key: cord-0041792-x997gryp
authors: nan
title: Molecular immunology of virus infections
date: 2004-02-19
journal: J Cell Biochem
DOI: 10.1002/jcb.240530803
sha: f5fdd09062a6e376788d7f55e170029a56cf5cd8
doc_id: 41792
cord_uid: x997gryp

nan

Institute, Boston. 3National Cancer Institute, Bethesda, 4Thomas Jefferson Cancer Institute, Philadelphia.

CD8+ T-cells(TCD8+) play an important role in controlling viral infections. T c D~+ recognize peptides of 8 to 10 residues derived from viral proteins located in the cytosol of infected cells. These peptides are recognized in a complex with class I molecules encoded by the major histocompatibility complex (MHC). Since the processing pathway begins in the cytosol and association with class I molecules occurs in an exocytic compartment proteolysis and peptide transport are thought to play key roles in presentation to T c D~+ Two MHC encoded molecules termed Tap1 and Tap2 seem to specifically trans-port peptides from the cytosol into the intracellular compartment that contains class I molecules. To characterize the structure and function of the Tap genes we have inserted them into vaccinia virus. We have been studying the biochemical, immunocytochemical, and antigen specificity of the transporters in mutant cells failing to express one or both Tap genes. In addition, the endoplasmic reticulum location of peptide association with class I molecules in the exocytic pathway has been demonstrated using a vaccinia recombinant encoding an endoplasmic reticulum retained class I molecule.

Class I1 MHC a-and 8-subunits associate with invariant chain trimers in the endoplasmic reticulum. Following assembly the upinvariant chain complex is transported through the Golgi apparatus and then diverted from the constitutive transport pathway into the endosomal system. Here the invariant chain is proteolytically degraded and ap dimers are released to bind peptides derived from endocytosed proteins. The precise mechanisms involved in peptide generahon and binding, and the route by which ap-peptide complexes are subsequently delivered to the plasma membrane, are unknown. A number of mutant cell MECHANISMS OF CLASS I1 MHC-RESTRICTED ANTIGEN PROCESSING, Peter Cresswell, Howard Hughes Medical Institute, YaleUniversity School of Medicine, New Haven, CT 06510. lines, defective in an uncharacterized MHC-linked gene or genes, are impaired in their ability to generate functional ap-peptide complexes. We have determined that many ap dimers from one such cell line, T2.DR.3, are associated with a nested set of invariant chain-derived peptides, and that such ap dimers can be efficiently loaded with antigenic peptides in oitro. Comparative analyses of class I1 MHC transport, invariant chain processing, and subcellular distribution in wild-type and mutant cell lines will be presented. Most of the seauence diversitv amone. MHC class I alleles is found in the amino acid residues that line the peptidebinding groove. This diversity alters :he chemical composition and spatial properties of the peptide binding groove and in turn dictates the characteristics of the peptide that can be accommodated.

In order to characterize those peptides that bind to specific class I molecules .n vivo, we identifed the major H2-K'restricted peptide from VSV as a unique Jctamer, VSV N52-59, Arg-Gly-Tyr-Val-Tyr-Gln-Gly-Leu. Alanine iubstitued peptide variants were used to define the role of each amino acid -esidue in the octapeptide in terms of its interaction with the H2-Kbmolecule m d with the TCR. As a result of binding studies we postulated that Tyr3, Tyr5 ind Leu8 were MHC anchor residues, while studies using a panel of T cell :lones recognizing VSV/K'complexes on cells suggested that Argl Val4 , Gln6 ind Gly7 were important in TCR recognition.

To evaluate these hypothesis at the structural level we exploited a high yield bacterial expression system and in vitro co-complex formation (protein folding) to prepare a homogeneous MHC class I molecule containing VSV N52-59 Jeptide. This complex was crystallized and its structure solved using molecular -eplacement techniques.

The structure of mouse H-2K revealed its similarity to three human class I HLA molecules, consistent with the high primary sequence homology and ;ommon function of these peptide-presenting molecules. Electron density was 'ocated in the peptide binding groove, to which a single peptide in a unique :onformation was unambiguously fit. The peptide extended the length of the qroove, parallel to the =-helices, and assumed an extended, mostly 0-strand :onformation. The peptide was constrained within the groove by hydrogen bonding of its main-chain atoms and by contacts of its side-chains with the H-2 K b molecule. Its amino terminal nitrogen atom formed a hydrogen bond with the hydroxyl group of Tyrl7l at one end of the groove, while the carboxyl terminal oxygen formed a hydrogen bond with the hydroxyl group of Tyr84 at the other end, amino acids which are conserved among human and mouse MHC molecules. This anchoring of each end of the peptide appears to be a general feature of peptide-MHC class 1 molecule binding and imposes restrictions on its length. The side-chains of residues Tyr3, Tyr5, and Leu8 of the peptide fit into the interior of the K 'molecule with no appreciable surface exposure, a finding in support of previous biological studies that showed the importance of these residues for binding. Thus the basis for binding of specific peptide sequences to the MHC class I molecule is the steric restriction imposed on the peptide side-chains by the architecture of the floor and sides of the groove. The sidechains of Argl, Val4 and Gln6 as well as the main-chain of Gly7 of the peptide are exposed on the surface of the complex, thus confirming their availability for T cell receptor contact, as previously suggested by experiments which demonstrated that a specific subset of these residues were interactive with specific TCRs. The overall picture that arises from our studies is that the TCR/MHC interaction is unique since only 3 to 4 residues of the peptide have sufficient solvent accessibility for TCR interaction, with the majority of the peptide residues being buried. T cell recognition thus depends on only a few of the residues of a peptide presented in the context of the much larger pattern of amino acid side chains of the 2ohelices of the antigen presenting domain of the MHC.

Department, New York University, New York, NY 10003, and 'Pathology Department, New York University School of Medicine, New York, NY 1001 6.

This has been probed with inhibitors of both microtubules (taxol, colchicine) and target of both the humoral and cell mediated immune response of mice and microfilaments (cytochalasianl. As the cold treatment of cells had previously other species t o this pathogen. The cell mediated immune response is been shown t o interfere with presentation, we expected microtubuleorincipally class II MHC restricted and includes both proliferative and cytolytic dissociation or -polymerization to substantially inhibit presentation. This was an effector cells. We have studied the processing and presentation of this protein inconsistent finding, even when the inhibitors were incorporated in assay by murine B cells by both exogenous and endogenous routes. Published studies medium to prevent reversible changes in the absence of drug; the tendency was have demonstrated 1 I an absolute dependence on newly synthesized la to greater sensitivity t o CTL effectors, not diminished recognition. Treatment molecules. a s emetine treatment completely prevents sensitization. 21 of processing and presenting cells with microfilament-interfering drugs did not Acidification of the endosomal compartment is required, whether infectious virus alter their recognition. However, cytochalasian treatment of T cells had or purified protein antigen is employed, a s chloroquine. methylamine. and profound inhibitory effects. This was not seen with either taxol or colchicine ammonium chloride pretreatment prevent uptake and degradation of purified treatment of our clones, however, thus implying a central role for microfilaments protein, inactivated virions, and infectious virus. 3) Integrity of the vesicular (but not microtubulesl in T cell cytoskeleton and lethal hit delivery. We have compartments of the cell and their normal fluidity is necessary, a s Brefeldin A also been examining alternative sites (to late endosomesfearly lysosomesl of treatment reversibly blocks sensitization of cells for CTL clone recognition and protein degradation using 2 forms of the glycoprotein synthesized in the there is a reversible "cold block" in cells treated at 1 5 C 41 A late post-Gobi presenting cells and limited, due t o either temperature sensitive mutation (ts045 block was observed in cells treated with acidification inhibitors, which virus) or t o genetic manipulation (vaccinia expressing a "poison tail' construct, presumably interfere with degradation of the invariant chain of Class I1 MHC.

generously provided by JK Rose, Yalel. These protein variants are degraded We have continued these studies by examining the role of the cytoskeleton in within the lumen of the pre-Golgi endoplasmic reticulum, and peptides generated antigen processing and presentation both for effects on the processing cell and readily sensitize A20 cells for T cell recognition. The site of peptide acquisition on the CTL clone recognizing la'-peptide complexes on the surface of A20 cells. and the maturation of the la complex are under intense study. However, most cancers are not cured by T cell based immunotherapy. We have recently shown that the nonimmunogenic murine tumor, 101 .WT, presented endogenously generated viral antigens in the context of MHC class I poorly, despite the presence of these antigens intracellularly in high quantities. Hypothesizing that this tumor might evade recognition by Tco,+ by failing to process and present tumor antigens in the context of class I molecules, we transduced lOl.WT, with interferon-y (IFN-)I) cDNA to create 101 .NAT. Gene-modification increase class I expression by > 100-fold and reversed the viral antigen presentation deficit of 101 .WT. Significantly, 101 .NAT could be used to generate CD8' TIL that were therapeutically active in vivo against established pulmonary metastases from the wild-type tumor.

To examine whether or not such antigen processing deficiencies were present in human tumor cells, we used a recombinant vaccinia virus expressing mouse the K" molecule.

Because lysis by mouse TCoB+ was our read out, our assay independent of both the HIA type of the tumor, and the presence, or absence, of specific cellular proteins. Human small cell lung carcinoma (SCLC) cell lines, amongst others, processed endogenous antigens poorly. Pulse-chase experiments showed that MHC class I molecules were not transported by SCLC from the ER to the cell surface, suggesting that peptides were not available for binding to nascent MHC and /3,-microglobulin molecules. Consistent with this interpretation, northern blot analysis revealed low to non-detectable levels of mRNAs for MHC encoded proteasome components LMP-7 and LMP-2, as well as the putative peptide transporters TAP-1 and TAP-2. Treatment of cells with IFN-y enhanced expression of these mRNAs. and completely reversed the observed functional and biochemical deficits. In order to bypass the need for the processing of intracellular antigens, we constructed a vaccinia virus capable of endogenously synthesizing a 9 amino acid long 'minimal determinant' from the nucleoprotein gene of influenza A/PR/8/34 preceded by an ER insertion signal sequence, (Anderson et a/, J Exp Med 174:489, 1991) thus eliminating the need for both protease activity and transporter activity. The NP 9-mer preceeded by a signal sequence was very efficiently presented by the SCLC lines to Tcoe+. Thus, antigens, of precisely the right size, targeted to the endoplasmic ( reticulum, can bypass much of the antigen processing machinery.

Coiitrol ofl. 'ir-nl liifecriuri: role of the minority 76 T cell population that is found in the Absence of either the CD4+ or the CDB' populations lungs of influenza-infected mice has been much more difficult to throughout the course of the host response does not greatly assess. Analysis of TCR mRNA profiles, by both in situ modify the pattern of recovery, though depletion of both T hybridization and PCR. has shown that a limited range of TCRl and cell subsets leads to persistence of the virus in lung and TCR6 phenotypes predominates in primarily and secondarily infected greater incidence of mortality. The CDB', H-2Db-restricted H -Z b mice. However, although these lymphocytes are activated to virus-immune T cells are predominantly Va8. 3', combined express mRNA for a variety of cytokine genes and are proliferating, with a range of Va phenotypes. Virus elimination proceeds, their function (if any) is not understood, nice that are however, with apparently normal kinetics when all CD4' and homozygous for a C6 mutant gene have no problems in eliminating the Vp8' T cells are removed by mAb treatment. Mice (H-2*) virus, so this T cell subset is not essential for recovery. Even transgenic for a V@8.l gene also clear influenza virus by s o , Co and Cp mutant mice. though more susceptible, show a variable a CD8' T cell-mediated mechanism, even though Va8.l does capacity for delayed (day 20) clearance of influenza virus from the not define the major CDB' T cell responder phenotype f o r respiratory tract. This is complicated in the 01 mutants set by influenza virus peptide in this haplotype.

There is the presence of a substantial population of TCR@+C-lymphocytes in apparently considerable redundancy in the spectrum of the inflammatory exudate. Hybridomas derived from these cells have possible TCR usage in the influenza-virus-specific CDB' T not been shown to be virus-specific, though they do respond to the cell response. The mechanism whereby CD4+ T cells clear appropriate superantigens. Evidence of a virus-specific response, virus is currently not understood, and could reflect either which discriminates cells infected with influenza A and influenza T cell help for antibody (local?) production or recognition B viruses has, however, been found for both lymph node populations of virus-infected type I1 alveolar cells which express !4HC and a hybridoma derived from the C@ mutant mice.

Superantigens (SAG) represent a new class of T cell stimulating antigens. Mls-1, encoded by the Mtv-7 sag gene, is the prototype of endogenous retroviral SAGS. Although Mls-1 has been functionally defined for over two decades, the biochemical nature of this molecule has remained elusive until recently. Mls-1 induces a strong in vitro proliferative response of T cells expressing the Vp6 or 8.1 chains. In vivo, MIS-1 causes the deletion of immature T cells bearing these particular T cell receptors, as well as the Vp7 or 9 chains. We have cloned and sequenced the Mtv-7 sag gene. Comparison of the deduced amino acid sequence with that of other MMTV sag genes suggests that the polymorphic 3' end encodes the T cell receptor terminus was, therefore, used to generate an Mls-1 specific antipeptide mAb. This mAb reacts with a recombinant baculovirus product of Mtv-7 sag and blocks proliferation of T ceils in response to Mls-1 specific Stimulation. In addition, it recognizes the Mls-1 protein on the surface of LBB.A and Mtv-7 sag transfected cell lines. We have demonstrated that Mls-1 is efficiently presented by the human HLA-DR1 molecule to murtne T cells. Similar to the murine system, a hierarchy has been observed in the ability of various HLA class I1 molecules to present Mls-1. We have generated a soluble Mls-1 molecule which will be used for binding studies to class II and T cell receptor molecules. 

usage with conservation of junctional region residues TCR with influenza A virus (H3 sLbtype) recognise variable regions specific for HA1 48-67, 56-76, 120-139, 186-205, or 177-199 used of the hemagglutinin HA1 subunit that have featured $n antsgenic v 1 in association with four distinct J elements and different drift, and a single Class I1 restriction element ( A (1) and by X-ray crystal structure analysis 01 the NC41 Fab in complex with neuraminidase (2). The crystal structure shows that the antibody contacts 19 amino acid residues on the NA surface which are localized on live polypeptide loops surrounding the enzyme active site. Since the coordinates 01 the structures were not available, we determined the crystal structure of the N9 influenza neuraminidase a priori, and used this structure as a starting point to design experiments to determine the relative importance of each of the contacting amino acids in the interaction with antibody and their role in virus escape from the antibody. Mutant NA genes were COnStNCted and the proteins expressed from these were tested for eflects of the replacement on NC41 binding. Our data revealed that NAs with changes at 368. 400, and 434 completely losl NC41 recognition. NAs with side chains replaced at residues 346 and 373 exhibited binding reduced to less than 50% of wildtype binding. Changes in seven other contacting residues, including substituted side chains which ditfered considerably lrom wildtype NA in size and charge, had IX) sgnaicant effect on NC41 binding. These resuns indicate that only a few of the many residues which make up an epitope are crucial lor interaction and provide the critical contacts required lor antibody recognition. This implies that antibody escape mutants are only selected if they contain changes at these crucial si1es.p) NUSS. J.. P. Whitaker. and G. Air. 1993 . Identification of critical contact residues in the NC41 epitope of a subtype N9 influenza virus neuraminidase. Proteins: Structure, Function and Genetics in press.

Rika Ishikawa. Helen C. Su, and Jordan Orange NK ceils respond early whereas T cells respond late dunng viral infections.

Our labaratory has been studying me regulation of these populations dunng experimental infections 01 mice with either lymphocytlc choriomeningitis virus (LCMV) or munne cytomegaiorirus (MCMV). The endogenous production of cylokines and lheir biological functions lor regulating acute immune responses are being characterized. Akhough interieukin-2 (IL-2) IS capable of tnduang NK ceil proliferation. ii does not appear to be a major mediator of the NK cell responses to nrai infecuon as the peak IL-2 production during infmon mrrelates wkh late T cell responses and as NK cell activaiion and proliieration are resistant to cydosporin A. in viw NK cell proliferation does correlate with the production of virus-induced inlerlerons (IFNs) and lhe NK cell response induced during LCMV infenion can be blocked by treatment wivl antibodies neunalizing IFNs a/B. NK and T celis are bcth induced to express high levels of 1FN-y at Qmes of peak acwalion during infecuon.

factors-I3 (TGF-E) are also produced a1 eariy times coinading with lhe decline of NK cell aclivatiw and lhe inaease in T mll acbvanon dunng miemon. Experiments evaluating the sendliviiviry of the virus-induced lymphocyte responses to TGF-pmediated inhibition have demonstrated that NK cell prolieration is 100-fold more sensilive lhan is T call proliferation. Taken topether. these stude~ damonswate lhat a variety of qlokines are induced dunng infenion and that l h e~e wokines can act 10 mordinate m e NK and T call prolierative responses.

Cellular interaclwns regulating these iymphocyles and their regional kalization were charanerized in tissue M m s prepared from infected mice. Histological Mice with a neor insertion in their p2 microglobulin genes lack p2 Antigen presentation in these animals may not be via the microglobulin, and hence cannot produce normal Major classical endogenous pathway. We have shown that class II+ Histocompatibility Complex class I proteins on their cell surfaces. target cells can be sensitized for lysis by incubation with This lack of class I expression during T cell ontogeny results in inactivated LCMV and that the presentation is sensitive to the failure to positively select substantial numbers of ap T cell chloroquine. suggesting that shed virus is important in providing receptor, CD8+ T cells in peripheral lymphoid organs. Normal targets for CTL and presentation may occur via the exogenous mice are susceptible to lymphocytic choriomeningitis virus when pathway.

IMMUNE RESPONSES OF P2-MICROGLOBULIN DEFICIENT MICE, Jeffrey A. Frelinger, Department of Microbiology and Immunology, University of North Carolina at Chapel Hill, Chapel Hill, NC 25599 challenged intracranially because they produce CD8+ cytotoxic T lymphocytes (CTL) which can react with virus infected cell in the central nervous system, and ultimately cause encephalitis and death. In contrast, p2 microglobulin deficient mice which do not oroduce sionificant number of CD8+ CTL. were exoected to be

We have also studied the response to Lisferia rnonocytogenes.

p2 microglobulin deficient mice can respond to Listeria. and are resistant to lethal infection in vivo, although these mice do have higher levels of bacteremia early.

ksistant toLCMV challenge as are acutely CD8 depleted mice. Surprisingly, these mice died from LCMV disease, although more slowly than normal mice. We further showed that these mice produced CD4+ CTL able to lyse LCMV infected targets only when the target cells expressed class II as well as class I on their cell surface. We extend these results to examine the kinetics of induction of CTL and of NK activity. In accord with the kinetics of LCMV disease induction, we have found that CD4+ CTL are induced in p2 microglobulin deficient mice more slowly than CD8+ CTL in normal mice. When both signals are provided, for example when dendritic cells present antigen, the T cells undergo multiple rounds of cell division in response to the IL-2 that they produce, returning eventually to a resting state in which they are ready to respond to antigenic stimulation once again. We and others have shown that the T cell surface molecule CD28 transduces a costimulatory signal either when occupied by monoclonal antibodies or by its ligand, the APC surface molecule B7. CD28 crosslinking greatly enhances the amount of IL-2 and other lymphokines p d u c e d by T cells stimulated through their TCRs. CD28 signal transduction involves a cyclosporin A-resistant biochemical pathway distinct from the hydrolysis of inositol phospholipids.

If Thl cells do not proliferate following TCR signaling they become unable to produce IL-2 in response to subsequent antigenic stimulation. Unresponsiveness (also called anergy) results when the TCR is occupied and CD28 signaling does not occur, for example when B7negative APCs present antigen or when the CD28/B7 interaction is blocked by an uncrosslinked anti-CD28 antibody. Anergy can also be induced when the TCR is occupied and subsequent cell division is

These experiments together argue for considerable plasticity of the immune system. CD4+ T cells which are only infrequently found in the cytotoxic population are recruited for CTL activity in mice where very few competent CDB+ T cells are present. This indicates that in the absence of CD8+ T cells, CD4+ clones, which would normally be displaced by higher affinity CD8+, cytotoxic T cells, can expand. These mice are thus extremely interesting for the determination of the ultimate functional repertoire of CD4+ cells.

T CELL ANERGY, Marc K. Jenkins, Julia Johnson, Besty Kearney, inhibited. This occurs when Thl cells are stimulated with peptide antigen and APC in the presence of anti-IL-2 and anti-IL-2 receptor antibody or agents that inhibit IL-2 responsiveness such as rapamycin, genistein, or PGE2. Chronic stimulation of Thl cells with anti-CD3 antibody when APC are present results in IL-2 production. however the capacity of the T cells to proliferate is inhibited by an unknown mechanism. Anergy is also induced under these conditions. Therefore, based on these results the optimal situation for lymphokine production, clonal expansion, and retention of subsequent responsiveness by T h l cells is aansient anngenpresentation by a B7-expressing APC.

Using a differential screening approach we have recently shown that although T cells stimulated with anti-CD3 antibody in the absence of APC produce very little IL-2 mRNA, they d o produce large numbers of uanscripts that encode macrophage inflammatory protein-la and another unknown member of the small cytokine family. The accumulation of neither of these transcripts is enhanced by CD28 costimulation. Although the functional significance of these findings is presently unclear it is possible that these cytokines may be involved in the unresponsive state either by preventing T cell responsiveness to IL-2 or by modifying the costimulatory properties of APC. To investigate the role of immunologic tolerance mechanisms in chronic infection of the newborn, we have generated HBeAg-expressing transgenic mice (BIO.S/e31). These mice were tolerant to both HBeAg and the nonsecreted HBcAg at the T-cell level. Furthermore, nontransgenic littermates born to HBeAg-expressing mothers showed reduced T-cell responses to HBc/HBe antigens, suggesting that tolerogenic HBeAg may transverse the placenta. Tg mice did not produce antibody to HBeAg but did produce IgM antibodies to HBcAg via a T cellindependent pathway. The coexistence of tolerance to HBc/HBe T cell determinants and production of antibody to HBcAg in vivo parallel the immunologic status of neonates born to carrier mothers. The maintenance of T cell tolerance to HBc/HBe antigens required the continued presence of the tolerogen and in the absence of HBeAg persisted for less than 16 weeks. The reversibility of T cell tolerance to HBc/HBe antigens may explain the inverse correlation between age of infection and rates of viral persistence. These observations suggest that a function of the HBeAg may be to induce immunologic tolerance in ulcro. Expression of HBeAg may represent a viral strategy to guarantee persistence subsequent to perinatal infection. Further studies in F I hybrid Tg mice (BIO x BIO.S/e31) illustrated that "self tolerance to HBeAg is variable depending on the MHC genotype. A dominant T cell site on HBeAg (~120-131 and I-AS-restricted) is tolerogenic, w ereas a proportion of T cells recognizing p129-140 in the context of I-A evade induction of tolerance, persist in the periphery, and can be activated in vivo by a single injection of the 12 residue T cell selfpeptide. Furthermore, the self-reactive T cells can cooperate with self-' b ~~ reactive, HBeAg-specific B cells to mediate in vivo production of autoantibody sufficient to neutralize detection of the autoantigen in serum. This model illustrates that T cells specific for an immunnogenic T-cell site on a nonsequestered autoantigen can escape induction of tolerance (i.e., are not deleted or anergic) and, more importantly, can mediate autoreactivity in vivu. These murine studies suggest that chronic HBV (CH-B) carriers may also possess "quiescent" T cells that have evaded tolerance induction. In order to examine the relevance of this murine model, 200 HBeAg-positive CH-B carriers with varying degrees of liver disease (ie. CAH, CPH, ASC) were analyzed with novel and sensitive serological assays capable of detecting serum anti-HBe and anti-HBs antibodies regardless of the simultaneous presence of their respective antigens. All 200 patients were seronegative for anti-HBs and anti-HBe by commercial assay. This analysis revealed:

(1) virtually all CH-B patients with liver disease and approximately 50% of CH-B patients without liver disease demonstrated ongoing humoral immune responses specific for HBeAg, HBsAg, and pre-SAg in addition to variable responses to HBcAg; (2) three serologic profiles of CH-B infection were identified; (3) these "silent" immune responses may occur for a number of years prior to liver disease or viral clearance; (4) the quantitative or qualitative characteristics of the immune responses correlate with clinical status; and ( 5 ) this array of humoral immune responses appears to be mediated exclusively by HBeAg-specific T helper cells. These results redefine the serology of CH-B infection. The serological data is also consistent with the hypothesis that T cell tolerance to HBeAg plays a role in chronicity, and that HBeAg-specific T cells emerging from the tolerant state mediate liver injury and HBV clearance in CH-B infection.

N 016 IMMUNE REGULATION BY T CELL CYTOKINES, Edmonton, Alberta T6G 2H7, Canada. The immune system can respond to infectious agents by a variety of effector mechanisms, each of which is appropriate for different types of pathogen. The decision between these responses is under saingent regulation, and part of this control is mediated by the cross-regulation of T cell subsets secreting different cytokine patterns. Two major T helper subsets are THl and TH2 cells, which differ markedly in the patterns of cytokines that they secrete after antigen stimulation. These patterns are mainly responsible for the different functions of the two subtypes. THI cells induce an inflammatory response, including activation of granulocytes and macrophages. In contrast, TH2 cells are excellent helpers for B cell antibody production, and in the absence of a significant Interferon y response, TH2 cells induce a strong allergic response due to secretion of IL4, ILS and ILlO. Other T cell subtypes also exist, but the functions of these cells, and the interrelationships between them, are less well understood.

The functions of THl and TH2 cells are often reciprocal, especially in strong immune responses such as occur during many parasite Tim Mosmann, Department of Immunology, University of Alberta, infections. Subset-specific cytokines are involved in this regulation: ILlO inhibits THI responses at the level of macrophage activation and indirectly, by inhibiting THl cytokine production; IL4 enhances differentiation of TH2 but not THI cells; and IFNy inhibits the proliferation of TH2 cells. We have recently been studying the functions of another TH2-specific cytokine, P600, which was initially characterized as an inductionspecific cDNA clone isolated from TH2 cells. We have now expressed recombinant P600 protein and characterized the functions of this cytokine. P600 does not have activity in many of the assays in which TH2-specific cytokines are active. However, P600 enhances the production of an adherent cell population from bone marrow precursors. The resulting cells have high expression of the macrophage markers MAC1 and F4/80. Although these cells are not phagocytic for antibody-coated erythrocytes, they are effective antigen-presenting cells for T cell clones specific for particulate, soluble or alloantigens.

The repertoire of the cytotoxic lymphocyte response to infection was studied in mice infected with lymphocytic choriomeningitis virus (LCMV). At early stages of infection, LCMV stimulated activated NK cells and non-MHC-restricted cytotoxic T cells (CTL), whereas at later stages both virus-specific and allospecific CTZ were induced. Antiviral activity of the NK cells and of non-MHC-restricted CTL was shown in vivo against murine cytomegalovirus but not against LCMV, whose replication in vivo was controlled by virus-specific CTL. Concomitant with the induction of the virus-specific CTL was a 1040-fold increase in the number of CD8' T cells per spleen and a deficiency in the ability of the T cells to proliferate in response to mitogens. The T cell number subsquently declined, and responsiveness to mitogens returned. Limiting dilution analyses of CTL generated in vivo revealed a high frequency of virus-specific CTL cross-reactive with alloantigens. Activated Cn and subsequent memory CTL expressed CDllb and IL-2 receptors. The activated CTL isolated from mice during acute infection were triggered to undergo apoptosis upon TcR/CD3 cross-linking. Culture of activated (CTL) response to HBV encoded antgens is thought to be responsible for liver cell injury and viral clearance during HBV infection. Using a strategy involving an initial stimulation of peripheral blood mnonuclear cells (PBMC) with HBVderied synthetic peptdes, followed by restimulation with HLA class I matched stable HBV transfectants. we have recently demonstrated. at the clonal level, that most patients with acute viral hepatitis develop a polyclonal HLA class I restricted CTL response to multiple HBV encoded antigens. In contrast. the response is not detectable in patients with chronic HBV infection. Our resuns may be summarized as follows. First, we have identilied a nucleocapsid epitope (STLPETWVRR) that is dually restricted by the HLA-A31 and A H alleles. We have shown that both of these class I alleles recognize precisely the same sequence and that they both bind to the same agretopic residues, while the T cell receptors of independently derived CTL clones recognize different epitopic residues. Second. we have identified ten HLA-A2 restricted CTL epitopes within the viral envelope. nucleocapsid. polymerase and X proteins. and we have shown that these epitopes contain the HLA-A2.1 allele specific binding mtif described by Rammensee's group. Third, we showed that the HBV specific CTL response in individual patients with acute viral hepatitis extends to multiple viral epitopes and different restriction elements; i.e., it is polyclonal. multispecifii and polytmrphically restricted. thereby minimizing the chances that CTL "escape mutants" will emerge in this setting. Fourth, we showed that there is a hierarchy of responsiveness to the HLA-A2 restricted epitopes in infected patients; ie two Of the epitopes (core FLPSDFFPSV and envelope WLSLLVPFV) are recognized by the majonty of infected patients while the remainder are seen by a minority of the patients we have studied thus far. This hierarchy does not correspond to the HLA-A2.1 binding affinity of the peptides, and therefore is probably determined at the level of v i m sequence, antigen processing or the T cell repertoire. Finh. we demonstrated that CTL nonresponse to a particular peptide sequence in an otherwise responsive patient may reflect infection by a virus that expresses a variant sequence in the region of the epitope. For example, several patients who failed to respond to the immunogenic envelope peptide GLSPTVWLSV were shown to be infected by a variant in which the C-terminal valine anchor residue was substituted by an alanine Interestingly, the variant peptide was found to bind as well to HLA-A2.1 as the prototype peptide, and the variant peptide itself was immunogenic in at least one patient who was infected by the variant virus. Thus, it would appear that the conservative substitution of an alanine for a valine at the C-terminal anchor residue may have paradoxically affected the ability of the peptide to be recognized by the CTL receptor mre than it affected its binding to the HLA-A2 molecule. Finally, several of these epitopes overlap with functionally and immunologically important domains wlthin the viral nucleocapsid and envelope proteins. For example, the H U -A 2 restricted envelope epitope FLLTRlLTl contains the core residues (RILTI) of an HLA DFw4 restricted helper T cell epitope previously described by Celis, et at. Additionally. 2 other envelope epitopes overlap wlth an important topogenic sequence within the transmembrane orientation of HBsAg. and the dually reStricted nucleocapsid epitope described above includes sequences involved in the nuclear localization and genome encapsidation functions of the nucleocapsid protein. The generation of a CTL response to such epitopes might yield a double benefit to the host: first, by destroying the infected cells and, second. by precluding the emergence of viable mutants. A possible by-product of this research may be the rational design of CTL-based therapeutic strategies to terminate persistent HBV infection. The same strategy can be applied to study the CTL response to any human pathogen whose genome has been cloned and sequenced. i Cytomegalovirus (CMV causes lethal disease in the immunocompromised host whereas infection of the immunocompetent is usually asymptomatic. Yet, even in presence of normal immune control, episodes of productive infection do occur and cause the horizontal transmission of the virus. Thus, the physiological antiviral immune response protects vilal tissues but executes a less stringent control over others. By studying the infection of the mouse with the murine CMV (MCMV) we wish to unravel the relative contribution of the specific immune effector mechanisms involved.

Cell transfer experiments showed that only CD8' T lymphocytes dominate in the protection against lethal infection This subset appears not to require CD4+ T lymphocytes for activation because CD4' ce!l deficient mice generate a comparable CD8+ response and control the infection. The productive infecticn in the salivary gland, however, remained completely refractory to the activity' of the CD8' T lymphocyte subset. Despite of high titers in this organ the CD8+ T lymphocytes prevent the spread to other tissues. Cell transfer studies and experiments with CD4+ deficient mice showed that in the salivary gland CD4' T lymphocytes are essential effector cells, and that they can operate in absence of the CDB' subset. Thus, in the normal host, clearance of vital tissues and control of horizontal spread is under control of different T lymphocyte subsets. CD4+ subset control of salivary gland infection requlres the activiv Of TH1 cells. This became evident from the blockade of effector function after administration of antibodies against IFN gamma. IFN gamma. in concert with TNF, represents a powerful inhibitor, active in the late phase of the Viral replication cycle. The molecular identification of the major targets for the CD4' subset are under study. Analysis of the specificity of the CD8+ subset led to the definition of the naturally processed peptide derived from the dominant antigen, an immediate early protein. This nonameric peptide, when flanked by appropriate sequences in a carrier protein can serve as a vaccine.

Remarkably, viral functions of the early phase prevent presentation of this peptide. This inhibition is due to the retention of the nascent peptide loaded MHC complex at the EWcis Golgi level. The progress of the immune response counteracts this evasion mechanism because antigen presentation by cells pretreated with IFN gamma are capable to export the peptidelMHC complex.

IMMUNOPATHOGENESIS OF OCULAR HERPETIC, B a r r y T. Rouse, The most dramatic immunologic effect of infection by the human immunodeficiency virus (HIV) is the severe depletion of CD4+ T helper cells (TH). However, a complex pattern of TH dysunction occurs before CD4+ cells fall to critical levels, such that TH function, assessed by interleukin 2 (IL-2) production, is first lost to recall antigens (including to HIV antigens), followed by loss of response to HLA alloantigens, and finally by lack of IL-2 production in response to PHA. The loss of recall antigen-stimulated IL-2 production occurs concomitant with the loss of PHA-stimulated IFNy production. suggesting a lack of TH-1 function. We also noted sequential increases in PHA-stimulated IL-4 and IL-10 production, suggesting an increase in TH2 function. Because these events are predictive for the onset of AIDS, it appears that there is a TH-1-TH-2 switch in the progression toward AIDS. Furthermore, we independently observed that PBMC from a high proportion (45-75%) of seronegative, HIV-exposed individuals exhibit strong TH 1 activity to HIV env-synthetic pepides. The seronegative, HIVexposed groups studied included: gay men; IV drug users; health care workers; and newborn infants of HIV+ mothers. Only 2% of presumed unexposed individuals exhibit similar TH activity. Considered together, these two seemingly unrelated sets of data suggest that immunoregulatory cytokines contribute to TH dysregulation after HIV infection, and that HIV-specific TH-1 type immunity may be protective against HIV infection.

N 023

Institutes of Health, Bethesda, MD 20892 USA. Recent advances in the understanding of antigen processing and presentation facilitate the identification of antigenic sites recognized by CD4+ helper T cells and by CD8+ cytotoxic T lymphocytes (n). and suggest methods to optimize the activity of these for consmction of synthetic peptide vaccines aimed at eliciting T-cell immunity. We have characterized the functional role of each residue in an HIV envelope peptide recognized by helper T cells from mice and humans. Only a few of the amino acid residues were necessary for binding to class I1 major histocompatibility complex (MHC) molecules and immunogenicity. as shown using a peptide consisting of only four of the original residues, with the rest replace by Ala The other residues were more remarkable for the negative impact they could have. Several of the non-critical residues could play a deleterious role if replaced with the wrong substitution, and removal of one negative charge in the natural sequence led to a peptide 100-fold more potent for stimulating specific T cells and for binding to MHC molecules, emphasizing the importance of adverse interactions in peptide-MHC binding. Similarly for a (TIZ site presented by multiple murine and human class I MHC molecules, (JIZ specificity was shown to focus on the dishnction between aromatic and aliphatic side chains at one position. This result, in turn, allowed the construction of chimeric peptides that induced broadly crossreactive responses for multiple strains of HIV-I, and that may reduce the risk of outgrowth of escape mutants, Thus, understanding the molecular basis for MHC binding and Another problem in the design of a synthetic vaccine is the polymorphism of the MHC, leading to differential recognition of different antigenic determinants of the same protein by individuals of different MHC types. One approach to overcome this problem that we have used for helper T-cells specific for the HIV-1 envelope protein was to identify multideterminant regions encompassing several overlapping determinants presented by different MHC molecules. Synhetic peptides spanning such multideterminant regions elicited T cell responses in multiple strains of mice and in humans of multiple HLA types. An alternative approach is to identify truly promiscuously recognized antigenic determinants. We have found one HIV CTL determinant to be presented by four different class I MHC molecules in mice and by at least two different class I molecules in humans. At least in mice, the same minimal core 10-residue sequence is presented by all four class 1 molecules. These antigenic sequences can then be coupled to produce a synthetic vaccine eliciting neutralizing antibodies and CTL in animals (or humans) of multiple MHC types.

Thus, analysis of the T-cell responses to viral proteins can conmbute to the rational design of new vaccines. Similar approaches can be applied to tumor antigens recognized by T cells, including mutant oncogene products and viral oncogene products as potential targets for vaccine prophylaxis and immunotherapy of cancer. ligands. Several interesting aspects have been noted and require nonapeptides with two positions conserved, will be discussed: for example. Qa-2 molecules are peptide whereas the other positions can be occupied by various receptors of higher stringency than ordinary class I amino acid residues. This information can relatively easily molecules, and A l l molecules obviously can accommodate be obtained by immunoprecipitation of the MHC molecule of peptides from 8 to 1 1 residues, always with a charged Cinterest. elution of the associated peptides. and sequencing terminus. For determining class I1 peptide motifs, a useful of those peptides all together as a pool. This laboratory has way appears to be to use both pool sequences as well as a used that approach for several MHC molecules, including H-few individual peptides. Participation in this work of the 2 Kd, Ld, Dd, Kb, Db, Kk, K b l ; Qa-2; HLA-M ('0201 and following persons is acknowledged: M. Takiguchi Normal moiise fibroblasts are resistant to the cytolpc effects of tumor necrosis factor iTNFi. Hoiseier. expression of the adenovirus EIA protein renders these normall! resistant cells censiti to TNF lysis. TNF sensitiiit) is obsened both when EI.4 IS introduced into cells h) acute iims infection as a e l l as b! transfection and production of stable cell lines. A variety of acti\ities has been ascribed to EIA. including transactivation of viral and some cellular genes and repression of expression of other cellular genes. EIA transforms cells i n cooperation with activated ras. The abilit) of EIA to effect transformation appears t o require its binding to several cellular proteins including retinohlasroma (pRB). Sequence anallsis of multiple adenovirus serotypes has identified three conserved regions of the molecule (termed CR I . CR2 and CR3 I. Mutational analjsis has located several EIA functions to one or more of these regions. Hence. identification of EIA regions responsible for induction of TNF sensitivity could provide information about the mechanism b! w hich EIA influences the complex series of intracellular eients leading to TNF cytolysis. Previous studies using acute infection with A panel of adenoiiruses containing deletion mutants within EIA had implicated onl! residues within CRI in induction of TNF sensitivity. Surprisingl>. when these same mutants were employed as plasmids and transfected to produce stable cell lines. deletion of either the CR! or CR2 region resulted in loss of TNF sensitibit!. indicating that both regions are required. This finding suooests that two EIA functions are necessary for TNF sensitiiit!. one of 2 k -h IS provided hj CRI. and the other by either CRZ or another gene in adenoiirus. E\idence that another geneis) in adenovirus can function to alter cellular susceptibility to TNF comes from the following obserration: Although neither SV40 T antigen nor the bovine papilloma virus (BPV) early region alone alters cellular susceptibility to TNF. both complement adenovirus mutants lacking all of EIA in inducing TNF sensitivity. Thus. T antigen. BPV. and a non-EIA adenovirus gene can all contribute to the cellular phenotype. with the suggestion that T antigen and BPV provide one function (the "CRI" function) that complements the "CRZ" function elsewhere in adenovirus. Furthermore, i t appears that the two regions of EIA can function independently to induce TNF sensitivity. This suggestion comes from transfection studies in which mutant proteins lacking either CRI or CR2 were introduced separately or together into mouse fibroblasts and stable cell lines developed. Although cells expressing either mutant alone remained as resistant to TNF as the parent NIH3T3 cell line, cells expressing both mutants become nearly as sensitive to TNF as cells transfected with the intact EIA gene. This complementation again suggests that two separate functions. acting in rrum, are responsible for induction of cellular sensitivity to TNF lysis. Dyson rr a/ (J. Virol. 66.406 i 1992)) have recently shown that EIA makes two distinct contacts with pRB. one involving residues within CRI and one within CR2. and that both regions are necessary for the highest affinity binding betueen EIA and pRB to occur. Furthermore. synthetic peptides derived from CRI and CR2 were found to hind independently to pRB. Thus, the ability of EIA to occup) two independent binding sites on pRB correlates with its abilit! to create the TNFsensitive cellular phenotype. The Netherlands. In C57BU6 (B6.H-2') ndnu mice large tumors arise after subcutaneous implantation of syngeneic Adenovirus (Ad) E l transformed embryo cells. In previous studies we showed that large tumor masses can be eradicated by i.v. infusion of cloned CTL directed against an immunodorninant AdBElA encoded peptide in Combination with interkukin-2 (1.2). The E1A peptide recognized is a 10 mer comprising amino acid sequence 234-243, SGPSNTPPEI (1,2). Thi s peptide is presented by the H-2Db MHC class I molecule and has a characteristic H-2D' binding ma8 with N and I anchor amino acids at postions 5 and 10, respectively (2,3). Cell lines transformed by mutant EIA constructs were generated with T -I, T -K and position 8 P -S mutations in the immunodorninant CTL epitope. Ten-mer peptides containing these mutations all bound equally well to the H-2Db molecule. CTL recognition and MHC binding studies w l h cell lines carrying these mutations in addaion to amino acid replacement studes with the Ad 234-243 peptide confirm the notion that position S(N) and lO(1) are cruclally important for Do binding. whereas CTL recognition is abolished by most alterations at positions 6-9. CTL clones generated against the cell lines with El transformed lines carrying the afore mentioned mutations in the immunodominant C n epitopedd not recognueE1 encodedpeptides. lmteadthese CTLs appear to recognize an E1A induced peptide encoded by a normal cellular gene. The expression of t h i s C n eptope in contrast to the E l A endoced epitope 234-243, is shut off by introduction of an activated H-ras gene. Interestingly ras transfened El transformed cells in contrast to parental El transformed cell without activated ras. are tumorigenic in immunocompetent B6 mce. This could be the result of lossof ElAinduced but not ElAencodedCTLepitopeexpression. combined with TGF beta production by the El + ras transfected cell lines (see also abstract of R.

Offringa et al. Virginia 22908, U.S.A. Epstein-Barr virus (EBV) transforms human B cells in vitro to give lymphoblastoid cell lines (LCLs) which constitutively express a limited number of viral proteins: these include the nuclear antigens EBNAs 1, 2 , 3A, 3B, 3C and -LP and the latent membrane proteins LMPs 1. 2. Such LCL-like cells also arise in vivo but their outgrowth is contained by the cytotoxic T lymphocyte (CTL) response. Thus healthy virus-carrying individuals possess strong CTL memory to EBV which can be reactivated in vitro by stimulating with autologous LCL cells: these reactivated cells recognise LCL targets specifically and in a HLA class I-restricted manner.

Using recombinant vaccinia viruses expressing the individual EBV latent genes as a means of sensitising target cells in CTL assays, the dominant target antigens presented by a variety of different HLA class I alleles have been mapped. Interestingly, the EBNA3A, 3B. 3C family of nuclear antigens provide the dominant epitopes on a number of HLA backgrounds and several of the relevant peptide epitopes (for instance, those presented by HLA-A11 and by several HLA-B27 subtypes) have been precisely defined.

Many of the EBNA proteins, including the EBNA3 family, are not expressed in some of the more important EBV-associated malignancies. In particular Burkitt's lymphoma cells only express EBNA1, whilst nasopharyngeal carcinoma cells and the malignant cells of EBV-positive Hodgkin's Disease express EBNAI, LMPl and LMP2. Current efforts therefore concentrate on CTL epitopes present in this subset of latent proteins. Of particular interest here is an epitope of LMP1, now defined at the peptide level, which is conserved in all EBVisolates studied to date and which is presented by HLA-A2.1, a common allele in many human populations. The frequency of CTL memory to this epitope in A2.1-positive individuals is now being assessed, as are means of selectively reactivating this memory population in vitro. Collins', Herbert Morse III', and Mark Connors', 'Laboratory of Infectious Diseases and *Laboratory of Respiratory syncytial virus (RSV) is the most serious viral respiratory tract pathogen of infancy and early childhood. A hallmark of RSV infection in humans is its ability to cause repeat infections. Three major mediators of immunity to RSV have been identified, namely serum IgG antibodies to the RSV fusion (F) and attachment (G) glycoproteins, nasal wash IgA antibodies, and CD8' cytotoxic T-cells. Reinfection with RSV results from inadequacies in one or more of these arms of the immune response. These inadequacies reflect one or more of the following: immunological immaturity at the time of first infection, antibody-mediated suppression of the antibody response to first infection by transplacentally transferred maternal IgG serum antibodies, or waning IgA or CD8* T-cell immunity. For example, the CD8* T-cell immune response is highly effective at restricting replication of RSV challenge virus on days 6-9 following immunization of mice with an immunogen inducing mainly CD8* T-cell immunity, but is ineffective at restricting replication or at accelerating clearance of virus 45 days after immunization. Repeat infection with RSV also reflects the inherent high infectivity and olte Abstracts virulence of this virus as well as the existence of two antigenic subgroups that are about 25% related by cross-neutralization analysis.

A central problem with RSV immunization has been the observation that immunization of infants and young children with formalin-inactivated RSV (FI-RSV) not only failed to protect against RSV infection but actually potentiated the illness in those vaccinees subsequently infected with RSV. An analysis of the immunological basis of disease potentiation revealed that FI-RSV vaccine created a series of specific immunological imbalances including induction of antibodies with decreased ability to neutralize RSV, augmentation of CD4* T-cell response, and a decreased or absent CD8' T-cell response. Furthermore, depletion of CD4*, but not CD8*, T-cells in PI-RSV immunized mice at the time of RSV challenge abrogated inflammatory cell infiltration in the lungs of mice. Considered together, our findings suggest that the potentiated disease in humans is a delayed type hypersensitivity reaction at the sites of pulmonary virus replication. Immunization with RSV subunit and vaccinia virus vectored vaccines will be discussed in this context. The site of envelopment at the inner nuclear membrane determines for herpes simplex virus (HSV) the need to be transported vectorially to the extracellular compartment. During this transit the envelope glycoproteins are processed to the mature forms. In addition, HSV has evolved a specific function at the plasma membrane that prevents reinfection of infected cells by progeny virus, a phenomenon designated as restriction to superinfection.

In the past few years it emerged that HSV transport can be dissected in two major steps that require viral and cellular functions. The first step consists in the virion sorting from the space between the inner and outer nuclear membranes to the cytoplasm and occurs by pinching off the virions in vacuoles derived from the outer nuclear membrane. This step requires the viral protein UL20, in that a HSV mutant deleted in the ORF UL20 remains blocked at the space between the inner and outer nuclear membranes. As the block only takes place in some cell lines, some cells can complement the ORF UL20 deletion in the viral genome. UL20 is a multiple transmembrane protein that appears to be located in the virion itself. The results support the view that UL20 belongs to a family of viral and cellular membrane proteins involved in intracellular transport. For the subsequent steps in the vectorial transport, HSV exploits the exocytic pathway. Studies with mutant cells defective in Golgi enzymes or with a variety of Golgi inhibitors indicate that virion particles interact with the Golgi apparatus, and this interaction is crucial for virus exit out of the cell and concomitantly for processing of the virion envelope glycoproteins to mature forms. Immunocytochemical location of Golgi-resident proteins coupled with electron microscopic examinations of HSV-infected cells reveals a fragmentation of Golgi apparatus and a dispersal of the fragmented elements throughout the cytoplasm. This occurs at the time of virus maturation and is peculiar to some cell lines. Restriction to superinfection is mediated by the presence of the viral glycoprotein D (gD) at the plasma membrane. In gD-cells the superinfecting virus is taken up by endocytosis and degraded. Surprisingly, HSV mutants able to overcome the gD-mediated restriction and to infect productively the gD-expressing cell lies were found to be mutated in gD itself. These findings raise the possibility that the block to superinfection observed in gDexpressing cells is exerted through an interaction between cellular gD and virion gD. Isolates or HIV-I or HIV-2 differ in their biologic behavior in cell culture with regard to cell tropism. replication rate. ability to induce cytopathic changes. and modulation of CD4 expression. We have compared the following series of molecularly cloned HIV-I isolates for their ability to causr deletion of CD4 T cells in hu-PBL-SCID mice:

HIV-Isnj -T-cell tropic, high replication. cytopathic. CD4 modulation:

HIV-1~~1 6 2 -Ma-tropic, low replication, non-cytopathic. no CD4 modulation; HIV-1 SF? -T-cell tropic, medium replication. cytopathic.

CD4 modulation: and HIV-1~~1 3 (a later isolate from the same patient as SF?) -T-cell and MO-tropic. high replication. cytopathic, CD4 modulation. We extended these studies to two HIV-2 isolates: H l v -2~~

-T-cell tropi. high replication; and HIV-Zucl -MO-tropic, low replication. non-cytopathic, no CD4 modulation. Hu-PBL-SCID mice were infected with 100 tissue culture infectious doses of each of these HIV, and C D 4 , CDX-. CD3-. and CD45-positive human cells were enumerated by flow cytometry at 2 and 4 weeks after infection. In addition. viral load was determined by quantitative PCR analysis of proviral copy number and by plasma p24 core antigen levels. The macrophage-tropic. non-cytopathic HIV-1SF162 isolate induced the most rapid and profound CD4 T cell depletion, and the most highly cytopathic strain in virro. H I v -1~~3 3 , induced the least and slowest CD4 T cell depletion. The relative ability of HIV-1 isolates to induce CD4 T cell death was SF162>SFI3>SFZ>SF33. This ability was not correlated with viral replication, as the relative viral burden at 4 wk after infection was SF33>SF162>SF13=SF2. These results were confirmed when we examined the HIV-2 isolates. Infection with HIV-~UCI led to rapid and extensive CD4 depletion (comparable to HlV-1~~162). whereas intection with H l v -2~~ did not result in a detectable decline in CD4 T cell numbers. We conclude that the behavior of HIV isolates in tissue culture may not predict their pathogenic potential in vivo, and suggest that macrophage-tropism may be a marker for a more highly cytopathic effect (either directly or indirectly mediated) for human CD4 T cells in HIV-infected individuals. In addition, viral load may not be an accurate predictor of disease progression. (Supported by NIH grants A129181 and AI30238). Negative strand RNA viruses are a group of viruses whose genomes consist of single-stranded RNA of negative polarity. Since the purified RNA of these viruses is not in itself infectious when introduced into cells, novel methods have had to be develcped to allow the genetic manipulation of these viruses. Recently! a ribonucleoprotein transfection protocol was designed which allows the rescue of influenza virus cDNA-derived RNA into infectious influenza virus. Thus, taking advantage of DNA technologies, it is now possible to alter the genome of at least one negative strand RNA virus and obtain viruses made from synthetic RNA. Introducing mutations into the coding region of influenza virus genes allows a detailed structure function analysis of influenza virus proteins. This approach is at times Limited due to the fact that lethal mutations do not provide insight into the mechanism of the inactivation. However, it can help us to learn whether a feature is essential for infectious virus formation. For example, studies involving a mutational analysis of the cytoplasmic tail of the hemagglutinin have shown that palmitoylation is a postuanslational modification essential for viral morphogenesis. When the mutation is not lethal, the approach may help us to correlate specific biological properties with specific sequences in proteins. For example, modifications of sequences in the hemagglutinin or the neuraminidase have allowed us to obtain novel influenza viruses with changes affecting host range, virulence and/or antigenicily. pentamers Crystallographic analysis has provided knowledge of shell structures to atomic resoluuon In 1956 it was proposed that a single antibody molecule, bindmg to a cnucal sile, is sufficient to neuualize poliovirus (I) We sought such a single-hit anubody from a panel of 46 IgC monoclonal anubodies (mAb) 32 against human rhinovirus 14 (HRV14) and 14 dgamst type 1 poliovuus (PV1) Most of these anubcdies neuualized by smple aggregauon 8. however, were ver) poor precipitators, producing no aggregates or clumps 01 two or threc vnons which d d not funher aggregate with ume suggesung that both arms o l the IgG molecule were stably attached to the same vinon Saturauon binding levels. 30 moleculeslvirion. supported this picture, which implies that the anchcimg arm of non-aggregating lgGs binds in an onentauon favorable for rapid binding of the second arm to an adjaccnt epitope probably across 2-fold symmetry joins. thereby lmking adjacent pentamers (2) This picture has now been visually confirmed by cryoelecmn microscopy which shows that bound Fab fragments from one of lhese anubodies mAbl7. mrecled against epitope IA of HRV14, pomt drecrly across a pentamer joint toward its partner

The mtact mAbl7 molecule bndges the canyon at the point where the cellular receptor molecular ICAM 1 and, as predicred. mAbl7 fflterfered with vual auxhmenL about 6 molecules of the IgG were requued to mhibit bindmg to HeLa cells by 50% Do Dimmaviruses have mls? The 8 non-precipitaung monoclonal a n u w e s in our collection were possible candidates for one-hit neutrahzauon To test this pssibihty, Ihe specific infecuvity of stable anubody-monomenc vuus complexes, separated from small amcans of viral aggregates by sucrose gradient sedirnentauon, was plotted agmnst the number of bound IgG molecules The number of IgG molecules requued to neutralize e p i w (3)

iniual mf&tiwty (zero an1ib;dies bound) to 3740 (1 hit) ranged from 6 to 20: none exhibited one-molecule neuualizauon (4).

The neutralization stoichiometry of the 6-molecule antibody can be reconciled with Dulbecco's single hit model if the cnucal site were one of the 12 penramers. rather than a single subunit (protomer). In this case only five of the 30 paus of sites would involve the critical pentamer; hence. on average, 6 antibodies would be required to bind one IgG to the cnucal pentamer. The criucal pentamer model is compauble with the idea that emergence of the RNA genome dunng uncoaung involves a speclfic locauon, analogous to a tiul. These speculauons point to Ihe need of a more duect means of testing the validity of mil model for uncoaung. Kropff, Universilit Erlangen-Nurnberg, Erlangen, Germany The vast majority of anti-envelope antibodies produced after infection with human cytomegalovirus(HCMV) are directed against the major envelope glycoprotein complex,gp55-l16(gB). This response includes virus neutralizing, non-neutralizing and infected cell surface reactive antibodies. We have previously shown that the immunodominant antibody binding site on gB consisted of a 70 a.a. continuous domain(AD-1), as defined with both murine and human monoclonal antibodies. We have further investigated the antigenic properties of this domain utilizing an expanded panel of murine monoclonal antibodies and human convalescent serum. In agreement with previous results, human antibody recognition of this region of gB required the intact 7Gaa domain. Deletions of either the amino-or carboxy-terminus of AD-I, as well as internal aa replacements abolished antibody recognition of the domain. In addition, we have shown that high affinity, AD-1 specific antibodies produced in patients undergoing primary infection with HCMV were restricted to the lgGl subclass. Mice immunized with the intact AD-I produced antibodies which were reactive with the full length gB but exhibited differing affinities and biological activities. A panel of monoclonal antibodies was derived following immunization with AD-1 and utilized to further characterize the antigenic structure of AD-I. Together with AD-1 mutants, this panel of monoclonal antibodies should provide additional insight into the antigenic structure of this unique region of gB. were assayed for their proliferative responses to peptide presented by a panel of transfected mouse L cells expressing HLA-DR7 heterodimers with single point mutations in either the DR alpha molecule or the DR7 beta molecule. Mutations in the region of the DR beta chain which forms the base of the peptide binding pocket had little effect on antigen presentation, while mutations in the region of DR beta which forms one side of the pocket could decrease the effectiveness of antigen presentation to one or both of the T cell clones. In contrast, mutations in the region of DR alpha forming the base of the binding pocket had the greatest effect on antigen presentation to both clones.

Use of a peptide which was slightly longer at the amino terminus overcame the effects of some of the DR alpha mutants, but none of the DR beta mutants. Binding studies indicate that failure to bind peptide cannot account for all of the negative effects of mutations in DR molecules.

T cell receptor sequences are being determined for these two clones, and for additional clones with the same peptide and HLA-DR specificities, to search for correlation between reactivity to the panel of mutant Class I1 molecules expressed by L cells and the T cell receptor sequence. Like other antiviral responses, antibodies elicited in mice by infection with lactate dehydrogenase-elevating virus (LDV) predominantly belong to the IgG2a subclass. This isotypic bias could be due to modulation of immune mechanisms by the live infectious agent, via for example secretion of soluble factors or a particular type of antigen presentation, or to specific properties of the viral antigenic determinants, such as the presence of repetitive motives.

To test the latter hypothesis, mice were immunized with whole inactivated virus and anti-LDV antibody isotypes were assayed in spleen cell supernatants after a second challenge with the same antigenic preparation. A large proportion of lgGl was found in these antibodies which therefore shared their isotypic pattern with responses elicited by immunization with a soluble protein antigen such as keyhole limpet haemocyanin. In addition, LDV infection concomitant with immunization inhibited the lgGl response triggered by inactivated virus. These results suggest that IgG2a restriction of anti-LDV antibodies elicited by live virus is a consequence of the infectious process. The results showed that the predominant antibody reactivity was to the capsid antigens and less frequently to early region genes. The Western blot assays detected antibodies to linear epitopes on the L? protein that were masked in the capsids, or not sufficiently represented. The capsid ELISA assay detected antibodies to the L1 protein that were primarily to conformational epitopes. Antibodies to E7 were strongly associated with cervical cancer patients and were detected by both ELISA and Western blot assays. The association of these HPV antibodies with various risk factors will be presented. CTlS. It is now well cslablished that interaction between viral antigenic peptides and MHC is greatly dcpendent on the length of the aniigtns. On the basis 3f CI'L lysis, three Dh-restricted epitopes of LCMV have been mapped in the viral glycoproteins GPI and GP2 and nuclcoprotein NP and the sequenca recognired by CI'Lq characterized. The goal of this study was to define the optimal MHC binding sequence of each epitope.

Synthetic peptides truncated at the N-or C-terminus flanking the crucial Db anchoring residue Asn of the viral antigens were tested in binding assays specific for Db MHC on viahle cells (the murinc lymphoma cell line RMAS). 7he results allow a precise delineation of the optimal binding sequences: the GPI and NP epitopes appcar as peptides: GP33-41 (KAVYNFATCG) and NP396-404 (FOPONGOFI), while the GP2 antigen is defined as a pept~de: GP276-286 (SGVENPGGYCL). Interestingly, dramatic alterations of the binding properties o l the epitopes are observed after C-, rather than N-, terminus modifications (shortening or lengthening). The peptide sequences optimal for binding u) MHC and for recognition by CK will he compared. Making use of a prokaryotic protein expression system, we have devised a method for screening peptide libraries with cytotoxic T lymphocytes (CTL). Our lab and others have previously shown that for H-2 Kb restricted octamers, residues in positions 3, 5 and 8 are important MHC binding anchor residues, while positions 1, 2, 4, 6 and 7 may contact peptide specific CTL. Oligonucleotides degenerate at codons 2, 4, 6 and 7 and encoding S-1 and anchor residues 1-3, F-5 and L-8 (SXIXFXXL) were cloned downstream of the ma/€ gene (encoding maltose binding protein, MBP) in an IPTG inducible expression vector (pMAL-c, New England BioLabs). Polyclonal pools of fusion protein were purified by affinity chromatography, and the peptides, released from MBP via factor Xa restriction protease digestion, were fractionated on a RP-HPLC gradient. A library of 190,000 clones contained several peptides that mimicked the H-2 Kb restricted epitopes OVA 257-264 and VSV-N 52-59, as well as self-peptides antigenic for alloreactive CTL. Such MHC allele specific libraries may have many applications in identifying natural epitopes and in characterizing the pre-immune repertoire. HLA class I restricted cytotoxic T-lymphocytes (CTL) play an important role in controlling the proliferative potential of Epstein-Barr virus (EBV) transformed B-lymphocytes. Processed products of one or more of the viral antigens expressed in transformed cells serve as their targets. Six EBV encoded nuclear proteins (EBNA1 to -6) and two membrane proteins (LMP1 and -2) have been identified in lymphoblastoid cell lines (LCL) of normal B-cell origin. We have investigated the target specificity of HLA All restricted CTLs derived by stimulation of lymphocytes from EBV seropositive individuals with autologous EBV transformed LCLs. Recombinant vaccinia viruses carrying the coding sequences for EBNAl, -2A. -2B.

-5, -3 , -4, -6 and LMPl were used to induce high levels of expression of the relevant EBV antigen in fibroblasts derived from appropriately HLA class I matched individuals. EBNA4 expressing fibroblasts were the predominant target of HLA All restricted CTLs in three out of three donors tested. A less pronounced and less regular EBNA6 specific cytotoxic component was found in two of the donors. Recombinant vaccinia virus carrying deletion mutants of the EBNA4 gene and synthetic peptides covering relevant regions of the prozein were used to identify the HLA All restricted EBNA4 epitopes recoznized in different donors. A 9 aa long synthetic peptide corresponding to one of such epitopes induces EBV specific lysis of HLA All positive targets at a concentration of lO-I4M and promotes assembly and surface expression of HLA All polypeptides in HLA All transfected peptide transporter mutant cell lines. composed of an a and p chain is responsible for recognizing antigen in the context of the MHC molecules. Restricted usage of either the V a or Vp chains of the TCR which has been observed in many immune mediated diseases and infections allows TCR-specific antibodies (Abs) to be used as a therapeutic approach. To assay for TCR a and p chain restriction, the srmcture of the TCR in LCMV specific cytotoxic lymphocytes (a) in Balb/c (H-2d) mice was investigated. Greater than 95% of the LCMV specific Cn. response at the clonal level is restricted to a single nine amino acid peptide from the viral nucleoprotein presented by Ld in H-2d mice infected with LCMV. Three CTL clones with specificity for this unique NP epitope presented by Ld were randomly selected for study of their TCR structure. Using "Anchor PCR", all three clones were found to have unique V a and Vp chains relative to each other and lack restriction to any particular variable chain. Additionally, none of the three (7n clones possess the Va4 chain previously reported to be restricted to the H-2b (Db) GP2 epitope in the LCMV model. Molecular analysis of two different MHC backgrounds ( H-2d, H-2b) which present four major epitopes on two different MHC molecules( Ld, Db), shows no resmction to one or two specific variable chains across these backgrounds. These results indicate that in this virus system, TCR-specific Abs directed at restricted variable chains will not be an effective therapy, and suggests that in situations with more complex patterns of MHC presentation like humans, TCR-specific immunotherapy is less likely to be useful. and peptide fragments in a pre-golgi companmenr Little is known about the factors which regulate the formation of these antigenic peptide fragments within the cell. To examine the role of residues within a core epitope and in the flanking sequences for the generation and presentation of the newly synthesized peptide fragment recognized by CD8+ CTL we have mutagenized the coding sequence for the CTL epitope spanning residues 202-221 in the influenza A/Japan/57 HA. In this study over 60 substitution mutations in the epitope were tested for their effects on target cell sensitization using a cytoplasmic viral expression system. The HA202-221 site contains two overlapping subsites defined by Cn. clones 11-1 and 40-2. Mutations in HA residues 204-213 or residues 210-219 often abolished target cell lysis by Cl'L clones 11-1 and 40-2, respectively. Although residues outside the core epitope did not usually affect the ability to be lysed by CTL clones, substitution of a Gly residue for Val-214 abolished lysis by clone 11-1. These data suggest that residues within a site which affect MHC binding and T cell receptor recognition appear to play the predominant role in dictating the formation of the antigenic complex recognized by CD8+ CTL and therefore the antigenicity of the protein antigen presented to CD8+ T cells. Most alterations in residues flanking the endogenously expressed epitope do not appreciably affect the generation and recognition of the site. (TMEV) into CBA mice, results in a biphasic disease ofthe central nervous system (CNS). The first phase resembles poliomyelitis with some of the mice exhibiting paralysis. The virus is cleared &om the CNS at about 4 weeks post-infection, and this clearance correlates with the development of a neutralising antibody response. The chronic phase of the disease occurs ffom 6 weeks post-infection and primary demyelinating lesions can be seen in the spinal cord. In order to define the B cell epitopes of TMEV, neutralising monoclonal antibodies were raised against B e h in Balb/c mice. Two of these antibodies, which react with the capsid protein VP1 in a W e s t e r n blot, will neutralise the demyelinating strains of TMEV as well as a virulent strain which causes acute encephalitis. The monoclonal antibodies were used to select two neutralisation-resistant variants of BeAn. These viruses can not be neutralised by the monoclonal antibodies, nor by hyperimmune sera raised against BeAn in rabbits and mice. However, sera raised a g a a the two variants will neutralise themselves and the wild type virus. The B cell epitope has been located to the carboxyl region of VPI by peptides and by sequencing the variants.

These results indicate that this B cell epitope, which is conserved through the difKerent subgroups of TMEV, is immunodominan4 but that other neutralising subsidiary sites also exist. The immunodominant epitope of FMDV, which is located on a loop comprising residues 132-159 of VPl, one of the four capsid proteins, also contains the putative cell attachment site Arg Gly Asp at aa 145-147. By growth in tissue culture, seven antigenic variants have been isolated from the tongue epithelium of a cow infected with a virus of serotype A12 and additional ones have been obtained by growing the virus in the presence of monoclonal neutralizing antibodies. These variants differ only at residues 148 and/or 153 of VP1.

The serological relationship of seven of these variants have been studied by indirect ELISA, radioimmunopreciptation and seroneutralization using the viruses and peptides corresponding to the 141-160 region of the immunodominant site.The suuctural basis of the antigenic variation was investigated by circular dichroism (CD) spectroscopy and molecular modeling of peptides corresponding to residues 132-159 of VP1. The results show that it is possible to place the variants into two groups: those with and those without Pro at position 153. Nevertheless, there is considerable cross-reactivity among the variants that can be accounted for by the presence of the highly conserved sequence Arg Gly Asp that corresponds to the cell attachment site on the virus particle. T cell repsonses to antigenic proteins are generally limited to recognition of a very small number of epitopes. The basis of this restriction was examined using the cytotoxic T lymphocyte (CTL) response to A/Japan/57 influenza hemagglutinin (HA) as a model.

Previous studies have s own that the class I MHC restricted response in H-2' haplotype mice is limited to recognition of two regions of the HA molecule. Additional regions of the HA molecule were found which bind to H-2Kd but do not serve as class I MHC-restricted T cell epitopes. To determine if intracellular competition among antigenic peptides for processing may prevent presentation of cryptic epitopes, an HA gene construct deleted of the two dominant H-2 Kd restricted epitopes was inserted into vaccinia virus and used to prime mice. Even in the absence of dominant epitopes, CTL recognizing cryptic HA epitopes were not detected. Because potential epitopes may be destroyed or may not be accessible for processing, a non-recognized H-2Kd binding synthetic HA peptide (HA 177-188) was used as a pre-processed antigen to elicit CTL in vitro. Peptide-specific CTL were not demonstrable, suggesting HA 177-188 specific CTL might be deleted during ontogeny. Forty-two variant peptides were therefore synthesized with single or double amino acid substitutions or deletions within the HA 177-188 sequence and were tested for their ability to elicit peptidespecific CTL in vitro. All of the variant peptides elicited peptide-specific CTL capable of lysing target cells sensitized with exogenous peptide, but only one peptide elicited CTL which could recognize endogenously expressed peptide. These data suggest antigenic peptides must bind MHC molecules such that peptide conformation within the Ag-binding groove of the MHC permits the high affinity interaction with TCR necessary for recognition of an enodgenously produced epitope.

A RADIOSENSITIVE AFT SIGNAL FOR HYBRIDOMAS. Shiv A. Prasadg, Steven P. Fling §, and Dale S. Gregersons'. Departments of §Microbiology and 'Ophthalmology, University of Minnesota, Minneapolis, MN.

Experimental Autoimmune Uveoretinitis (EAU) is a T cell mediated disease induced in LEW rats by immunization with bovine retinal S-antigen (S-Ag), or pathogenic peptides. Self-reactive T cell hybridomas recognize sequences on S-Ag that are conserved between both species whereas non-self reactive hybridomas recognize only bovine sequences. These hybridomas are also distinguished by their activation requirements. Ag presentation to the self-reactive hybridomas requires a radiosensitive APC signal. Ag presented by non-irradiated splenic APC stimulates the self-reactive hybridomas, however low doses of y-radiation reduce the ability of the APC to present Ag to the hybridomas. This radio-sensitive APC signal is protected by treatment with PMA or LPS and IL-4. Neither the radio-sensitivity of this signal nor the protection corresponds to changes in MHC class I1 expression. The autoreactive hybridomas also d o not respond to Ag presented by thymic APC. In contrast to the autoreactive hybridomas, a nonself-reactive hybridoma does not require a radiosensitive APC signal, and responds to Ag presented by splenic or thymic APC. These data suggest a radio-sensitive, tissuespecific co-signal required to activate autoreactive T cell hybridomas, and may aid in understanding peripheral tolerance of autoreactive T cells. Several reports in literature described identification of peptides representing CTL epitopes in a number of human immunodeficiency virus type 1 (HIV-1) proteins. We have developed a new screening method that allowed us to identify those peptide epitopes that actually induce viral specific CTLs. Using this method we have identified V3 loop peptides from several HIV-1 strains capable of inducing HIV-specific CTLs. Because of the known role of V3 loop peptides in inducing HIV neutralizing antibodies, we also tested these peptides for inhibition of HIV infection of human T cells. We observed that V3 loop peptides from several HIV-1 strains when used at nanogram quantities exhibit efficient inhibition of virus infection of several human T cell lines as well as freshly isolated normal human T cells. Further, our studies revealed that these V3 loop peptides were capable of inhibiting syncytia formation between HIV-1 infected cells and normal CD4expressing cells indicating that these peptides can prevent cell-to-cell spread of virus, a phenomenon known to play a major role in HIVinduced pathology in AIDS patients. Even though the V3 loop region is known to be highly variable among different virus isolates, a comparison of several HIV-1 strains revealed that most strains can be grouped into five or less categories. We therefore propose that V3 loop peptides as a mixture will provide a potent immunotherapeutic formulation for: (1) Thomas. Division of Virology, National Institute for Medical Research, London NW7 lAA, U.K. Sequence analysis of monoclonal antibody selected laboratory variants of influenza have previously identified five major antigenic sites on the 3D structure of the hemagglutinin molecule. Hovever. there has been no attempt hitherto to investigate the influence of host genetic background andlor natural route of infection on the protective antibody repertoire.

Ye have found that, folloving natural infection of inbred mice (CBAICa. O K BALBlk or BALBlc) with X 3 1 virus (H3NZ subtype) the neutralising antibody response is highly restricted and focused on HA1 158, o r HA1 198 o r HA1 6 3 .

The structural basis for such restriction is being investigated by sequence analysis of IgG heavy and light chain gene usage. 7 6 T cells have been shown to rscognke aewral dwerse anugens; hwover. their mechanlsm d recognbn remalns umlmr. We studled tha raognklon of a Herpes Simpler "INS (nSv) Type 1 glycopmteln I (gl) spednc -76 T d l done (Tgl4.4). Mdenlar doning and lmmunopreclpltstlon sxperlments showed that Tg14.4 expresruu, a k Z c 6 and a v11.2012 heterodlmer. Prevlous data lndlcated that antloen recognNan wa8 MHC Indeperdent. In order to exnmlne how the g1 aMlgen In preaenld, mdecdar mutants of gi were constructed. A truncated farm d gl, thn ha6 both the transmembrane and cytoplasmk reglons deleted @It). was uandected IRo L cells. glt WM not expressed on the cell sutface and was not recoqnlzed by Tg14.4. In contrasl, e chimeric form c4 gl, consklng of the cntracellular poruOn d g1 and the lntrecellular podon of CDB (gl:CDe) r-ued cell sutfau, expresslon of gl and was reccgnlzed by Tg14.4. mls dernonsvated that the presemtlon cd gl was dependent on cdl wrtan, agrerslon. In addklon, the peptMe transporter mutant cell, RWU-5, trarufectad wRh gl, presented antlgen efRclemUy. These reWns suggest that the gl prolaln does na mike the cyloprasmlc protesie machlnery for antigen presentation and may be reccgnbed as an unprocmed antlgen. In oupporl d thls lamer poslblllty. anll-gl antlbcdle6 blocked 6 W k lyah by Tg14.4. Fldly. and most strlklngly. Tg14.4 recognized a lldubla form of g1 when Immobilized on plastic. These tlndlngs IndloBre a novel form of amlgen recognklon s h e thlr T cell recognlzes unprocessed antlgan In an MHC Independent manner and unllke superantlgens, Is anrlgen specwlc. In Ilgh of HSY'a abllky to colonize MHC deflclent CNS cells, T cells that recognize unproceosad sntlgen might be en Impoftant eftedor In HSV pmectlon. These procedures have relied on the use of monoclonal antibodies to isolate COS cells expressing the desired antigen. The significant advantage of the COS cell system is that plasmids replicate episomally to high copy number after transfection into the cells. This results in high levels of expression of introduced genes, and allows rapid and easy recovery of the gene of interest from identified cells. To date, the COS cell system has not been employed to identify antigens recognised by other components of the immune system, such as cytotoxic T lymphocytes.

Our research interests lie in the identification of those antigens of the protozoan parasite, 7heilena pan'a, which are recognised bv CTL from immune cattle. We wished to take advantage of the COOS cell system to establish a procedure to enable screening of a library of parasite genes directly with CTL. In initial, model expenments. we have used bovine alloreactive CTL and COS cells transiently transfected with a plasmid comtaining cDNA encoding the corresponding bovine class I MHC molecule. We have demonstrated that, in a standard microcytotoxicity assay, bovine alloreactive CTL can detect a bovine MHC molecule transiently expressed in a COS cell population. The sensitivity of the assay is such that the alloreactive T2 cells. derived from the EBV transformed B lymphoblastoid cell line ,174, has a profound defect in the class I mediated antigen presentation pathway due to a chromosomal deletion encompassing -1 megabase in the MHC class I1 region. This deletion includs pepbde transporter genes associated with antigen presentation (TAP1 and TAP2) and genes coding for proteasome subunits. Surprisingly, we demonstrated that T2 cells, after infection with Sendai virus, were readily lulled by H-2Kb restricted CD8+ T cells, whereas T2 or T2Db were not killed. Control erperiments revealed that P815A2 was not killed whereas killing against P815Kb was readily observed. Killing of Sendai virus infected T2Kb was dependent of CD8+ T C R a P T cells. Treatment of effector cells with anti-CD4, anti-NK1.l and complement (C') or C' alone gave significant lysis levels, while lysis levels were markedly reduced after treatment with anti-CD8 or anti-TCRap and C'. In accordance with previous findings, T2Kb was not able to present vesicular stomatitis virus (VSV) antigens to VSV-specific CTL in the present study. Furthermore, presentation of influenza virus A/PR/8/34 and AINT160I68 by T?Db after virus infection was greatly impaired. In addition, the RMA-S mutant line also presented Sendai virus specific antigen after v i m infection. Interestingly, Brefeldin A (BFA) treatment completely blccked presentation of Sendai virus antigen in F'815Kb control cells. Unexpectedly, presentation of Sendai virus antigen in T2Kb was not blocked by BFA, even when a high concentration of BFA was used. The present finding suggested that a non-classical antlgen presentation pathway may function in T? cells. The time between virus infection of antigen presenting cells (APC) and presentation of peptide associated MHC molecules to T-cells might vary among different APC, such as monocytes and EBV-transformed 8-cells (B-KL). We investigated the kinetic7 of antigen presentation by measuring [Ca '1 changes in influenza (FLU) specific T-cell clones induced by FLU infected monocytes or B-LCL.

B-LCL presented class-I1 restricted peptides 2 hours after infection, reaching plateau values after 4 hours. Monocytes, however, presented these peptides much more rapidly as they reached plateau values within 1 hour. Like, B-LCL, monocytes were inhibited in class I1 presentation by chloroquine and presentation in both cells was unaffected in case of W irradiated FLU.

Monocytes also presented class I restricted peptide of FLU in accelerated fashion, in comparison to B-LCL. Both APC types were inhibited in class I presentation in case of W irradiated FLU.

In conclusion: mOnOCyteS presented class I and I1 FLU-derived peptides more rapidly than B-LCL. Since both class I and I1 presentation was accelerated, it seems that both rapidity of antigeen processing and factors involved in viral entry (adherence, entry or uncoating) determine the kinetics of antigen presentation. superantigen, also known as MIS-la, which in viva stimulates almost exclusively CD4+ Tcells expressing V06. Spleen cells from BALB.D2 (Mls-la) are injected either locally (in the hind foot pad) or systemically (i.v.) into BALBk (Mls-lb) mice. The changes in the V06+CD4+ population are recorded using flow cytometry. Lymphokine secretion patterns are anlysed using the ELISPOT technique as well as PCR analysis of lymphokine mRNA expression. Ig-production is analysed using an isotype-specific ELISPOT assay. Furthermore, the appearance of anergy to MIS-la is followed. Local and systemic injection of Mls-la-expressing cells both induce a vigorous B cell response as well as an initial expansion of the V06+CD4+ population. We are now correlating the isotype distribution of the B cell response with the lymphokine secretion pattern and the development of anergy and deletion of the responsive Tells. Using a new mAb specific for T cells expressing the human Vp2 gene segment we haxe discovered that Vp2 gene use is bimodal in a normal human population. About 80% of normal individuals have 8.22 1.1% VpZ'CD?' T cells. The other 20% have 1.2?0.6% VpZ'CD?' T cells. Both Vp2 high and Vp2 low expression are stable over time. Vp2 low expression is inherited, probably as an autosomal dominant trait, and is not linked to the TCR locus. Vp2 low individuals have Vp2 genes and the T cells which express them can be expanded by the Vg2 selective toxin TSST-1 and by the Vp2 specific mAh indicating that these cells are not anergic. Clearance of influenza A virus infected cells is predominantly accomplished by CD8+-vim specific cytotoxic T cells. In order to characterize this effector population, we have analysed the TcR usage in lymphocytes isolated from mediastinal lymph nodes (MLN) and bronchoaveolar lavage (BAL) of m K x 3 I-infected C57BV6 (H-2b) mice. We identified a strong enrichment of Vp8.3' CD8' T cells in both MLN and BAL of secondarily-infected mice, and BAL taken from mice after primary infection. Depletion of Vg8.3' T cells in vivo did not compromise the ability of mice to clear the virus, suggesting that other TcR elements are also involved in viral clearance. The VB8.3' CD8' T cells are directed to an immunodominant epitope of flu nucleoprotein (NP) since expansion of these cells was also seen in mice infected with a recombinant vaccinia virus construct containing the NP epitope 366-380. Analysis of T cell hybridomas, generated from freshly isolated BAL of infected mice, confirmed and extended these observations. The majority of flu-specific T cell hybridomas were Vp8.3', though other TcR Vg elements were represented. The Vp8.3' hybridomas recognized the core H-2Db-restricted NP epitope, 366-374. These hybridomas expressed similar Dp and JP elements of the TcR whereas V a usage was more variable. It was concluded that I) the preferential expansion of Vp8.3+ CD8'T cells in response to influenza infection is due to recognition of an immunodominant epitope within NP, and 2) the response to influenza A virus infection is dominated by, but not exclusively, Vp8.3+T cells. We have shown that T cell receptor (TcR) beta chain transgenic mice, expressing Vg8.1 on essentially all T cells, respond immunologically to influenza viral challenge, clearing virus from the lungs with kinetics comparable to the non-transgenic (CBNCa, H-2k) counterpart. Here we characterize the TcR repertoire of, and the ligands recognized by, influenza specific T cells from transgenic and wild type mice. To determine the MHC restriction pattern, antigen specificity and TcR variable region gene usage in influenza specific T cells in wild type and TcRg transgenic mice, we have generated a large panel of hybridomas from CD4+ and CD8+ T cells derived from the mediastinal lymph nodes of these strains. Most of the Class I restricted hybrids derived from the transgenic and wild type mice recognize nucleoprotein (NP) epitopes. Furthermore, these hybrids are restricted by Kk, indicating that the same epitope of NP is likely recognized in both strains of mice. The specificities of Class I1 restricted hybrids from transgenic and wild type mice will be compared to determine if similar TcR ligands are generated and recognized. Phenotypic analysis of the TcRVB usage has shown that while all of the transgenic hybrids express Vg8.1, cells bearing this Vg element are poorly represented in the non-transgenic repertoire (20%). This finding suggests that while Class I restricted T cells in transgenic and non-transgenic mice may recognize identical ligands, they do so by using different TcR repertoires. This has implications for the plasticity of the T cell receptor repertoire. Cytotoxic T-lymphocytes (CTL) recognize peptides derived from endogenously expressed proteins in association with dass I MHC. In contrast, exogenous antigens associate with class I1 M H C following endocytosis to a n endosomal compartment. W e u s e d the b i n d i n g and translocating d o m a i n s of Pseudomoms exotoxin A (PE) fused with CTL epitopes from influenza matrix protein and nucleoprotein as a m e a n s of delivering exogenous antigens to intracellular class I MHC. These recombinant fusion proteins (PEMa, PENP) were capable of sensitizing class I M H C + target cells for lysis by the appropriate CTL. In contrast, a n irrelevant protein, glutathione-s-transferase, fused to the matrix epitope did n o t sensitize target cells. A point mutant of PEMa, corresponding to an endosomal proteolytic processing mutant of PE, h a d a substantially reduced ability to sensitize target cells. Thus the presentation of the CTL epitope by class I M H C required internalization and processing of PEMa similar to that of the toxin. However, unlike PE, PEMa m a y not require translocation to the cytoplasm to exert its effect. Two inhibitors of PE intoxication, NH4C1, which raises endosomal pH a n d inhibits PE a t a step subsequent to proteolysis, and brefeldin A, which m a y inhibit PE by disrupting the Golgi complex, did not inhibit sensitization of target cells by PEMa. Also, PEMa was capable of sensitizing T2 cells for lysis. The T2 mutant cells are defective in transport of peptides from the cytosol to the lumen of the ER for presentation by class I MHC. These results suggest that PEMa is proteolytically processed in a n endosomal compartment leading to association of the epitope with class I MHC, possibly by internalization of class I M H C and recycling to the cell surface. Fusion proteins such as PEMa m a y be useful as vaccines intended to elicit a CTL response. When the rnurine respiratory tract is infected with influenza virus, there is an influx of 7 6 T cells into the lung. In order to understand the role of these cells, it has been important to establish whether they have specificity for viral proteins. We have used Cg-mutant mice kindly provided to us by Drs Mombaerts and Tonegawa (MIT, Boston MA) in order to work in a system deficient of virus-specific ap T cells. Cell suspensions from the mediastinal lymph nodes of N H K (X31) infected mice were cultured with conA for 3 days before fusion with BWn-3cells. The 76 T cell hybridomas were screened for virus specificity by measuring IL2 released from cells in the presence of influenza-infected cells. A selected hybridorna could be stimulated by macrophages presenting the infecting virus N H K (H3N2), as well as N P R 8 (HlN1) but not BIHK. H-2 compatible or H-2 dtfferent virus-infected macrophage cell lines could be used to stimulate the hybridoma. This suggests that the 76 T cell hybridoma recognizes a conserved internal viral protein which is not presented by classical class I or II molecules. In the mouse, self superantigen (Mls) which influence the T cell repertoire and stimulate strong MLR responses have been described. These superantigens (sAg) are encoded by the ORE in the 3'LTR of MMTV virus integrated in the mouse genome. They are responsible for clonal deletion and inactivation of T cells expressing specific Vps.

No such self-sAgs have been described in man. Although human T cells respond to bacterial SAG, it is not known whether human T cells or human MHC class 11 molecules can interact with MMTV sAg.

In order to, determine if human T cells are able to recognize murine endogenous sAg, we transfected the mtv7 sAg in DAP DR1 cells. These transfectants have been used to stimulate human T cells and we characterized this T cell response.

The human T cell response to mtv7 was inhibited by anti-CD4 and anti-MHC class 11, but not by anti-class I and anti-CD8 mAbs.

Moreover, T cells responding to mtvl sAg expressed a restricted number of Vps. Cells expressing Vp12, 13, 14 1 5 and 23 were amplified following mtv7 stimulation. These Vps are the most homologous to the mouse Vps that recognize mtv7.

These results clearly indicate that human T cells are able to recognize endogenous retroviral sAg ar.d suggest that such endogenous sAg exist in man. We are presently investigating if human T cells are able to respond to other MMTV sAgs and we are characterizing their interactions with human MHC class I1 molecules. MA 02129, ** Dept. of Pathology, Mass. General Hospital

with processed viral antigen. We wished to assess the spectrum of TCR gene usage in HIV-1 specific CTL isolated from infected individuals.

Peripheral blood mononuclear cells (PBMC) obtained from a seropositive individual were cloned at limiting dilution in the presence of IL-2, feeder cells, and a CD3-specific monoclonal antibody or PHA as a stimulus to T cell proliferation. CTL epitopes were defined using autologous EBV lymphoblasts incubated with synthetic HIV-1 peptides. cDNA was prepared from approximately 5x106 cloned T cells. PCR was then performed with 5' specific oligonucleotides from the variable regions of the 29 known Va genes and 24 known Vp genes in conjunction with a 3' primer from the constant region of the respective TCR a or p gene. PCR products were then sequenced directly using the dideoxy chain termination technique.

infected individual over a 29 month span. All clones are HLA-B14 restricted and recognize the same nine amino acid minimal epitope. All clones thus far analyzed have utilized Val4 and Vp4 T cell receptor (TCR) genes. Sequencing of these genes has revealed identical diversity and joining regions, indicating a restricted TCR usage for this epitope which is maintained over time. Another HLA-B14 restricted clone isolated from this same patient, which is specific for an epitope in the reverse transcriptase (RT) molecule, utilizes the Va 21 and Vp 14 genes. Studies are in progress to determine whether CTL clones specific for other HW-l epitopes from this individual show limited TCR gene usage.

This study demonstrates that HIV-1 specific CTL of a given epitope specificity and HLA restriction can exhibit limited T cell receptor gene usage in a given individual which is maintained over time. Analysis of clones with the same epitope specificity and HLA restriction from different individuals will aid in the understanding of the role TCR gene rearrangement plays in CTL recognition of HIV-1 epitopes.

Envelope specific CTL clones have been isolated from an HIV-1 Genetic factors influencing the host Immune response, Including T cells, contribute to the generation of HR. T cell response was studied in BALB/c congenlc mice that have different susceptlbilitles to HR. These congenlc strains differ only at the Igh-1 locus, which Influences T cell repertoire. In dralnlng lymph nodes T cell subsets were analyzed by the polymerase chain reaction using oligonucleotide primers specific for 18 T cell subsets, defined by their Vp gene usage of TCR, following AC Inoculation with KOS strain of HSV-1. TCR Vp gene usage In submaxilliary lymph node T cells was comparable In these congenic strains prior to inoculation. However, the kinetics of each T cell subset responding to herpes antigen differed between susceptible (BALB/c) and resistant (C.B-17) mice. Vp 4, 6, and 14 were the major subsets responding to AC inoculation In C.B 17, while Vp 9 and 13 were the major subsets In BALBlc. Higher levels of TCR Vp 1, 4, 13, and 18 expression were observed In C.B 17 compared to BALBlc mice at earlier time points post inoculation. These results suggest that TCR Vp gene usage was Influenced by Igh-1-linked gene products and may contribute to the HR susceptibility patterns observed In BALBlc Igh-1 disparate congenic mice. superantigen is found on the cell surface as a proteolytic fragment. To investigate whether this proteolytic cleavage is a requirement for viral superantigen presentation, protease recognition motifs in the vSAG-6 amino acid sequence were destroyed using site directed mutagenesis. These constructs were then expressed in a B-cell line and tested for their ability to stimulate VP3+ T-cell hybrids in vitro. To the same end, the effects of specific protease inhibitors on vSAG processing are also being explored. In addition, vSAG-6 chimeric constructs are being generated using trans-membrane regions of Invariant chain that result in secreted or plasma membrane-bound type I1 proteins.

Expression of these constructs should elucidate whether a vSAG-6 proteolytic fragment remains associated with the trans-membrane region of the protein, or whether the fragment alone can associate with MHC class I1 molecules with high affinity. Further mutational analysis of vSAG-6 should delineate regions of the superantigen that bind MHC class I1 products. In a previous study it was shown that HLA-A2.1 restricted influenza-specific cytotoxic T lymphocytes (CTL) derived from HLA-A2.1 transgenic mice recognized the same M1 peptide as HLA-A2.1 restricted human CTL (J.Imm =1226 (1991)). Analysis of 10 independently derived murine CTL lines using a panel of mouse T cell receptor (TCR) VB specific monoclonal antibodies demonstrated that VB5.1, VB6 or VB8.1 dominated this response. TCR from representative clones expressing either VB5.1,6 or 8.1 were sequenced using VB specific and degenerate Va primers to amplify cDNA by PCR. Both the VB5.1 and VB8.1 CTL were found to utilize the JB2.6 segment whereas the VB6 utilized JL31.1. Three different Va:Ja combinations were found among the clones and no consensus in amino acid sequence or length was seen for junctional regions of either a or p chains.

Thus at least 3 different murine TCR are used to recognize the same human MHC:peptide complex. Comparison of these murine TCR sequences with sequences of human HLA-A2.1 restricted M1 peptide specific CTL revealed no common features of a or B junctional regions. Interestingly, however murine VB6 has striking homology to human VB17 which is overrepresented in the human CTL response. Human V a sequences that are the most homologous to the murine V a sequences were also represented in the human T cell response, suggesting some conservation of TCR components recognizing the same h4HC:peptide complex. Interestingly, the amino acid sequences of VSAG-1 and vSAG-6 are identical, yet deletion of Vp3 bearing thymocytes varies between the two strains. This phenomenon could be explained by differences in: expression levels of vSAG on presenting cells, ability of presenting cells to efficiently present antigen and/or induce deletion, and/or susceptibility of reactive thymocytes to deletion. Data will be presented from our current study of the mRNA expression levels of vSAGs 1 and 6 within thymic microenvironments, thymocyte subsets, and in the periphery of these mice, using RNase protection assays. Concomitantly, surface vSAGs will be detected by their ability to stimulate Vp3 bearing T cell hybridomas and their effect on thymocyte deletion will be analyzed by fluorescent antibody staining. These experiments will expand our understanding of the location, timing and cell types involved in the negative selection of "selfreactive thymocytes.

EXPRESSION OF VIRAL SUPERANTIGENS-1 AND -6, The BM5 mixture of replication competent and defective renoviruses causes a dramatic and fatal immunodeficiency disease (MAIDS) in susceptible strains of mice. The disease is characterized by lymphadenopathy and splenomegaly including an early increase in CD4 T cell number, but loss of CD4 T cell function. Importantly, both the initiation and the progression of MAIDS is dependent on CD4 T cells. Recently, Hiigin et al. suggested that the BM5 remviral mixture encodes a product which stimulates CD4 T cells bearing several Vps in vitro. Here we report that the mRNA synthesis of Vp5.2, but not other Vps. is dramatically and selectively depressed in the thymus and in peripheral CD4 T cells of C57BL/6 mice with MAIDS as early as four weeks after infection. These results indicate that the BM5 mixture of retroviruses most likely encodes a functional superantigen or stimulates the expression of an endogenous superantigen specific for vps.2.

To further evaluate the importance of this deletion in disease' induction and progression, we infected C57WJ mice which lack a panel of Vp genes including Vp5.2. The progression of the disease was strikingly slower in C57L/J as compared to age matched C57BU6 mice which develop the characteristics of MAIDS within the finr few weeks of infection. The extent of splenomegaly. lymphadenopathy and function of CD4 T cells were compared between the two strains of mice during the course of the disease. At eight weeks of infection, splenomegaly and lymphadenopathy of C57UJ mice were less pronounced and the CD4 T cells still responded to mitogenic stimuli by proliferation and IL-2 production, whereas the CD4 cells from infected C57BU6 mice failed to do so.

These results indicate that Vb5.2 deletion in BM5 induced MAIDS could play a role in disease progression and induction of anergy. We have previously described an MIS-lalike clonal deletion of mature CD4+ T cells which express Vp6 and Vp8.1 chains of the 'r cell receptor in half of the mice of a BALBlc. MIS-lb colony (BALBlc. IC). This occurs in the absence of the Mtv-7 provirus which is responsible for thc clonal deletion in MIS-la mice (I). Using a PCR assay, we showed that Mtv-7 homologous transcripts were present in the mammary glands of lactating BALBlc, IC mice and in the thymuses andlor spleens of BALBlc, IC virgin mice with deletion of Vp6+ lymph node T cells, and not in BALB/c. IC with normal Vp6 expression (2). These results indicate that this BALB/c colony is infected with an exogenous MMTV retrovirus whose vSAG is similar to that of Mtv-7 provirus, as recently reported. Transmission of this virus is achieved through the milk from the laccating mother to the suckling pups. Mother forstering with mice of different H-2 backgrounds showed that the requirement for I-E expression is more stringent when the clonal deletion is induced by the exogenous than the endogenous virus. Thymectomies performed at 4-5 weeks of age (at least 4 weeks before detection of clonal deletion), did not affect the occurrence of clonal deletion in peripheral lymph nodes when tested twenty weeks later. This suggesu that clonal deletion can be achieved without further intrathymic contact with the antigen. Since MMTV is uansmitwd through milk and is likely to be present in the gut, we evaluated the percentage of Vp6+CD4+ T cells within the gut intraepithelial lymphocyte (IEL) population. Mice with normal Vp6 expression in lymph nodes may show partial deletion of Vp6+CD4+ IEL. This is explained by exclusive localisation of Mtv-7-like RNA in the gut of some mice. Finally, despite the similarity of the vSAG encoded by the endogenous Mtv-7 with the exogenous MMTV in BALBlc, IC mice, the latter has a much lower stirnulatory capacity when compared to the former.

(1) Papiemik M. Pontoux C, Gisselbrecht S (1992). J.Exp.Med. 175:

(2) Desaymard C, Tucek C, Rocha B, Korman A, Papiernik M (1992) .

Int.Immuno1. (in press).

Charles D. Surh, D a n P. Gold, a n d Jonathan Sprent, Department of Immunology, The Scripps Research Institute, La Jolla, C A 92037, and San Diego Regional Cancer Center, San Diego, C A 92121 Through the use of xenogeneic rat -> mouse chimeras and a polymerase chain reaction method t o quantitate rat T cell receptor Vg m R N A composition we demonstrate in t w o ways that rat T cells can recognize mouse Mls antigens. The rat -> mouse chimeras possessing de n o w generated rat T cells were established b y transplanting Lewis rat fetal liver cells into neonatal C.B-17 SCID (H-Zd, Mlsc) mice. Firstly, the lymph nodes from the rat -> mouse chimeras contained considerably reduced levels of rat Vg 3, 5, 12 and lbpositive T cellsthese Vg+ T cells a r e easily detectable in normal Lewis rats although Vg11+ T cells are inherently very low. (Note that rat TCR Vg chains a r e numbered according to the homologous m o u s e conterparts). The deletion of rat Vg16+ T cells in the chimeras w a s confirmed with a recently produced rat Vg16-specific m A b HIS 42. Secondly, the T cells from the rat -> mouse chimeras proliferated very strongly to H-2 matched Mls-positive DBA/Z (H-Zd, Mlsa) stimulator cells i n mixed lymphocyte culture, although these T cells showed complete tolerance to syngeneic BALB/c (H-2d, Mlsa-) stimulator cells. The rat T cell population responding to DBA/2 stimulator cells was highly enriched for rat Vg6+ and 8.2+ T cells. Similar results were obtained with chimeras established by tranplanting fetal liver from another rat strain (Fisher 344) into C.B-17 SCID mice. These data suggest that rat T cells a r e able to react to mouse MIS antigens, a n d that rat T cells are be subjected to MMTV antigen-mediated clonal deletion during T cells differentiation in the rat -> mouse chimeras. Mouse Mammary Tumour Virus (MMTV) has long been known to contain an open reading frame (ORF) within its unusually long U3 region. This ORF has the potential to encode a set of proteins of varying sizes. Recently we have shown that the ORF encodes a negative transcriptional regulatory factor (nafl which functions in trans to decrease the rate of transcription from an MMTV based indicator construct [J.Virol.,l990;64.6355]. Downregulatory naf effects can also be seen on homologous and heterologous promoters linked to the luciferase gene. It has also been shown that the MMTV ORF encodes a superantigen (sag), the expression of which during ontogeny, results in the VR specific deletion of Tlymphocytes. This appears to occur by a specific interaction between the sag and the 13 chain of the T-cell receptor that is common among a given class of T-cells. Although sag and naf are both products of the ORF, it is not clear how these t w o activities are related. However using a series of constructs we have been able to distinguish between the required coding regions for nafand sag. Additionally, we have identified a novel promoter located within the U3 part of the long terminal repeat region (LTR). Both the previously described promoter and the novel U3 promoter are active in mammary cells and Blymphocytes. However the U3 promoter is preferentially active in LPS stimulated B-lymphocytes and is able to direct expression of the sag product, making it a candidate for the lyphocyte specific expression of this gene in the mouse. The role of the immune response in RSV infection is not well understood. Re infection is common in both children and adults despite the presence of RSV specific antibodies and a T cell response to RSV antigens. It is intriguing that inactivated RSV (IRSV), but not infectious RSV (Inf-RSV) stimulates the T cell proliferative response in vitro. We have analyzed whether stimulation of blood mononuclear cells with IRSV and Inf-RSV results in differences in expression of surface markers and cytokines important in the immune response. HLA-DR, ICAM-1, CDlla, CDllb, CDl lc, CD14, CD4 and CD8 expression was determined by flow cytometry on unexposed and RSV-exposed monocytes and lymphocytes 24 and 48 hours after stimulation. A dramatic effect on monocyte surface markers was found with both Inf-RSV and IRSV: a 4-8 fold increase in HLA-DR, a 2-4 fold increase in ICAM-1, a 2-4 fold increase in CDl lc, a 1.5-2 fold increase in CDllb, and a 2 4 fold decrease in CD14; Inf-RSV was generally twice as effective as IRSV. There was no change in CDl la expression. The only effect of RSV on the lymphocytes was an increase in CDl lb positive, HLA-DR negative cells found with Inf-RSV but not with IRSV. Exposure of the monocytes to virus in the absence of T cells gave none of the changes in surface antigen expression. Various cytokines such as IFN-y, IL-2, IL-4, IL-10, and GM-CSF have each been shown to induce some of the changes seen in monocytes after RSV stimulation. These cytokines are being analyzed in the supernatants of IRSV and Inf-RSV stimulated cells. The high HLA-DR and ICAM-1 expression on RSV infected monocytes would suggest effective antigen presentation, but these cells do not support RSV-specific T cell proliferation, as measured by 3Hthymidine incorporation. Instead their state of activation may result in increased production of mediators which counteract immune responses, as well as influence T cell functions other than proliferation. This abmact of a proposed presentation does not necessarily reflect EPA policy. T cell clonal anergy is a form of tolerance in which mature T cells are physically present but are not able to mount an immune response. We induced T cell unresponsiveness to MIs-la in mice transgenic for T cell receptor VP8.1 by injection of MIs-la spleen cells, by mating with Mls-la mice, and by generating bone marrow chimeras in which Mls-la is present only on nonhematopoietic cells. CD4+8-VP8.1+ cells from all these groups did not proliferate in response to irradiated spleen cells from MISl a mice. However, the mechanisms underlying the unresponsiveness appear to differ. CD4+8-Vp8.1+ cells from Mlsl a spleen cell injected Vp8.1-transgenic mice mobilized cytoplasmic Ca2+ but proliferated at a reduced level in response to crosslinking with anti-TCR mAb. However, these cells formed conjugates, mobilized Ca*+ and proliferated in response to Mlsl a when activated B cells were used as stimulators. In Mls-lam VP8.1-transgenic mice, a subpopulation of CD4+8-Vp8.1+ cells did not mobilize cytoplasmic Can+ after TCR crosslinking. They did not form conjugates, mobilize Ca2+ or proliferate in response to Mls-la on activated B cells. Finally, CD4+8-VP8.1+ cells from the bone marrow chimeras proliferated to TCR crosslinking at a partially reduced level; and formed conjugates, mobilized Ca2+. and proliferated in response to MIS-1 a on activated B cells.

These features indicate that there are at least two distinct mechanisms underlying T cell clonal anergy in vivo, and that some forrn(s) of clonal anergy can be reversed when antigen is presented by an appropriate antigen presenting cells. Vaccinia virus, a member of the poxvirus family of large cytoplasmic DNA viruses, has been used extensively in studying the mechanism of antigen processing and presentation. However, as with some other viruses, vaccinia encodes a number of gene products which are involved in interfering with the host cell immune response. These include an interleukin-lp receptor, complement control proteins and serine protease inhibitors (serpins). Previous data from recombinant vaccinia viruses expressing the influenza virus haemagglutinin (HA) or nucleoprotein (NP) have shown that vaccinia infection interferes with the presentation of some epitopes from these proteins to class I-restricted cytotoxic T lymphocytes (CTL). Recombinant vaccinia viruses expressing the influenza HA or NP have been constructed in which either or both vaccinia serpins (B13R and B22R) have been deleted. The presentation of influenza epitopes from cells infected with influenza or vaccinia virus recombinants with or without the serpins was monitored by 5lCr release after incubation with epitope-specific CTL. An epitope presented from influenza virus infected cells but not from serpin positive vaccinia virus infected cells was re-presented after serpin deletion. This restoration was epitope and serpin specific and occurred at early but not late times during vaccinia infection. Further experiments to address this block at late times are being undertaken.

N 202 T h e e f f e c t o f t h e A d e n o v i r u s E 3 / 1 9 k protein o n a n t i g e n p r e s e n t a t i o n in vivo. J o s e p h i n e H. C o x , J a c k R . B e n n i n k , J o n a t h a n W. Yewdell, R . M a r k L. B u l l e r a n d G u n a s e g a r a n Karupiah. L a b o r a t o r y of V i r a l Diseases, N a t i o n a l Institutes of Health, Rethesda, MD 20892. leads to a poor CTL response resultmg ln persistent infection Those mce infected as adults aith Clone 13 sho
 a generalized munosuppression, wth IOU to negligible CTL responses generated in response to mfection wth other viruses such as vaccinia, intluenza, and herpes sunplev Genetic and sequence studies map the unmunosuppressive phenorlpe of Clone 13 to a smgle ammo acid change at position 260 in the gl

Cantab Pharmaceuticals Research, Cambndge Sclence Park, Cambndge. U.K.', Department of Experimental and Clinical Microbiology

In HIV vaccine trials in infected individuals, there has been speculation that repeated antigenic stimulation might convert naive cells into memory cells thus enhancing HIV infection and subsequent destruction of these cells. Therefore we compared the naive and memory phenotypes of patients who were participating in a gp160 vaccine protocol at three time periods over approximately a one year time period.Methods: Twelve (five presented in this abswct) HIV-infected patients from the phase I gp160 mal were selected based on the following criteria: ininalCD4 > 500, new antibody response to H N epitopes, and new peak LSI > 15 in a proliferation assay to the gp160 vaccine. Fresh heparinized blood specimens were stained w i t h panels of monoclonal antibodies using a three color approach (FITC, PE, PerCP) and processed using a whole blood lysis protocol. Comparison of results was performed with a Wilcoxin signed-rank test.There were no statistically significant differences (p<.05) in anv Darameter between the three time points. All values in the Conclusion: Comparison of naive and memory phenotypes in a group of early stage HIV-infected patients receiving the gp160 vaccine at three time points over approximately one year did not reveal statistically significant differences in any of the phenotypes listed. Specifically. there did not appear to be a depletion of memory CD4 or CDX cells. Analysis of the larger group of 12 patients will be completed shortly. A large blinded, prospective trial is underway with the gp160 vaccine and will better answer whether there are any changes in the naive and memory phenotypes with repeated immunization. Immunization with soluble protein antigens usually stimulates CD4+ but not CD8+ T cells. Immunization of mice with low doses of two viral proteins (i.e., HBV-derived S-antigen [HBsAg] and SV4O-denved T-Ag) primed CD8+ cytotoxic T lymphocytes (CTL) in vivo.responders. Target cells pulsed in virro with HBsAg particles were specifically lysed by anti-HBsAg CD8+ CTL. In contrast to the single immunization with "naked" S-antigen particles that efficiently primed CD8+ CTL, immunization with Santigen adjuvanted with either aluminium hydroxide, or incomplete Freundvs adjuvants did not induce a CTL respons. This protein fragment hence contains structural determinanui required for this protein to access the 'endogenous' processing pathway for H-2 class I-resh-icted antigen presentation to CD8+ CTL. can function a s a Tc e l l independent antigen i n mice.This capacity of the €!I!cAg has been suggested t o be responsible for the enhanced immogenicity e h b i t e d by Hk-particle based chimeric vaccines i n non-human species l i k e rabbits and guineapigs,where HB is not a natural i n f e c t i 0 n . h man, however,little i s known regarding the contribution of a T-cell independent B-cell reactivity induced by HkAg for the developrent of protective imnunity to HB.Therefore,we analyzed the HkAg induced IgG anti-HBc production i n v i t r o i n supernatants from T-cell depleted Bc e l l cultures obtained from 10 chronic HBsAg c a r r i e r s , 8 individuals with natural acquired inrnunity to HE! and 7 HB-susceptible controls by an ELISA developed a t our 1aboratory.Spontaneous production of IgG anti-HBc was only observed i n some B-cell cultures from chronic HBs-Ag carriers,but not from H B -i m e donors or HB-suscept i b l e contro1s.h HB9.g stimulated B-cell cultures IgG anti-HBc production increased significantly i n the major i t y of chronic HBsAg carriers,whereas HB-inmrune donors were non-responders i n vitro.The results indicate that, i n man,the capacity for T-cell independent B-cell activation exhibited by HECAg i s related to the chronic HBs-Ag carrierstate and i s not a feature of established immunity t o HB. (Sylvan et al.,Archives of Virology,1992,Suppl 4:29-35.