sufficient evidence consists of a positive mammalian germ-cell test. In addition, positive responses in at least two different test systems, at least one of which is in mammalian cells, and evidence of germ-cell interaction, together con— stitute sufficient evidence. Evidence of lesser quan- tity and/ or quality of both mutagenic response and germ-cell activity constitute suggestive evi- dence. Limited evidence consists of positive re- sults in either mutagenicity assays or tests for chemical interactions in the gonad, but not both. EPA’s guidelines became final in September 1985. Currently and for the foreseeable future, the greatest value of EPA's guidelines is the recogni— tion of germ-cell mutagenicity as a legitimate end— point to consider in assessing the potential adverse effects of substances in the environment. Ch. 9—Mutagens: Regulatory Considerations 0 117 Quantitative Extrapolation If the levels of risk from suspected human germ- cell mutagens is to be estimated in the absence of direct evidence of harm in human beings, data from experimental systems must be used in a ”quantitative extrapolation.” The experimental systems are basically those mentioned in the EPA guidelines discussed in the previous section, and those discussed elsewhere in this report. The three categories of tests are: 1) whole animal heritable mutation studies; 2) animal somatic-cell mutation studies, either in vivo or in vitro; and 3) human somatic—cell mutation studies, either in vivo or in vitro. Unfortunately, the kind of information (i.e., measures of human mutations) that would link results from these three categories of tests to hu- man heritable mutations is scanty. It is encourag— ing, however, that using tests available now, such information can be generated at least for some substances. For EtO and EDB, for instance, mu— tagenicity data are available in both somatic and germ-cell systems in animals, and some somatic cell (cytogenetic) data are available from human beings exposed at known occupational levels. Obtaining more information to fill in the ar- rows of the extrapolation ”parallelograms” pre— sented in chapter 7 of this report should be a high priority for regulators. In fact, EPA's Reproduc- tive Effects Assessment Group has collaborated with other groups in the Federal Government to fund such studies (168). Without the kind of in- formation that would come from coordinated studies in several test systems, there is little chance of writing a successful regulation that limits ex- posure to a specific level (short of a complete ban for an agent acknowledged to be unacceptably risky at any level). Given the experience with car- cinogens, a regulation that states an exposure level without adequate experimental evidence and the- ory behind it will not survive a court challenge, which, in the United States today, appears to be the final test of a regulation. Even with good experimental data, some of the same problems that plague extrapolating from ani- mals to humans to determine acceptable exposure levels for carcinogens are certain to hinder quan— titative extrapolation for estimating levels of mutagenic risk at specific levels of exposure. In carcinogen extrapolation, there still are unresolved controversies about the appropriate conversion factors between species and about the shape of dose-response curves. The latter is important be- cause most animal bioassays test extremely high doses in relation to the animals' body weights, while humans are generally exposed at lower levels over longer periods of time. Though the de- tails of extrapolation for mutagenicity differ from those for carcinogenicity, the problems will un- doubtedly be similar. Right now, there is not enough information about the relationships be- tween results in various test systems to address intelligently the practical problems of actually per- forming extrapolations. Given the appropriate information, it may be— come possible to carry out quantitative extrapo- lations for mutagenicity. When the time comes, the regulatory agencies will need to require that the appropriate tests are done, either by manu- facturers or by the Federal Government. The test- ing requirement may take various forms, which may vary by statute. Mutagenicity and Carcinogenicity As mentioned previously, while the regulatory apparatus exists for acting against human germ- cell mutagens, in fact no regulations based on germ-cell mutagenesis exist except for radiation exposure limits. There are several reasons for this. First, there are no proven human germ-cell muta- gens, and only a limited number of presumptive human mutagens known from animal tests. Sec- ond, it has often been thought that regulations based on demonstrated carcinogenicity would automatically protect against mutagenicity as well. In fact, however, this may not always be true. Voytek (167) reported a preliminary assess- ment indicating that the risk of heritable genetic disease in the first generation after exposure to EDB was greater than the lifetime risk of cancer in the exposed individuals, based on an extrapo- lation from animal data. It is widely held that a somatic mutation is a necessary step in the development of cancer. Many substances that are mutagenic in bacteria, in cells in culture, and in Drosophila also are car- cinogenic in laboratory rats, mice, or both. Short- 118 - Technologies for Detecting Heritable Mutations in Human Beings term tests based on mutagenicity in these lower organisms are therefore used as screens for car- cinogenicity. The most widely used screening test for carcinogenicity is the Ames test, which meas- ures mutagenicity in strains of the bacterium Salmonella. Under some statutes (e.g., TSCA), negative results in short-term tests, meaning that the substance is not mutagenic in those systems, can obviate the need for a long-term bioassay, a savings of up to $1 million to a manufacturer (145). Positive results in short-term tests, mean- ing that the substance is mutagenic in these sys— tems, have not been accepted as a basis for regu- lation under any statute, but they have probably halted the development of new chemicals. Man- ufacturers know that positive mutagenicity tests may trigger the requirement for a long-term bio- assay, which in a high proportion of cases will turn out positive. The product, whatever it is, might never get to market. Rather than risk a fi- nancial loss, many manufacturers simply will not proceed with that product. New Methods for Measuring Mutation Rates and Their Potential Effects on Regulation The new techniques for detecting heritable mu- tations and the somatic-cell techniques that even- tually may be used to predict germ-cell mutagen- esis will lead to the consideration of effects that are increasingly removed from measurable or even hypothetical adverse health effects in humans. Mutagenicity endpoints may be detected that could be more sensitive than those currently used to predict carcinogenicity. In the regulatory con- text, judgments will have to be made, most likely in the absence of certain knowledge of effects, about the appropriate actions to be triggered by demonstrations of various kinds of changes in DNA, detectable by various analytic methods, that can be convincingly linked to specific ex- posures. From a public health standpoint, it is most appropriate to act under the assumption that mutations of any kind are deleterious, and that environmental agents at levels that cause any reliably detected changes in the DNA should be subject to available regulatory controls. This does not get around the problem of defining “safe" or ”acceptable” exposure levels, however. Animal ex- periments will be needed to explore the quantita— tive relation between subtle changes detectable anywhere in DNA and the levels of adverse ef- fects that might be observed in the animal. At present, testing for safety is a significant part of the research and development investment in new products, whether they are chemicals, drugs, or food additives; determining the risks of sub— stances already in the environment is a significant task for the Federal Government. If they become available, new mutagenicity testing technologies using experimental animals could either impose significant new testing requirements in addition to those already in place, or could replace some expensive and not entirely reliable tests for car- cinogens. Regardless of the methods used to de- tect mutations, the relation between specific types of changes in DNA and health effects will have to be studied experimentally to shed light on the meaning of a ”positive result.” Until this knowl- edge is available, the ability to detect an effect without knowing the likelihood of any health con— sequences for human beings will remain a thorny public policy question which scientists, regulators, and politicians must address. Appendixes Appendix A Federal Spending for Mutation Research A quick review of the origins of biotechnology il- lustrates the difficulty of identifying the specific areas of research that would further our knowledge of hu- man mutations, and, in turn, the problems of estimat- ing the current amount of Federal expenditures for such research. In 1968 and 1969, Hamilton 0. Smith, a microbiol— ogist at Johns Hopkins University, had no thought of studying human heritable mutations. No venture cap- ital firm was looking for biotechnology companies to invest in, and had it been, it would have been dis- appointed. No such company existed. Smith was studying how some bacteria can take up molecules of DNA from their growth medium and recombine it into their genetic material. He knew that the bacteria would not take up DNA from other spe- cies of bacteria or from viruses, and he added some viral DNA to the experiment expecting that it would remain inert and intact while the bacterial DNA un— derwent recombination. Instead, the viral DNA was rapidly degraded. Looking at that result, Smith hy- pothesized that a bacterial enzyme(s) could discrimi- nate between foreign DNA (from the virus), which it degraded and self DNA (from the same kind of bacte— ria), which it left alone. He purified the enzyme and showed that it recognized particular sequences of nucleotides on the viral DNA and cut it at those sites (128). These experiments, which identified and character- ized the first known restriction enzyme, opened the door to the discovery of hundreds of others. The use of restriction enzymes, allowing precise cutting of DNA and precise joining of DNA pieces from differ- ent parts of an organism's genome as well as joining of pieces from different organisms, led to fundamen- tal advances in the basic approach of research in molecular biology and genetics (146,147). Accord- ingly, these enzymes are basic ingredients of the new technologies for studying human heritable mutations described in this report. However, someone in 1969 who had tried to iden- tify research projects that would contribute to our un- derstanding of human mutation rates would almost certainly not have included Smith’s. His was basic re- search directed at understanding the mechanism of recombination in an organism far removed from hu- mans. This OTA report provides additional examples of the difficulty of estimating how much money is be- ing spent on research that may be important to un- derstanding human mutations. Current Federal Expenditures for Mutation Research OTA queried Federal research and regulatory agen- cies about their support of research directed at under- standing human mutations. Questions were asked about the amount of money spent specifically on hu- man mutation research and the amount spent on ”re- lated” areas of biological research. The first category was tightly defined—the research had to focus on de- velopment or applications of methods for detecting and/ or counting human somatic or heritable mutations in Vivo. The category of “related” research was much broader, including studies examining genetic contri- butions to such common diseases as diabetes and ar- thritis, studies of mutagenesis and repair mechanisms, development of instruments for measuring cellular ef- fects of various kinds, and tests for mutagenic activ- ity in lower organisms as well as in cultured human cells. The expenditures shown in table A-1 must be treated with some reservation about their accuracy, the total of $14.3 million estimated as the amount spent on hu- man mutation research being more precise than the ap- proximately $207 million estimated for related genetic research. The $14.3 million may be an overestimate— Table A-1.—Federal Expenditures for Research Related to Human Mutations Human Related . mutation genetic Federal agency research research Department of Energy ............. $ 6,220,0003 $ 30,123,000 Department of Health and Human Services: Centers for Disease Control ..... 277,000 1,387,000 National Institutes of Health ..... 7,244,000b 156959.000" Food and Drug Administration . . . 1,200,000 Environmental Protection Agency . . . . 606,000 1,189,000 National Science Foundation ........ 0 16,336,000 Total ......................... 14,347,000 207,194,000 3This includes $1 ,850.000 (or halt of the total $3,700,000 at current exchange rates) spent annu- ally for the U.S. contribution to the Radiation Effects Research Foundation, the group that coor- dinates ongoing genetic studies of the survivors of the Nagasaki and Hiroshima bombings and their offspring. bThis total was obtained by inspecting grant titles and reading grant abstracts to identify those that focused on human mutation research. cThis figure is the budget for the Genetics Program in the National Institute of General Medical Scrences. It does not include all genetics research at NIH. some of which goes on in other institutes. SOURCE; Office of Technology Assessment. 121 122 - Technologies for Detecting Heritab/e Mutations in Human Beings some projects studying clinical effects of mutations may provide no information about detection—and the $207 million for related genetic research is probably an underestimate, since much genetic research is sup- ported by parts of the National Institutes of Health (NIH) not queried. For comparison, the total NIH budget for the fiscal year 1985 was about $4 billion. Most of the NIH funding for mutation research is di- rected at identifying mutations in at—risk families (where specific tests for specific mutations are appro- priate), whereas DOE supports a search for more gen- eral methods to detect mutations. Successes of the research efforts in human mutation research and related genetic research will probably generate a requirement for additional funding in the near future. It will be a costly venture to take any new technique (such as those described in ch. 4) and apply it to studies of populations sufficiently large to pro- vide a good possibility of yielding information about human mutations. For instance, the ongoing monitor- ing of the Nagasaki and Hiroshima populations using clinical examinations and protein analyses (see ch. 3) costs about $3.7 million annually to the US and Jap- anese governments together (48). Appendix B Acknowledgments and Health Program Advisory Committee HEALTH PROGRAM ADVISORY COMMITTEE Sidney S. Lee, Chair President, Milbank Memorial Fund H. David Banta Project Director STG Project on Future Health Technology The Netherlands Rashi Fein Professor Department of Social Medicine and Health Policy Harvard Medical School Harvey Fineberg Dean School of Public Health Harvard University Patricia King Professor Georgetown Law Center Joyce C. Lashof Dean School of Public Health University of California, Berkeley Alexander Leaf Professor of Medicine Harvard Medical School Massachusetts General Hospital Frederick Mosteller Professor and Chair Department of Health Policy and Management School of Public Health Harvard University Norton Nelson Professor Department of Environmental Medicine New York University Medical School Robert Oseasohn Associate Dean School of Public Health University of Texas Nora Piore Senior Fellow and Advisor to the President United Hospital Fund of New York Dorothy Rice Regents Lecturer Department of Social and Behavioral Sciences School of Nursing University of California, San Francisco Richard Riegelman Associate Professor George Washington University School of Medicine Walter Robb Vice President 8: General Manager Medical Systems Operations General Electric Milwaukee, WI Frederick C. Robbins University Professor Department of Epidemiology and Biostatistics School of Medicine Case Western Reserve University Cleveland, OH Frank E. Samuel, Jr. President Health Industry Manufacturers’ Association Rosemary Stevens Professor Department of History and Sociology of Science University of Pennsylvania 123 124 - Technologies for Detecting Heritable Mutations in Human Beings ACKNOWLEDGMENTS OTA would like to thank the members of the advisory panel and the contractors who provided information for this assessment. In addition, OTA acknowledges the following individuals for their assistance in reviewing drafts or supplying information. F. Alan Andersen U.S. Food and Drug Administration Benjamin J. Barnhart U.S. Department of Energy Fred Bergmann National Institute of General Medical Sciences David Botstein Massachusetts Institute of Technology No'él Boucl< Northwestern University School of Medicine Michele Calos Stanford University School of Medicine L.L. Cavalli—Sforza Stanford University School of Medicine George Church University of California at San Francisco Mary Ann Danello U.S. Food and Drug Administration Amiram Daniel Health Industry Manufacturers Association Frederick J. DeSerres National Institute of Environmental Health Sciences Robert Dixon U.S. Environmental Protection Agency William Fitzsimmons National Institute of General Medical Sciences W. Gary Flamm U.S. Food and Drug Administration Joseph Fraumeni National Cancer Institute Helen Gee National Institutes of Health Margaret Gordon National Institutes of Health Ronald Hart National Center for Toxicological Research John Heddle Ludwig Institute for Cancer Research (Canada) Richard Hill U.S. Environmental Protection Agency Abraham Hsie Oak Ridge National Laboratory Seymour Jablon National Research Council Irene Jones Lawrence Livermore National Laboratory Moin Khoury Johns Hopkins University Norman Mansfield National Institutes of Health Karen Messing University of Quebec at Montreal Alexander Morley Flinders Medical Center (Australia) John J. Mulvihill National Cancer Institute DeLill Nasser National Science Foundation Neal 5. Nelson U.S. Environmental Protection Agency Per Oftedal University of Oslo Doug Ohlendorf E.I. du Pont de Nemours 8: Co. Dilys Parry National Cancer Institute Claude Pickelsimer Centers for Disease Control William L. Russell Oak Ridge National Laboratory F. Raymond Salemme E.I. du Pont de Nemours & Co. Michael M. Simpson Congressional Research Service David A. Smith U.S. Department of Energy App. B—Acknowledgments and Health Program Advisory Committee - 125 Gail Stetten Johns Hopkins Hospital J. Timothy Stout Baylor College of Medicine Gary Strauss US. Environmental Protection Agency Larry R. Valcovic US. Environmental Protection Agency Peter E. Voytek U.S. Environmental Protection Agency References References 10. . Albertini, R.]., “Studies With T-Lymphocytes: An Approach to Human Mutagenicity Monitor- ing,” Indicators of Genotoxic Exposure, Banbury Report 13, B.A. Bridges, B.E. Butterworth, I.B. Weinstein (eds.) (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1982), pp. 393-410. . Albertini, R.J., “Somatic Gene Mutations In Vivo as Indicated by the 6-Thioguanine Resis- tant T-Lymphocytes in Human Blood,” Mutat. Res. 150:411-422, 1985. . Albertini, R.)., Castle, KL, and Borcherding, W.R., “T-Cell Cloning To Detect the Mutant 6- Thioguanine-Resistant Lymphocytes Present in Human Peripheral Blood,” Proc. Nat. Acad. Sci. (USA) 79:6617-6621, 1982. . Albertini, R.]., O’Neill, ].P., Nicklas, ].A., et al., ”Alterations of the hprt Gene in Human 6- Thioguanine Resistant T-Lymphocytes Arising In Vivo,” Nature 316:369-371, 1985. . Altland, K., Kaempfer, M., Forssbohm, M., et al., ”Monitoring for Changing Mutation Rates Using Blood Samples Submitted for PKU Screen- ing,” Human Genetics, Part A: The Unfolding Genome (New York: Alan R. Liss, Inc., 1982), pp. 277-287. . Awa, A.A., “Cytogenetic Study," J. Radiat. Res. (Supp1.) 16:75—81, 1975. Awa, A.A., Sofuni, T., Honda, T., et al., "Pre- liminary Reanalysis of Radiation-Induced Chro— mosome Aberrations in Relation to Past and Newly Revised Dose Estimates for Hiroshima and Nagasaki A—Bomb Survivors,” Radiation Ef- fects Research Foundation Technical Report Ser- ies 8—83, 1983. . Awa, A.A., Honda, 8., and Neriishi, et al., “An Interim Report of the Cytogenetic Study of the Offspring of Atomic Bomb Survivors in Hiro- shima and Nagasaki,” Human Genetics, Part A: The UnfoIding Genome (New York: Alan R. Liss, Inc., 1982). . Bigbee, W.L., Langlois, R.G., Vanderlaan, M., et al., ”Binding Specificities of Eight Monoclonal Antibodies to Human Glycophorin A—Studies With MCM and MkEn(UK) Variant Human Erythrocytes and M- and MNv-type Chimpan- zee Erythrocytes,” J. Immunol. 133:3149-3155, 1984. Bigbee, W.L., Vanderlaan, M., Fong, S.S.N., et al., “Monoclonal Antibodies Specific for the M- and N-Forms of Human Glycophorin A,” MoIec. Immunol. 20:1353-1362, 1983. 11. 12. 13. 14. 15, 16. 17. 18. 19. 19a. 20. 21. 22. Boice, ].D., Ir., Day, NE, and Andersen, A., et al., “Second Cancers Following Radiation Treatment for Cervical Cancer/'1. Nat. Cancer Inst. 74(5):955-975, 1985. Boué, A., Boué, J., and Gropp, A., ”Cytogenet- ics of Pregnancy Wastage,” Adv. Hum. Genet. 14:1-57, 1985. Branscomb, E., ”Detecting Genetic Injury in Body Cells: The Red Blood Cell as Sentinel," typescript, contract report prepared for the Of— fice of Technology Assessment, U.S. Congress, Washington, DC, 1985. Bridges, B.A., “An Approach to the Assessment of the Risk to Man From DNA Damaging Agents," Arch. Toxico]. Suppl.3:271-281, 1980. Bridges, B.A., ”Status and Prospects, in Indica- tors of Genotoxic Exposure," Indicators of Geno— toxic Exposure, Banbury Report 13, B.A. Bridges, B.E. Butterworth, and LB. Weinstein (eds.) (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1982), pp. 555-558. Cariello, N., ”Methods To Characterize Muta- tions in Human T-Lymphocytes,” typescript, contract report prepared for the Office of Tech- nology Assessment, U.S. Congress, Washington, DC, 1985. Cavalli—Sforza, LL, and Bodmer, W.F., The Genetics of Human PopuIations (San Francisco: W.H. Freeman 8: Co., 1971). Chandley, A.D., “The Origin of Aneuploidy," Human Genetics, Part B: MedicaI Aspects (New York: Alan R. Liss, Inc., 1982), pp. 337-347. Charles, D., and Pretsch, W., “A Mutation Affecting the Lactate Dehydrogenase Locus LDH—l in the Mouse. 1: Genetical and Elec- trophoretic Characterization,” Biochem. Genet. 19:301-307, 1981. Charles, D., and Pretsch, W., “Linear Dose- Response Relationship of Erythrocyte Enzyme Activity Mutations in Offspring of Ethylnitro- sourea-Treated Mice," Mutat. Res., in press. Church, G.M., and Gilbert, W., “Genomic Se— quencing,” Proc. Nat. Acad. Sci. (USA) 81:1991— 1995, 1984. Costa, T., Scriver CR, and Childs, B., ”The Ef- fect of Mendelian Disease on Human Health: A Measurement,” Amer. J. Med. Genet. 21:231- 242, 1985. Crow, ].F., and Denniston, C., “Mutation in Hu- man Populations,"AdV. Hum. Genet. 1459—123, 1985. 129 130 - Technologies for Detecting Heritable Mutations in Human Beings 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. Czeizel, A., “Epidemiological Follow-Up Study on Mutagenic Effects in Self-Poisoning Persons,” Environmental Mutagens and Carcinogens, T. Sugimura, S. Kondo, and H. Takebe (eds.) (New York and Tokyo: Alan R. Liss, Inc., and Univer- sity of Tokyo Press, 1982). Darby, S., Nakashima, E., and Kato, H., “A Parallel Analysis of Cancer Mortality Among Atomic Bomb Survivors and Patients With Ankylosing Spondylitis Given X-Ray Therapy," J. Nat. Cancer Inst. 75(1):1-19, 1985. Delehanty, J., White, R.L., and Mendelsohn, M.L., “Approaches To Determining Mutation Rates in Human DNA," Mutat. Res. 167:215- 232, 1986. Dempsey, J.L., Seshadri, RS, and Morley, A.A., ”Incresased Mutation Frequency Follow- ing Treatment With Cancer Chemotherapy," Cancer Res. 45:2873-2877, 1985. Dobson, R.L., and Felton, ].S., “Female Germ Cell Loss From Radiation and Chemical Ex- posures,” Am. J. Industrial Med. 4:175-190, 1983. Dobson, R.L., Kwan, T.C. and Straume, T., “Tritium Effects on Germ Cells and Fertility,” European Seminar on the Risks From Tritium Ex- posure, G. Gerber and M. Myttenaere (eds.) (Brussels: Commission of the European Commu- nities, 1984). Dreyer, N.A., Kohn, H.I., Clapp, R., et al., The Feasibility of Epidemiologic Investigations of the Health Effects of Low-Level Ionizing Radiation, NUREG/CR-1728 (Washington, DC: US. Gov- ernment Printing Office, 1980). Eber, S.W., Dunnwald, M., and Heinemann, G., et al., ”Prevalence of Partial Deficiency of Red Cell Triosephosphate Isomerase in Germany-A Study of 3,000 People,” Hum. Genet. 67:336- 339, 1984. Ehling, U.H., “Methods To Estimate the Genetic Risk,” Mutations in Man, Gunter Obe (ed.) (Ber- lin: Springer Verlag, 1984). Emery, A.E.H., and Rimoin, D., ”Nature and Incidence of Genetic Disease,” Principles and Practice of Medical Genetics, A.E.H. Emery and D. Rimoin (eds.) (Edinburgh: Churchill Living- stone, 1983). Emery, A.E.H., An Introduction to Recombinant DNA (Chichester: John Wiley, 1984). Fischer, S.G., and Lerman L.S., ”DNA Frag- ments Differing by Single Base-Pair Substitutions are Separated in Denaturing Gradient Gels: Cor- respondence With Melting Theory,” Proc. Nat. Acad. Sci. (USA): 80: 1579—83, 1983. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. Generoso, W.M., Cain, K.T., Cacheiro, N.L.A., et al., ”Response of Mouse Spermatogonial Stem Cells to X-Ray Induction of Heritable Recipro- cal Translocations," Mutat. Res. 126:177-187, 1984. Gunther, M., and Penrose, L.S., ”The Genetics of Epiloia," ]. Genet. 31:413-430, 1935. Harris, H., The Principles of Human Biochemi- cal Genetics, 4th ed. (Amsterdam: North- Holland, 1980). Harris, H., Hopkinson, D.A., and Robson, E.B., “The Incidence of Rare Alleles Determining Elec- trophoretic Variants: Data on 43 Enzyme Loci in Man,” Ann. Hum. Genet. 37:237-253, 1974. Hassold, T.I., and Jacobs, P.A., “Trisomy in Man,” Ann. Rev. Genet. 18:69-97, 1984. Hayes, A., Costa, T., Scriver, C.R., et al., “The Effect of Mendelian Disease on Human Health, 11: Response to Treatment,” Amer. J. Med. Ge- net. 21:243-255, 1985. Hitotsumachi, S., Carpenter, D.A., and Russell, W.L., “Dose-Repetition Increases the Mutagenic Effectiveness of N-ethyl-N-nitrosourea in Mouse Spermatogonia," Proc. Nat. Acad. Sci. (USA) 82:6619-6621, 1985. 7 Hook, E.B., ”Prevalence of Chromosome Ab- normalities During Human Gestation and Impli— cations for Studies of Environmental Mutagens," Lancet 2: 169-172, 1981. Hook, E.B., “Contribution of Chromosome Ab- normalities to Human Morbidity and Mortality and Some Comments Upon Surveillance of Chromosome Mutation Rates,” Prog. Mat. Res. 3:9-38, 1982. Hook, E.B., and Cross, P.K., "Surveillance of Human Populations for Germinal Cytogenetic Mutations,” Environmental Mutagens and Car- cinogens, T. Sugimura, S. Kondo, and H. Takebe (eds.) (New York and Tokyo: Alan R. Liss, Inc., and University of Tokyo Press, 1982). Hook, E.B., Schreinemachers, D.M., and Wil- ley, A.M., et al., “Rates of Mutant Structural Chromosome Rearrangements in Human Fe- tuses: Data From Prenatal Cytogenetic Studies and Associations With Maternal Age and Paren- tal Mutagen Exposure,” Amer. ]. Hum. Genet. 35296-109, 1983. International Commssion for Protection Against Environmental Mutagens and Carcinogens, Committee 4, Final Report, “Estimation of Genetic Risks and Increased Incidence of Genetic Disease Due to Environmental Mutagens," Mutat. Res. 115:225—291, 1983. International Commission on Radiological Pro— References 0 131 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. tection, Recommendations of the International Commission on Radiological Protection, ICRP Publication 26 (Oxford: Pergamon Press, 1977). Jablon, 5., National Research Council, personal communication, April 1986. Jackson, E.R., Fowle, J.R., III, and Voytek, P., “Some Research Needs To Support Mutagenic- ity Risk Assessments From Whole Mammal Studies,” Utilization of Mammalian Specific Lo- cus Studies in Hazard Evaluation and Estimation of Genetic Risk, F.J. deSerres and W. Sheridan (eds.) (New York: Plenum Press, 1983), pp. 315-322. Jensen, R.H., Bigbee, W., and Branscomb, E.W., ”Somatic Mutations Detected by Immunofluo- rescence and Flow Cytometry,” Biological Dosim- etry Cytometric Approaches to Mammalian Sys- tems, W.G. Eisert and ML. Mendelsohn (eds.) (Berlin: Springer-Verlag, 1984), pp. 161-170. Johnson, F.M., and Lewis, S.E., “Electrophoreti— cally Detected Germinal Mutations Induced in the Mouse by Ethylnitrosourea,” Proc. Nat. Acad. Sci. (USA) 78:3138-3141, 1981. Johnson, F.M., and Lewis, S.E., ”The Detection of ENU-Induced Mutants in Mice by Electropho- resis and the Problem of Evaluating the Muta- tion Rate Increase," Utilization of Mammalian Specific Locus Studies in Hazard Evaluation and Estimation of Genetic Risk, F.J. deSerres and W. Sheridan (eds.) (New York: Plenum Press, 1983), pp. 95-123. Johnson, F.M., Roberts, G.T., Sharma, R.K., et al., ”The Detection of Mutants in Mice by Elec- trophoresis: Results of a Model Induction Exper- iment With Procarbazine," Genetics 97:113-124, 1981. Jones, l.M., Burkhart—Schultz, K., and Carrano, A.V., “A Study of the Frequency of Sister Chro- matid Exchange and of Thioguanine Resistant Cells in Mouse Spleen Lymphocytes After in vivo Exposure to Ethylnitrosourea,” Mutat. Res. 143:245-249, 1985. Kato, H., and Schull W.J., “Cancer Mortality Among Atomic Bomb Survivors,” Life Span Study Report 9, Part 1, Radiation Effects Re- search Foundation Technical Report 12—80, 1982. Kato, H., Schull, W.J., and Neel, J.V., “Survival in Children of Parents Exposed to the Atomic Bomb, a Cohort-Type Study,” Amer. J. Hum. Genet. 18:339, 1966. Lerman, L.S., ”A Proposal for the Detection of Mutations in a Large-Scale Sampling of the Hu- man Genome Using Two-Dimensional Denatur- ing Gradient Separations," typescript, contract 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. report prepared for the Office of Technology Assessment, US. Congress, Washington, D.C., 1985. Livingston, G.K., ”An Overview of Chromo- somal, Micronucleus, Heritable Effects, Domi— nant Lethal, Heritable Translocation, and Spe- cific Locus Tests," Reproduction: The New Frontier in Occupational and Environmental Health Research, J.E. Lockey, et al. (eds.) (New York: Alan R. Liss, Inc., 1984), pp. 417-427. Lubs, H.A., and Ing, P.S., "Human Cytogenetic Nomenclature," Principles and Practice of Med- ical Genetics, A.E.H. Emery and D. Rimoin (eds.) (Edinburgh: Churchill Livingstone, 1983), p. 162. Lyon, M.F., “Sensitivity of Various Germ Cell Stages to Environmental Mutagens,” Mutat. Res. 872323-345, 1981. Lyon, M.F., ”Problems in Extrapolation of Ani- mal Data to Humans,” Utilization of Mammalian Specific Locus Studies in Hazard Evaluation and Estimation of Genetic Risk, F.J. deSerres and W. Sheridan (eds.) (New York: Plenum Press, 1983), pp. 55-69. Lyon, M.F., ”Measuring Mutation in Man," Na- ture 318: 315-316, 1985. Malling, H.V., and Valcovic, L.R., ”Biochemi- cal Specific Locus Mutation System in Mice,” Arch. Toxicol. 38:45-51, 1977. Malling, H.V., ”Perspectives in Mutagenesis," Environ. Mutag. 32103-108, 1981. Matsunaga, E., ”Perspectives in Mutation Epi- demiology. 1: Incidence and Prevalence of Genetic Disease (Excluding Chromosomal Aber- rations) in Human Populations," Mutat. Res. 99:95-128, 1982. Matsunaga, E., ”Perspectives in Mutation Epi- demiology. 5: Modern Medical Practice Versus Environmental Mutagens: Their Possible Dys- genic Impact,” Mutat. Res. 11:449-457, 1983. Maxam, A.M., and Gilbert, W., ”A New Method for Sequencing DNA,” Proc. Nat. Acad. Sci. (USA) 74:560-564, 1977. Mendelsohn, M., ”Prospects for Cellular Muta- tional Assays in Human Populations," Assess- ment of Risk From Low-Level Exposure to Ra- diation and Chemicals, A.D. Woodhead, et al. (eds.) (New York: Plenum Press, 1985), pp. 415-427. Mendelsohn, M., Lawrence Livermore National Laboratory, personal communication, May 6, 1986. Messing, K., and Bradley, W.E.C., ”In Vivo Mu— tant Frequency Rises Amoung Breast Cancer Pa- 132 - Technologies for Detecting Heritable Mutations in Human Beings 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. tients After Exposure to High Doses of Gamma Radiation," Mutat. Res. 152:107-112, 1985. Miller, I.R., ”Perspectives in Mutation Epidemi- ology. 4: General Principles and Considerations,” Mutat. Res. 114:425-447, 1983. Miller, 0.]., "Chromosomal Basis of Inheri— tance,” Principles and Practice of Medical Ge- netics, A.E.H. Emery and D. Rimoin (eds.) (Edinburgh: Churchill Livingstone, 1983), pp. 49-64. Mohrenweiser, H.W., “Feasibility Study: Detec- tion of Chemically Induced Mutation by Assay of Metabolic Characteristics,” Substitute Chem- ical Program. 11: Toxicological Methods and Genetics Effects Workshop (Washington, DC: US. Environmental Protection Agency, 1975). Mohrenweiser, H.W., "Biochemical Approaches To Monitoring Human Populations for Germi- nal Mutation Rates. II: Enzyme Deficiency Var- iants as a Component of the Estimated Genetic Risk," Utilization of Mammalian Specific Locus Studies in Hazard Evaluation and Estimation of Genetic Risk, E]. deSerres and W. Sheridan (eds.) (New York: Plenum Press, 1983) pp. 55-69. Mohrenweiser, H.W., “Enzyme—Deficiency Var— iants: Frequency and Potential Significance in Human Populations," lsozymes: Current Topics in Biological and Medical Research 10:51-68, 1983. Mohrenweiser, H.W., unpublished data, cited in Neel, J.V., Satoh, C., Goriki, K., et al., “The Rate With Which Spontaneous Mutation Alters the Electrophoretic Mobility of Polypeptides,” Proc. Nat. Acad. Sci. (USA) 83:389-393, 1986. Mohrenweiser, H.W., “One-Dimensional Elec- trophoresis and Quantitative Enzyme Assays as Technologies To Detect Mutation Rates in Hu- man Beings,” typescript, contract report pre- pared for the Office of Technology Assessment, U.S. Congress, Washington, DC, 1985. Mohrenweiser, H.W., and Neel, ].V., "A 'Dis- proportion’ Between the Frequency of Rare Elec- tromorphs and Enzyme Deficiency Variants in Amerindians,” Amer. ]. Hum. Genet. 36:655- 662, 1984. Morley, A.A., Trainor, K.J., Seshadri, R., and Ryall, R.G., ”Measurement of In Vivo Mutations in Human Lymphocytes," Nature 302:155-156, 1983. Mukai, T., and Cockerham, C.C., “Spontane- ous Mutation Rates at Enzyme Loci in Droso— phila melanogaster," Proc. Nat. Acad. Sci. (USA) 74:2514-2517, 1977. Muller, H.I., ”Artificial Transmutation of the Gene," Science 66:84-87, 1927. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. Mulvihill, ].J., and Byrne, 1., “Offspring of Long- Time Survivors of Childhood Cancer," Child- hood Cancer: Late Effects, Clinics in Oncology 4(2):333-343, 1985. Mulvihill, J.J., and Czeizel, A., “Perspectives in Mutation Epidemology. 6: A 1983 View of Sen- tinel Phenotypes,"Mutat. Res. 123:345-61, 1983. Myers, R.M., “Determination of the Mutation Rate in Humans by Denaturing Gradient Gel Electrophoresis," typescript, contract report pre- pared for the Office of Technology Assessment, US. Congress, Washington, DC, 1985. Myers, R.M., “Determination of the Mutation Rate in Humans by Ribonuclease Cleavage at Mismatches in RNAzDNA Heteroduplexes," typescript, contract report prepared for the Of- fice of Technology Assessment, US. Congress, Washington, DC, 1985. Myers, R.M., Larin, Z., and Maniatis, T., “De- tection of Single Base Substitutions by Ribo- nuclease Cleavage at Mismatches in RNAzDNA Duplexes," Science 230:1242-1246, 1985. National Research Council, Committee on the Biological Effects of Ionizing Radiation, The Ef- fects on Populations of Exposure to Low Levels of Ionizing Radiation (Washington, DC: Na- tional Academy Press, 1980). National Research Council, Identifying and Esti- mating the Genetic lmpact of Chemical Muta- gens (Washington, DC: National Academy Press, 1983). Neel, ].V., ”The Detection of Increased Muta- tion Rates in Human Populations,” Persp. Biol. Med. 14:522—537, 1971. Neel, J.V., “Some Trends in the Study of Spon- taneous and Induced Mutation in Man," Hum. Genet, Proc. 5th Int]. Cong., Mexico (Amster- dam: Excerpta Medica, 1977), pp. 19-32. Neel, J.V., ”Frequency of Spontaneous and In- duced 'Point’ Mutations in Higher Eukaryotes,” I. Hered. 74:2-15, 1983. Neel, J.V., University of Michigan Medical School, personal communication, Aug. 21, 1985. Neel, I.V., Mohrenweiser, H., and Hanash S., et al., “Biochemical Approaches To Monitoring Human Populations for Germinal Mutation Rates. 1: Electrophoresis,” Utilization of Mam- malian Specific Locus Studies in Hazard Evalu- ation and Estimation of Genetic Risk, F.J. deSerres and W. Sheridan (eds.) (New York: Ple- num Press, 1983), pp. 71-93. Neel, ].V., Mohrenweiser, H.W., and Meisler, M.H., ”Rate of Spontaneous Mutation at Human Loci Encoding Protein Structure,” Proc. Nat. Acad. Sci. (USA) 77(10)6037—6041, 1980. References - 133 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. Neel, J.V., Rosenblum, B.B., Sing, C.F., et al., “Adapting Two-Dimensional Gel Electrophore- sis to the Study of Human Germ-Line Mutation Rates,” ch. 9, Two-Dimensional Gel Electropho- resis of Proteins, ].E. Celis and R. Bravo (eds.) (London: Academic Press, 1984), pp. 259-306. Neel, J.V., Satoh, C., Goriki, K., et al., ”The Rate With Which Spontaneous Mutation Alters the Electrophoretic Mobility of Polypeptides,” Proc. Nat. Acad. Sci. (USA) 83:389-393, 1986. Nicklas, 1., ”The Use of T-Lymphocytes for Mutagenicity Monitoring,” typescript, contract report prepared for the Office of Technology Assessment, US Congress, Washington, DC, 1985. Olson, M.V., “Detection of New Germinal Mu- tations by Surveying Human DNA for the Gain or Loss of Restriction-Enzyme Cleavage Sites,” typescript, contract report prepared for the Of- fice of Technology Assessment, US. Congress, Washington, DC, 1985. Otake, M., and Schull, W.]., ”In Utero Exposure to A-Bomb Radiation and Mental Retardation: A Reassessment,” Brit. J. Radiol. 57:409-414, 1984. Phillips, ].A., and Kazazian, H.l-I., “Principles and Practice of Medical Genetics," Haemoglo- binopathies and Thalassaemias, A.E.H. Emery and D.L. Rimoin (eds.) (Edinburgh: Churchill Livingstone, 1983), pp. 1019-1043. Porter, I.H., ”Control of Hereditary Disorders," Ann. Rev. Public Health 3:277-319, 1982. President's Commission for the Study of Ethical Problems in Medicine and Biomedical Research and Behavioral Research, Screening and Coun— seling for Genetic Conditions (Washington, DC: U.S. Government Printing Office, 1983). Racine, R.R., Langley, C.H., and Voelker, R.A., ”Enzyme Mutants Induced by Low-Dose Rate Gamma-Irradiation in Drosophila: Frequency and Characterization,” Environ. Mutag. 2:167- 177, 1980. Rosenblum, B.B., Neel, J.V., and Hanash, S.M., ”Two—Dimensional Electrophoresis of Plasma Polypeptides Reveals 'High' Heterozygosity In- dices," Proc. Nat. Acad. Sci. (USA) 805002- 5006, 1983. Rosenblum, B.B., Neel, ].V., Hanash, S.M., et al., “Identification of Genetic Variants in Eryth- rocyte Lysate by Two-Dimensional Gel Elec- trophoresis,” Amer. J. Hum. Genet. 36:601-612, 1984. Russell, L.B., Aaron, C.S., de Serres, F., et al., ”Evaluation of Mutagenicity Assays for Purposes 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. of Genetic Risk Assessment,” Mutat. Res. 134:143—157, 1984. Russell, L.B., Russell, W.L., Popp, R.A., et al., ”Radiation Induced Mutations at Mouse Hemo- globin Loci,” Proc. Nat. Acad. Sci. (USA) 73:2843-2846, 1976. Russell, L.B., and Shelby, M.D., “Tests for Heritable Genetic Damage and for Evidence of Gonadal Exposure in Mammals,” Mutat. Res., 154:69-84, 1985. Russell, W.L., “X-Ray-Induced Mutations in Mice," Cold Spring Harbor Symposium in Quantitative Biology XVII:327-335, 1951. Russell, W.L., “Mutation Frequencies in Female Mice and the Estimation of Genetic Hazards of Radiation in Women," Proc. Nat. Acad. Sci. (USA) 74:3521-3527, 1977. Russell, W.L., “Dose Response, Repair, and No- Effect Dose Levels in Mouse Germ-Cell Muta- genesis,” Problems of Threshold in Chemical Mutagenesis, Proceedings of a Nissan Sympo- sium on Dose-Response Relationship for Genetic Effects of Environmental Chemicals, Tokyo, May 7-9, 1984, pp. 43-50. Russell, W.L., Hunsicker, P.R., Carpenter, D.A., et al., ”Effect of Dose Fractionation on the Ethylnitrosourea Induction Specific-Locus Mu- tations in Mouse Spermatogonia,” Proc. Nat. Acad. Sci. (USA) 79:3592-3593, 1982. Russell, W.L., and Kelly, E.M., ”Mutation Fre- quencies in Male Mice and the Estimation of Genetic Hazards of Radiation in Men,” Proc. Nat. Acad. Sci. (USA) 79:542-544, 1982. Russell, W.L., and Kelly, E.M., “Specific-Locus Mutation Frequencies in Mouse Stem-Cell Sper- matogonia at Very Low Radiation Dose Rates,” Proc. Nat. Acad. Sci. (USA) 79:539-541, 1982. Russell, W.L., Russell, L.B., and Kelly, E.M., ”Radiation Dose Rate and Mutation Frequency,” Science 128:1546-1550, 1958. Sachs, E.S., Van Hemel, 1.0., and Galjaard, H., et al., “First Trimester Chromosomal Analysis of Complex Structural Rearrangements With RHA Banding on Chorionic Villi” (letter), Lan- cet 2:1426, 1983. Sanderson, B.].S., Dempsey, ].L., and Morley, A.A., “Mutations in Human Lymphocytes: Ef- fect of X- and UV-Irradiation," Mutat. Res. 140:223-227, 1984. Sanger, F., Nicklen, 5., and Coulson, A.R., “DNA Sequencing With Chain-Terminating In— hibitors," Proc. Nat. Acad. Sci. (USA) 7425463- 5467, 1977. Sankaranarayanan, K., “The Role of Non— 134 - Technologies for Detecting Heritable Mutations in Human Beings 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. Disjunction in Aneuploidy in Man: An Over- view,” Mutat. Res. 61:1-28, 1979. Sankaranarayanan, K., Genetic Effects of Ioniz- ing Radiation in M ulticellular E ukaryotes and the Assessment of Genetic Radiation Hazards in Man (Amsterdam: Elsevier Biomedical Press, 1982). Satoh, C., Awa, A.A., and Neel, ].V., et al., “Genetic Effects of Atomic Bombs," Human Genetics, Part A: The Unfolding Genome (New York: Alan R. Liss, Inc., 1982), pp. 267-276. Satoh, C., Neel, ].V., and Yamashita, A., et al., ”The Frequency Among Japanese of Heter- ozygotes for Deficiency Variants of 11 Enzymes," Amer. J. Hum. Genet. 35:656-674, 1983. Schull, W.]., Neel, J.V., and Hashizume, A., “Further Observations on Sex Ratio Among In- fants Born to Survivors of the Atomic Bombs,” Amer. ]. Hum. Genet. 18:328, 1966. Schull, W.J., Otake, M., and Neel, ].V., ”Genetic Effects of the Atomic Bombs: A Re- appraisal," Science 213:1220-1227, 1981. Schull, W.J., Otake, M.,and Neel, ].V., ”Hiroshima and Nagasaki: A Reassessment of the Mutagenic Effect of Exposure to Ionizing Radia— tion,” Population and Biological Aspects of Hu- man Mutation, E.B. Hook and I.H. Porter (eds) (New York: Academic Press, 1981). Smith, C.L., ”Purification, Specific Fragmenta— tion, and Separation of Large Molecular Weight DNA Including Intact Chromosomes," type- script, contract report prepared for the Office of Technology Assessment, US. Congress, Wash- ington, DC, 1985. Smith, C.L., and Cantor, C.R., “Pulsed-Field Gel Electrophoresis of Large DNA Molecules,” Na- ture 319:701-702, 1986. Smith, H.O., ”Nucleotide Sequence Specificity of Restriction Endonucleases," Science 205:455- 462, 1979. Sobels, F.H., ”Evaluating the Mutagenic Poten- tial of Chemicals—The Minimal Battery and Ex- trapolation Problems,” Arch. Toxicol. 46:21-30, 1980. Sobels, F.H., ”The Parallelogram: An Indirect Approach for the Assessment of Genetic Risks From Chemical Mutagens,” Prog. Mut. Res. 3:323-327, 1982. Stamatoyannopoulos, G., and Nute, P.E., ”De Novo Mutations Producing Unstable Hbs or HbsM. II: Direct Estimates of Minimum Nucleo- tide Mutation Rates in Man," Hum. Genet. 60:181-88, 1982. Stamatoyannopoulos, G., and Nute, P., "Detec- tion of Somatic Mutants of Hemoglobin,” Utili— zation of Mammalian Specific Locus Studies in 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. Hazard Evaluation and Estimation of Genetic Risk, F.]. de Serres and W. Sheridan (eds) (New York: Plenum Press, 1983). Stamatoyannopoulos, G., Nute, P., Lindsley, D., et al., ”Somatic-Cell Mutation Monitoring System Based on Human Hemoglobin Mutants,” Single-Cell Mutation Monitoring Systems— Methodologies and Applications, A.A. Ansari and F]. deSerres (eds) (New York: Plenum Press, 1984), pp. 1-35. Stamatoyannopoulos, G., Nute, PE, and Miller, M., “De Novo Mutations Producing Un- stable Hbs or Hbs M. I: Establishment of a De- pository and Use of Data To Test for an Asso- ciation of De Novo Mutation With Advanced Parental Age,” Hum. Genet. 58:396-404, 1981. Stein, Z., Kline, J., and Susser, E., et al., ”Mater- nal Age and Spontaneous Abortion,” Embryonic and Fetal Death, l.H. Porter and EB. Hook (eds) (New York: Academic Press, 1980). Stein, Z., Susser, M., and Warburton, D., et al., ”Spontaneous Abortion as a Screening Device," Amer. ]. Epidem. 102:275-290, 1975. Strauss, G.H.S., ”Direct Mutagenicity Testing: The Development of a Clonal Assay To Detect and Quantitate Mutant Lymphocytes Arising in Vivo," Indicators of Genotxic Exposure, Banbury Report 13, BA. Bridges, B.E. Butterworth, and LB. Weinstein (eds) (Cold Spring Harbor, NY: Cold Spring Laboratories, 1982), pp. 423-439. Streisinger, G., ”Extrapolations From Species to Species and From Various Cell Types in Assess— ing Risks From Chemical Mutagens," Mutat. Res. 114:93-105, 1983. Thilly, W.G., ”Potential Use of Gradient Denaturing Gel Electrophoresis in Obtaining Mutational Spectra From Human Cells,” Car- cinogenesis: The Role of Chemicals and Radia- tion in the Etiology of Cancer, vol. 10, E. Huber- man (ed.) (New York: Raven Press, 1985), pp. 511-528. Tobari, Y.N., and Kojima, K., ”A Study of Spontaneous Mutation Rates at Ten Loci Detect- able by Starch Gel Electrophoresis in Drosophila melanogaster,” Genetics 70:397-403, 1972. Trainor, K.]., Wigmore, D.]., Chrysostomou, A., et al., “Mutation Frequency in Human Lymphocytes Increases With Age,” Mech. Age. Develop. 27:83-86, 1984. Tucker, M.A., Meadows, A.T., Boice, J.D., Jr., ”Cancer Risk Following Treatment of Childhood Cancer,” Radiation Carcinogenesis: Epidemiol- ogy and Biological Significance, ].D. Boice, Ir., and ].F. Fraumeni, Jr. (eds) (New York: Raven Press, 1984). References 0 135 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. Turner, D.R., Morley, A.A., Haliandros, M., et al., “In Vivo Somatic Mutations in Human Lymphocytes Are Frequently Due to Major Gene Alterations," Nature 315:343-345, 1985. UNSCEAR Report, Sources and Effects of Ioniz- ing Radiation, Report of the United Nations Sci- entific Committee on the Effects of Atomic Ra- diation, United Nations, New York, 1982. U.S. Congress, Office of Technology Assess- ment, Assessment of Technologies for Determin- ing Cancer Risks From the Environment, OTA- H-138 (Washington, DC: U.S. Government Printing Office, June 1981). U.S. Congress, Office of Technology Assess- ment, Impacts of Applied Genetics: Micro- Organisms, Plants, and Animals, OTA-HR-132 (Washington, DC: U.S. Government Printing Office, April 1981). U.S. Congress, Office of Technology Assess- ment, Commercial Biotechnology: An Interna- tional Analysis, OTA-BA-218 (Washington, DC: U.S. Government Printing Office, January 1984). U.S. Department of Health and Human Services, Public Health Service, Food and Drug Adminis- tration, The Selection of Patients for X-Ra y Ex- amination, HHS Publication (FDA) 80-8104 (Washington, DC: U.S. Government Printing Office, January 1980). U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, Multiple Primary Cancers in Connect- icut and Denmark, National Cancer Insititute Monograph 68, NIH Publication 85-2714, J.C. Boice, et. al. (eds) (Bethesda, MD: 1985.) U.S. Department of Health and Human Services, National Institutes of Health, Genetic Sequence Data Bank (Cambridge, MA: Bolt, Beranek & Newman, Inc., 1985). U.S. Department of Labor, Occupational Safety and Health Administration, “Occupational Ex- posure to Ethylene Dibromide: Notice of Pro- posed Rulemaking,” Federal Register 48(196): 45956-46003, Oct. 7, 1983. U.S. Department of Labor, Occupational Safety and Health Administration, “Occupational Ex— posure to Ethylene Oxide: Final Standard,” Fed- eral Register 49(122):25734-25809, June 22, 1984. U.S. Environmental Protection Agency, “Federal Radiation Protection Guidance for Occupational Exposure," Federal Register 46(15), Ian. 23, 1981. U.S. Environmental Protection Agency, ”Pro- posed Guidelines for Mutagenicity Risk Assess- ment," Federal Register 9(227):46314-46321, Nov. 23, 1984. U.S. Environmental Protection Agency, Work— 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. shop Proceedings: Approaches for Improving the Assessment of Human Genetic Risk-Human Bi- omonitoring (Washington, DC: 1984). U.S. Environmental Protection Agency, Office of Radiation Programs, Radionuclides: Back- ground Information Document for Final Rules, EPA 520/ 1-84-022-1 (Washington, DC: October 1984). Van der Ploeg, L.H.T., ”Determining Mutation Frequencies in Humans: The Use of Pulsed Field Gel-Electrophoresis for the Analysis of Chromo- somes," typescript, contract report prepared for the Office of Technology Assessment, U.S. Con— gress, Washington, DC, 1985. Van der Ploeg, L.H.T., Smits, M., Ponnudurai, T., et al., “Chromosome-Sized DNA Molecules of Plasmodium Falciparum,” Science 229:658- 661, 1985. Van Dyke, D.L., Weiss, L., and Roberson, ].R., et al., ”The Frequency and Mutation Rate of Bal- anced Autosomal Rearrangements in Man Esti- mated From Prenatal Genetic Studies for Ad- vanced Maternal Age," Amer. J. Hum. Genet. 352301-308, 1983. Vijayalaxmi, and Evans, H.J., “Measurement of Spontaneous and X—Irradiation-Induced 6- Thioguanine-Resistant Human Blood Lympho- cytes Using a T—Cell Cloning Technique," Mutat. Res. 125:87—94, 1984. Voelker, R.A., Schaffer, H.E., and Mukai, T., “Spontaneous Allozyme Mutations in Droso— phila: Rate of Occurrence and Nature on the Mu— tants,” Genetics 94:961-968, 1980. Vogel, F., “Mutation in Man,” Principles and Practice of Medical Genetics, A.E.H. Emery and D. Rimoin (eds.) (Edinburgh: Churchill Living— stone, 1983), pp. 26-48. Vogel, F., ”Gene or Point Mutations," Mutations in Man, G. Obe (ed.) (Berlin: Springer-Verlag, 1984), pp. 101-127. Vogel, F., and Altland, K., ”Utilization of Ma— terial From PKU-Screening Programs for Muta- tion Screening,” Prog. Mut. Res. 3:143-52, 1982. Vogel, F., and Motulsky, A.G., Human Ge- netics: Problems and Approaches (Berlin: Springer-Verlag, 1982). Vogel, F., and Rathenberg, R., “Spontaneous Mutation in Man," Adv. Hum. Genet. 5:223- 318, 1975. Voytek, P., ”Approaches for Regulating Muta- genic Agents,” Reproduction: The New Frontier in Occupational and Environmental Health Re- search (New York: Alan R. Liss, Inc., 1984), pp. 429-438. Voytek, P., Reproductive Effects Assessment 136 - Technologies for Detecting Heritable Mutations in Human Beings 169. 170. Group, US. Environmental Protection Agency, personal communication, Aug. 7, 1985. Warburton, D., Stein, 2., and Kline, J., et al., ”Chromosome Abnormalities in Spontaneous Abortion: Data From the New York City Study,” Embryonic and Fetal Death, I.H. Porter and EB. Hook (eds.) (New York: Academic Press, 1980), pp. 261-287. Waters, M., Allen, J., Doerr, C., et al., ”Appli- cation of the Parallelogram Approach To Pre- 171. 172. dict Genotoxic Effects in Humans,” typescript, 1985. Weatherall, D.I., The New Genetics and Clini- cal Practice (London: Nuffield Provincial Hos- pitals Trust, 1982). Yang, T.P., Patel, P.I., Chinault, A.C., et al., “Molecular Evidence for New Mutation at the hprt Locus in Lesch-Nyhan Patients,” Nature 310:412-414, 1984. Index Index Achondroplasia (dwarfism), 7, 30, 31 Ad hoc studies, epidemiologic, 14, 103, 104, 107 Adenine (A), 23, 24, 27, 64, 66 Adenosine deaminase deficiency, 105 Agarose gel electrophoresis, 59, 60, 61, 63, 68 Agency for Toxic Substances and Disease Registry (ATSDR), 18-19 Agent Orange, 18 Alanine, 27 ALARA principle, 114 Alkylating agents, 93-94 Alleles, 27, 44, 77, 82 Alpha-globin chain, 47 Alpha-thalassemia, 29 American Society of Human Genetics, 19 Ames test, 15, 118 Amino acids, 9, 25-27, 29, 44, 46, 47, 49, 57, 58, 78; see also specific amino acids Amniocentesis, 41 Animals, mutation studies in, 3—5, 6-7, 19, 48, 78, 81-87, 91-100, 117-118 Aniridia, 7 Antibodies, in somatic mutation studies, 77-78 Apert's syndrome, 31 Arginine, 27 Asparagine, 27 Aspartic acid, 27 Atomic Bomb Casualty Commission, 48 Atomic bomb survivors, effects of radiation on, 9-10, 32, 48-50, 77, 86, 106, 108 Atomic Energy Act, 114 Atomic Energy Commission (AEC), 82, 114 Autoradiographic assay, in somatic mutation studies, 74-75 Autoradiography, 56, 59, 61, 64, 65, 68, 74-75 Autosomes, 27, 38, 42, 44, 46, 49 Bacteria, 81, 91, 92, 94, 116, 117, 118 Benzola]pyrene, 97 Beta—globin chain, 47, 77 Biochemical analyses, 37, 44—48, 49-50, 51, 52 Biologic extrapolation, see Extrapolation, qualitative Biological effects of ionizing radiation, 113 Biological samples, banking of, 17-18 Blood cells, somatic mutation studies on, 73-78 Blood proteins detection of variants in, 44—48, 49-50 electrophoretic analysis of, 8-9, 10, 37, 44—46, 50, 51, 83 molecular analysis of, 46-48 see also, Enzymes; Hemoglobin proteins Cancer, 3, 7, 9, 15, 31—32, 38, 50, 73, 77, 104—105, 108-109, 117 Cancer registries, 105, 109 Case-control studies, in epidemiologic testing, 107 Cells, see Blood cells; Chorionic villi cells; Germ cells; Red blood cells; Somatic cells; Stem cells; TG’ cells; White blood cells Centers for Disease Control, 18 Chemicals denaturing, 12, 59, 61 effects on DNA, 23, 59, 81 effects on germinal mutations, 7, 9 effects on heritable mutations, 6-7 effects on somatic mutations, 7, 74-78 in extrapolation models, 93-97, 100 as mutagenic agents, 6—7, 9, 32, 48, 91-97, 106 ”standards" for, 17 Chemotherapy, 108 Children, genetic diseases in, 7-8, 30-32, 37, 38-40, 47-48, 49, 50, 109, 110 Chorionic villi cells, 41 Chromosomes abnormal at birth, 41—43 banding methods, 8, 28, 41, 42, 43, 50 biochemical analyses of, 49-50 cytogenetic analysis of, 37, 40-43, 49, 52, 55, 98 DNA and, 23-25, 27-28, 41, 43 health effects of abnormal, 40—43, 49 mutations in, 5, 6, 8, 9, 13, 29-30, 31, 32, 40-43, 49, 51, 68, 82, 92, 95 numerical abnormalities, 8, 29-30, 31, 41, 49, 116 pulsed field gel electrophoresis and, 68 staining methods, 8, 28, 41, 42, 43, 50 structural abnormalities, 8, 29, 31, 41, 49 studies in mice, 82 surveillance of abnormal, 105 translation of, 30, 82 Chronic hemolytic anemia, 46, 47 Clonal assay, in somatic mutation studies, 75 Cloning, 58-59, 60, 64-65, 75 Codons, 6, 25-27, 29 Cohort studies, in epidemiologic testing, 107, 109 Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA), 3, 15, 114 Congenital hypothyroidism, 105 Consumer Product Safety Act, 114 Consumer Product Safety Commission, 114 Cyanosis, 47 Cyclophosphamide, 97 Cysteine, 27 139 140 - Technologies for Detecting Her/table Mutations in Human Beings Cytogenetic analysis, of chromosome abnormalities, 37, 40-43, 49, 52, 55, 98 Cytosine (C), 23, 24, 27, 64, 66 Cytotoxic drugs, 108 Deoxyribonucleic acid (DNA) chromosomes and, 23-25, 27—28, 41, 43 current methods for studying, 6-10 denaturing of, 11-12, 59-64 double helix, 24, 25, 27 double-stranded, 12—13, 23, 59-68 EPA guidelines and, 116 extrapolation studies and, 91-100 Federal expenditures on research, 121-122 function and organization of, 23-28 genetic nature of, 3, 4, 5-6 genomic, 6, 11, 55, 57-68 heteroduplexes, 12, 55, 61, 62, 63, 64-65, 67 homoduplexes, 61, 63, 67 hybridization, 66-67 nontranscribed, 45 nucleotide chains in, 23-25 organizational hierarchy of, 6 replication of, 23-25, 74 sequencing, 23-25, 55, 57-58, 60, 61-64, 66-67, 104 single—stranded, 12-13, 59-68 structure of, 23-25, 81 use of new technologies in examining, 10—14, 17, 55-69, 74-76 Department of Defense, 114 Department of Energy (DOE), 16, 17, 19, 55, 114, 122 Department of Transportation, 114 Diabetes, 3 Dioxin, 106 Disease registries, 104-106, 107 Diseases dominant, 44 environmental factors in, 31 genetic, 3, 5, 7-8, 9, 30-33, 37, 38-40, 42, 44-45, 47-48, 49, 50, 93, 94, 103-110, 108-109, 110, 115 “multifactorial,” 31 recessive, 44 see also Sentinel phenotypes and specific diseases DNA-RNA heteroduplex analysis, 12, 64-65 Dominant lethal test, 81-82 Dominant mutation tests, 83 Dose-response relationships, in quantitative extrapolation studies, 97—99 Down syndrome, 3, 8, 30, 31, 42 Drosophila, 48, 81, 82, 96, 116, 117 Duchenne muscular dystrophy, 3, 30 Electrophoresis, 8-9, 10, 11-12, 13, 37, 44-46, 49, 50, 51, 55, 68, 83, 84 see also Gel electrophoresis and specific types of gel electrophoresis Embryos, in rodent tests, 81-82 Endonucleases, restriction, use in detecting mutations, 11 Environmental Protection Agency (EPA), 15, 16, 19, 114, 115, 116, 117 Enzymes biochemical analyses of, 49-50 HPRT, 74 kinetic studies of, deficiency variants, 44, 45-46 quantitative studies of deficiency variants, 45—46, 83, 84 restriction, 11, 13, 58-64, 66, 67, 68, 75 RNaseA, 12 role in protein synthesis, 25-27 somatic mutation studies on, 73 Epidemiologic studies, 9-10, 14, 18, 49, 103-110 Epstein-Barr virus, 104 Escherichia coli, 58, 60 Ethylene dibromide (EDB), 115, 117 Ethylene oxide (EtO), 115, 117 Ethylnitrosourea (ENU), in quantitative extrapolation studies, 97, 98-99 Eukaryotes, 92, 94, 95, 116 Exposure registries, 106, 107 Extrapolation aims of, 91-92 from animal to human data, 86-87 defined, 14-15 limits of, 99-100 qualitative, 15, 91, 94-97, 99-100 quantitative, 15, 91, 97—100, 115, 117 theoretical models for, 15, 93-94 Federal Food, Drug, and Cosmetics Act, 114 Federal Insecticide, Fungicide, and Rodenticide Act, 115 Federal Radiation Council (FRC), 113, 114 Fetuses, chromosome abnormalities in, 40, 42, 82 Fluorochrome, 77 Food and Drug Administration (FDA), 19, 114 Fruit flies, see Drosophila Fungi, 92 Galactosemia, 105 Gamma radiation, 97 Gel electrophoresis, 11-12, 13, 55, 56, 58, 59, 65 see also Agarose gel electrophoresis; One- dimensional denaturing gradient gel electrophoresis; Pulsed field gel electrophoresis; Two-dimensional denaturing Index - 141 gradient gel electrophoresis; Two-dimensional polyacrylamide gel electrophoresis Genes dominant, 27, 44, 51 globin, 29, 46, 47-48, 55 hprt, 55, 74-76 mutations in, 5, 6—7, 29, 38-40, 42-43, 44, 47-48, 50, 51, 55, 57-68, 74—76, 92, 95 recessive, 27, 44 Genomes and genomic sequencing, 6, 11, 57-58, 66-67, 68 Germ cells, 6, 7, 9, 13, 27, 30, 31, 37, 38, 39, 48, 49, 80-84, 91-100, 108, 115-116, 117; see also Germinal mutations Germinal mutations in cells, 7 defined, 6 electrophoretic studies on, 51 new, 37—38 radiation-induced, 9—10, 48-49 relationship to somatic mutations, 73 see also Mutations Glossary of Acronyms and Terms, vii-xii Glutamic acid, 26, 27, 29 Glutamine, 27 Glycine, 27 Glycophorin A, 76-77 Glycophorin assay, in somatic mutation studies, 76-77 Guanine (G), 23, 24, 27, 64, 66 “Guidelines for Mutagenicity Risk Assessment,” 116 Health effects of chromosome abnormalities, 40-43 of mutations, 4, 30-32, 108-109 see also Children; Diseases; Humans; Infants Heart disease, 3, 25, 32 Hemoglobin A (HbA), 46 Hemoglobin assays, in somatic mutation studies, 77-78 Hemoglobin C, 77 Hemoglobin M (HbM), molecular analysis of, 44, 46—48, 51 Hemoglobin S, 77 Hemoglobins, unstable, molecular analysis of, 44, 46-48, 51 Hemophilia, 3 Heritable mutations chemicals and, 6—7, 9, 32, 48, 91-97 defined, 6 detection and measurement of, 11-13 epidemiologic studies in, 9-10, 14, 18, 49, 103-110 estimating effects in humans from animal data, 86-87 extrapolation of data to predict risk in humans, 91-100 factors that influence induced rates in mice, 84-86 health effects of, 30-32, 108-109 induced by chemicals, 6—7 kinds and rates of, in humans, 32-33, 37—52 laboratory studies on, 81-87 new technologies and, 3, 10—14, 55—69 rodent tests for, 81-84 spontaneous, studies in humans, 7-10 see also Mutations Heritable translocation tests (HTTs), 82, 94-96, 98 Heteroduplexes, 12, 55, 61, 62, 63, 64-65, 67 Heterozygotes, 44, 45, 77, 82 Histidine, 27, 47 Histidinemia, 105 Hodgkin's disease, 108 Homocystinuria, 105 Homoduplexes, 61, 63, 67 Homologous sequences, 64 Homozygotes, 44, 82 hprt, 55, 74-76 Humans animal studies and, 84-87 epidemiologic studies in, 9—10, 14, 18, 49, 103—110 Federal expenditures on mutation research in, 121-122 frequencies of hprt" T-lymphocytes in, 76 health effects of mutations in, 30-32, 108-109 kinds and effects of mutations in, 4, 5—6, 29-32 kinds and rates of heritable mutations in, 32-33, 37—52 mutagenic risk predictions by extrapolation, 91-100 mutation studies in, 3-4, 7-14, 19, 44—52, 55-69, 73-78 new technologies for detecting mutations in, 10—14, 55-69, 73—78 persistence of new mutations in, 32-33, 37 sentinel phenotypes in 7—8, 37, 38—40, 45, 50, 51, 52, 55, 105, 106, 109 spontaneous mutation rates in, 32-33, 51 see also Children; Diseases; Infants Huntington disease, 31 Hypoxanthineguanine phosphoribosyl transferase (HPRT), 74 Infants, genetic diseases in, 7—8, 30-32, 37, 38-40, 42, 44—45, 49, 50, 109 Insects, 91, 92, 116 142 - Technologies for Detecting Heritab/e Mutations in Human Beings International Commission for Protection Against Environmental Mutagens and Carcinogens (ICPEMC), 55 International Commission on Radiological Protection (ICRP), 113, 114 International dioxin registry, 106 Isoleucine, 27 Karyotypes, 28, 41, 42 Kinetic studies, of enzyme deficiency variants, 45-46 Laboratory studies, of heritable mutation rates, 81-87; see also Animals Lambda (virus), 59, 60 Leucine, 27 Leukemia, 9, 32, 109 Lysine, 26, 27 Maple syrup urine disease, 105 Marfan’s syndrome, 31, 109 Measurement techniques, validation of, 104 Messenger RNA (mRNA), 25, 26 Methemoglobinemia, 47 Methemoglobinemic cyanosis, 46 Methionine, 27 Methyl isocyanate, 106 Mice effects of radiation on, 82-86, 97 in extrapolation tests, 93, 96 factors affecting mutations in, 84-86 heritable mutation tests in, 82-84 mutation effects in females, 85-86 mutation effects in males, 84-85 spermatogonial studies in, 84-85, 97-98 unscheduled DNA synthesis in spermatids of, 97-98 Mine Safety and Health Administration (MSHA), 114 Monitoring, in epidemiologic testing, 14, 103-104, 106, 107 Mortality and morbidity, 31, 41, 49, 82 Mouse spot test, 96 Mutagenicity assays, 91-100; see also Mutations and specific tests for mutagenicity Mutants, in somatic mutation studies, 74-78; see also Mutations Mutations animal studies, 3-5, 6-7, 19, 48, 78, 81-87, 117—118 at birth, 41 in chromosomes, 5, 6, 7, 8, 9, 13, 29-30, 31, 32, 40-43, 49, 51, 68, 82, 92 at conception, 41 current methods for studying, 6-10, 37-52 dominant, 32, 38, 44 environmental factors in, 3, 4, 37, 73, 96, 106, 108-109 Federal expenditures on research, 15, 18, 117, 121-122 in genes, 5, 6—7, 29, 38-40, 42-43, 44, 47—48, 50, 51, 55, 57—68, 74-76, 92, 95 genetic inheritance and, 23-33 germinal, see Germinal mutations health effects of 4, 30—32, 108-109 heritable, see Heritable mutations human studies, 3—4, 7-14, 19, 44—52, 55-69, 73-78, 103-110 kinds of, and effects in humans, 4, 5-6, 29-32, 55 new, persistence of in humans, 32-33, 37, 41 new technologies for detecting in humans, 10-19, 55-69 options for research on, 16-19 in proteins, 3, 8—9, 25—27, 29, 44—48, 49, 56-57 recessive, 32, 33 regulatory considerations, 15, 17, 113-118 somatic, see Somatic mutations spontaneous, 5, 7, 32-33, 51, 83 test methods for research in, 14—16 X-linked (sex-linked), 32, 37, 38 National Cancer Institute, 104 National Center for Toxicological Research, 19 National Council on Radiation Protection and Measurements (NCRP), 113, 114 National Environmental Protection Act of 1970, 114 National Institutes of Health (NIH), 122 National Research Council (NRC), 32—33 National Toxicology Program (NTP), 19 Neuroblastoma, 109 Nuclear Regulatory Commission (NRC), 106, 108, 114 Nucleotides and nucleotide sequencing, 5-6, 8, 11, 12, 23-25, 29, 44, 46, 47, 57-58, 59, 61, 66, 68, 93 Occupational Safety and Health Act, 114, 115 Occupational Safety and Health Administration, 114, 115 Office of Technology Assessment (OTA), 4, 15, 121 One-dimensional denaturing gradient gel electrophoresis, 55, 59-61, 63 Oocytes, in mice, 85-86 Osteogenesis imperfecta, 30 Index 0 143 Parallelogram, as extrapolation model, 93, 97, 117 Peripheral blood lymphocytes (PBLs), 97 PFGE, see Pulsed field gel electrophoresis Phenotypic data limitations of, 40, 45 mutation rates estimated from, 39-40 surveillance of, 38-40 see also Sentinel phenotypes Phenylalanine, 26, 27 Phenylketonuria (PKU), 45, 105 Polyacrylamide, use in protein separation, 55, 56-57 Polymorphisms, 58-59, 64 Polypeptide chain, 26, 37, 46; see also Proteins Prokaryotes, 94, 95 Proline, 27 Proteins biochemical analysis of, 37, 44-48, 49-50, 51, 52 DNA and, 23-27, 45, 46, 50, 57 mutations in, 3, 8-9, 25-27, 29, 44-48, 49, 56-57 one-dimensional separation of, 8-9, 44-45, 49, 56 synthesis of, 25-27 variant, mutation studies on, 8-9, 44-50, 55, 56-57, 73-74 two—dimensional separation of, 9, 56-57 see also Amino acids; Blood proteins; Enzymes; Hemoglobins, Nucleotides and nucleotide sequencing Public Health Service, 18 Pulsed field gel electrophoresis (PFGE), 13, 55, 68 Qualitative studies, 15, 91, 94-97, 99-100 Quantitative studies, 15, 45-46, 55, 91, 97-100, 115, 117 Radiation effects on atomic bomb survivors, 9-10, 32, 48-50, 77, 86, 106, 108 effects on DNA, 23 effects on fruit flies, 82 effects on human germ cells, 9 effects on mice, 82-86, 97 effects on pregnant women, 9 effects on US. citizens, 86-87, 106 in extrapolation studies, 97-100 in Marshall Islands populations test, 50 in somatic mutation studies, 74 mutagenic effects of, 73 US. Government efforts to protect humans from, 113-114 Radiation Effects Research Foundation, 16, 48, 104 Radioactive thymidine, 74 Red blood cells, somatic mutation studies on, 73, 76-78 Regulatory considerations definition of germ-cell mutagen, 115-117 mutagenicity and carcinogenicity, 117—118 new methods for measuring mutation rates, 118 protection against agents other than radiation, 114 quantitative extrapolation, 117 radiation protection, 113-114 role of U.S. Government, 15, 17, 113-118 Research and public policy, use of new technologies in, 14-16 Restriction fragment length polymorphisms (RFLPS), 55, 58-59, 60 Retinoblastoma, 7, 31, 109 Ribonuclease A (RNaseA), 64, 65 Ribonuclease cleavage of mismatches in RNA/DNA heteroduplexes, 12, 55, 64-65 Ribonucleic acid (RNA) heteroduplexes, 55, 64—65 role in protein synthesis, 26, 45 Risk assessment, EPA guidelines for, 116 RNA, transcribed, 45 RNA/ DNA heteroduplexes, ribonuclease cleavage of mismatches in, 12, 55, 64-65 Rodents, heritable mutation tests in, 81—84 Salmonella, 118 Sentinel phenotypes, 7-8, 37, 38-40, 45, 50, 51, 52, 55, 105, 106, 109; see also Phenotypic data Serine, 27 Sickle-cell anemia, 105 Sister—chromatid exchange (SCE), 99, 116 6-TG, use in somatic mutation studies, 74-75 Somatic cells, 7, 13-14, 27, 31-32, 73, 91-100, 115, 116, 117; see also Somatic mutations Somatic mutations defined, 6 detection and measurement of, 4, 13-14 effects of, 31—32 new methods for measuring in humans, 73-78 see also Mutations Specific locus tests (SLTs), 50, 82-83, 85, 94—96, 97-98, 116 Spermatogonial studies in mice, 84-85, 97-98 Spontaneous mutation rates, 5, 7, 32-33, 51, 83 Squirrel monkeys, 85 Stem cells, in mice, 84, 85 Subtractive hybridization, 12, 55, 66—67 ”Superfund,” 18, 114 Surveillance, in epidemiologic testing, 14, 103, 104-106, 107 T-lymphocytes, 74—76 Technologies, new 144 0 Technologies for Detecting Heritable Mutations in Human Beings for detecting mutations in humans, 10-19, 55-69 development of, and research on, 16—19 and heritable mutations, 3, 10-14, 55-69 use in research and public policy, 14-16 Tester mice, 82-83 Thioguanine resistant (TG’) cells, 99 ‘ Threonine, 27 Thymine (T), 23, 24, 27, 64, 66 Toxic Control Substances Act of 1976 (TSCA), 3, 15, 114, 118 Transfer RNA (tRNA), 26 Tryptophan, 27 Turner's syndrome, 30, 42 Two-dimensional denaturing gradient gel electrophoresis, 55, 61-64 Two—dimensional polyacrylamide gel electrophoresis, 56-57 ZDGGE, see Two-dimensional denaturing gradient gel electrophoresis Z-D PAGE, see Two-dimensional polyacrylamide gel electrophoresis Tyrosine, 27, 47 United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), 113-114 U.S. Congress, 4, 15, 16-19, 113; see also US. Government U.S. Government efforts to protect humans from radiation, 113-114 expenditures on mutation research, 15, 18, 117, 121-122 regulation by, 15, 17, 113-118 role in mutation research, 4, S, 15, 16-19, 113-118 see also US. Congress and specific Federal agencies Validation of measurement techniques in epidemiologic testing, 104 Valine, 26, 27, 29 Viruses, effects on DNA, 23 White blood cells, somatic mutation studies on, 73, 74-76 Whole mammal tests, see Specific locus tests; Heritable translocation tests Wild-type mice, 82-83 Wilms' tumor, 7, 31, 109 Xeroderma pigmentosum, 32 Yeast, 81, 92, 116 502 GENERAL UBBAflY-ILC.BEBKELEY lllllllllllllfllllllll 4‘ BDDDLSQSLN