key: cord-023143-fcno330z
authors: nan
title: Molecular aspects of viral immunity
date: 2004-02-19
journal: J Cell Biochem
DOI: 10.1002/jcb.240591009
sha: 
doc_id: 23143
cord_uid: fcno330z

nan

Mechanisms of T-cell mediated clearance of viruses from the central nervous system are poorly understood, but likely to differ from those employed in the periphery because the CNS lacks lymphatic drainage and constitutive expression of MHC class I antigen, and the unique structure of the CNS vasculature imposes constraints on access by leukocytes and soluble immune mediators. To study the mechanism by which viruses are cleared from neurons in the central nervous system, we have developed a mouse model involving infection with a neurotropic variant of mouse hepatitis virus (OBLV60).

grew preferentially in the olfactory bulbs of Balbk mice.

Using in situ hybridization, we found viral RNA localized primarily in the outer layers of the olfactory bulb, including neurons of the mitral cell layer. Virus was cleared rapidly from the olfactory bulb between 5 and 11 days. Athymic nude mice failed to eliminate the virus demonstrating a requirement for T lymphocytes. Immunosuppression of normal mice with cyclophosphamide also prevented clearance.

Both CD4+ and CD8+ T-cell subsets were important as depletion of either of these subsets delayed viral clearance. Gliosis and infiltrates of CD4+ and CD8+ cells were detected by immunohistochemistry at 6 days.

The role of cytokines in clearance was investigated using an RNase protection assay for IL-la, IL-lp, IL-2, IL-3, IL-4, IL-5, IL-6, TNFa, TNFP and IFNy. In immunocompetent mice there was upregulation of RNA for IL-la, IL-lp, IL-6, TNFa and IFNy at the time of clearance.

Nude mice had comparable increases in these cytokine messages with the exception of IFNy.

Induction of MHC-I molecules on cells in infected brains was demonstrated by immunohistochemistry in normal and nude mice, suggesting that IFNy may not be necessary for induction of MHC-I on neural cells in vivo.

Luca G. Guidotti, Kazuki Ando, Tetsuya Ishikawa, Lisa Tsui and Francis V. Chisari. The Scripps Research Institute, La Jolla, CA 92037

Although cytotoxic T lymphocytes (CTL) are known to clear viral infections by killing infected cells, recent studies suggest that they can also suppress the replication of certain viruses by noncytolytic mechanisms. We have examined this area by monitoring the immunopathological and antiviral consequences of antigen recognition by hepatitis B virus (HBV) specific CTL in HBV transgenic mice that express the viral gene products in their hepatocytes. We have shown that intravenously injected CTL rapidly trigger their target hepatocytes to undergo apoptosis, but that the direct cytopathic effect of the CTL is minimal in comparison with the cytopathic effects of the antigen-nonspecific intrahepatic inflammatoly response that they activate. In addition to killing the hepatocyte, the same CTL also downregulate HBV gene expression and completely abolish HBV replication in the hepatocytes that they don't destroy. This noncytolytic antiviral CTL effect is mediated by at least two distinct processes in these animals. First, the CTL cause a quantitative reduction in the steady state content of all HBV mRNA species in the hepatocyte, and this is followed by disappearance of all of the corresponding viral proteins in the liver and serum. The CTL initiate this process by secreting IFNy and TNFa when they are activated by antigen recognition. since the regulatory effect of the CTL can he prevented completely by prior administration of the corresponding antibodies. Nuclear run-on experiments reveal that viral mRNA transcription is unaffected despite the profound reduction in HBV mRNA content in the liver, suggesting that the CTL-derived cytokines accelerate viral mRNA degradation in the hepatocyte. A second noncytolytic antiviral pathway is also activated by the CTL. We have recently shown that HBV nucleocapsid particles, and the replicative HBV DNA intermediates that they contain, disappear from the transgenic mouse liver following either CTL administration or partial hepatectomy. the latter of which triggers hepatocellular regeneration without any change in hepatocellular HBV mRNA content. These results suggest that preformed HBV nucleocapsid particles may be actively degraded during hepatocyte turnover, and they raise the possibility that similar events might also occur in nondividing hepatocytes that are activated by noncytolytic signals delivered by the CTL. We propose that, in addition to their pathogenetic effect, the comhined effects of the CTL response at die HBV mRNA. nucleocapsid and rcplicative DNA levels may represent a curative antiviral stimulus during HBV infection. Since the virus must contain molecular elements that IEspond to these CTL-induced antiviral signals. inactivating mutations at these loci could be very efficiently selected by immune pressure, because a single mutation could abrogate the antiviral effect of a wide spectrum of T cell responses, irrespectrve of epitope specificity. Identification of these viral response elements and the intracellular pathways that interact with them may lead to the development of new strategies for antiviral drug design. Human fibroblasts infected with HSV are resistant to lysis by CD8+ cytotoxic T lymphocytes (CTL), yet human B cell lines can be efficiently lysed by these CTL. The effect on human fibroblasts is rapid (within 2 hr of infection of cells), occurring before synthesis of MHC class I is altered by virus infection. A recombinant HSV, F-USBMHC, which expresses mouse MHC class I proteins does not render human fibroblasts sensitive to lysis by mouse CTL. MHC class I molecules are retained in the ER of HSV-infected fibroblasts i n a misfolded, unstable form and stability of the MHC complex can be restored by addition of exogenous peptides. Using a panel of HSV mutants and Ad expression vectors we demonstrated that the HSV IE protein ICP47 was both necessary and s f i c i e n t to cause retention of class I and ICP47 expression in fibroblasts caused the cells to resist lysis by CD8+ T lymphocytes. ICP47 is a soluble, cytosolic protein and we have found no evidence of membrane association.

Therefore, it appears that ICP47 inhibits cytosolic stages of the antigen presentation pathway so that antigenic peptides do not reach the ER. To date, polyclonal and monoclonal antibodies directed to ICP47 have not specifically precipitated any of the previously described components of the antigen presentation pathway and we have not found ICP47 associated with TAP transporter proteins or proteosomes i n these experiments. The effects of ICP47 are being assessed in proteosome and TAP transporter assays. GST-ICP47 fusion proteins tightly bind a 8.5 kDa cytosolic cellular protein which is found in a number of adherent human cell lines but not lymphocytes. The protein has been purified and sequencing is in progress. In addition, radiolabelled ICP47 binds to a single cellular protein of =55 kDa on ligand blots. These proteins are good candidates as cellular targets of ICP47 and as novel components of the antigen presentation pathway. Preliminary experiments support the hypothesis that ICP47 is very effective i n blocking CD8+ T lymphocyte responses in vivo, perhaps explaining the predominance of CD4+ vs. CD8+ anti-HSV CTL i n vivo. We expect that ICP47 may be very useful, not only to elucidate antigen presentation pathways, but also to prevent immune recognition of gene transfer vectors and a s a immunosuppressive agent.

SUSCEFTIBILITY TO POLYOMA VIRUS-INDUCED TUMORS IS CONFERRED BY AN ENDOGENOUS MMTV SUPERANTIGEN. Aron E. Lukacherl, Yupo Ma2, John P. Carroll2, Sara R. Abromson-Leeman2, Joseph C. Laning2, Martin E. Dorf2, and Thomas L. Benjamin2. IDepartment of Pathology, Emory University School of Medicine, Atlanta, GA 30322, and 2Department of Pathology, Harvard Medical School, Boston, MA 02115. Susceptibility to tumors induced by mouse polyoma virus varies among inbred mouse strains. We have previously shown that polyoma tumor susceptibility is controlled by products of MHC as well as non-MHC genes. In crosses between MHC-nonidentical strains differing in tumor susceptibility, resistance correlates with dominantkodominant inheritance of the resistant H-2 haplotype. We have observed the opposite pattern of inheritance of susceptibility in crosses between MHC-identical strains. In crosses between the highly susceptible C3WBiDa mouse and the highly resistant but MHC-identical (H-2k) C57BWcd.I mouse, polyoma tumor susceptibility is conferred by a single autosomal dominant gene, which we have designated Pyvs. PyJ does not encode cell receptors for the virus, affect viral dissemination or anti-viral antibody responses, or affect intracellular events essential for productive infection or cell transformation by the virus. Whole-body irradiation renders CS7BWcd.I mice fully susceptible to polyoma-induced tumors, indicating an immunological basis for this strain's resistance. We hypothesized that P y J encodes an Mtv superantigen (SAG) that confers susceptibility to C3WBiDa mice by deleting precursors of polyoma-specific T cells. We found that tumor susceptibility in (C3WBiDa x C57BWcd.I) x C57BWcdJ backcross mice cosegregated with Mtv-7. Inheritance of Mtv-7 showed perfect concordance with absence of peripheral Vp6+ T cells. Genotyping of backcrossmice using markers of simple sequence repeat polymorphisms flanking Mtv-7 showed no evidence of recombination between Pyvs and Mtv-7. Strongly biased usage of Vp6 by (a) polyoma-specific CD8+ CTL from virus-infected C57BWcdJ mice and by @) CD8+ T cells infiltrating a polyoma tumor in a virus-immune C57BWcd.I host provide further evidence that T cells bearing this Mtv-7 SAG-reactive Vp domain are critical anti-polyoma tumor effector cells. These results indicate identity between P y J and Mtv-7 sag, and demonstrate a novel mechanism of inherited susceptibility to virus-induced tumors based on effects of an endogenous superantigen on the host's T cell repertoire. Infection of mice with lymphocytic choriomeningitis virus (LCMV) causes a transient to longlasting immunosuppression dependent upon virus-isolate dose of virus and age, H-2, non H-2, level of CD4+ T cells, of CD8+ T cells and kinetics of neutralizing antibodies of the host. The immunohistological analysis suggests that CD8+ T cell dependent disappearence of marginal zone macrophages of follicular dendritic cells and of virus infected cells in general correlates with immunosuppression. The details of mechanisms responsible for these findings are now being analysed. A role of this CD8+ T cell dependent immunosuppression in the establishment of a LCMV carrier state in immunocompetent mice is suggested by the following experiments: The otherwise slow and low neutralizing antibody response agamst LCMV is accelerated and enhanced by CD8+ T cell depletion at the time of infection, suggesting virus-specific immunopathology being responsible at least partially. The ELISA antibody response is not significantly altered under the same conditions but is abrogated if LCMV-specific T cell receptor transgenic mice are infected with high doses of LCMV, indicating, that suppression of the specific antibody response depends upon the relative kinetics of CTL versus antibody responses. Whether exhaustion of specific CTL responses is enhanced by similar mechanisms remains to be tested. The role of interleukins of the relative distribution of virus in the mouse and in the various aspects of immunosuppression are now being studied. Immunosuppression, caused by CD8+ T cell-dependent immunopathology, may also be operational in HIV infection in humans. Such a pathogenesis of HIV-triggered AIDS could explain several aspects of the disease process not readily fitting the (unproven) conventional idea that HIV is causing immunodeficiency via direct viral pathogenicity. The cellular immunity against two DNA tumor viruses (i.e. human adenovirus type 5 (Ad 5) and human papillomavirus type 16 (HPV16)) was studied with respect to possible immune escape mechanisms and to the development of CTL epitope based peptide vaccines. After identifying an immunorelevant CTL epitope in the Ad 5E1A protein to which CTL clones were directed that could eradicate Ad 5E1 induced tumors in nude mice, an amino acid replacement study of this epitope revealed a point mutation that totally eliminated the possibility to recognize the mutant peptide by the CTL clones directed against the wild-type peptide sequence. New viral constructs were made that contained this point mutation and used to transform mouse embryo cells. However, these mutant tumor cells were still immunogenic and CTL clones specific for these mutant tumor cells were shown to react with a peptide derived from the Ad 5E1B protein. These Ad 5EIB specific CTL clones, however, were as effective as the Ad 5E1A specific CTL clones in the eradication of Ad 5E1 induced tumors in nude animals, indicating that a choice can be made of immunorelevant epitopes to which an immunization strategy could be developed. In addition, we discovered that by supertransfection of Ad 5E1 induced tumor cells with the activated ras oncogen the possibility of Ad 5E1B specific CTL to recognize the Ad 5E1 induced tumors was eliminated whereas the Ad 5E1A specific CTL could still kill these tumor cells. This might indicate a new mechanism of tumors to escape CTL.

In an HPV16 induced mouse tumor model an immunosubdominant CTL epitope was identified in the E7 protein that could, upon immunization with that peptide,protect mice against a subsequent challenge with HPV16 induced tumor cells. By changing the anchor residues in that peptide an even more immunoprotective peptide could be generated. Combined, these data indicated a successful use of a CTL epitope based peptide vaccine in the prevention of HPV16 induced tumors in mice. Subsequently this led us to identify relevant CTL epitopes of HPV16, that is highly associated with cervical carcinoma in humans, for the major HLA-A alleles (i.e. HLA-A *0101, A *0201, A"0301, A*1101 and A *2401). Together these alleles cover a majority of all humans. CTL epitopes were identified through peptide-MHC binding assays followed by in v i m peptide immunizations with high affinity binding peptides to induce primary CTL responses and immunogenicity studies in HLA-A transgenic mice. Thereafter, memory CTL responses were measured in cervical cancer patients against selected peptides. Combined, these data led us to develop a CTL epitope based peptide vaccine that could be of use in HPV16 induced cervical cancer patients. A clinical trial for this disease is scheduled to start in the fall of 1994.

CLASS II PRESENTATION OF AN ENDOGENOUSLY SYNTHESIZED GLYCOPROTEIN. Carol S. re is^'.'.^, Shirley M. Bartido', Miriam Stein', and Stephanie Diment3.4, Biology Department', and Center In contrast to Class I presentation which is well characterized to use peptide fragements of proteins sythesized in the cytoplasm, exogenously administered experimental antigens enter the class II MHC pathway through endocytosis. We have been studying the recognition of the glycoprotein of vesicular stomatitis virus (VSV) which can enter either the exogenous or endogenous pathways for presentation to CD4 + T cells. Investigations of the intracellular sites involved, the proteolytic processes involved in the epitope generation, will be discussed.

The glycoprotein studied in detail is a truncated form of the wt type 1 glycoprotein, termed poison tail (Gpt) . Expressed with a vaccinia virus vector, the Gpt remains Endo H sensitive and never becomes Endo D sensitive, indicating that it is restricted to the endoplasmic reticulum. Gpt is degraded in the ER, and w e believe the degradation products include the immunogenic epitopes recognized by a panel of LAd and I-Ed T cell clones and hybridomas. lmmunofluorescence studies have confirmed the ER localization. Flow cytometric evaluations s h o w that the Gpt never appears on the cell surface, in contrast to the wt G.

The peptides generated are not secreted; using an innocent bystander assay, Gpt-infected cells are incapable of sensitizing 5'Cr-labeled uninfected APC. This contrasts with the rapid ability of supernatants from wt G-vaccinia virus-infected cells to sensitize APC for T cell recognition.

Investigations of the characteristics of the enzymes contributing to the degradation of the Gpt have shown that a reducing environment is essential, as diamide treatment of cells prevents degradation. Lysosomotropic drugs (eg. NH,CI and leupeptin) d o not alter the half-life of the protein, but do prevent presentation of the peptides; this is inconsistent with an autophagic component to the proteolysis. pH optima are physiological, as pH8 environment inhibits the enzyme activity. Inhibitors of enzyme classes are consistent with a trypsin-like, and not cystein-, cathepsin B-, or chymotrypsin-like class.

Supported by NIH grant Al 18083 to CSR. (EMCV) and Mengovirus are related members of the cardiovirus genus of picomaviruses. Their RNA genomes encode a large polyprotein which is cleaved proteolytically in co-and post-translational reactions to yield all mature viral proteins necessary to establish an infection. Although originally thought to be exclusively murine in host range, both viruses actually infect a wide range of mammals. EMCV has caused devastating epizootics in captive primates (eg: macaques, chimps and baboons), domestic pigs and exotic zoo collections (elephants, lions and tigers). Death, following ingestion of virus-contaminated material, is rapid, and caused by extensive meningoencephalitis and virus-induced damage to the CNS. Myocarditic lesions are common in older animals. When administered intracerebrally, the LD,, for EMCV strain R is about 1 pfu. We are studying the pathogenesis of EMCV and Mengo with engineered cDNA plasmids containing infectious viral sequences. Many plasmids contain h'uncated versions of the unusual 5' noncoding homopolymeric poly(C) tract that is a hallmark of these cardioviruses. Short poly(C) Mengoviruses grow very well in tissue culture but are 106-10'z fold less pathogenic to mice than the wild-type strains. Animals receiving sublethal doses of short-tract Mengo strains develop high titers of neutralizing antibodies, exhibit potent CTL responses and acquire lifelong protective immunity against challenge with wild-type virus. The genetic stability of the short-tract strains, even upon serial brain passage, mark them as safe, efficacious live vaccines. Currently, we believe the poly(C) phenomenon is due to interference by the wild-type virus sequences (long poly(C) tract) with normal cellular cytokine induction mechanisms (ie: IFa and IFP) during the initial stages of animal infection. The targeted cells are probably macrophages, and their singular ability to correctly respond or not respond to poly(C) tract length during the first few hours of infection determines whether an inoculated animal will live (protectively vaccinated) or die. The short-tract viruses probably induce IF in the macrophages, and are consequently killed then rapidly cleared from the host In related experiments we've found that attenuated Mengo strains can easily carry large heterologous insertions within their genomes, and express these sequences into protein them during replication in animals. The resulting immune response (B cell and CTL) to the chimeras is directed towards the foreign sequences (epitopes) as well as towards the Mengo proteins. A chimeric HIV vaccine, a rabies vaccine and an LCMV vaccine have been developed and tested. The LCMV chimera seems especially effective, as a single pfu of this engineered Mengo strain, administered orally to a mouse, is sufficient for complete immunogenic protection against intracerebral challenge with wild-type LCMV virus. RSV is the most common cause of serious viral lower respiratory tract disease in infants and children. We have recently renewed our efforts to generate a safe and effective live attenuated RSV vaccine for topical administration that will overcome the deficiencies of previously studied live and non-living RSV vaccines. This vaccine will be a bivalent vaccine consisting of subgroup A and B live attenuated virus components. Since the peak incidence of severe disease caused by RSV is in the 2-month old infant, an RSV vaccine will need to be effective when given to 1-month old infants. Based on the success of live poliovirus vaccines given early in infancy, it is anticipated that the intranasally administered live virus vaccine will infect and induce a protective local and systemic immune response even in infants with passively acquired maternal antibodies. The main approach that we have taken in this effort to develop the live RSV vaccine is to introduce one or more ts mutations by chemical mutagenesis into a cold-passaged virus (cpRSV) that had been partially attenuated by the acquisition of host-range mutations selected by passage in cells of a heterologous host species. We have developed a large set of cpRSV subgroup A rs mutants (termed cprs mutants) that contain the host-range mutations selected during cold passage and two or more ts mutations introduced by chemical mutagenesis. These mutants have been evaluated in virro for their level of temperature sensitivity and in vivo in rodents, chimpanzees, and humans. A large set of RSV subgroup B cpts mutants has been similarly produced and evaluated.

The immunogenicity and protective efficacy of three candidate live attenuated RSV vaccine strains that represent a specaum of attenuation were evaluated for protective efficacy in chimpanzees. Prior to infection some of these animals were given RSV immune globulin by the IV route to simulate the condition of the very young infant who possesses passively-acquired maternal RSV antibodies. The three candidate vaccine strains were immunogenic and induced significant resistance to RSV challenge in both groups of chimpanzees. Interestingly, the chimpanzees infused with RSV antibodies prior to immunization were primed more effectively for an unusually high serum neutralizing antibody response to infection with challenge virus than chimpanzees which did not receive such antibodies. This high booster response occurred despite marked reshiction of replication of the challenge virus. The evaluation of two candidate vaccines in seronegative human infants will also be described. RS virus is immunologically interesting for at least t w o reasons: 1) Upper respiratory reinfection occurs despite previous exposure and demonstrable immunological memory: 2) Humans or rodents previously immunised against virus infection can show enhanced disease during reinfection. Others have shown that passive transfer of antiviral antibody either protects against virus infection or has no effect, and there is no evidence of antibody enhancement of disease in vivo. By contrast, T cell immunity appears closely associated with disease augmentation. We have focused on examining the immunological mechanisms of disease enhancement in mice.

Initial studies showed that transfer of CD8+ cytotoxic T lymphocytes (CTL) causes rapid virus clearance from the lungs of RS virusinfected mice, but also increased disease severity with alveolar haemorrhage and polymorphonuclear (PMN) cell recruitment to the lung. This disease (reminiscent of shock lung) could sometimes be fatal, whereas normal mice recover well from similar doses of RS virus. Next, we compared the effects of CD4' and CD8+ T cells, using polyclonal T cells separated immunomagnetically from mixed lines grown in vitro with viral antigen. CD4' T cells were more pathogenic than CD8+ T cells in a dose-for-dose comparison, but that the type of pathology varied depending on the type of cell injected.

While testing recombinant vaccinia viruses expressing single RS viral proteins for their ability to protect mice against infection, we observed that animals sensitised t o the major surface glycoprotein G (attachment protein) developed lung eosinophilia after challenge with RS virus intranasally. T cell lines from the spleens of mice sensitised with various recombinant vaccinia viruses were established.

Those form mice primed with the M 2 (22K) protein were predominantly CD8' CTL, and that produced few cytokines. Those from mice primed with fusion protein (F) generated mixed T cell lines with both T h l CD4+ T cells, and CTL. Mice primed to G protein gave rise to predominantly CD4' T cells producing Th2 cytokines. ln vivo transfer of these cell lines into na'ive RSV infected mice reproduces the patterns of disease seen in mice sensitised in vivo with the respective antigens. The mouse model of RS virus disease therefore has excellent potential for illustrating mechanisms of lung immunopathology. The eye is a complex organ whose function is to transmit light images through different cell and tissue layers and liquid media to a neurosensory retina. elements as could occur when invading pathogens arrive and an inflammatory response with its swelling, plasma protein extravasation, leukocyte infiltration and tissue damage results. inflammatory responses when possible and rely on immune defenses which do not involve tissue distortion and damage. Restricting tissue damaging responses is not always effective and the process is best developed in response to agents delivered to locations such as the anterior chamber. where an inflammatory response is initiated which may result in ocular impairment. Such herpetic stromal keratitis (HSK) is a common cause of blindness in man. Animal model studies indicate that HSK is a multi-step process initiated by virus in an avascular structure. HSK fails to occur in the absence of T cells or replicating virus. disappears several days before a visible inflammatory response becomes evident. Evidence will be presented that the secondary agonists which drive the inflammatory response may not be viral antigen(s) per s e . Multiple cell types are involved in HSK, with the respective role of functional sets of lymphocytes changing according to the clinical phase of the disease. In addition, nonspecific inflammatory cells such as neutrophils and NK cells also influence the severity of lesions. Basically the reaction begins with T cells that produce type one cytokines, particularly IFN-y, dominating the scene, but during remission type 2 cytokines, notably IL-10, appear as mechanistically involved. From the use of knockout mice for various immunological parameters, evidence will be presented that numerous mechanisms of pathogenesis may be at play during HSK. Damage to corneal tissues in all systems appear to involve TNFo.

A second ocular damaging event in which immunopathology is at least partially involved is herpetic retinal necrosis. Evidence that this disease may involve the immunopathological role of CD4 T cells and protective effects by CD8' T cells will be presented, as will be suggestions by which the pathological events are mediated at the molecular level.

Thus it is in the eye's functional interest to limit Acute viral infections and live vaccines often confer long-term immunity The nature of T and B cell memory is different. B cell memory is manifested not only by the presence of memory B cells but also by continuous antibody production In contrast, the effector phase of the T cell response i s shortlived and long-term T cell memory is due to the presence of 'quiescent' antigen specific memory T cells that are present at higher frequencies and are able to respond faster upon re-exposure to virus due to increased levels of adhesion molecules In this talk I w i l l present our results on. (i) the bone marrow as a major site of long-term antibody production after acute viral infection, (ii) the role of C D 4 ' T cells and B cells (immune complexes) in maintaining CD8+ T cell memory, (iii) the role offas antigen in regulating T cell responses, and (iv) the efficacy ofvarious antigen delivery systems in inducing long-term T cell memory Sendai virus is a natural respiratory viral pathogen of mice. Intranasal infection of mice with the virus provokes a virus specific antibody-forming-cell reaction that exhibits a distinct kinetic pattern in the lymph nodes that drain the respiratory tract, in the spleen, and in the bone marrow. The bone marrow AFC population is extremely long-sustained, and supports an active humoral response that essentially persists for the lifetime of the infected animal. Thus the conventional categories of "primary" and "secondary" response may not apply to the humoral response of mice naturally exposed to respiratory viruses.

Paradoxically, the population of B cells that reacts most rapidly to Sendai virus infection does not itself secrete antibody, but can he demonstrated by the recovery of hyhridomas that secrete "polyspecific" antibodies. The activation of this polyspecific B cell population is, like the humoral response, extremely persistant. Viral infection thus sets in train multiple B cell "memory" processes. Variation in the rules of development and turnover of different B cell populations constrains the mechanisms that may operate to generate these different forms of memory.

ESTABLISHMENT AND MAINTENANCE OF T CELL MEMORY TO RESPIRATORY VIRUSES, Peter C. Doherty, Sam Hou, Christine Ewing, David Topham, Anthony McMickle, James Houston, and Ralph Tripp, Department of Immunology, St. Jude Children's Research Hospital, Memphis, TN 38105. The analysis of the development and memory phases of the CD4+ "helper" n h ) and CD8+ cytotoxic T lymphocyte (CTL) responses to the respiratory pathogens, influenza virus and Sendai virus (parainfluenza type 1) have been characterized by a combination of limiting dilution analysis (LDA) for determining Th and CTL precursor @) frequency and FACS separation of lymphocytes with different activation phenotypes. The interpretation at this stage, largely based on the analysis of the CTL response, is that the development phase of T cell memory and the primary response are synonymous.

Virus-specific CTLp are produced in considerable excess of the numbers required to provide the effector CTL that terminate the primary infection, with only a fairly small proportion localizing to the target organ (the lung) that supports virus growth. Even when many of the proliferating CTLp are killed by administration of a small dose (20 mgKg) of the DNA-targeted drug cyclophosphamide (Cy), there is no indication of immune exhaustion. The CD4+ Th response has, at this stage, not been analyzed through the course of the primary infection. Use of the LDA approach to determine Thp frequencies is inherently more difficult, as the "read-out'' is lymphokine production and there is considerable "bystander" activation in these primary responses to respiratory viruses. Memory Thp and CTLp are characterized initially by the expression of an "activated" phenotype: CD44-high, L-selectin-low, CD49d

(VLA-4) high. After some months, an increasing proportion of the memory T cells revert to the L-selectin-high CD49d-low form typical of naive CTLp. The change, which is never absolute, seems to occur first with CD49d and the rate varies for different viruses. Current experiments are addressing the possibility that intercurrent infection, particularly with the mouse y-herpesvirus 68 which causes persistent infection of lymphoid tissue, may be inducing a switch back to the activated pattern, as a consequence of "bystander" effects, or "low affmity" stimulation via the clonotypic TCR in responding lymphoid tissue. The question of such cross-reactivity and/or exposure to "high lymphokine" environments for the long-term maintenance of memory is also being addressed. To study the factors which regulate the generation and persistence of specific T cell memory we have used model systems utilizing T cell receptor transgenic mice as a source of enriched naive cells which can be either cultured in vitro to generate effector populations or restimulated in adoptive hosts. In either case one can visualize the development of an expanded effector population. We have documented that the proliferation and IL-2 production of the naive T cells depends on their activation by APC expressing high levels of co-stimulatory molecules. We find that B7.1 and ICAM-I as costimulators strongly synergize and that increased T cell receptor triggering can both increase the magnitude of the response and decrease its dependence on costimulation.

When cytokines IL-4 Vs IL-I2/IFNy are present at the initiation of the response of either CD8 or CD4 cells they dictate that the effectors generated will be polarized either towards IL-4 and IL-5 secretion or IL-2 and IFNy secretion, respectively. The fate of the effector population generated and followed in vitro, also is tightly regulated by Ag, cytokines and probably by costimulation. CD4 effector cells not re-exposed to Ag, produce no cytokines and they die within 3-4 days. Effectors restimulated with Ag make massive amounts of cytokines, regardless of the presence of cytokines, at low densities of Ag and with little dependence on costimulation. When there is little IL-2 produced and no cytokines added, effectors die rapidly by apoptosis. However the combination of IL-2 and TGFP block apoptosis and support expansion of the effector population which is greatly enhanced by periodic Ag stimulation. Some conditions favor the reversion of effector-like cells to a more resting memory phenotype and these are being further explored.

We have also examined the development and maintenance of memory after transfer of effector cells to adoptive hosts. Long-lived polarized memory populations are generated from the polarized effectors and these persist for prolonged period in the absence of apparent Ag stimulation. This supports the idea that factors other than antigenic stimulation, present in situ can support the expansion and maintenance of memory cells. The rabies glycoprotein (G) is the only external protein of the virion and is therefore responsible for any interaction that rabies makes with the host cell during the first steps of the virus cycle. The G protein is also the target of neutralizing antibodies. There are around 450 trimers of G at the virion surface which constitute the spikes visible by electron microscopy. Upon exposure to slightly acidic pH, the glycoprotein undergoes a conformational change which results in Ion er and less regular spikes. Strikingly and quite differently from influenza hemagglutinin, this conformational change is reversible: if the p d is risen back to 7.0, the s ikes re ain their neutral configuration (1). Probably as a consequence, the viral infectivity is totally preserved after an exposure of 2 hours at p 8 6.4 an cf 37T, which induces the conformational change, followed by an incubation at neutral pH. Since the conformational change is reversible, there is a pH-dependant equilibrium between the native and the low-pH conformation: the higher the pH, the more spikes are in their native configuration. Two main antigenic sites and several minor sites have been identified on the native rabies glycoprotein (2). Specific amino acids belonging to each of the two major antigenic sites are important or essential for viral virulence. For instance a lysine in position 198, which is part of antigenic site 11, is important, although not essential, for the viral virulence. Similarly, the arginine 333, which belongs to antigenic site 111, is essential for pathogenicity while dispensable for multiplication in cell culture (reviewed in 3). Viral strains mutated at arginine 333 have lost the capability to penetrate certain categories of neurons, suggesting that this mutation affected the recognition of specific receptors or subsequent interactions necessaly for the penetration of the virus at nerve terminals. Therefore the two main antigenic sites are regions of the glycoprotein which also interact specifically with neurons in the animals. We have found that neutralization requires the fixation of at least one or two IgG for every three spikes, irrelevant of the anti enic site recognized by the antibody (4). Most neutralizing antibodies recognize conformational epitopes which are accessible on the native configuration of the protein. Some epitopes remain accessible also on the acidic configuration while others are not. In addition, a minority of antibodies recognize epitopes which are only accessible on the acidic conformation. This is not unlikely in view that each spike has a certain probability to undergo a conformational change, even at neutral pH. In consequence the surface of the virus probably fluctuates and G epitopes which are not accessible on the native glycoprotein could be transiently exposed. Conformational flexibility at neutral pH and physiological temperatures has also been observed for poliovirus (5). Structural flexibility of external proteins could have important implications in virus-host interactions. Katpus, Norlhwestern University Medical School, Chicago, IL 6061 1 Theiler's murine encephalomyelitis viruses (TMEV) are endemic enteric pathogens of wild and colony-reared mice. lntracerebral inoculation of susceptible mouse strains leads to a chronic, progressive inflammatory demyelinating disease of the central nervous system (CNS) characterized clinically by an abnormal gait, progressive spastic hind limb paralysis and urinary incontinence, and histologically by parenchymal and perivascular mononuclear cell infiltration and demyelination of CNS white matter tracts. Demyelination is related to persistent CNS viral infection. Due to the similarity in clinical and histological presentation, TMEV-induced demyelination is considered to be a highly relevant model of multiple sclerosis (MS). Our current interests are in determining the phenotype, fine specificity, lymphokine profile and TCR usage of CNS-infiltrating cells involved in the effector stages of TMEV-induced demyelination. Based on a variety of experimental evidence, it is clear that demyelination induced in SJUJ mice by infection with the BeAn strain of TMEV is a Thl-mediated event: (a) disease induction is suppressed in T cell-deprived mice and by in vivo treatment with anti-I-A and anti-CD4 antibodies; (b) disease susceptibility correlates temporally with the development of TMEV-specific, MHC-class Il-restricted DTH responses and with a predominance of anti-viral lgG2a antibody; (c) activated (Le., lL-2RC) T cells infiltrating the CNS are exclusively of the CD4+ phenotype, and (d) proinflammatory cytokines (IFNq and TNF-p) are predominantly produced in the CNS. We have mapped the predominant Thl epitope on the virion to amino acids 74-86 of the VP2 capsid protein. A Thl line specific for VP274-86 exacerbates the onset of demyelination in recipient mice infected with a suboptimal dose of TMEV. TMEV-infected SJUJ mice fail to exhibit peripheral DTH and T cell proliferative responses to the major myelin proteins, MBP and PLP, and pre-tolerization with neuroantigens has no affect on the incidence or severity of TMEV-induced demyelinating disease, whereas neuroantigen-specific tolerance prevents the induction of relapsing experimental autoimmune encephalomyelitis (EAE). In contrast, tolerance induced with intact TMEV virions specifically anergizes virus-specific Thl responses and results in a dramatic reduction of the incidence and severity of clinical disease and CNS dernyelination in SJUJ mice subsequently infected with TMEV. These results have important implications for a possible viral trigger in MS as they indicate that chronic demyelination in TMEV-infected mice is initiated in the absence of demonstrable neuroantigen-specific autoimmune responses and are consistent with a model wherein early myelin damage is mediated via primarily by mononuclear phagocytes recruited to the CNS and activated by pro-inflammatory cytokines produced by TMEV-specific Thl cells. The concept that prions m novel pathogens which are different fium both viroids and viruses has received increasing support from many avenues of investigation over the past decade. Enriching fractions from Syrian hamster (SHa) brain for scrap= prion infectivity led to the discovery of the prion protein 0. Prion diseases of animals include scrapie and mad cow disease; those of humans present as inherited, sporadic and W o r n neurodegenemive disorders. The inhecited human pion diseases m genetically linked to mutatim in the PrP gene that result in non-conswative amino acid substitutions. Transgenic v g ) mice expressing both SHa and m o w @lo) PrP genes were used to demonstrate that the "specie9 bank?' for -pie prions resides in the primary structure of PIP. This concept was strengthened by the results of studies with mice expressing chimeric MdSHa transgenes &om which "artificial" prions have been synthesized. Similar chimeric Mdhuman (Hu) RP transgenes were constructed which differ from M O W by 9 amino acids between residues 96 and 167. AU of the Tg(MHu2M) mice developed neurologic drsease -200 days after inmulation with brain homogenates from three patients who died of Creutzfeldt-Jakob disease (CJD). Inoculation of Tg(MHu2M) mice with CJD prims produced MHu2Mprpsc, inoculation with Mo prions produced MoPrW. Ihe patterns of MEluZMPrPc and MOM% accumulation in the brains of Tg(MHu2M) mice wen differenl About 10% of Tg(HuPrP) mice expressing HUF" and non-Tg mice developed neurologic diseane >500 days after inoculation with Cn, prions. The different susce@Uies of Tg(HW) and Tg(MHu2M) mice to human prions indiate that additional species specific factors such as chaperone proteins are involved in prion replicaton. Diagnosis, prevention and treament of human @on diseases should be faciliated by Tg(MHu2M) mice. In other sindies, Tg mice were compared expressing wt and mutant MoPrP. Overexpression of the wtMoPrP-A aansgene -8-fold was not deleterious to themiw but it did shorten scrapie incubation times from -145 d to -45 d after inoculation with murine m p i e pnons. In contrast, overexpression at the same level of L MoRP-A transgene mutated at codon 101 (corresponding to codon 102 in HuRP) pmdnced spontaneous, fatal neurcdegeneration between 150 and 300 d of age in two lines of Tg(MohP-PIO1L) mice designated 2866 and 2247. Genetic crosses of Tg(MoPrP-P101L)2866 mice with gene targeted mice lacking both RP alleles ( P m -p ) produced anhats with a highly synchronous onset of illness between 150 and 160 days of age. The T g~o P r P -P l O l L ) 2 8 6~~ mice had numerous PrP plaques and widespread spongiform degeneration in contrast 10 the Tg2866 and 2247 mice that exhibited spongifonn degeneration but only a few PrP amyloid plaques. Another line of mice designated Tg2862 overexpress the mutant transgene -32-fold and develop fatal neurodegeneration behveen 200 and 400 d of age. Tg2862 mice exhibited the most severe spongiform degeneration and had numerous, large PIP amyloid plaques. While mutant MoPrPCCplOlL) clearly produces neurodegeneration, wtMoprpC profoundly modifies both the age of onset of illness md the mumpathology for a given level of transgene expression. Our tidigs and those from other smdies suggest that mutant and wtPrP interact, phaps through achaperone-like protein as noted above in SNdieS of Tg(MHu2M) mice, to modify the pathogenesis of the dominantly inhe&ed prion diseases.

Anton, Heidi T. Link, and Jonathan W. Yewdell, Laboratory of Viral Diseases, NIAID, Bethesda, MD 20892-0440. CD8' lymphocytes (TCD8+) play an important role in host immunity to viruses and other intracellular parasites. Virus-specific TCDI+ recognize MHC class I molecules in association with peptides of 8 to 10 residues derived from viral proteins. This presentation will focus on how and where antigenic peptides are generated by cells. To begin to characterize the nature of proteases involved in the generation of antigenic peptides from cytosolic proteins, we used a panel of recombinant vaccinia viruses expressing different forms of influenza virus nucleoprotein (NP). We found that the efficiency of generation of two NP peptides is related to the metabolic stability of the source gene product. There has been considerable speculation that such short lived proteins are degraded by proteasomes in a ubiquitin-targeted process. Our observations, however, call into question the importance of ubiquitin targeted-proteolysis in generating antigenic peptides from exogenously provided or endogenously synthesized viral proteins. We also examined the extent to which antigenic peptides can be generated in the endoplasmic reticulum (ER). We found that antigenic peptides could be produced from short precursors (17 residues) hut not from a number of full length proteins (influenza virus hemagglutinin, NP, ovalbumin) that are targeted to the ER by a NH2-terminal signal sequence. Peptides were generated much more efficiently from the COOH-terminus of the 17 residue precursor than from the NH2-terminus. These findings indicate that the ER has a much more limited capacity than the cytosol to generate antigenic peptides, but that ER proteases (particularly aminopeptidases) could perform the final proteolytic steps in the generation of class I binding peptides from precursors imported from the cytosol by TAP, the MHC encoded peptide transporter.

Potential advantages of synthetic peptide or engineered recombinant vaccines are that they can be limited to contain only the specific antigenic determinants for desired responses without other determinants that elicit unwanted responses, and that the sequences of the determinants themselves can be modified to enhance potency or breadth of crossreactivity. However, they can have the disadvantage that any single determinant may be presented by only a limited selection of major histocompatibility complex (MHC) molecules of the species. To overcome the problem of MHC polymorphism, we have identified determinants presented by multiple MHC molecules, and have also located multideterminant regions of the HIV-1 envelope protein that contain overlapping determinants each presented by different class I1 MHC molecules, so that the whole multideterminant region is presented by multiple MHC molecules of both mouse and human. We have made use of "cluster peptides" spanning these multideterminant regions of the HIV-1 envelope to provide help for neutralizing antibody (Ab) and CD8+ cytotoxic T lymphocyte (CTL) responses to peptides attached to these helper regions. These synthetic peptide vaccine constructs containing the P18 peptide from the V3 loop of HIV-1 IIIB or MN, elicited both neutralizing Ab and CTL in multiple strains of mice. The cluster peptides inducing helper T cells were essential for elicitation of Ab and CTL to the P18 segment of both IIIB and MN strains of HIV-1 in mice of several MHC haplotypes. Several adjuvants were compared for their ability to elicit both CTL and Ab simultaneously, without one response inhibiting the other. A single formulation in incomplete Freund's adjuvant (IFA) could elicit all 3 responses, neutralizing Ab, CTL, and TH1 helper cells. The CTL specific for the MN strain P18 peptide crossreacted with strains SC, SF2,2321, and CDC4. The peptides in F A also elicit high titers of antibodies in rabbits. Boosting was found to enhance CTL responses as well as Ab responses. These constructs are being prepared for a human immunotherapy trial. These vaccine constructs are potent and also avoid sites on gp160 that are known to elicit enhancing antibodies or autoimmune responses that might conmbute to disease pathogenesis. However, we can potentially improve on these by tinkering with the internal structure of the individual epitopes. We have found that replacing a negatively charged glutamic acid residue with an uncharged amino acid in one of the helper determinants makes it 10 to 100-fold more potent in binding to the class I1 MHC molecule and in eliciting murine helper T cells that still recognize the natural HIV-1 sequence. Thus, such a modified peptide should be more potent as a vaccine, while retaining the ability to elicit T cells that will respond to HIV proteins that of course do not have the altered sequence. We are currently mapping the critical residues for presentation of one of these peptides by human HLA-A2, with the intent of developing modified peptides that will be more potent as components of a human vaccine. Thus, by leaming how these peptides bind to MHC molecules and Tcell receptors, we can design internally modified determinants to construct more potent or more crossreactive second generation vaccines. We are testing these vaccine approaches in a mouse model in which mice can be protected against tumor cells expressing HIV proteins as would an HIVinfected cell. DNA vaccines, comprised of non-replicating plasmids encoding viral proteins, are capable of generating protective immunity in animal models of several viral diseases. In preclinical models of influenza infection, reduced viral shedding was observed in DNA-vaccinated ferrets after challenge with the human clinical virus strain, A/Georgia/93. Cross-strain protection was conferred by DNA encoding the major internal proteins (nucleoprotein, NP, and matrix, M1) and the surface protein haemagglutinin (HA) from the antigenically-distinct previous virus strains, A/Beijing/89 and A/Hawaii/91. This protective efficacy was greater than that seen by immunization with the widely-used clinical vaccine composed of killed A/Beijing/89 virus. Thus, compared to a killed virus vaccine, protection seen with the DNA vacane against a drifted virus strain was greater. We previously demonstrated that immunization of mice with NP DNA generated MHC Class I-restricted cytotoxic T lymphocytes.

Mice likewise were protected from death and morbidity following cross-strain challenge'. HA DNA vaccines generated neutralizing antibodies in mice, ferrets and primates, and provided protection in m i d and ferret models of influenza.

In animal models of other viral diseases, immune responses and protection against viral challenge have been seen after immunization with DNA encoding viral proteins. DNA encoding HIV gp120 generated CTL and neutralizing antibodies in monkeys. Antigen-specific proliferative responses and, in mice, secretion of high levels of yIFN relative to levels of IL-4, months after immunization were also observed . Immunization of rabbits with DNA encoding L1, the major viral capsid protein of Cotton tail Rabbit Papilloma Virus (CRPV), resulted in neutralizing antibodies and protected against the development of warts after inoculation with CRPV. Mice immunized with DNA encoding the glycoprotein gD from Herpes Simplex Virus type 2 (HSV-2), developed neutralizing antibodies and were protected from death when subsequently challenged with HSV-2.

DNA vaccines were protective in animal models of various viral diseases. Neutralizing antibodies, helper T cells (Thl) and cytotoxic T cells were generated. Cross-strain protection due to cellular immunity was demonstrated.

' Science, 1993 2593745-1749 , 2DNA Cell Biol, 1993 The profile of a neurovirulent virus is determined by its mechanism of entry into the CNS (neuroinvasion), the type of CNS cell in which it replicates (neurotropism) and its ability to cause pathologic effects in the brain (neurovimlence). Whereas neuroectodermal cells, especially neurons, are the target cells of most neurovirulent viruses, the main target cell in the brain for SIV and other lentiviruses is the macrophage. Infection in, expression of viral antigens by and products of SIV replication exported from these cells result in inflammation and degenerative changes in the brain and concomitant loss of neurons. SIV strains that are mainly T-cell tropic cause transient activation of T-cells and during this period, infected T-cells cross the blood brain barrier and localize in the brain causing persistent hut minimally productive infection and minimal neuro pathologic effects. Viral proteins but not virions are produced continuously. By virtue of the tropism of the virus for CD4 T cells, many infected animals eventually become immunosuppressed and develop AIDS, but not classical ueurological disease. Viruses which are macrophage tropic invade the brain presumably also in T lymphocytes and the viruses infect macrophages in the brain. However, productive virus replication is minimized by antiviral CD8 T cells which suppress (kill?) all virus producing cells throughout the body, including the CNS. Productive virus replication in brain macrophages and accompanying inflammatory changes develop only when CD8 cells fail i.e. after profound immunosuppression sets in. The neurological disease that results from productive virus replication in macrophages in the brain therefore depends on presence of an appropriate macrophage-tropic viral phenotype invading the neuropil and development of immunosuppression in the host. The neurological disease could therefore be defined as one of the AIDS syndromes. The adenovirus (Ad) early transcription region (E3) codes for more than 7 polypeptides, four of which have already been shown to alter the immune response to Ad infection. The amount of the class I major histocompatibility complex (MHC) on the plasma membrane can be reduced by the binding of the Ad E3 gpl9K protein to the MHC heavy chain, which prevents transport of the complex out of the endoplasmic reticulum. This process interferes with presentation of viral peptides to cytotoxic T lymphocytes. Cytolysis by tumor necrosis factor-o (TNF) is inhibited by 4 distinct viral polypeptides, 3 of which (the Ad E3 14.7K or the complex of the 10.4K and 14.5K proteins) are coded in the E3 region. The E3 polypeptides are translated from a family of viral mRNAs, that are synthesized from a single viral promoter and processed by alternative splicing. We have studied the functions of the E3 polypeptides in several murine models. The goals of these experiments were to determine the effects of the Ad E3 polypeptides in acute and persistent viral infections as well as in a transplantation model designed to measure whether these viral immunoregulatory proteins would abrogate allogeneic graft rejection. In a vaccinia virus (V.V.) pneumonia model, in which the isolated Ad E3 14.7K or Ad E3 gpl9K genes were inserted into the V.V. pathogen, the Ad anti TNF polypeptide increased viral virulence but the Ad anti MHC had no effects.

In addition to manipulating the Ad E3 genes in viral constructs, several transgenic mouse lines containing the Ad E3 genes have been constructed for these experiments. The E3 genomic DNA behind the rat insulin promoter (RIP) has been used to generate transgenic animals. Islets from RIP-E3 transgenic animals (H-2b'd) have been transplanted allogeneically to H-2d recipients and remained viable, secreting insulin until the end of the experiment at 94 days; in contrast, control nontransgenic islets of the same genotype were rejected by 21-28 days. The E3 genes behind the native E3 promoter have been inserted into mouse embryos to generate transgenic animals, and the expression of the transgene monitored in multiple organs. The E3 promoter of the transgene is responsive to stimulation by the Ad E1A following infection with an E3 minus Ad 7001 and can also be upregulated by administration of bacterial lipopolysaccharide. The effects of this transgene on Ad pathogenesis are currently being studied. Thus, these viral immunoregulatory genes have been shown to alter viral pathogenicity during acute infection and to downregulate the host immune response sufficiently to permit islet cell transplantation. These results on manipulating the Ad E3 genes for the control of the host immune response also have implications for designing adenovirus vectors for gene therapy. "Emerging" infections can be defined as infectious diseases that either have newly appeared in the population, or that are rapidly increasing their incidence or expanding their geographic range. Recent viral examples include AIDS, Ebola, and Hantavirus Pulmonary Syndrome (fnst identified in a 1993 outbreak in the southwestern U.S.). Emerging viral infections show a number of common features. Most "new" viruses derive from existing viruses that move into new areas or acquire new hosts ("viral traffic"). Many are zoonotic (originating from animal sources) (even pandemic influenza appears usually to be a reassortant originating in wildfowl). Ecological or environmental changes (either natural, or, often, man-made) may precipitate emergence of new diseases by placing people in contact with a previously unfamiliar zoonotic reservoir or by increasing the density of a MtUd host or vector of a pathogen, increasing the chances of human exposure. Upon introduction into a human population from a zoonotic reservoir, the newly introduced virus may cause localized outbreaks of disease. Some may show rapid variation and evolution upon introduction, and some evidence suggests a role for immune selection in this process. A few viruses (such as HIV) may succeed in establishing themselves and disseminating in the human population, becoming truly "human" infections. Human activities can also play an important role in establishment and dissemination. Migrations from rural areas to cities, now an accelerating worldwide phenomenon, or other displacements, can introduce remote viruses to a larger population; the virus may then spread along highways and (globally) by air travel. The development of an effective system of surveillance and rapid response is essential, but resources for this are presently inadequate. Vaccine development, production, and deployment problems also need to be addressed. Immunopathology may be a key feature of many of these infections, a number of which manifest as hemorrhagic fevers. Many of the life threatening complications are due to increased vascular permeability. The resemblances to septic shock suggest that cytokines (such as TNF) are likely to be important in the pathogenesis of these infections. The response of cells, such as the macrophage, that induce or synthesize key cytokines, may be an important element, and the ability to infect these cells may be one common denominator. Why some viruses elicit this response, while other closely related viruses do not, cannot yet be predicted from molecular data. Better understanding of these aspects of the immune response should lead to additional therapeutic strategies. (Supported by NIH grant ROI RR03121.) Genetic approaches have been used to detect and characterize numerous previously unidentified hantaviruses. PuumalaRrospect HilVSin Nombre-like viruses or virus variants are present throughout North and South America, Europe and Russia. Several of the American viruses identified are associated with the newly recognized Hantavirus Pulmonary Syndrome (HPS), a severe respiratory illness with high mortality. The genetic relationships of these and previously characterized hantaviruses have been studied by phylogenetic analysis of the nucleotide sequence differences located in PCR bgments amplified from the G2 encoding region of the virus M segments. The relationships observed are consistent with a long-term association of viruses with their primary rodent reservoirs and suggestive of coevolution of host and virus. A Sin Nombre virus isolate is now available and its genetic characterization has been completed. Various virus antigens have been expressed and are being used to probe the interaction of the virus with the host immune system.

Hantaviruses cause significant morbidity and mortality throughout the world. More than 200,000 cases of hemorrhagic fever with renal syndrome (HFRS) are reported annually in Asia, Europe and Scandinavia. The etiologic agents of HFRS are Hantaan, Seoul and Puumala viruses, with Hantaan virus causing the most severe form of the disease. In 1993, a new hantavirus was discovered in the United States (initially termed Four Comers virus), and was identified as the etiologic agent of hantavirus pulmonary syndrome (HI'S). Vaccines for hantaviruses are not readily available, although a number of inactivated viral preparations have been made and tested in Asia. Recurrent problems with inactivated hantaviral vaccines have been lot to lot variability, the need for repeated immunizations, and their inability to elicit long-lasting neutralizing antibody responses in immunized volunteers.

Because of such limitations on traditional vaccine development for these viruses, as well as the viruses' hazardous nature and slow, low-titer replication in cell culture, we used a recombinant DNA approach to develop a vaccine for I-IFRS. Our vaccine is a recombinant vaccinia virus expressing the M segment of Hantaan virus under control of the vaccinia virus 7.5 K promoter and the S segment under control of the 11 K promoter. The M segment, which encodes the G1 and G2 envelope proteins, was included because of our findings that: (1) immunization with vaccinia or baculovirus-expressed G1 and G2 induced a neutralizing and protective immune response in hamsters; and, (2) neutralizing antibodies to G1 or G2 could passively protect hamsters from challenge with virulent virus. The S segment, which encodes the nucleocapsid protein (N), was included because of our finding that hamsters immunized with baculovirus-expressed N also were protected from subsequent infection. Although the protective immune response to N is probably cell-mediated, the importance of such a response is presently not well defined. Assessment of our vaccine in preclinical studies, indicated that immunized hamsters developed neutralizing antibodies and were protected from displaying viral antigen in their lungs after challenge. In a Phase I, dose escalation, clinical study, the vaccine induced neutralizing antibodies in individuals immunized subcutaneously with approximately lo7 PFU of the recombinant virus. In addition to humoral responses, immunized volunteers developed a cell-mediated immune response as indicated in lymphocyte proliferation assays. Larger clinical studies, including alternate routes or booster immunizations, are planned. Based on these studies, we anticipate that the vaccine will be efficacious for preventing HFRS caused by Hantaan and the antigenically closely related Seoul virus. We are studying the cross-protective properties of this vaccine with more distantly related hantaviruses such as Puumala virus. Although we expect this vaccine to be safe as well as effective, we also are investigating the use of more attenuated pox-viruses as vaccine vectors. Infection of mice with lymphocytic choriomenigitis virus (LCMV) results in a profound expansion in the number of spleen CD8 T cells and in the induction of virus-specific CTL activity. Thereafter, the CD8 T cell number declines, and the CTL activity diminishes, though the frequency of LCMVspecific precursor CTL per CD8 cell, as assessed by limiting dilution assays (LDA), is remarkably stable throughout long-term immunity. The decline in T cell and total spleen leukocyte number at the late stages of acute infection is associated with high levels of apoptosis, as detected by the in situ nucleotidyl transferase assay. Apoptosis occurred in both the T cell and B cell populations, with the B cells dying in clusters. This apoptosis was also seen in tfansgenic mice ectopically expressing Bcl-2 in the T and B cells and in C57BL/6 Ipr/@r mice, which have a mutation in the fas gene. T cells from the infected animal underwent apoptosis in vitro when stimulated through the TcR with anti-CD3, thereby explaining some of the immunosuppression seen during acute viral infections. Memory cells persisted for over a year and could be found in blast-size cell populations. Challenge of LCMV-immune mice with either Pichinde virus, vaccinia virus, or murine cytomegalovirus led to the reactivation of the LCMV-specific CTL response. LDA analyses showed unexpectedly that these heterologous viruses crossreacted with subpopulations of LCMV-specific memory T cells. This memory T cell response to virus from an earlier infection was associated with enhanced immunopathology and enhanced clearance of virus during a heterologous virus challenge. Over the course of the acute infection, CTL specific for the second virus were preferentially expanded over the crossreactive CTL, and after the acute infection, when the T cell response had subsided, CTL memory to the first infection had decreased. There is therefore a network of memory T cells which contribute to and are modulated by infections with putatively unrelated viruses, and apoptosis plays a homeostatic role in the course of these T cell responses.

IMMUNE RESPONSES TO LIVE ATTENUATED RETROVIRAL VACCINES, R. Paul Johnson*?, Cara Wilsont, Kelledy Mansons, Michael Wyands, Bruce Walker?, Ronald C. Desrosiers* *New England Regional Primate Research Center, Southborough, MA 01772 thfectious Disease Unit, Massachusetts General Hospital, Boston, MA 021 14 §TSI/Mason, Worcester, MA Immunization of rhesus macaques with live attenuated retroviruses deleted in nef can induce protective immunity against challenge with pathogenic SIV. Development of protective immunity in these vaccinated animals occurs only after several months of infection, with maximal protection observed after one year. The specific immune responses responsible for mediating protection have not been defined, and little is known about the cellular immune responses in animals vaccinated with these live attenuated retroviruses. We have analyzed cellular and humoral immune responses in rhesus macaques and chimpanzees infected with live attenuated retroviruses. SIV-specific neutralizing antibodies were present in vaccinated animals, but did not clearly correlate with protection against challenge. CTL specific for envelope and gag were identified in vaccinated macaques studied 12 or more months after vaccination. Quantitation of SIV-specific CTL activity in one of these animals using limiting dilution analysis revealed a relatively high precursor frequency of cytotoxic T lymphocytes, up to MOO0 for gag and 1/8500 for envelope. CD8+ lymphocytes obtained from vaccinated macaques were also able to suppress SIV replication in autologous CD4+ cells. Suppression mediated by unstimulated CD8 autologous cells was maximal when cells were in direct contact with SIV-infected lymphocytes, but CD8+ cells activated by an anti-CD3-specific monoclonal antibody were able to release. a potent soluble inhibitor of SIV replication. In contraSt to the relatively vigorous CTL response present in vaccinated macaques, we were not able to detect consistent CTL activity in chimpanzees infected with a HIV-1 molecular clone (NL43) or attenuated viruses at periods up to one year after infection, despite the use of a variety of stimulation techniques. Proliferative responses to HIV p24 and gp160 were observed in chimpanzees infected with N U 3 and attenuated variants. Although the relative contribution of these immune responses to protective immunity is not known, the relative vigor of the cellular immune responses observed in vaccinated macaques suggest they may play a role in mediating resistance to challenge. Obiectives: To analyze the magnitude and specificity of the CTL response to HIV-1, and to determine the TCR usage by clonal CTL responses in infected persons, including persons with documented infection of up to 15 years with CD4 cells > 500/mm3. Methods: HIV-l-specific CTL activity was evaluated in PBMC as well as in PBMC stimulated in vitro with HN-1 infected autologous CD4 cells, using target cells infected with recombinant vaccinia viruses expressing HN-1 proteins. CTL epitopes recognized by these individuals were determined using cloned effector cells. Quantitative cultures were performed by endpoint dilution, and viral quantitation was determined by QC-PCR TCR analysis was performed by PCR, using both family-specific primers and anchored PCR, followed by sequencing. Sequence analysis of CTL epitopes in autologous viruses was determined by PCR amplification and sequencing. Clonal frequency was analyzed in PBMC by oligonucleotide probe to the CDR3 region of the TCR.

Studies performed in long-term non-progressing persons indicate the presence of a vigorous and broadly directed CTL response. Detailed epitope mapping in a person infected for 15 years, who by QC-PCR had <I67 viral RNA molecules per ml and who tested negative by bDNA assay, revealed CTL clones directed at six different epitopes, with detectable recognition of three of these epitopes using freshly isolated PBMC as effector cells. The CTL clones were broadly reactive against many hown HIV-1 variants, and both the p17 and p24 epitopes were cross reactive with SIV. Sequence analysis of autologous virus within the three dominant CTL epitopes revealed the lack of immune escape variants in PBMC, plasma, and virus obtained from coculhue. supernatant. Analysis of TCR usage by clones to defined epitopes in persons at different disease stages revealed heterogeneity in TCR usage and heterogeneity in recognition of virus variants withii these epitopes. Analysis of clonal frequency using an oligonucleotide probe to the CDR3 region of the TCR in an individual with a vigorous response to gp41 suggested that approximately 6% of circulating T cells were a single HIV-l-specific clone. Conclusions: Our results indicate that cytotoxic T lymphocytes are part of the immune response in persistent non-progressive HIV-1 infection, and that prolonged infection can occur without the development of viral immune escape variation within defined CTL epitopes. In addition, vigorous circulating CTL responses can occur in the setting of an e x m e l y low viral load. Heterogeneity in the ability of clones from different individuals to recognize sequence variation within CTL epitopes may be important in the pathogenesis of infection. CTL to conserved epitopes or CTL which are broadly reactive with multiple viral mains may be important in conmbuting to maintenance of the asymptomatic state. factor-I (IRF-1) gene is essential for the transcriptional activation of inducible nitric oxide synthase (NOS) in murine macrophages. It also plays an important role in the activation of Interferon and IL-6 and IL-7 receptor genes. To investigate the role of IRF-1 related systems of defense against viral myocarditis, the transgenic mice were innoculated with l o 5 pfu of coxsackie virus 83 (CVB3). Homozygous IRF-1 (+) and heterozygous IRF-1 (+/-) mice were randomized into groups to determine heart and spleen viral titres, cardiac pathology and mortality rates at early and late stages of the disease.

The mortality was dramatic in the IRF-1 -Imice. The hearts of the homozygous animals also had increased mononuclear cell infiltration, necrosis and viral replication. We condude that IRF-1 regulated gene products such as iNOS and IFN are essential for the early defense against CVB3-induced myocarditis by attenuating viral replication and inflammatory responses. The flu season is a yeariy problem, especially in the elderly population. Annual vaccination lacks broad protection and antibody titers induced in the elderly are frequently lower than those acheived by younger vaccinees. Using the mouse model, we are investigating the differences in immune response between old and young mice to influenza vaccine and at the same time testing the ability of a potent adjuvant to boost the response in the the elderly.

Our studies show defects in the immune response of the elderly. Antibody titers after immunization are lower in the elderly for both seronegative and seropositive mice. CMI studies show that helper T-cell response (proliferation) to in-vitro antigen stimulation is also decreased in the elderly. FACS analysis of WBCs show the elderly mice have lower leukocyte and lymphocyte populations, lower CD4 and B-cell populations and a higher proportion of macrophages and CD8 cells than young mice. It was also noted that the young mice had very consistent WBC profiles while the elderly profiles had greater variability. Serum cytokine studies show lower levels of IL2, IL4 and IL5, but higher levels of IL6 and IFNg in the elderly as compared to the young mice.

The addition of MF59 to the immunization increased the antibody titers of both naive and seropositive young and old mice. The helper T-cell response of the young mice was unaffected, while the response of the elderly group was increased. Cytokine profiles show an increase in IL2, IL4 and IL5 levels for both young and old, while IL6 and IFNg levels went up for young animals but remained constant for the elderly mice. Coxsackievirus 8 3 (CVB3), a member of the Picornavirus family, is a common cause of acute or chronic human myocarditis. However, whether cell damage results primarily from direct virusmediated effects or from immune-mediated destruction of the heart tissue during the infection remains unclear. We used a cardiovirulent variant of CVB3 which is able to infect several strains of mice. The infection results in a disease which strongly implicates immunopathogenetic mechanisms in myofiber necrosis. To characterize the role of the immune system in this disease we used various transgenic knockout (ko) mice with different immune defects. CVB3 is able to infect normal C57BL/6 mice and immunocompromised (CD4 ko and 62M ko) mice and causes a lethal disease after i.p. injection in a viral dose dependent manner. We found elevated LD5Os in immunocompromised mice, which also die slightly later than normal mice. This virus replicates in the hearts of all mice used with maximal titers 3-5 days pi.. Large pathological changes were detectable only in heart tissue of CD4 ko mice, and were maximal at 1-2 weeks after infection. These results suggest the possibility that CD4' T cells may limit tissue damage by an as yet undefined mechanism; the roles of CD4+ and CD8' T cells in this disease are under investigation.

Research supported by a grant of the DAAD, Germany.

MICE LACKING p56lck ARE RESISTANT TO COXSACKIEVIRUS B3 INDUCED HEART DISEASE.

Tamara Martino, Karen Aitken, Joseph Penninger, Jan0 Nikhilanandhan, Tak Mak, Fayez Dawood, Wen-Hu Wen, Lily Wee, Martin Petric, and Peter Liu, The Toronto and Princess Margaret Hospitals, University of Toronto, Canada. Coxsackievirus 83 (CVB3) causes severe heart disease in adult mice, and is a useful model for the human cardiac conditions of myocarditis and dilated cardiomyopathy. In this study, we have demonstrated that transgenic mice lacking the T-cell signalling molecule p56kk are resistant to CVB3 induced heart disease. The virus replicates in the heart, and is cleared from the myocardium at a rate comparable to "normal" littermates, but there is only minimal mortality, or damage to the cardiac tissue. However, in Ick-deficient mice depleted of natural killer (NK) cells prior to viral infection, there is increased CVB3 replication, and some infiltration by mononuclear cells in the myocardium. To further elucidate the role of the immune system in CVB3induced heart disease, cytokine expression in the hearts of CVB3 infected normal, Ick-deficient, and NK-cell depleted mice is under investigation. We conclude that T-cell activation is critical to produce substantive myofiber necrosis after CVB3 infection, that NK-cells play an fundamental role in protecting the mice from severe virus infection, and that Ick transgenic mice serve as an important model for understandina the Dathogenic mechanisms involved. We now show that CD4+ T cells from 62m-/-mice exhibit antigen specific release of serine esterase and interferon-gamma. Further, as quantitated by flow cytometry, the early decrease in CD4+ T cells seen in normal mice 3 days after LCMV infection, previously presumed to be due to the activity of CD8+ CTL, still occurs in 62m-/-mice. However, 621114mice show an earlier rise in percentage and absolute numbers of CD4+ T cells by 7 days after infection. CD4+ T cells from am-/mice rise to above baseline levels with kinetics paralleling the early rise in CD8+ T cells in normal mice after LCMV infection. CD4+ CIZ activity is lower in 62m-/mice and have a later peak than CD8+ CTL from normal mice. These differences may explain the lower mortality and longer time course of immunopathologic meningitis in 621114-mice. mice can assume some of the cytotoxic functions of the CD8+ cell population from normal mice, including immune-mediated pathology. Further, CD4+ T cells from 62m-/-mice have the capacity to release serine esterase, and show kinetics similar to CD8+ T cells from normal mice.

Mice genetically engineered to be deficient in 62-microglobulin

In conclusion, these data suggest that CD4+ T cells from am-/- Australia.

molecule is an essential component of the T cell help required for an antibody response. The interaction between CD40 and its ligand, in the presence of B cell acting cytokines, such as interleukin 4 (IL-4), is sufficient to trigger B cells to proliferate and differentiate and to induce immunoglobulin class switching. A recombinant vaccinia virus encoding both factors, CD40L and IL-4, did not, however, stimulate an enhanced antibody response in mice infected with the construct. Instead, the antibody levels were lower than those of mice infected with a control virus. The reduced antibody response reflected the diminished growth of CD4OL-expressing viruses, to the extent that immunocompromised mice were able to resolve these infections. The kinetics of virus clearance indicate that the anti-viral mechanism of CWOL is rapidly activated and has generalized tissue distribution. The CWOLmediated clearance of virus is independent of the anti-viral cytokines, TNF and IFN-y. The anti-viral activity of C W L may represent a surprising and potent effector mechanism of T cells activated during a virus infection. Mary's Medical School, Norfolk Place, London W2 lPG, U. K. and *Nationallnstirute forMedica1 Research, London, NW7 1AA.

Respiratory Syncytial (RS) virus is the most important respiratory pathogen of infants, causing the majority of cases of bronchiolitis, it can be life threatening in very young infants and those with underlying cardiopulmonary disease or immunodeficiencies. T cell immunodeficiency can lead to prolonged infection in children and T cell transfers clear virus in the murine disease model. Mice can be readily infected with RS virus and have proved a valuable model of its pulmonary pathology and the T cell response to it.

IL-3 has a broad spectrum of activities on proliferation and terminal differentiation of early blast cells and committed progenitors of many lineages. Less is known about the effects of IL-3 on T cell development and proliferation and its effect on the host response to infection by common pathogens. The IL-3 gene was disrupted in 129/Sv embryonic stem (ES) cells by homologous recombination after transfection with omega-type constructs flanked with a herpes simplex thymidine kinase gene. These cells were used to produce a mouse strain deficient of IL-3. Mice were infected intranasally with A 2 strain RS virus; pathological changes in the lung were monitored by cytology of bronchial lavage, and virus persistence by nested RT-PCR of lavage fluid and lung homogenate. No differences were found in pathology or virus persistence between knockouts and normal mice. IL-3 deficiency is therefore compatible with apparently normal responses to viral infection. Recombinant adenoviruses are an atuactive vehicle for gene therapy to lung in the treatment of cystic fibrosis (CF). First generation viruses deleted of Ela and Elb transduce genes into airway epithelial cells in vivo, however, expression of the transgene is transient and associated with substantial inflammatory responses, and gene transfer is significantly reduced following a second administration of the virus. In this study, we have used mice deficient in immunological effector functions in combination with adoptive and passive nansfer techniques to define antigen-specific cellular and humoral immune responses that underlie these important limitations.

Our studies indicate that MHC class I resmcted CD8+ cytotoxic T lymphocytes (CTLs) are activated in response to viral antigens leading to destruction of virus infected cells and loss of transgene expression. MHC class II associated presentation of viral antigens activates CD4+T helper (TH) cells of the -1 subset and, to a lesser extent, of the Tm subset. LDV establishes a life long viremic infection in mice which is not controlled by anti-LDV immune responses. Anti-LDV antibodies are rapidly generated but they neutralize LDV infectivity only poorly. Here we show that LDV-specific cytotoxic T cells (CTLs)are generated within 7 days pi. C T L s were detected by H-2 specific lysis of LDV-specific target cells. These target cells were generated by transfection of 3T3-NM cells with a retrovirus vector carrying the gene (OW 7) for the LDV nucleocapsid (N) protein. Spleen cells from LDV-infected mice were incubated in vitro with LDV-infected macrophages or transfected 3T3-NIH cells and cultured for 5 days in the presence of IL-2. The CTLs had largely disappeared in mice by 30 days pi. The following experiments were conducted to further explore the mechanism of immune escape. In situ hybridization was used to localize LDVinfected cells in tissues of infected mice. These were found in persistently infected mice only in the liver. testis, and lymphoidal tissues. In addition large amounts of virion or virion debris accumulated in the germinal centers of the spleen and lymph nodes beginning about 3 days p.i. The accumulation of large amounts of LDV antigen in the germinal centers may be causally related to the suppression of the CTL response as well as to the polyclonal activation of B cells associated with LDV infection. To determine whether immune selection plays a role in LDV immune escape, LDV was reisolated from nude and immunocompetent BALB/c mice at various times pi. The OWs for the Nand the two envelope glycoproteins were R T K R amplified and sequenced. The role of cytokines in initiating the immune response to Venezuelan Equine Encephalitis (VEE) was investigated. Mice were infected in the left rear foot pad with either cloned virulent VEE (V3000) or an isogenic single-site attenuated VEE mutant (V3032) (single amino acid change at E2 glycoprotein position 209 from glu-lys) and at 1,6,24,48,72, and 96 hours post-infection, the popliteal lymph nodes and spleen were harvested and examined for TNF-a, IL-12, IFN-y, IL-2, IL-4, IL-6, and IL-10 gene expression using a quantitative RT-PCR. In both strains elevations of TNF-a, IL-12, IFN-y, IL-10, and IL-6 were detected in the lymph node, with no changes in either IL-2 or IL-4; however, whereas cytokine elevations were detected by 6 hours after injection of V3000, elevations were delayed until 24 hours following infection with V3032. Cytokine mRNA elevations in the spleen were less substantial, but elevated levels of IFN-y, IL-10, and IL-12 were also observed. These findings suggest that VEE infection induces an early highly specific cytokine pattern associated with simultaneous elevations in both IFN-y and IL-10 and that a single amino acid mutation can induce temporal changes in this pattern without alterring either the actual cytokines induced or the degree of their elevation. We are currently performing intervention experiments to investigate the effect of intravenous anti-IL-12 andor anti-IFN-alp antibody administration on this early in vivo cytokine response to examine in particular whether these cytokines may be required for the observed elevations in either IFN-y or IL-10. To confirm the importance of the immune system in clearing rotavirus infection we infected Rag 2 knockout mice with murine rotaviruses. Both knockout p u p s and adults developed chronic infections.

To determine if a T cell independent antibody could be mediating the clearance of rotavirus infection in nude mice we measured IgG, IgM and IgA in faeces and serum of rotavirus infected mice. A low and transient IgM response was found in serum but not faeces of infected nude mice but not in age matched naive nude mice. The protein specificity of these antibodies is being determined.

To determine if a primary infection in n u d e mice would induce protection (memory) against a secondary infection, we challenged six week old mice that had been infected as pups with the murine EW strain of rotavirus with the murine Cambridge strain of rotavirus. The previously infected mice, as well as naive nude mice, both shed very low but equal levels of rotavirus in the stools. Interestingly adult naive heterozygous littermates shed more virus than the nudes. The factors responsible for this relative resistance of adult nude mice to rotavirus infection as well as the mechanism for viral clearence in nude mice will be discussed. Symp. 1990, 121:149-160) . The great majority of the poxvirus encoded proteins actively mediating the evasion of host defense are secretory proteins termed "virokines". Viruses produce proteins which can either compete with or antagonizz molecules critically involved in generating immune responses. The vaccinia virus complement control protein (VCP) was the first virokine to be reported that has structural similarity to humadmouse complement control proteins, with the greatest similarity to the human complement 4b binding protein (bC4-BP) and the first protein to have a postulated role in the viral evasion of host defense (Kotwal and Moss, Nature 1988, 355:176-179) . Bioactive VCP, purified from serumfree medium has been shown to be more potent than the hC4bBP ( Kotwal, Am. Biotech. Lab., 1994) and has also been shown to bind to two important complement components, C3 and C4, and block the complement cascade at multiple sites in vitro (Kotwal et al., Science 1990, 2502327-830; Mckenzie, Kotwal et al., J. Inf. Dis. 1992 , 1661245-1250 . The importance of this protein was demonstrated in vivo using recombinant virus lacking an intact DNA coding for VCP (Isaacs, Kotwal and Moss, PNAS 1992, 89:628-672) and further underscored by the Smallpox virus genomic sequence analysis, indicating the presence of a highly conserved homolog of VCP ( Massung et al., Nature 1993,366:748-751) . In order to determine the precise effects of VCP at the site of infection, a mouse model has been developed. Measurement of the specific swelling response and microscopic examination of the histological changes in BALB/c and DBN2 mice has revealed that VCP is capable of regulating inflammatory response in vivo (Jayaraman and Kotwal, 1993 Mol. Immunol. 30:20) . In order to ensure that the possiblity of the unrelated mouse strains sharing identical H-2 haplotypes but differing at numerous other loci was not overlooked, we performed similar experiments in congenic strains. Well characterized C5deficient mice injected with poxviruses in comparison to normal mice, produce a significant inflammatory response, accompanied with severe ulceration that persists for several days. The data suggests that the host complement response is a critical factor in Egulating poxvirus pathogenesis. Previous studies have suggested that immune response was one of the major factors that contributed to the host resistantlsusceptible mechanism to Sendai virus infection (Brownstein, 1981) . Various immune regulatory molecules were investigated in inbred susceptible 129/J (129) mice and resistant C57BL/6J (B6) mice which share the same MHC haplotype. When they were given 200 EID, of Sendai virus intranasally, 129 mice showed histologically diffuse interstitial pneumonia compared to the local moderate inflammation in 86 mice. Virus was eliminated from the infected lung 3 days earlier in B6 mice than in 129 mice. The cytokine levels and cytokine-producing cells were investigated in the acute pneumonic lung. Also the recall cytokine production in the draining mediastinal lymph node (MLN) were studied. The maximal numbers of IL-2, IFN-y, IL-10, and IL-4 producing cells were comparable for each strains of mice, while more IL-6 producing cells were present in the lung of 129 mice. However, the numbers of these cytokine producing cells peaked in 129 mice 3 days later than in B6 mice. Analysis of soluble cytokines in bronchoalveolar lavage fluid showed that at least two-fold higher levels of IFN-y, IL-10, and IL-6 were present in 129 mice than in 86 mice. Recall cytokine production in the draining MLN showed only the presence of IL-2, IFN-y, IL-10, and IL-6, while no IL-4 was detected. Generally, higher levels of these cytokines were detected throughout the whole infection process in 129 mice. Therefore, the susceptibility of 129 mice to Sendai virus is not due to insufficient cytokine production in these two anatomical sites, yet heavier and prolonged virus burden in susceptible mice may drive the more rigorous and protracted induction of immune regulatory molecules. Cytokines play a decisive role in determining the type of immune response elicited by an antigen, both in driving Th cell differentiation along Thl or Th2 pathway and, ultimately, the effector and regulatory activity of these subsets. We have studied the role of cytokines in virus infections by constructing recombinant virus vectors encoding cytokine genes. During replication in mice, these vectors produce the encoded factor which is secreted from the infected cells with profound immunological consequences. Cytokines studied todate fall into two categories; either altering viral pathogenesis or selectively stimulating specific immune responses. The expression of IL-2, IFN-y, TNF-a or IL-7 in recombinant vaccinia virus (rVV) dramatically alters pathogenicity, such that immunodeficient mice resolve the normally lethal infection. On the other hand, expression of IL-4 by rVV suppresses the induction of cytotoxic T cells (CTL) inhibiting the clearance of virus which leads to an increase in viral pathogenicity. The expression of IL-4 by rVV was shown to inhibit the host IL-2 response required for the development of CTL's. NO production was also significantly suppressed by IL-4. rVVs expressing IL-5 or IL-6, although not affecting pathogenicity, preferentially stimulates IgA reactivity to coexpressed antigens. Encoding cytokines in recombinant viruses has clearly demonstrated the important role of cytokines as anti-viral effector molecules and in regulating appropriate immune responses. In mice infected with lymphocytic choriomeningitis virus (LCMV), interleukin-I2 (IL-12) doses of >I00 ngld induce profound immunotoxicities characterized as almost complete inhibition of virus-induced CD8+ T cell expansion and CTL activation, and up to 2 log increases in viral replication [Orange, Wolf, and Biron, J. Immunol. 152:1253, 19941 . Serum tumor necrosis factor (TNF) is also observed under these conditions. The studies reported here further characterize the expression and function of TNF in this context. Northern blot and in sifu hybridization analyses demonstrated that IL-12 induced TNF-cx expression and that LCMV infection synergized with IL-12 for this induction. Administration of antibodies neutralizing TNF reversed the IL-12-induced immunotoxicities in LCMV-infected mice and restored anti-viral defenses. The TNF-mediated immunotoxicities appeared to result from an induced cellular sensitivity to the factor, as splenic leukocytes and CD8+ T cells isolated from LCMV-infected mice were more sensitive to TNFmediated cytotoxicity in culture than were equivalent populations prepared from uninfected mice. Additional physiological changes were observed in IL-12-treated uninfected mice and were dramatically elevated in IL-12-treated virus-infected mice, including: 1) decreases in body weights; 2) elevation of circulating glucocorticoid levels: and 3) decreases in thymic mass. These changes were also reversed by anti-TNF.

The results delineate a unique TNF-mediated immunotoxicity and have significant implications concerning detrimental consequences of in vivo TNF andlor IL-12 for protective anti-viral responses. 

Lactate dehydrogenase-elevating virus (LDV), a naturally occurring virus, causes a persistent infection in mice and presents an ideal model for the study of immune modulation during acute and persistent virus infections. Within a few days following infection with LDV there is a pronounced polyclonal activation of B cells followed by the suppression of primary B cell responses to T-dependent Ag. We investigated the effect of acute and persistent LDV infection on the development of a memory B cell response to the model protein antigen, horse cytochrome c (cyt), by employing a modification of the splenic fragment assay. About a 50% decrease in the frequency of responding Agspecific memory B cells was observed in BALB/c mice infected with LDV, whether the mice were immunized with cyt at the time of LDV infection or three weeks later. This may be due in part to a defect in T cell help, since in cultures of normal memory B cells and T cells derived from LDV acutelyinfected mice the frequency of responding B cells was also decreased two-fold. In situ hybridization using a cDNA probe specific for LDV revealed two patterns of LDV RNA within the spleen. Twenty-four hr p.i. LDV RNA was located within the marginal zone, surrounding each follicle. This pattern is consistent with permissive macrophages. During persistence viral RNA could no longer be detected in the marginal zone, but was located within the follicles. The absence of LDV-permissive cells within the follicular region suggests that the source of LDV RNA is not due to ongoing viral replication. One possibility is that circulating virus is trapped by a specific cell population within the follicle. The effect of virus trapping within the spleen provides a mechanism by which LDV and other viruses can modulate immune cell function during persistent infections. IFN-y can be produced by activated NK cells. This cytokine enhances immune responses by augmenting macrophage antigen presentation. Viral infection induces IFN-dp and NK cell activation. Changes in splenic architecture, cell trafficking, and cytokine expression were examined during viral infections of C57BU6 mice. At times coinciding with IFN-dp production and NK cell activation, there was a redistribution of nucleated cells from red pulp to white pulp regions in spleens isolated from mice infected with either lymphocytic choriomeningitis virus (LCMV) or murine cytomegalovirus (MCMV). Cell transfer experiments with dioctadecyl-3,3,3',3'-tetramethyl indocarbocyanine perchlorate-or PKH26-GL-labeled bone marrow cells isolated from normal mice demonstrated an infection-induced accumulation of non-T/non-B cell populations along recipient splenic marginal zones.

Flow cytometric analyses demonstrated that approximately 10% of the transferred bone marrow cells accumulated in spleens after 20 hrs and 30% of these expressed the NK cell marker, NKl.l+. In vivo antibody treatment procedures, to eliminate cell subsets in donor mice, demonstrated that the cells localizing at the marginal zone were derived from AGMI+ and N K l . l + populations. A small subpopulation of marginal zone cells in infected mice were shown to be expressing high levels of IFN-7 mRNA by in sifu hybridization. Treatment with anti-AGMI or anti-NK1 .I antibodies eliminated both endogenous NK cells and the IFN-y mRNA positive cells. These data demonstrate that newly derived NK cells accumulate along marginal zones. The results also suggest that this trafficking pattern may act to enhance immune responses by facilitating delivery of cytokines to specialized antigen presenting cells.

David Segal, Janet Ruby, Alistair Ramsay and Ian Ramshaw.

Depamnent of Cell Biology, John Curtin School of Medical Research, PO Box 334, Canberra, ACT, 2601 Australia Cytokine expression has been shown to correlate with protective or ineffective immune responses in a number of disease models. Recently there has been the suggestion that immunity to some retroviruses is associated with the production of ceaain patterns of cytokines. To explore this further we have have used Rauscher murine leukemia virus (R-MuLV) infection of C57BU6 (resistant) and BALB/c (susceptible) mice to elucidate the role of cytokines immunity to retroviruses. Initially the in viho proliferation of spleen and lymph node cells from infected mice was examined. In response to stimulation with immohilised anti-CD3 antibodies the proliferation of spleen hut not lymph node cells from infected mice. was found to he rapidly suppressed. Suceptible BALB/c mice exhibited a much greater suppression than resistant C57BU6 mice. The cause of this suppression is under investigation however, the immunosuppressive molecules nitric oxide and prostaglandins are not involved. In vitro cytokine production by spleen and lymph node cells from R-MuLV infected mice was determined. In response to stimulation with immobilised anti-CD3 antibodies, spleen cells from infected BALB/c mice produced diminishing amounts of IFN-7 and IL-2. In contrast spleen cells from infected C57BU6 mice produced IFN-y and L-2 to levels that were only slightly less than uninfected controls. a-6 production by spleen cells from infected mice of both strains was at levels higher uninfected controls. Anti-CD3 stimulated lymph node cells from infected mice produced elevated IFN-1 suggesting that suppressed cytokine production is spleen specific. Expression of cytokine genes in vivo is currently being investigated using RT-PCR to detect cytokine mRNA in the spleens of infected mice. We have previously shown that primary resting murine B lymphocytes are non-permissive for vesicular stomatitis virus (VSV), however, a productive infection can be induced when infected B cells are activated with anti-immunoglobulin (a-lg) plus IL-4 or lipopolysaccharide (LPS). We posit VSV in unactivated primary B cells provides a paradigm of persistently infected lymphocytes and activation dependent recall of an active infection. Analysis of the behavior of virus in unstimulated B cells during long term culture and the requirements for subsequent induction of productive infection has been limited by the poor survival of primary cells in culture. We circumvented this limitation by using highly purified small B cells from mice transgenic for the bcl-2 proto-oncogene, expression markedly extends in vitro survival of unstimulated primary B cells. Overexpression of bcl-2 does not alter B cell infection or induction of a productive infection by activators during acute infection. Infection does not effect B cell survival in culture. Unstimulated virus infected B cells produce primary viral mRNAs but not viral proteins or infectious particles (PFU) during culture. Persistently infected B cells stimulated with a-lg plus IL-4 produced a fully productive VSV infection at all times analyzed, up to 3 weeks post infection. In contrast, VSV production in persistently infected B cells activated with LPS markedly declined relative to acutely infected activated cells (50-1 00 fold by week 1 and 1,000 fold by week 2). Cells were not completely refractory to LPS activation as VSV protein was produced. The selective LPS deficiency is unique to persistently infected cells as uninfected cultured B cells proliferate and differentiate to produce antibody upon LPS activation. These data show that a persistent infection may selectively alter the host cell response to previously productive activators which may as a consequence interfere with immune regulation.

RSV-G GLYCOPROTEIN SPECIFIC T CELLS PREFERENTIALLY SECRETE IL-5 AND PREDISPOSE TO PULMONARY EOSINOPHILLIA., Anon Srikiatkhachorn. and Thomas J. Braciale, The Beirne B. Carter Center for Immunology Research and the Departments of Microbiology, Pathology, and Pediatrics, University of Virginia Health Sciences Center, Charlottesville, VA 22908 We studied the immune responses to two different glycoproteins of respiratory syncytial virus (RSV) in a murine model. BALB/C mice were immunized with recombinant vaccinia virus expressing either RSV-fusion glycoprotein ( VAC-F), attachment glycoprotein (VAC-G) or 8-galactosidase (as a control). These mice were given RSV intranasally three weeks after priming and then sacrificed 5 or 14 days later. Spleens and bronchial lymph nodes were harvested for in vitro culture and lungs were harvested for histologic studies . We found that bulk cultures obtained from both VAC-F and VAC-G immunized animals secreted both Thl and Th2 type cytokines when stimulated with RSV infected spleen cells . However, the levels of 11-5 and 1FN-y were higher in bulk cultures derived from VAC-G primed animals while the levels of IL-2 were higher in the bulk culture from VAC-F primed animals. The IL-4 and IL-5 production was relatively short lived since spleen cells and bronchial lymph node cells obtaind form mice sacrificed 14 days after intranasal inoculation produced much lower levels of IL-4 and 11-5 while the levels of IL-2 and IFN-y production were comparable to bulk cultures obtained from mice at the peak of infection. There was little inflammatory response in the lungs obtained from mice immunized with the control vaccinia.

In contrast , lungs from mice immunized with VAC-F or VAC-G showed significant infiltration of inflammatory cells. There was a striking infiltration of eosinophils in the lungs from mice primed with VAC-G. These eosinophils could be detected aroud major bronchi and blood vessels, as well as, in some cases, in lung parenchyma. This study suggests that the immune responses to different viral glycoproteins may be distinct and may play important roles in viral pathogenesis. During infection of normal mice with lymphocytic choriomeningitis virus (LCMV), NK cell responses peak on day 3 and subside as CD8+ T cell responses are activated at day 7 post-infection. In contrast, 02M-/-mice, lacking CD8+ T cells, have dramatically elevated NK cell responses on day 7 postinfection. The 02M-/-response is evidenced by increased NK cell activity, as well as up to 5-fold increases in blast and total NKI.I+CD3-cell numbers. NK cell responses in normal mice are cyclosporin A (CsA)-resistant and interleukin (IL)-2independent, whereas day 7 NK cell responses in 02M-/-mice are CsA-sensitive and IL-2-dependent. To investigate the role of additional cytokines in regulating cellular responses during acute viral infections, production and function of IL-4 and transforming growth factor4 (TGF-0) were examined. Induction of IL-4 mRNA, at late times post-infection of normal mice, was shown by in situ hybridization of T cell-enriched splenic leukocytes and polymerase chain reaction (PCR) amplification of cDNA from RNA. ELlSAs of media cor.aitioned with cells isolated on days 0, 3, 5, 7, 9, and 14 post-infection demonstrated delayed induction of IL-4 protein as compared to CTL activation. TGF-0, evaluated in biological and ELISA assays, was induced maximally at days 7 to 9 post-infection. The kinetics of TGF-0 production by cells from infected 82M-/mice was similar to that of normal mice. However, cells from 02M-/-mice produced IL-4 at early but not at late times postinfection. Together, these results suggest that either IL-4 is a critical cytokine for shutting off NK cells during normal responses to viral infection, or that the 02M-l context modulates responsiveness of NK cell subsets to other late cytokines. Studies are in progress to distinguish between these two possible mechanisms. The induction of fever in response to infection is an important host defense mechanism that enhances aspects of the immune response and restricts the replication of some microorganism. Vaccinia virus, a member of the poxvirus family, is a complex cytoplasmic DNA virus that encodes a variety of proteins that interfere with host immune functions, such as complement regulatory factors and soluble receptors for IL-lp, TNF and IFNy.

Here we show that expression of the vaccinia virus IL-1p receptor (VIL-lpR) in the W R strain prevents the febrile response and reduces the severity of infection in intranasally inoculated mice. Fever was recorded on days 1-6 after infection of mice with a vIL-lpR deletion mutant, but not in animals infected with wild type WR or a virus revertant. These studies were extended to other virus strains that were used as smallpox vaccines, and expression of the vIL-lpR was consistently found to prevent the onset of fever. Vaccinia virus induced a severe hypothermia after 6 days in infected mice that was independent on VIL-lPR expression and correlated with virus replication in the brain, the organ that controls body temperature.

These results represent the first example of a virus mechanism to inhibit the host febrile response and suggest a central role for soluble IL-lp in the induction of fever in poxvirus infections. Measles virus (MV) infection can depress cell-mediated immune responses for months following clinical disease. MV is known to infect the thymus during human illness and this may contribute to immune suppression. We have used the SCID-hu m o w with co-implants of human fetal thymus and liver to determine the effect of virulent and avirulent strains of MV on the thymus. SCID-hu mice were. infected by direct inoculation of the graft with 103 PFU of either a wild type strain of MV(Chicago-1,Chi-1) or an attenu-ated strain (Moraten, Mor) and sacrificed at intervals over 28 days. Peak viral titers, as judged by plaque assay on Vero cells, were reached by Chi-1 on d4 (105.7 PFU/ third of implant), and Moron d21 (103.2 PFU/ third of implant). Hematoxylideosin stained sections of Chi-1-infected thymuses showed marked distortion of the cortex and medulla by d4 with thymocyte poilolosis and decreased cellularity. By d14, these. implants were mostly devoid of normal thymocytes. Mor-infected thymuses showed relatively preserved architecture and cellularity. Suspensions of the cells from implants stained with mAbs to CD3,CD4 and CD8 were analyzed by flow cytomehy. There were significant decreases in the CD4+CD8+ cell pop-ulation by d10 with complete loss of all such cells by d28 with Chi-1, and only modest reductions with Mor. Immune fluorescence staining of sections with a MV mAb to hemagluttinin(HA) and Abs for either human cytokera-tins(AEl/AE3) or CD15 co-localized MV predominantly to epithelial and monocytic cells. Additionally, MV antigen was present diffusely by d4 in both cortex and medulla in Chi-I infection whereas Mor-infected implants had only patchy distribution by d21. Only rare cells stained both with MV HA and CD2 or CD4. MV HA was not expressed over background on any CD4+ cells judged by FACS. We conclude that MV replicates in the SCID-hu thymic implant primarily in epithelial and monocytic cells, and that the attenuated virus reproduces more slowly and with less cellular disruption. Little MV HA could be demonstrated in thymocytes, therefore the data suggest that significant infection of the thymic epithelial stroma disrupts the thymic microenvironment which normally supports and aids in selection of immature T cells. Part of the long-term immune suppression seen in MV infection may be due to infection of the thymic epithelial stroma with subsequent loss of thymocytes. It is becoming increasingly evident that many poxviruses contain genes that enable the virus to evade the host's immune system. Myxoma virus is a leporipoxvirus and is the causative agent of myxomatosis, a rapidly lethal disease in the European rabbit (Oryctolagus cuniculus). One possible mechanism of immune evasion is virus-induced downregulation of cell-surface receptors important for an immune response. Cell-surface levels of several receptors on a rabbit T cell lymphoma cell line (RL-5) were monitored by flow cytometry. Following infection with Myxoma virus, cellsurface levels of CD4 were found to drop dramatically. Other cell surface antigens such as CD18, CD43, and CD45 were unaffected during infection with myxoma virus. Further more, the downregulation of CD4 by myxoma virus could be inhibited by treating cells for an extended period of time with PMA, suggesting that the downregulation was not simply a masking of the epitope via viral antigens. Analysis of CD4 levels in the presence of cytosine arabinoside indicates that late gene expression is not necessary for the modulation. Since the tyrosine specific protein kinase p56lck associates with the cytoplasmic domain of CD4 we have also examined the association of p56lck with CD4 as well as steady state levels of p56lCk during viral infection. The modulation of surface CD4 has also been described in HIV infected T cells suggesting that the loss of cell-surface CD4 may be a common viral immune evasion tactic by lymphotrophic viruses. I n addition, stably-transfected cell l i n e s expressing e i t h e r U S 1 1 o r US2-6 gene products s i g n i f i c a n t l y reduced l e v e l s of MHC class I heavy chain. Studies are i n progress t o f u r t h e r d e f i n e t h e mechanism by which t h e s e v i r a l gene products a l t e r immune recognition. Cytotoxic T lymphocytes (CTL) may play a significant role in containing the spread of HIV in infected individuals. Although HIV-infection is associated with immune suppression, a vigorous CTL response has been detected in infected adults. HIV can be transmitted from mother to child. One third of vertically infected children has a rapid evolution toward disease, with onset of AIDS before 18 months. The other two thirds remain asymptomatic for years. The bimodal course of disease evolution in HIV-infected children could be related to differences in the host immune control of viral replication.

HIV-specific CTL response from fresh and in vitro activated PBMC of HIV-infected children was measured. The vast majority of infected chidren had detectable HIV-specific CTL, which where CDS+CD8+. We previously showed that among children with a slow disease progression, fresh CTL were more frequent in the P2A(paucisymptomatic) group than in the Pl(asymptomatic) and the P2B-F groups (symptomatic group).

The cohort of children has now been followed during 4 years, and 46 children have been tested at least once. We found that CTL responses were less frequent in the children with a rapid disease progression than in the children with a slow disease progression at the same age. Our data suggest that CTL response is an important factor in delaying disease evolution. We, as well as others. have proposed that Sag function is critical to the ability of milk-borne M M N to infect mice. To determine whether this is the case, we created transgenic mice (HYB PRO/Cla) with a frameshift mutation int the sag gene. Young HYB PRO/Cla mice (c 10 weeks of age) showed no deletion of their cognate Vp14* T cells, unlike transgenic mice carrying a functional sag gene However, a slow, progressive loss was seen in the HYB PROlCla mice as they aged, indicating that it was due to expression of wild type Sag protein. Thus, as the HYB PRO/Cla mice aged, there was production of virus that appeared to lose the Cla mutation. The HYB PRO/Cla mice produced transgene RNA in their lactating mammary gland and shed virus in their milk. Their nontransgenic offspring of showed infection with transgene-encoded MMTV because they had the typical slow deletion of Vp14+ T cells characteristic of C3H MMTV infection and because we detected transgene-derived M M N RNA in their mammary glands. Cloning and sequencing of the viral RNA produced by the nontransgenic offspring of the HYB PRO/Cla mice showed that recombination between the Mtv-1 endogenous viral RNA and the transgene-encoded RNA occurred, such that the frameshift introduced by the Cia mutation was repaired.

These results show that there is selection of infectious virus that contains a functional sag gene. Thus, it appears that the only virus that is capable of being transmitted by the milk borne infection pathway is that which encodes a functional Sag protein. 

Hepatitis B virus (HBV) causes acute and chronic liver diseases and is closely associated with hepatocellular carcinoma. In order to understand the cellular immune response against HBV in chronic HBV infection, T cell proliferation, cytotoxicity and cytokine production were studied. We found that although the majority of asymptomatic HBsAg carriers and patients of chronic hepatitis B (CHB) had no proliferative response to HBsAg, some individuals in both groups showed significant T cell proliferation against HBsAg. In contrast, the proliferative T cell response to HBcAg in asyrnpatomatic HBsAg carriers was significantly stronger than that in patients of CHB with acute exacerbation. In addition, the frequency of HBcAg-reactive T cell precursors measured by limiting dilution assay was much higher in asymptomatic HBsAg carriers than in patients of CHB. Therefore, T cell responses against HBsAg and HBcAg are regulated differently in chronic HBV infection. Furthermore, we demonstrated HBsAg-and HBcAg-specific cytotoxic T lymphocyte (CTL) activity in asymptomatic HBsAg carriers, using autologous HBsAg-and HBcAg-expressing lymphoblastoid cell lines (LCL) as target cells, respectively. The cloned CTL were able to produce IFN-y, TNF-a or GM-CSF after stimulation. These findings demonstrate that T cell response to HBV is not completely suppressed in asymptomatic HBsAg carriers. Most of them have strong HBcAg-specific response and some of them have HBsAg-specific response.

Transcription and Tax The human T-lymphotropic virus type I (HTLV-I) promoter contains the structural features of a typical RNA polymerase I1 (pol 11) template. The promoter contains a TATA box 30 bp upstream of the transcription initiation site, binding sites for several pol I1 transcription factors, and long poly A+ RNA is synthesized from the integrated HTLV-I proviral DNA in vivo. Consistent with these characteristics, HTLV-I transcription activity was reconstituted in v i m using TBP, TFIIA, rTFIIB, rTFIIE, rTFIIF, TFIIH and pol 11. In HeLa whole cell extracts, however, the HTLV-I LTR also contains an overlapping transcription unit (OTU). HTLV-I OTU transcription is initiated at the same nucleotide site as the RNA isolated from the HTLV-I-infected cell line, MT-2, but was not inhibited by the presence of a-amanitin at concentrations which inhibited the adenovirus major late pol I1 promoter (6 pglml). HTLV-I transcription was inhibited when higher concentrations of a-amanitin were used (60 pglml), in the range of a typical polymerase In (pol 111) promoter (VA-I). Purified Tax, transactivates this promoter 5-to 10-fold in v i m . Interestingly, basal and Tax,-transactivated transcriptional activity of the HTLV-I LTR could be reconstituted with the 0.5 M phosphocellulose fraction. These observations suggest that the HTLV-I LTR contains overlapping Tax,responsive promoters, a typical pol I1 promoter and a unique pol I11 promoter which requires a distinct set of transcription factors. Tax, further in vifro transactivates a polymerase I1 template containing the 21 base pair repeats cloned upstream of the ovalbumin promoter and G-free cassette. Tax,-transactivated transcription was concentration dependent and inhibited by low concentrations of a-amanitin. Flaviviruses are arthropod-borne viruses whose route of infection is via the skin. They are mostly neurotropic and responsible for significant human morbidity and mortality. The classic cell-mediated immune response to a viral infection may be influenced by the ability of these viruses to modify expression of cell-surface molecules involved in the presentation of antigen to, and activation of, T cells. The skin Langerhans cell is the prototypic nonlymphoid dendritic cell and as such is uniquely placed to participate in a response against epidermally-acquired viral infections. The migratory properties of these cells contribute to their role as initiators of T cell-mediated immune responses within the draining lymph node. We have previously shown infection of epidermal cells in vifro by the flavivirus West Nile (WNV) results in an increase in MHC class I and I1

expression on the majority of epidermal cells and Langerhans cells respectively. In this study a technique for infecting the epidermis with WNV in vivo was developed. Tme-dependent increases in the surface expression of a number of antigens which are involved either directly or in a co-stimulatory capacity in initiating a cell-mediated immune response, were detected on both the majority of epidermal cells and the Langerhans cell population using flow cytometry. These increases were detectable as early as 16 hours after infection. A significant decrease in the percentage of Langerhans cells remaining in the epidermis was observed within 48 hours of infection. The phenotypic changes observed in vivo are analogous to those described following in vifro culture of Langerhans cells. These results, together with the reduction in Langerhans cell numbers, may represent the in situ maturation and concomitant migration of these. cells as a consequence of virus-induced cytokines within the skin microenvironment. which cause a wide variety of illnesses with high morbidity and mortality in humans throughout the world. Their high genomic stability argues for a survival strategy related more to interaction with the vertebrate host immune response, than a dependence on viral genetic mutation. Our previous work has shown that West Nile Virus (WNV) infection of many cell types directly induces functional increases in class I and 11 MHC expression. We report here that WNV infection of human embryonic fibroblasts (HEF) results in the increased expression of CD54 by two distinct mechanisms. An early, direct cytokine-independent mechanism operates within 2 h of virus infection, while an indirect mechanism, regulated by type 1 interferon (IFN), operates within 24 h of virus infection. CD54 expression increased by 4-5 fold within 2h of WNV infection on HEF, and by 6-7-fold within 24h. WNV-inactivated, conditioned supematants removed from infected HEF cultures after 4 h incubation did not alter CD54 expression on unqimulated HEF. whereas conditioned supernatants from 24 h-infected Cultures increased CD54 expression by about 1.5-2-fold after incubation for 24 h, but not after 4 h, similar to CD54 induction by 200Ulml of IFN-p. Increased CD54 expression on HEF by WNV was also cell-cycle dependent. CD54 increased only in quiescent, contact-inhibited infected HEF in Go phase. In contrast, induction of CD54 by types 1 and 2 IFN was not cell-cycle dependent. Other viruses, including double-stranded DNA viruses, vaccinia, and adenovirus 2 and 5, and the single, positive-stranded RNA alphavirus, Semiliki Forest virus, did not induce CD54 expression on HEF after 24 h. Another alphavirus, Ross river, was able to induce CD54 but only by the indirect mechanism of type 1 IFN-dependent release. Poly I.C, also, increased CD54 expression to the same extent as IFN-p after 24 h, making it unlikely that the early increase was due to a nonspecific viral effect. The closely related flavivirus, Kunjin, induced increased CD54 expression in a manner similar to WNV. The ability of flavivhses to induce increased CD54 expression directly within a few hours of infection may be an important virus-host survival strategy promoting cell-cell adhesion and hence possible further viral infectiodreplication. Recognition of viral peptides presented on the cell surface in association with class I MHC molecules leads to lysis by cytotoxic T cells (CTL) and forms an important part of the immune response to HIV infection. HIV virus has a high mutation rate and variation in the region of the viral epitope may allow evasion of this immune response. Variation could theoretically affect processing of the antigen, binding of the epitope to the HLA molecule or recognition of the presented epitope on the cell surface. We have studied proviral sequence variation in gag and CTL responses in a number of HLA B8 patients infected with HIV. Amino acid substitutions, such as a lysine to arginine change at position 3 of the PI7 gag nonamer CCKKKYKLK, lead to loss of recognition of the peptide by CTL from the patient whose provirus contained this sequence. These variant peptides bind to HLA 68 with comparable affinity to the index peptide suggesting that this loss of recognition is likely to be caused by changes in the interaction between the HLA-peptide complex and the T cell receptor. Other changes, such as lysine to arginine or glutamine at position 7, not only cause loss of recognition, but also lead to inhibition of lysis of targets bearing the index peptide. Thus it appears that in addition to loss of recognition by cytotoxic T cells, naturally occurring epitope variants may act as "antagonists", as has been demonstrated in MHC class II systems. Antagonism may be an important mechanism allowing immune escape by the HIV virus. genes. Subsequent complex formation between peptide, class I and p2microglobulin in the ER results in stable cell surface expression of the trimeric MHC-1 molecule. In previous studies we showed that in HPV-16 positive cervical carcinomas there was a loss of MHC-1 protein expression, which correlated at the single cell level with loss of TAP protein.

In this study we investigated whether loss of TAP and MHC-1 is mediated by an HPV-16 encoded protein. Human keratinocytes were transfected withvarious HPV-16 constructs including pAT16, the full length genome, pAT16ESX the full length genome with a premature stop codon in E5, pUC.ET16, the E6 and E7 oncogenes only, and pKVE5, expressing E5 from mouse Moloney LTR The different constructs were transfected into primary keratinocytes, cloned cells grown in medium supplemented with and without y-interferon ( y -a r ) for 48 hours. Cells were harvested and total RNA and protein harvested for Northern and Western blots respectively. Western blots showed very low steady state levels of TAP-1 and MHC-1 heavy chains in the cells with pAT16 as well as those containing ES alone, which was marginally increased by y-lFN. In contrast, primary keratinocytes, pAT16ESX and pUC.ET16 lines showed comparable TAP-1 and MHC-1 protein levels, which increased a & y-IFN treatment. Northem blots showed no differences in the amounts of TAP-1 and MHC-1 mRNA between the different cell lines. The data indicate that expression ofHF'V-16 E5 leads to post-transcriptional loss of MHC-1, presumably by interfering with TAP. To map and characterize functional differences between E1A of Ad5 and Adl2, we previously constructed a series of hybrid Ad5/12 E1A genes and used them with Ad12 E1B to transform primary Hooded Lister rat kidney cells. At least two regions within the first exon of Ad12 E1A were identified which influenced tumorigenicity. This study further examines the role of these regions in tumorigenicity by analyzing their affect on cell surface MHC class I expression and sensitivity to class I-restricted CD8+ as well as to non-class Irestricted NKS. The BCRFl open reading frame of Epstein-Barr virus exhibits remarkable sequence homology with the coding sequences of interleukin-10 from a variety of organisms. Many of the numerous immunological properties ascribed to interleukin-10 are shared by the product of BCRFl and this has led to it being termed viral interleukin-10. In order to investigate the activity of viral interleukin-I0 (vIL-10) and its interactions with the human interleukin-10 receptor we have expressed the protein in a bacterial and the eukaryotic COS-7 expression systems. The bacterially expressed vIL-I0 was partially purified and used to set up two assays to measure I L l O activity: i)the increase in IgM secretion from an EBV transformed B cell line -MT4.L and ii)the downregulation of class II HLA expression on the human monocytic cell line THP-1.

A series of deletion mutants (both N-and C-terminal as well as an internal deletion to remove a putative heparin binding domain) were constructed to identify possible domains within the vIL-10 protein that interact with the hIL-10 receptor and confer its biological activity. A number of these mutants have been expressed in the COS-7 expression system and their structure and biological activity are currently being assessed. The identification of the domains within vIL-10 that interact with the receptor or accessory proteins may aid in the understanding of the possible role of vIL-I0 within the EBV life cycle and in the pathogenesis of the numerous diseases associated with the virus. generation. To further test the role of CTL in Ad pathogenesis, viruses

lacking the Cll epitopes were tested When mutants that lack the immunodominate CTL epitope in EIA where used, a second immun-ssive epitope in ElB becorns the predominate target of ClU. These findings arc important since human Ad is currently being tested as a vector for gene therapy of cystic fibrosis. Our data suggest that when consuucting Ad vectors to be. used for gene therapy, one must retain either the 10.4K or 14.7K genes to decrease pathology and that Meting the genes that encode the antigens that a n recognized by ClU does not prevent the generation of Ad specific ClU. The interferons (IFNs) a n ? a family of cytokines whose functions include the protection of cells against viral infection. Type I IFNs include the 15 IFNa subtypes and IFNp that compete for binding to the same cell surface receptor, while type II IFN (IFNy) binds to a different receptor.

The orthopoxviruses, of which vaccinia virus (VV) is the prototypic member, have developed a number of anti-IFN strategies. The VV E3L protein competitively binds dsRNA and prevents the activation of IFNinduced and dsRNA-activated protein kinase (PKR), while the VV K3L protein shows sequence similarity to the eukaryotic initiation factor 2a (eIF2a) that is phosphorylated and inactivated by PKR. The K3L protein competitively binds the kinase and blocks host eIF2a phosphorylation and hence IFN-induced inhibition of host protein synthesis.

Onhopoxviruses also suppress cytokine action by expressing soluble cytokine receptors that bind and sequester the ligand; to date soluble receptors for interleukin-18, tumour necrosis factor and IFNy have been described. Supernatants from VV-infected cells were found to contain a soluble inhibitor of type I IFN that was conserved in most of the orthopoxviruses tested. The inhibitor was produced early in infection and did not inhibit IFNy. The IFNa/p inhibitor was mapped and the gene expressed from recombinant baculovirus. The inhibitor blocked the binding of 125I-IFNa to U937 cells and binding of 125I-IFNa to supernatants from baculovirus and VV-infected cells demonstrated that the inhibitor functioned as a soluble receptor for 1FNc1fp. Direct binding of 1251-IFNa to VV WR supernatants revealed that the soluble IFNa/p receptor had a high affinity for type I IFN. Deletion of the gene from the VV genome and ligand blotting of the soluble receptor demonstrated that IFN binding was encoded by a single protein. Competitive binding curves using IFNa from other species revealed that the poxvirus soluble IFNdp receptor bound human and bovine IFN with high affinity but murine IFN with relatively low affmity. Interestingly, the soluble IFNcrip receptor is highly conserved in variola virus. Given the importance of IFN in antiviral defense it is likely that the soluble IFNdp receptor plays an important role in the virulence of the orthopoxviruses. Endogenous processing of a viral glycoprotein for presentation t o CD4+ T cells has defined a previously under-investigated pathway in antigen processing and presentation. It may be important not only for pathogens, but also for self-proteins, and thus may be involved in self-tolerance. We have been characterizing the processing o f the ER-restricted Gpt glycoprotein of vesicular stomatitis virus (VSV) biochemically and enzymatically, by cellular localization using confocal immunofluorescence, cellular fractionation, and by T cell recognition assays. By flow cytometry, Gpt is undetected on the plasma membrane; in contrast, the wild type protein (G) is readily found following infection of A20 cells with a vaccinia virus vector, leading t o endogenous synthesis. The Gpt can be found exclusively in the ER compartment using co-localization with markers for ER (signal peptide binding protein, calnexin), and not in the Golgi compartment (a-mannosidase 11, wheat germ agglutinin), endosome, lysosome, or surface plasma membrane. This is consistent with the characteristics o f the localization of the proteases which appear to be responsible for its degradation. Work is in progress to localize the site of peptide binding to MHC heterodimers.

Supported by NIH grant A118083 t o CSR.

PRESENTATION OF AN OUT-OF-FRAME CLASS I RESTRICTED EPITOPE. T.N.J.Bullock and L.C.Eisenlohr, Department of Immunology, Thomas Jefferson University, Philadelphia, PA 19107. Antigen presentation by class I MHC molecules is thought to require the degradation of fully formed proteins in the cytosol. This degradative process supplies oligopeptide epitopes for transport into the endoplasmic reticulum (ER) where they can interact with and stabilize class I molecules. Stable class I molecules, associated with P2-microglobulin, can then proceed to the cell surface where they present the epitopes to T cell receptors.

The generally accepted model for protein translation, the scanning hypothesis proposed by KO&, is thought to describe the traditional method of translation for the majority of proteins. We wished to test the hypothesis that any internal methionine that is in good translation initiation context can be a source of short peptides, which may then be processed into class I epitopes.

nucleoprotein gene (NP), the target of the CTL response of several inbred mouse strains. NP contains three class I restricted epitopes at amino acids 50-57 (H2-Kk), 147-155 (H2-Kd) and 366-374 (H2-Db). The frameshift was introduced 26 amino acids upstream of the H2-Kd epitope. The mutated genes were then recombined with vaccinia virus and tested for presentation using CTL restricted to each of the epito s described above. We found that, whilst presentation of the H2-I@ epitope was unaffected by the frame shift, the epitope proximal to the frameshift (H2-Kd) was no longer presented to appro riately restricted CTL. However, presentation of the distal H2-Dg epitope was retained. Therefore we have shown, using a viral protein and a viral expression system, that out-of-frame epitopes can be processed and presented to CTL. Work is ongoing to c o n f m that internal methionines are capable of providing a platform for the initiation of translation for in-frame and out-of-frame epitopes.

We have created a frameshift mutation in the influenza PR8 The fine specificity of T cell recognition of peptide analogues of the influenza nucleoprotein epitope NP 383-391 SRYWAIRTR was studied using HLA B27-restricted influenza-specific cytotoxic T cell (CTL) clones, of defined T cell receptor (TcR) usage, derived from unrelated individuals following natural infection.

Synthetic analogue peptides were synthesized containing single amino acid substitutions, and tested both for binding to HLA B'2705 in vitro, and for presentation to CTL clones by HLA 827positive targets. Even conservative amino acid substitutions of the peptide residues P 4 , 7, and 8 profoundly influenced CTL recognition, without affecting binding to HLA 8'2705.

These amino acid side chains are thus probably directly contacted by the TcR. CTL clones which used the TcR V a l 4 gene segment (but not those using TcR Va12) were also sensitive to P1 substitutions, suggesting that the TcR alpha chain of these clones lies over the N terminus of bound peptide, and that the "footprint" of certain TcRs can span all exposed residues of a peptide bound to MHC class 1.

These results, taken together with previous structural and functional data, suggest that, for nonarner peptides bound to HLA 827, P i , P4 and P8 are "flag" residues with TcR accessible side chains. The E3/19K protein of human adenovirus type 2 (Ad2) is a resident transmembrane glycoprotein of the endoplasmic reticulum. Its capacity to associate with class I histocompatibility (MHC) antigens abrogates cell surface expression and the antigen presentation function of MHC antigens. At present, it is unclear exactly which structure of the E3/19K protein mediates binding to MHC molecules. Apart from a stretch of approximately 20 conserved amino acids in front of the transmembrane segment, E3/19K molecules from different adenovirus subgroups (B and C) share little homology. Remarkably, the majority of cysteines is conserved. In this report, we examined the importance of cysteine residues (Cys) for structure and function of the Ad2 E3/19K protein. We show that E3/19K contains intramolecular disulfide bonds. By using sitedirected rnutagenesis, individual cysteines were substituted by serines and alanines, and mutant proteins were stably expressed in 293 cells. Based on the differential binding of monoclonal antibody Tw1.3 and cyanogen bromide cleavage experiments, a structural model of E3/19K is proposed, in which Cys 11 and Cys 28 as well as Cys 22 and Cys 83 are linked by disulfide bonds. Both disulfide bonds (all four cysteines) are absolutely critical for the interaction with human MHC antigens. This was demonstrated by three criteria: loss of E3/19K coprecipitation, lack of transport inhibition and normal cell surface expression of MHC molecules in cells expressing mutant E3/19K molecules. Mutation of the three other cysteines at position 101, 109 and 122 had no effect. This indicates that a conformational determinant based on two disulfide bonds is crucial for the function of the E3/19K molecule, namely, to bind and to inhibit transport of MHC antigens. Previous studies have suggested that several abundant CMV proteins are major immunogenic targets in seropositive adults. We are interested in defining the major viral protein targets of a CD8' CTL response, in order to derive a vaccine strategy for individuals who are unable to mount immune responses which are lymphokinedependent because of immunosuppression. HLA-typed and CMV-pgsitive normal volunteers who have HLA-A alleles that represent -75% of the U.S. population are being tested to determine which of 5 abundant CMV proteins they recognize by a CD8' CTL response: p28, p65, p150, IE, and gB. T cell lines will be derived in order to unambiguously determine the HLA restriction of the CD8' CTL response to each of these proteins. Proteins which are recognized by the most HLA diverse population will be further characterized in terms of mapping of Class 1 epitopes through the use of T cell clones derived from the polyclonal cell lines by limiting dilution. The defined epitopes will form the basis of a vaccine strategy to augment the memory responses of seropositive volunteers against CMV. These epitopes will be used to boost the CTL precursor frequency of bone marrow transplant donors as a means to transfer cellular immunity to immunosuppressed hematologic transplant recipients. An alternative strategy is to immunize seropositive individuals with recombinant viral proteins as a means to boost immunologic memory. We are pursuing that strategy in a transgenic murine model of HLA-A2.1 developed by Dr. L. Sherman (Scripps Institute, La Jolla). We are vaccinating the transgenic mice with two well defined CMV proteins, p65 and gB together with either of two lipid-based adjuvants, commercially available D0TAPm (Bcehringer-Mannheim) or MF5gTH (Chiron, Emeryville, CA). Our preliminary studies with HSV-2 gB demonstrate that both adjuvants are effective at eliciting murine Class I restricted responses against the protein. Current studies are evaluating the recognition properties of the adjuvant-CMV protein complexes by HWA2 as a restriction element in the transgenic model. The CTL response to Sendai virus in C57BY6 mice is directed almost exclusively to a single H-2Kb-restricted epitope derived from the virus nucleoprotein, NPj24-332 (SEV-9). Analysis of 18 independent T cell hybridomas generated from C57BY6 mice following primary Sendai virus infection has shown that a very diverse repertoire of TCR is selected in response to this epitope. Crystallographic analysis of SEV-9 bound to Kb has shown that the side chaiis of peptide residues PhPI, Gl484,

A d s , and AlaPs protrude towards the solvent and are potentially available for recognition by the TCR Notably, residues GI484 and A d 5

protrude prominently from the peptide binding site due to their l o c a l i o n on a bulge in the center of SEV-9. To determine the importance of each of these residues for T cell recognition, we analyzed hybridoma responses to SEV-9 analogs substituted at each of these four positions. Preliminary data showed there generally appeared to be dominant recognition of Glyp4 and Asnm. However, individual hybridomas exhibited distinct patterns of fine specificity for residues PheP1 and AlaPs. Thus, individual hybridomas were dependent on one, both, or neither of these residues for recognition of SEV-9. These data are consistent with a critical role for the GI94 and A d 5 in governing TCR-SEV-9Eb recognition and suggest a structural basis for the diversity of the TCR repertoire selected by this @tope. Previous results from this laboratoty demonstrated that the dominant influenza A epitope recognized by HLA42.1 restricted CTL from HLA-A2.1 uansgenic mice was the M1 peptide epitope that is immunodominant in human CTL responses. However, analysis of a large number of CTL lines revealed a subset of influenza A/pR/8/34-specific murine CTL that recognized an HLA-A2.1 restricted epitope distinct from M1. Using recombinant vaccinia viruses encoding Werent influenza gene segments, the epitope recognized by these CTL was shown to be derived from the A/PR/8 NSl protein. Because these CTL did not recognize targets infected with the A/Alaska/6/77 saain of influenza, candidate peptide epitopes were synthesized based on sequences that included an HLA-A2.1 specific binding motif and that differed between A/PW and NAlaska All of these CTL recognized a nonamer and a decamer peptide which contained a common 8 amino acid sequence and two distinct sets of bmding mtif residues. However, the n0name.r peptide was able to sensitize CTL for half maximal lysis at 80-2500 fold lower doses than either the octamer or decamer. The homologous peptide derived from NAlaska NSl contained conservative amino acid changes at positions 4 and 8 and was not recognized at any tested concentration, although it bound with higher &ity to HLA-A2.1 than the peptide from A/PW8. The A/pR/8 NSl nonamer epitope was also recognized by human influenza A specific CTL derived from two individuals. These results substantiate the general utility of HLA class I aansgenic mice for the identification of human Cn epitopes for other pathogens. Furthemore, the recombinant DHFR was functional in the induction of gB epitope-specific CTL response upon immunization of C57BV6 mice. These results indicate that an viral epitope expressed in a cellular protein can be.

efficiently processed, presented and recognized by epitope-specific CTL, and suggest that the cellular proteins can be used to express CTL epitopes for induction of CD8+ immune responses. Virus-specific cytotoxic T lymphocytes (CTL.) were generated a day later at this site. To determine which APC was capable of stimulating virusspecific CTL precursors in the MLN, B, T and dendritic cells from the MLN of influenza-infkcted mice were separated and examined for the presence of virus. The predominant cell type which contained infectious virus was the dendritic cell. B and T cells from the MLN contained little, ifany, virus. The APC capacity ofthese populations was tested by their ability to stimulate vir~~-~pecific T cell hybridomas. Only dendritic cells from the MLN of influenza-infected mice were able to stimulate virusspecific T cell hybridomas, althwgh all APC populations from both naive and influenza-infected mice were effective stimulators after in y h pulsing with the appropriate intluenza peptide. Potential APC populations were also separated from the lung. V i s was detected in bronchioalveolar macrophages and dendritic cells but not B or T cells.

Both macrophages and dendritic cells isolated from intlum-infected lungs could stimulate virus-specific T cell hybridomas. The ability of the MLN and lung APC populations to stimulate naive CD8' T cells and generate virus-specific CTL is currently being examined. Virus infected cells present only a very limited number of peptides intracellularly processed from a viral protein to CTL even when many peptides hearing the MHC class I-restricted binding motif are present in the protein. Infection of H-2b mice w i t h lymphqtic choriomeningitis virus (LCMV) induces a CD8+ CTL response directed against three wellcharacterized epitopes presented by H-2Db molecules: "396-404 (FQPQ-NGQFI), GP33-43 (KAVYNFATCGI) and GP276-286 (SGVEN-PGGYCL). The H-2Db motif is characterized by a sequence of 9 to 11 a.a. with two anchor residues: Asn at position 5 and hydrophobic (Met, Ile, Leu) at the C-terminus. The LCMV NP and GP proteins contain thirly-one other peptides exhibiting the Db motif. However, no CTL response against one (or more) of these peptides has been characterized. Peptide binding to MHC is a critical step in antigen presentation. The aim of this study was therefore to analyze the binding properties of the potential Db LCMV peptides. The 34 LCMV peptides and 11 known Db-selective peptides were synthesized and their MHC binding affinities measured in two Db-specific binding assays.

Most of the LCMV peptides (28/34) did not bind to Db. The other 6 (including the 3 epitopes) and all the known Db peptides showed good affinity. Comparison of the sequences (good vs. non binders) allowed the identification of auxilliary anchors required for high binding affinity or of negative elements hampering MHC binding. In addition to the main anchors, the positive and negative factors at secondary residues play a crucial role in governing peptideMHC interactions. Knowledge of such factors might he of importance for the prediction of MHCrestricted CTL epitopes. Etienne Joly, Andrea Gonzalez, Carol Clarkson, Jonathan C. Howard and Geoffrey W. Butcher. Laboratory of Immunogenetics, Department of Immunology, The Babraham Institute, Cambs CB2 4AT, UK. Tap transporters from rats can be divided into two allelic groups, depending on their capacity to provide the RT1.Aa molecule with an appropriate level of suitable peptidesl. Recent results suggest that this might correlate with the RT1.Aa molecule requiring arginine-ended peptides (Powis et al., manuscript submitted), which the Tapb allele of the transporter is unable to translocate across the ER membrane efficiently2~3. RT1.A alleles are naturally linked with the Tapa or the Tapb allelic group4. We have set out to characterise various alleles for the RT1.A molecule, and find that, for the majority of TAPaassociated RT1.A molecules, 3 acidic residues line the C/E pocket, dictating Arginine as C-terminal anchor residue for the bound peptides. On the other hand, in Tapb-associated RT1.A molecules, one acidic residue at the most is found in the C/E pocket, which certainly results in a different anchor residue for the bound peptides. The selective pressure of viral infections must have driven this coevolution which affects dramatically the array of peptides presented to cytotoxic T lymphocytes. Cytotoxic T lymphocyte responses in HIV infection can be impaired due to variation in the epitope regions of viral proteins such as gag. We show here an analysis of variant epitope peptides in three gag epitopes presented by HLA B8. Seventeen variant peptides were examined for their binding to HLA B8; all but one bind at concentrations comparable to known epitopes. All except two could be seen by CTL clones grown from HLA B8 positive HIV-1 infected patients and were therefore immunogenic. However, in one haemophiliac patient studied in detail, there was a failure to respond to some of the peptides that represented virus present as provirus in his peripheral blood. In one case his CTL had previously responded to the peptide. Thus there was a selective failure of the CTLresponse to variant epitopes. This impaired reaction to new variants and failure to maintain responses to some epitopes late in HIV infection could contribute to the loss of immune control of the infection.

Pira, Anna Ferraris, Daniele Saverino, Peifang Sun and Annalisa Kunkl; Dept. Immunology, San Martino Hosp. Univ. of Genoa, 16132 Genoa, Italy.

Th epitopes present on viral proteins can be recognized by specific Th cells if appropriately expressed by antigen presenting cells (APC) as a result of uptake and processing. Since viral epitopes are not simply present in the context of viral proteins, but also in the context of whole viral particles, it is important to determine the role of the molecular and/or structural context on antigen uptake-processing-presentation.

Therefore we have generated panels of CD4+ human T cell lines and clones specific for different HIV antigens (gp120, p66, p24), in order to test their ability to respond to the Same epitopes present within synthetic peptides, recombinant proteins or inactivated virions (provided by G. Lewis, Dept. Microbiology, Univ. Maryland, Baltimore).

We could identify T cell lines and clones that were able to discriminate the molecular and structural context of the epitops. Certain T cells, in fact, responded to peptides and proteins, but not to viral particles, whereas other T cells were also able to proliferate when challanged in vitro with autologous APC and viral particles.

The data suggest that in the human Th cell repertoire specific for viral antigens T cells exist that can discriminate the molecularstructural context of Th epitopes. It will be interesting to ascertain whether T cells specific for epitopes that can only be recognized when provided in the context of a soluble molecule, but not of a viral particle, have any relevance in viva protection, or are a simple by-product of the cellular immune response.

Eric G. Pamer, Merceditas S. Villanueva, Section of Infectious Diseases, Yale University School of Medicine, New Haven, CT 06520 Listeria monocytogenes is a Gram positive bacterium that infects macrophages and secretes proteins into host cell cytosol.

The murein hydrolase p60 is secreted by L. monocytogenes and is required for complete bacterial septation. In the infected macrophage secreted p60 is processed by the host cell into the nonamer peptide p60 217-225 and is presented to cytotoxic T lymphocytes by the H-2Kd

MHC class I molecule. We have used strains of L.

monocytogenes that secrete different amounts of p60 to show that the rate of p60 217-225 production is proportional to the amount of antigen secreted into the host cell cytosol. p60 is degraded in the host cell cytosol with a half life of 90 minutes. The appearance of p60 217-225 is coupled to the degradation of newly synthesized p60.

We have determined the rate of intracellular p60 secretion and by accounting for the rate of p60 degradation we estimate that approximately 35 p60 molecules are degraded to produce one p60 217-225 epitope. This ratio is maintained over a range of intracellular antigen concentrations.

Our findings provide an estimate of the efficiency of antigen processing and demonstrate the remarkable capacity of the MHC class I antigen processing pathway to accommodate new epitopes. We have isolated and characterized three cytotoxic T lymphocyte (CTL) clones from the peripheral blood of two acute seroconversion patients and one patient in the first trimester of pregnancy. These clones were CD8+ and class I HLArestricted by the B7 molecule. All three clones recognized lllB and RF but not MN strains of HIV-1. Using vaccinia vectors expressing truncated versions of the HIV-1 envelope, the clones were found to recognize an epitope within amino acids 287-364, but not including 312-328 of gp120. Further mapping of the epitope with synthetic 20-mer peptides overlapping by 10, or 25-mers overlapping by 8, was unsuccessful. The sequence of the region of gp120 recognized by these clones was compared to the predicted HLA-87 peptide binding motif and a possible matching region was found. Using shorter peptides corresponding to this potential epitope recognition site, the minimum epitope recognized by the clones was determined to be the 10 aa sequence RPNNNTRKSI spanning amino acids We have further pursued a strategy to define a minimal cytotoxic epitope for a vaccine against CMV infection using T cell clones derived from individuals who have the MHC 835 gene (kind gifts of Drs. Riddell and Greenberg, Fred Hutchinson Cancer Research Center and Dr. Robert Siliciano, Johns Hopkins University Medical Center). We tested by chromium release assay (CRA) the recognition of a series of 835 allelic variants of EBV-LCL. by 835 restricted and CMV or HIV-specific T cell clones. Several conclusions quickly became apparent. The previously described 8'3501 peptide epitope from pp65 was not able to prime the autologous 835 EBV-LCL for killing by the pp65-specific CTL, whereas a recombinant vaccinia virus expressing whole pp65 could cause the same cell line to be recognized and killed in the same experiment. In addition, an HIV gp41-specific CD8' CTL which has a defined minimal cytotoxic epitope will only recognize and kill a subset of 835 EBV-LCL. The two T cell clones will not recognize each other's autologous EBV-LCL. The resolution of this interesting phenomena comes from sequence analysis of the HLA Class I B genes from both EBV-LCL. EBV-LCL which contain the B'3502 allele are recognized and killed by the pp65-specific T cell clone, and cell lines carrying 8'3501 alleles are recognized by the HIV gp41-T cell clone. We conclude that the reported CMV pp65 B"3501 restricted epitope is not correct, since the CTL in question will only recognize 6'3502 alleles in combination with the correct pp65 epitope. fragments with or without a signal sequence sensitize RMA-S/Kd to a similar limited extent. This data i s consistent with an inefficient movement of peptides from the cytoplasm into the ER by a Tap independent mechanism and does not reveal a processing competent compartment within the secretory pathway. Peptide transport by the transporter associated with antigen processing (TAP) was studied using a microsome system as previously reported by Heemels et. al.. In this system, a radiolabeled synthetic peptide which can be N-link glycosylated is used as the indicator peptide for the transport studies. The transport efficiency of synthetic peptides corresponding to antigenic peptides restricted to the murine Kd molecule was measured by inhibition of labeled peptide transported into the microsomes. The transport efficiency of three Kd epitopes in the type A influenza virus "147-155, HA204-212 and HA210-219 was found to be similar. An 11 amino acid peptide corresponding to HA204-214 which contains the 204-212 epitope was transported at a similar efficiency as the 9 amino acid minimum epitope. However, when the peptide sequence is further extended by one amino acid to residue 215, this peptide is poorly transported. These results suggest that the flanking region of an epitope can dramatically influence the transport of the epitope. When the transport kinetics of TAP was studied using the microsome system, the Vmax for transporting the indicator peptide (a variant of NP epitope that has the sequence TYNRTRALI) was found at 260.8 fmolelminute (+/-30.5). The Km for this peptide was found to be 231.9nM(+/-31.8).

BYPASSING A BLOCK IN ANTIGEN PROCESSING FOR CLASS I-RESTRICTED CYTOTOXIC T CELL RECOGNITION. Amy J. Yellen-Shaw and Laurence C. Eisenlohr. Thomas Jeferson Universitv. hiladelphia, PA., 19107. Previous work from our laboratory showed that processing of an influenza nucleoprotein (NP) epitope (amino acids 147-155) expressed endogenously from a recombinant vaccinia virus "minigene" is severely impaired when a flanking sequence (the dipeptide threonine-glycine) is appended to the Cterminus of the construct (147-158/R-). The inhibition of processing is overcome by placing the unprocessed peptide in the context of the fulllength NP molecule, demonstrating that regions of a protein outside the epitope itself critically affect the ability of the proteolytic machinery to fragment the protein appropriately. To determine the requirements for bypassing the block in antigen processing, we have constructed an array of "minigene"-expressing vaccinia recombinants in which the unprocessed epitope is extended by varying lengths toward either the C-terminus or the N-terminus of the NP molecule. Our results show that while an extension of the C-terminus by only one amino acid restores processability, a much longer extension of the N-terminus (75 < n < 100 amino acids) will also allow the substrate to be processed. It is therefore clear that a full-length, properly folded molecule is not required for liberation of the blocked epitope, and that probably more than one mechanism can contribute to enhancement of substrate proteolysis. We hypothesize that the C-terminal extension allows recruitment of an endopeptidase versus exopeptidase ("trimming") activity which is capable of cleaving the difficult bond. We considered the possibility that the N-terminal extension rescues processing by recruitment of the ubiquitin-dependent degradation system. To address this possibility we replaced all available ubiquitination sites (lysine residues) in one of the rescued constructs (50-158/R-) to see if the construct would still be processed and presented. The six available lysine residues were changed to arginine using PCR-based mutagenesis. The resulting construct (termed 6R) was recombined into vaccinia virus and tested for presentation to NP-specific CTL. The 6R construct was presented at a level equivalent to that seen with the wild-type 50-158/Rconstruct. This result provides clear evidence that entry into the ubiquitindependent degradation pathway is not responsible for rescue of presentation in this system and more importantly, that ubiquitination is not required for processing of all large substrates.

Chia-Chi Ku, Li-Jung Chien ,and Chwan-Chuen King, Institute of Epidemiology, National Taiwan University, Taipei, Taiwan, R.O.C. Dengue virus (DEN) can cause dengue fever (DF) and dengue hemorrhagic fever (DHF) I dengue shock syndrome (DSS) and DEN-2 was the most common serotype found in DHF outbreaks globally. Current hypotheses suggested that DHF may be associated either with antibodydependent enhancement (ADE) or with viral virulence. DEN can replicate predominantly in monocytedmacrophages (MIM), but whether peripheral blood lymhocytes (PBLs) are the target cells of DEN still remain controversial. In order to compare whether various clinically derived DEN-2 will interact with MIM and lymphocytes in different manners, we used two isolates --PLO46 strain (obtained from a DF patient during Taiwan 1981 outbreaks) and 16681 strain (isolated from a DHF patient in Thailand by CDC, USA) to infect primary MIM and lymphocytes as well as several types of cell lines. Primary lymphocyte culture was nonadherent cells obtained after 24 hr adherence of PBMCs, whereas the primary MIM culture was collected by depletion of lymphocytes using anti-CD3ICDI9 mAb and complement prior to adherence procedure and the purity of MIM culture was checked by CD14 surface marker staining. Supernatants (SN) of virus were harvested at various time points post infection after with several or without treatments. Our prelimanary data showed that DHF-associated DEN-2 strain had higher viral yield in certain age of MIM and a promonocytic cell line (HL-CZ) than Taiwan DF-associated DEN2 strain. In addition, this DHF-DEN2 strain was more likely to infect the promonocytic (HL-CZ) than well differentiated monocytic (CTV-1) and lymphocytic (H9) cell lines and also had higher peak yields than DEN-I virus in HL-CZ cells. Interestingly, DHF-DEN2 strain replicated much more efficiently in primary lymphocytes no matter these cells were activated with PHA or not, whereas Taiwan DF-DEN2 strain virus was hardly detectable in SN of both activated and non-activated lymphocyte cultures. Therefore we conclude that (1) different strains of dengue virus could orchestrate quite differently with immune cells, (2) different stage of MIM differentiation might be an important permissive determinants for dengue virus infection and replication, and (3) DEN virus strain virulence -a more important factor than lymphocyte activation status -seemed to determine whether this strain would infect human PBLs. Further studies should be focused on searching for detaied mechanisms of virus and immune cell interactions. (2) when viral yields were enhanced early than day5 post infection, it provided tremendous opportunity to attack the immune system and finally may lead to severe disease.

HIV-1 USING RECOMBINANT IMMUNOGLOBULIN MOLECULES, Marie-Claire Gauduin, Graham P. Allaway, Paul J. Maddon, Carlos F. Barbas, Dennis R. Burton, and Richard A.

University School of Medicine, New York. NY 10016. Primary isolates of HIV-1 have been shown to be less sensitive to neutralization by immune sera, monoclonal antibodies and CD4-based molecules than T cell line-adapted strains of HIV-I. We studied two immunoglobulin molecules for ability to neutralize primary isolates of HIV-I. lgG12 is an immunoglobulin molecule created from a combinatorial phage expression library and reacts with the CD4 binding site (CD4-BS) on gp120. CD4-lgG2 is a recombinant molecule in which the variable domains of both heavy and light chains of lgG2 were replaced with the first and second immunoglobulin-like domains of human CD4. Both molecules have been previously shown to effectively neutralize HIV-I in vitro. Ex vivo neutralizations were performed as follows: lgG12 and CD4-lgG2 were added at 25 pg/ml to wells containing serial dilutions of plasma from HIV-I-infected patients and PHAstimulated peripheral blood mononuclear cells from seronegative donors. P24 production was measured over 14 days of culture and an end-point titer of HIV-1 in the presence and absence of added antibody was determined. Both IgG12 and CD4-lgG2 were found to reduce the original HIV titer from seven plasma samples with high virus titer (>250 TCID50/ml) by up to 625-fold. This is in comparison to soluble CD4 which only reduced viral infectivity by 55-fold at the same concentration. In vitro binding and neutralization assays on isolates recovered from plasma confirm the potency and breadth of neutralization by these two molecules. These studies suggest that recombinant antibodies directed at the CD4-BS of HIV-1 gp120 are able to effectively neutralize primary isolates of HIV-1 and may be useful in dissecting the mechanisms of resistance to neutralization by other antibodies.

Dillner and P. Heino, Microbiology & Tumor Biology Center, Karolinska Institute, Stockholm, Sweden HPV 16 the major cause of anogenital precancers in man. The search for neutralizing epitopes that could form the basis for a preventive vaccine has shown that the surface-exposed imunodominant epitopes of the capsid are strongly conformationdependent, which has precluded detailed epitope analysis. Similarly, immunization with whole, denatured capsid proteins has only identified linear immunodominant epitopes positioned on the inside of the capsid. Reasoning that linear surface-exposed epitopes should exist, but might be cryptic, a set of 66 overlapping synthetic peptides corresponding to the entire HPV16 capsid proteins was used to generate hyperimmune sera. Several antisera against 3 different peptides were reactive with intact HPV16 capsids at titers up to 1:150.000.

HIV-1 SERUM ANTIBODIES AND MUCOSAL IGA. Basil Golding, John Inman, Paul Beining, Jody Manischewitz, Robert Blackburn and Hana Golding. Div. of Hematology and Viral Products, CBER, FDA, and Lab. of Immunology, NIAID, Bethesda MD 20892. Previously, we showed that HIV-1 proteins conjugated to 8. abortus (BA) could generate anti-HIV-1 neutralizing antibodies in mice even after depletion of CD4* T cells. In this study a 14-mer peptide from the V3 loop of HIV-1 (MN) was synthesized 013) and coupled to BA and KLH. BALB/c mice were immunized twice i.p. with these conjugates at two week intervals. V3-KLH induced mainly IgG1, whereas V3-BA induced all IgG isotypes but lgG2a predominated. Fecal extracts from mice immunized with V3-BA were shown by ELSA to contain IgA antibodies. Sera from these mice bound gp120, expressed on the surface of infected cells. Sera from mice immunized with V3-BA inhibited syncytia formed between CD4' T cells and chronically infected [HIV-I (MN)] H9 cells. Inhibition of syncytia, formed by other HIV-1 lab. strains correlated with the degree of their homology with the V3 region of HIV-I (MN). To mimic the efffect of HIV-1, mice were depleted of CD4' cells using anti-L3T4 at the time of primary or secondary immunization. Following primary immunization, CD4+ T cell depletion abrogated V3-KLH antibody responses, whereas responses to V3-BA were retained and sera from these mice were able to inhibit gp-120mediated syncytia. In secondary responses, CD4' T cell-depletion prevented boosting to V3-KLH, but V3-BA increased anti43 and syncytia-inhibiting antibodies. These results suggest that: 1. 8. abortus, can provide carrier function for a peptide and induce both serum and mucosal antibody responses, and 2. that infection with HIV-1 with subsequent impairment of CD4' T cell function would not abrogate anti-HIV-1 antibody responses if 8. abortus is used as a carrier to stimulate memory responses. Nucleotide sequence analysis of the VH genes revealed the usage of one particular VH germline element (VH61-1P) in all clones. This finding allowed the determination of somatically mutated positions in the VH regions. Two VSV-IND neutralizing antibodies expressed VH and VL genes in complete germline configuration whereas the rest of the clones showed somatic mutations which obviously were antigen dependently selected for. However, binding affinities of mutated and unmutated antibodies were comparably high. In order to determine the influence of somatic point mutations on one single antibody we generated a monovalent single chain antibody (FV-CK) of a mutated clone and reversed it stepwise to germline configuration by means of site directed mutagenesis. Surprisingly, already the germline configuration of FV-CK could neutralize VSV-IND, even though the binding affinty was lower than that of the mutated FV-CK. Every single somatic point mutation tested improved the binding avidity although some mutations reduced affinity. Thus, during the course of VSV-IND infection some antibodies are subjected to avidity maturation although this is not required for the generation of high affme, efficiently virus neutralizing antibodies,

Lisa Hyland'", Sam Hou'.~, and Peter C. Doherty'. 'Department of Immunology, St. Jude Children's Research Hospital, Memphis, TN 38 10 I, 2Departments of Immunology and Microbiology, and 'Pathology, University of Otago,Dunedin, New Zealand. The B and T cell responses in C57BL/6J(B6) mice treated with the mAb Mel-14 to L-selectin have been analysed following i.n. infection with Sendai virus. Mel-14 treatment caused a 70-90% decrease in the lymphocyte recruitment to the mediastinal (h4LN) and cervical (CLN) lymph nodes following infection with Sendai virus. The cellularity of the spleen was unchanged. The clonal expansion of CD8+ CTL precursors in the MLN was slightly delayed, but potent CTL effectors were present in the virusinfected lung by day 10 after infection and the overall magnitude of the response was not compromised. The prevalence of IgA antibody forming cells (AFCs) was greatly increased in both the MLN and the CLN of the mice given the Mel-14 antibody. The IgM response was prolonged and the IgG response, particularly IgGl, was delayed compared to controls. The altered pattern of the antibody response may reflect the limited availability in Mel-14-treated mice of Th cells secreting lymphokines which are involved in Ig class switching, by blocking the entry of CD4+ Th precursor cells into lymph nodes. FACS sorting for L-selectin+, 8220+, and L-selectin-, B220+ cell populations from the MLN and the CLN of normal B6 mice 9 days post Sendai virus infection, showed that the AFCs were from the L-selectin-, B220+ cell population, a population which comprised 6-10% ofthe total cell population.

We have distinguished targets of broadly neutralizing antibodies present in HIV-1 infected individuals by imunoselection in vitro and by the use of chimeric virus. One target of neutralizing antibodies, defined by an escape mutant with an ala to thr substitution at position 582 in gp41, is resistant to human monoclonal antibodies that map to a site closely congruent with that for CD4 binding. Substitution of gly, ser, and val fail to confer resistance. A second, defined by an ala to val substitution at position 281, upstream from the V3 loop, does not involve the same site and does not involve V3. Substitution of thr or ile also confers resistance. Replacement of the V3 loop of HIV-l(MN) into a clone of HIV-l(IIIB) allows the detection of two other broadly neutralizing targets. One recognizes the V3 peptide of MN but is affected by regions outside V3. The other appears to be conformational and outside V3, but its functional recognition is influenced by the V3 loop. All of these sites seem to depend on the overall conformation of the envelope protein rather than a single discrete linear epitope. Antibodies against amino acids 579-613 of the HIV transmembrane (TM) glycoprotein have been shown to enhance HIV infection in vitro in the presence of complement. There has been no study demonstrating that enhancing antibodies to this region of HIV, despite increasing levels of infectious virus 10 to 100 fold in vitro, adversely affect disease pathogenesis. In two separate studies reported herein, it is shown that animals which have high levels of antibody against this region of SIV, amino acids 603-622 of the envelope, fair poorly compared to animals with lower antibody levels against this region when subsequently challenged with SIV. When actively immunized with a synthetic peptide from this region of SIV, animals died earlier and failed to clear antigen at two weeks after infection compared to animals that received a control peptide (p<0.05). When animals were passively immunized with antibodies from a longterm survivor of SIV infection, those animals that received higher levels of antibody against the TM peptide died within six months compared to longer intervals for those animals that had lower levels of antibody to this region. When taken together, these data suggest that antibody to the TM region of SIV and HIV in general, and to this highly conserved peptide in particular, are detrimental to the host. Therefore, immunization strategies that minimize the immune response against TM or treatment protocols that decrease antibody levels against TM may lead to prolonged survival following exposure to lentiviruses. We have developed a mouse model to examine the immune response to HPV 16 proteins when these proteins are presented to the immune system via the epithelial route. In this model animals are grafted with keratinocytes expressing HPV E6 a n d E7 genes using a transplantation procedure which permits epithelial reformation. Animals so grafted when challenged intradermally with E7 either as protein or via a recombinant vaccinia virus exhibit a delayed type hypersensitivity response which is E7-specific and CD4+ T cell mediated. Animals grafted with a sub optimal priming inoculum of cells develop immune non-responsiveness and have an abrogated DTH response when challenged subsequently with a priming cell graft. In the present study w e have examined the antibody status in these animals. The E7 protein of HPV 16 was expressed in E. coli as a maltose binding fusion protein using the plasmid vector pMalc. After cleavage and affinity purification this protein was used in a n ELISA assay to measure antibody levels in 4 groups of mice (1) those not challenged with E7 (2) mice not grafted but challenged with E7 protein in the ear (3) mice primed by grafting with 107 HPV E7 expressing cells and challenged with E7 protein (4) mice primed by grafting with 5 x 105 HPV 16 E7 cells on day 7, grafted again with lo7 HPV 16 E7 cells on day 14 and challenged with E7 protein in the ear. Mice optimally grafted and challenged (group 3) exhibited high titres of IgG antibodies, particularly elevated levels of IgGza. Mice sub-optimally grafted (Group 4) exhibited IgG antibody levels comparable to the control group (1). The possible mechanisms of this immune attenuation are discussed. The Hepatitis C virus is a frequent cause of chronic liver disease. A proposed mechanism responsible for virus persistence is evasion of the host immune response through a high mutation rate of crucial regions of the viral genome. The portion of HCV genome coding for the amino-terminal part of the putative envelope protein (gp70) undergoes frequent mutation during the course of infection. We have cloned and sequenced the hypervariable region (HVRI) of the virus isolated from an HCV asymptomatic patient at three time points during 18 months follow up. Sequence analysis has allowed the identification of variants of this region and multiple antigenic peptides (MAP), corresponding to three HVRl variants, sequentially foundin the blood stream of the patient, have been synthesized. MAPS have been used as antigens for detection of specific antibodies in ELISA. Our results show that anti-HVRI antibodies and their cognate viral sequence coexist in the blood stream but a viral sequence becomes undetectable when the specific antibodies reach maximum levels of reactivity. Thus humoral immunity against the HVRl may play a role for virus clearence. The presence of anti-HVR1 antibodies was also investigated in 100 hepatitis C viremic individuals and 25 non-viremic patients. A high frequency of positive reaction (90%) against at least one of the three HVRl variants analysed in this study was detected in the viremic patients. Finally, competition experiments show that antibodies crossreacting with more than one HVRl variant are produced by HCV infected individuals. This results suggest that complex cross-reactivity exist between HCV isolates for antibodies against the HVRl region as described for antibodies against the gp120 V3 loop of HIV. We propose as mechanism for viral escape in HCV chronic infections the one described as the "original antigenic sin", observed firstly in influenza, in togavirus, paramixovirus, enterovirus, and recently in HIV infection. Using an adult mouse model to study active immunity against rotavirus infection, it was previously shown that oral immunization with some, but not all, animal rotavirus strains induced protection against subsequent infection following oral challenge witb the murine rotavirus strain EDIM (Ward et al., 1992) . To determine i f a specific rotavirus protein could be associated with protection in this model, mice were immunized with a series of 18 reassortants between the fully protective EDIM strain and a partially protective heterologous rotavirus strain (RRV-G). Reassortants that contained genes for EDIM proteins responsible for protection were anticipated to provide complete protection; however, no EDIM proteins were found to be both necessary and sd3cient for full protection. Instead, protection was found to be highly correlated with viral shedding (P = ,005) and with serum rotavirus IgA titers stimulated by the different reassortants (P < ,001). This indicated that protection was related to the intestinal replication properties of the different reassortants rather than to specific immunogenic properties of EDIM proteins. This conclusion was supported by the finding that the titers of serum rotavirus I& but not IgG, stimulated in mice following oral immunization with a series of animal rotaviruses was directly related to protection against EDIM. If these findings can be extended to humans, they suggest that the efficiency of intestinal replication following oral inoculation with a live rotavirus vaccine candidate may be the primary determinant of successful immunization. H a l l Medical Center, 55 individuals with adequate serum samples were identified as either rapidly progressing (RP) or slowly progressing (SP) by clinical and surrogate marker criteria. Anti-V3 profiles were determined using synthetic proteins derived from the amino acid sequences of the V3 region of 5 laboratory strains of HIV-1 in standard capture ELISA format. Serum obtained from each patient at multiple different time points was screened against these peptides. The majority of individuals in both groups demonstrated broad recognition, with reactivity to peptides corresponding to the V3 regions of MN, SF2, NY5 and HAN/SC. Less than 50% of individual in each group recognized the V3 peptide derived from IIIB, @=NS, between groups). As the RP progressed to AIDS there was significant nonspecific narrowing of response, while the SP remained broadly reactivity (P< .001). In v i m neutralizing activity of the homologous laboratory isolates was determined with cytotoxicity, cytopathic effect and p24 Ag inhibition assays. Although most patient serum was capable of inhibiting p24 Ag production in homologous lab strains while AIDS-free, there was no relationship with the ability to inhibit homologous virus effects on target cells and anti-V3 profiles. model, we show that after resolution of the acute infection, when antiviral plasma cells in the spleen decline, a population of virus-specific plasma cells appear in the bone m w and constitute the major sou~ce of longterm antibody production. Following infection of adult mice, Wspecific antibody secreting cells (ASC) peaked in the spleen at 8 days postinfection, but were at this time undetectable in the bone marmw. The infection was essentially cleared by 15 days and the ASC numbers in the spleen rapidly declined while an increasing population of LCMV-specific ASC appeared in the bone marrow. When compared to the peak response at 8 days post-infection, timepoints from 30 days to more than one year later demonstrated greater than a l@fold reduction in splenic ASC. In contrast, JLTvlV-specific plasma cells in the bone marrow remained at high numbers and correlated with the high levels of antiviral serum antibody.

The prewnce of antiviral plasma cells in the bone marrow was not due to a persistent infection at this site, since virus was cleared from both the spleen and bone marrow with similar kinetics as determined by infectivity and PCR assays. The IgG subclass profile of antibody m e t i n g cells derived from bone manuw and spleen correlated with the IgG subclass distribution of LCMV-specific antibody in the serum. Upon rechallenge with W , the spleen exhibited a substantial increase in virus-specific plasma cell numbers during the early phase of the secondary response, followed by an equally sharp decline. Bone marrow ASC populations and LCMV-specific antibody levels in the serum did not change during the early phase of the reinfection but both increased about 2-fold by 15 days post-challenge.

After both primary and secondary viral infection, LCMV-specific plasma cells were maintained in the bone marrow showing that the bone marrow is a major site of long-term antibody production after acute viral infection. "memory" T cells, and associated with responsiveness to soluble and recall antigens. CD4+ lymphocytes staining bright, dim, or negative (equivalent to an isotype control) for CD29 were evaluated in 49 uninfected controls (group l), 84 HIV-1 positive patients with 220% CD4+ T cells (group 2), and 47 HIV-I-infected patients with ~2 0 % CD4+ T cells (group 3). Most of these subjects also had 3-color staining for CD4\CD45RO\CD45RA.

The appearance of positive CD29 and CD45RO on HIVinfected and uninfected cells correlated well (R=.82 P<.OOl). The percentage of cells staining CD4+\CD29+.(bright plus dim) was 43.3 (95%Cl 37.3-49.4) in group 1, 28.9 (27.5-30.4 ) in group 2, and 10.2(8.6-11.9) in group 3. The respective values for these groups that were CD4+\CD2gbWM was 30.6 (26.9-34.3), 20.7 (1 9.3-22.2), and 7.4(6.3-8.6). Values for CD4+\CD45RO+ were 33.7 (31.8-35.5), 21.8 (20.5-23.1), and 9.9 (8.5-11.3), respectively. In single factor discriminate function tests, the %CD4+\CD29+ cells best predicted subject group (87% correct), proving to be a better discriminator than %CD4+\CD29b'h' (77% ~orrect),CD4+\CD29~"" (5 l%), CD4+\CD45RO+ (75%) and CD4+\CD45RO+\CD45RA-(63%).

Overall, no advantage was seen to splitting the CD4+\CD29+ cells into bright and dim positive subsets in the subjects studied for the purpose of stratifying early vs. late HIV infection. Likewise, splitting the CD4+\CD45RO+ compartment into CD45RA+ subsets did not improve the ability to distinguish between uninfected and early or late HIV-1 infected patients. The relationship between the virus-specific cytotoxic response in HIV infected patients and disease progression support the concept that a vaccine candidate should also induce a virus-specific CTL activity. Immunization of uninfected adult volunteers by a HIV-gpl60 recombinant canarypox virus was carried out in a phase I trial.Two injections of a recombinant canarypox expressing the HIV-l/MN gp160 were performed at month 0 and 1 and two boosts of recombinant gpl60MN/LAI at month 3 and 6 in alum or incomplete freund adjuvant(1FA). HIV-envelope specific cytotoxic activities were detected from CTL lines derived from PBMC stimulated by specific stimulation with autologous HIV infected blasts. CTL lines were obtained from 18 out of 20 donors : seven out of eighteen (39%) were found to present envelope specific cytotoxic activity at months 2, 4, 7 or 12 post immunization ; this activity was characterized as a CD3+,CD8+, MHC class-I restricted cytotoxic activity, and for at least two volunteers, this activity was still present two years after the first canari-pox/env injection.

Because avian poxviruses are incapable of complete replication and undergo abortive replication in mammalian cells , this is a n example of the persistence of long term memory CD8+ cytotoxic T lymphocytes in the absence of the priming antigen, indicating that T-cell memory might be independent of continued antigenic exposure. The University of Alabama at Birmingham, AL 35294. MHC Class I restricted CD8' CTL activity plays an important role in the control of influenza virus infection as indicated in studies in mice and humans. Cytokines such as IL-2 and IFN-y regulate the generation of virus-specific CTL responses. We recently demonstrated a good correlation between the induction of influenza virus-specific CTL activity and the production of IFN-y by the CD8' T cells at the single cell level using an IF?-specific ELISPOT assay, secreted IFN-y by an ELISA, and IFN-y specific mRNA expression by RT-PCR. Several recent studies have characterized CD4+ and CD8' T cells by their expression on the surface of distinct D45R isoforms. CD45RA is expressed on naive or virgin T cells, while CD45RO is expressed on memory T cells.

In the present study, PBMC of healthy young adult subjects were stimulated with influenza A virus and then enriched for CD8+ T cells. The CD8' cells were stained for CD45RO' (PE) and CD45RA' (FITC) cells and sorted. CTL activity against virus-infected autologous target cells was determined in a 4 hour 'lCr release assay while IFN-y production and expression was assessed by ELISPOT and quantitative RT-PCR, respectively. CDS+/CD45RO+ (memory) cells exhibited significant MHC class I CTL while CDS+/CD45RA+ cells exhibited no lytic activity. No activity was exhibited by freshly isolated or unstimulated CD8+/CD45RO+ T cells. Similarly, CD8+/CD45ROt T cells contained significantly higher numbers of IFN-y spot forming cells and higher quantity of IFN-y-specific mRNA than CD8+/CD45RAC cells. These data support our previous findings that IFN-y may serve as a useful surrogate marker for influenza virus-specific CTL activity in humans. In studying the kinetics of the CD8+ T cell response in LCMV infection we have observed a profound activation and proliferation of CD8+ T cells with a 10-40 fold increase in total number peaking at day 8-9 post infection. In C57BW6 mice, most of the viral antigen is cleared by day seven, and after day 9 the total CD8+ number per spleen drops about 10-fold. However, the relative specificity of the viral peptidespecific precursor CTL frequencies @CTWf) per CD8+ cell remains remarkably stable between day 7-8 of the acute infection and for many months thereafter. Thus, the decline in the CD8' T cell number is not a function of the TcR specificities but is rather an across-the-board event.

In contrast, we found that subsequent to the decline of the CTL response to a second heterologous virus infection such that the mouse was in a "resting, immune'' state, there often was a reduction in pCTL/f to the first virus. For example, infections with W or MCMV substantially reduced the pCTUf to LCMV or PV in all memory compartments, including spleen, lymph nodes, peritoneal exudate cells. Reinfection with the original virus substantially elevated its pCTUf and restored the pCTUf that had been reduced by a heterologous viral infection. Analyses of the progression of CTL responses during a heterologous virus challenge of a virus-immune mouse indicated a high frequency of crossreactive CTL appearing early during infection, but as the infection progressed there was a higher proportion of CTL specific only for the second virus. Thus, we believe that when the across-the-board apoptosis of T cells occurs late in the infection, CTL specific for the first virus are diluted by those responding to the second virus. This may cause the reduction in memory to the first virus and may be one of the mechanisms contributing to the waning of secondary immune responses to certain viruses over time if there is no re-exposure to the original infectious agent. T-cells which arise after virus infection will aid our understanding of Tcell memory and be useful in the design of vaccines which augment the memory response. To estimate the Sendai virus specific precursor frequency in memory mice, CD4+ cells from C57BL6 female mice which had been infected with Sendai virus intranasally (i.n.) more than two months earlier were subjected to limiting dilution analysis. Responder cell populations were enriched for CD4+ cells either by magnetic bead depletion of non-CD4+ cells, or by FACS after staining with anti-CD4 monoclonal antibody These enriched (>90% CD4+) responders were cultured with Sendai virus-infected, irradiated, T-cell depleted splenic antigen presenting cells (APC). Supernatants from these cultures were tested for activity on the cytokine-dependent CTLL cell line. Duplicate cultures of responders on uninfected APC were used to set the level of rejection (mean CPM + 3x std. dev.). Using this type of analysis we were able to demonstrate a frequency of memory Thp at 111600 CD4+ cells, compared to a frequency greater than 1/1OOOOO in naive controls. The memory CD4+ cells were further characterized as CD45RB-low (1/472) , CD44-high (1/294), Lselectin-low (1/364), and CD49d-high (VLA-4-high) (V102). This is close agreement with other phenotyping studies on CD4+ memory cell specific for soluble antigens.

T CELLS, Ralph A. Tripp, Sam Hou, Anthony McMickle, James Houston and Peter C. Doherty, Department of Immunology, St. Jude Children's Research Hospital, Memphis, TN 38105. The immune response of influenza A and Sendai-virusspecific, memory CD8' cytotoxic T lymphocyte precursors (CTLp) have been analyzed in C57BU6 mice infected intranasally with unrelated or cross-reactive respiratory viruses. The numbers of influenza A-specific memory T cells increased in the regional lymph nodes (LN), spleen and bronchoalveolar lavage through the course of an irrelevant infection (influenza B). Memory T cells showed evidence of enhanced steady-state activation. Profiles of CTLp recruitment were analyzed in association with T cell proliferation and activation to determine whether signaling via the T cell receptor is necessary to induce "bystander" stimulation of the memory T cell pool. The extent of T cell proliferation was addressed by treating mice with low doses of cyclophosphamide (Cy). "Resting" Sendai virus-specific memory T cells were unaffected by Cy treatment, however upon challenge with influenza and treated 5 or 6 days later, the emergence of influenzaspecific CTLp was severely diminished. Cell cycle analysis showed that Cy eliminated the majority of CD8' T cells from the LN and spleen resulting in DNA fragmentation of 12-18% ofthis lymphocyte subset. A decrease (though smaller) in the numbers of Sendai virus-specific CTLp indicated that some of the cycling cells killed by Cy were memory T cells, presumably activated in a "bystander" manner. The decrease in CTLp numbers for both influenza and Sendai virus-specific CTLp was still apparent 9 days after Cy treatment, long after the viral elimination. Thus, immune responses to unrelated antigens may be a mechanism involved in maintaining the pool of memory T cells. Experimentally VSV can result in an acute CNS infection of mice. Data from our in vitro experiments indicate that NO has inhibitory effect on productive VSV infection. VSV infection at Neuroblastoma NB41 A3 cells was significantly inhibited by lOOpM of a NO donor S-Nitro-N-acetylpencillamine (SNAP), while 1OOpM of the control compound N-acetylpencillamine (NAP) had no effect. When VSV infected NB41A3 cells were treated with 500pM of a constitutive NO synthase (cNOS) activator N-Methyl-D-Aspartate (NMDA), a significant inhibition of VSV production was observed. Inhibition by 500pM of NMDA was reversed by 300pM of NOS inhibitor N-Methyl-L-Arginine (L-NMA).

Work is in progress to determine the effects of inducible NOS (iNOS) in a glioma cell line C6 on VSV infection. Levels of NO and expressions of both cNOS in neurons and iNOS in glial cells in the CNS following VSV will be further investlgated. Supported by NIH grant A118083 to Carol S. Reiss.

Pediatrics, University of Iowa, Iowa City, IA. 52242 Mouse hepatitis virus, strain JHM (MHV-JHM), is a neurotmpic coronavirus which causes acute encephalitis and acute and chronic demyelinating encephalomyelitis in susceptible rodents. 40.90% of suckling C57BU6 (KbDb) mice inoculated intranasally with MHV-JHM at 10 days and nursed by dams immunized against the virus develop a chronic demyelinating encephalomyelitis characterized clinically b hindlimb paralysis, at 3-8 weeks postinoculation. the chronic demyelinating encephalomyelitis nor the clinical symptoms. Recently, it was shown that lymphocytes isolated from the central nervous system (CNS) of C57BU6 mice both acutely and persistently infected with MHV-JHM display a cytotoxic T lymphocyte (CTL) response to the S protein of MHV-JHM. This response was further characterized by identifying the CTL epitopes that are recognized by a bulk population of CTLs from the CNS of MHV-JHM infected C57BV6 mice. Three epitopes were identified using synthetic peptides and truncated forms of the S protein in primary CIZ assays. The epitopes recognized were amino acids 510-518 (CSLWNGPHL, Db), 598-605 (RCQIFANI, Kb), and 1143-1151 (NFCGNGNHI, Db). Thus, the results indicate that cytotoxic T lymphocytes responsive to the S protein of MHV-JHM in C57BU6 mice recognize both Kb and Db-restricted CIL epitopes. CTL lines and clones specific to these peptides and the entire S protein are being developed to test their biological significance in vivo with respect to the acute encephalitis and chronic demyelinating disease caused by MHV-JHM. A marked change in susceptibility to some neurotropic viruses during the first few postnatal weeks has long been recognised in rodents. Infection of neonatal or suckling mice with the neurotropic alphavirus, Semliii Forest virus results in lethal encephalitis. Infection of weaned animals is not lethal. Earlier investigations focusing on changes in specific immunity have shown this not to be the explanation. Infection of 3-4 week old mice with severe combined immunodeficiency does not result in acute rapidly fatal encephalitis. We have studied mortality, neuroanatomical distribution and spread of infection in mice of different ages and the effect of gold compounds on rendering infection of 3-4 week old mice lethal. Neuroanatomical distribution of infection correlates with synaptogenesis. As this is completed in different systems within the first two weeks postnatal, systems no longer transmit virus and infection switches from disseminated to focal and restricted. Complete productive replication and transmission of infection require smooth membrane synthesis which is present in neurones undergoing synaptogenesis, absent in mature neurones but inducible by administration of gold compounds. Infection of neurones undergoing synaptogenesis is productive and virus is transmitted along neuralpathways, infection spreads rapidly around the brain, destroys cells and animlas die of a fulminant encephalitis. In mice infected after 14 days of age replication in mature neurones is restricted, nonproductive, cannot be transmitted, does not spread, is non-destructive and non-lethal. As a consequence, in the absence of immune responses virus can persist in isolated CNS cells for life and can even be detected by reverse transcriptase PCR in immunocompetent mice months after infection. In the presence of an immune response, CD8+ T-cells recognise and destroy infected glial cells leading to dem yelination. A~k e r m a n n ,~ Virology Swine,' Virology Cattle,' and Avian Diseases3 Research Units, National Animal Disease Center, USOA, Agricultural Research Service, Ames, IA 5001 0 A recombinant pseudorabies virus (PRV) (LLTBAP) was constructed which contains a 3.0 kb deletion spanning the standard recombination junction of the unique long and internal repeat sequences replaced by e lacZ expression cassette. This deletion interrupted the large latency transcript gene (LLT) and truncated one copy of the diploid immediate early IEl80 gene. Replication and viral gene expression of LLTBAZ in Madin-Darby bovine kidney cells was similar to that of the parental virus and a virus rescued for the deleted sequences (LLTBres). When inoculated intranasally in 4-week-old or 4-day-old pigs, LLTBA2 replicated efficiently at the site of inoculation yet caused markedly reduced fatality when compared to the parent or LLTBres viruses. In particular, the LLTBA2-infected pigs did not exhibit neurological symptoms characteristic of PRV infection. To further examine the pathogenesis of LLTBA2, 4-day-old pigs were infected intranasally with LLTpA2 or LLTBres and necropsied at various times postinfection. Virus isolation from the nasal turbinate, tonsils, and trigeminal ganglia was comparable between the two viruses. Although both viruses spread to the brain and induced an inflammatory response in CNS tissues, virus isolation from brain tissues was reduced about 20-fold for LLTpA2. Abundant PRV antigen was detected in the cerebrum and cerebellum of LLTpresinfected pigs, but only a few antigen positive neurons were observed in the cerebrum of LLTBA2-infected pigs. While replication of LLTBres in the brain progressed until death at 7 days post-infection, replication of LLTpA2 in the brain ceased by 9 days post-infection and the pigs exhibited only mild clinical signs. Since LLTBA2 is capable of spread to the CNS, reduced neurovirulence of LLTBAZ is likely the result of its decreased ability to replicate in CNS tissues. The CNS is a target for HIV infection, and in individuals with AIDS this can lead to a devastatin dementia. Only certain viral variants appear capable 07 invading the CNS and infecting microglia and brain macrophages. In order to determine whether the virus entering the brain may be particularly pathogenic to the CNS, we isolated microglia from the brains of SIV-infected rhesus monkeys. Transfer of these cells into naive animals indicated that productive SIV infection could indeed be transferred. Furthermore, CNS infection occurred within a relatively short time span, and was associated with viral gene expression in the brain and pathology characteristic of HIV encephalitis. Serial transfer of microglia into additional animals also resulted in successful transfer of infection, neuroinvasion, and neuropathology. Behavioral analysis in a trained group of animals is ongoing.

This result demonstrates that neuropathogenic virions partition into the CNS during natural SIV infection, likely driven by mutational events that occur during the course of infection. Molecular characterization of the microglia-associated virus has revealed that a distinct pattern of sequence changes in the envelope gene occurs concomitantly with this in vivo selection. Our approach will allow the dissection of functional neuropathogenic elements present in these viruses. 

in non-specific host defense mechanisms. IFN-y-induced nitric oxide (NO) in murine macrophages was previously shown to inhibit the replication of poxviruses and herpes simplex virus type 1 (HSV-1) . We now demonstrate that murine macrophages activated as a consequence of vaccinia virus (VV) infection in viva express inducible nitric oxide synthase (iOS). The VVelicited macrophages were resistant to infection with W and efficiently blocked the replication of W and HSV-1 in infected bystander cells of epithelial and fibroblast origin. This inhibition was arginine dependent, correlated with NO production in cultures and was reversible by the NOS inhibitor NQJmonomethyl-L-arginine. The mechanism of NO mediated inhibition of virus replication was studied by treating VV-infected 293 cells with the NOproducing compound, S-niuoso-N-afetyl-penicillamine. Antibodies specific for temporally expressed viral proteins, a VV-specific DNA probe and transmission electron microscopy were employed to show that NO inhibited late gene protein synthesis, viral DNA replication and virus particle formation, but not expression of the early proteins analyzed. Further, we have also identified putative enzymatic targets of inactivation by NO that results in inhibition VV replication. Although antiviral CTL are important for virus elimination. they can only halt further virus spread, and cannot reduce the number of infectious particles already present. The beneficial effect of CTL-mediated lysis is apparent only if infected cells are lysed before assembly of progeny virus. If infectious virus was released from infected cells in solid tissues before the generation of neutralizing antibody or in sites where antibody did not readily penetrate, then recruitment of mononuclear phagocytes, which phagocytose and destroy infectious material and/or become non-productively infected, would definitely help control virus dissemination. In this context, iNOS induction in macrophages may be an important antiviral strategy. In addition, the inhibition of virus replication in infected contiguous cells by iNOS-expressing macrophages at infectious foci would prevent release of mature viral particles after lysis by NK cells and CTL. Since viral early proteins are expressed in such infected cells, their recognition and subsequent lysis by CTLs will not be hindered.

CNS PERSISTENCE, TROPISM AND GENETIC J. Pedro S i s , Anthony A. Nash and John K. Fazakerley, Department of Pathology, University of Cambridge, CB2 IQP, UK Theiler's murine enchephalomyelitis virus, a natural occuring enteric pathogen of mice, is a picomavirus belonging to the Curdovirus genus. Following intracerebral inoculation of 3-4 week old CBA or BALB/c mice, the BeAn strain causes a chronic persistent CNS demyelinating infection in a proportion of the CBA that survive acute infection. BALB/c mice are resistant to chronic demyeliating disease.

We have studied the tropism, persistence and genetic variability of BeAn, in CBA and BALB/c mice in the chronic phase of this disease.

By in situ hybridisation and reverse transcription (RT) PCR and Southern blot analysis, no viral RNA could be detected in the CNS of any BALB/c mice later than day 60 post-infection. In contrast, in a large group of CBA mice studied up until 393 days post-infwtion, viral RNA could be detected by both techniques in 50% of mice until as late as 268 days post-infection. By employing a combination of, in situ hybridisation for viral genome followed by immunocytochemistry for cell phenotypic markers, BeAn RNA was observed predominantly in oligodendrocytes and occasionally in astrocytes during persistent infection, in both brain and spinal cord. In the persistently infected mice, the striking total destruction of the pyramidal layer of the hippocampus, substantia nigra and anterior thalamic nuclei indicated that these were the mice that had had greatest dissemination of virus and highest virus titers during the preceeding acute phase of infection. Direct PCR t h d cycle sequencing of uncloned RT-PCR products, revealed that during persistent infection, loops I and II of the VPI capsid protein gene did not undergo any genetic variability. Furthermore, no changes were detected in this region in sequenced PCR products amplified from the CNS of mice with severe combined immunodeficiency in which no selective immunological pressure would have been operative.

INFECTION, Thomas E. Lane, Michael J. Buchmeier, Dorota Jakubowski, Debbie D. Watry, and Howard S. Fox, Department of Neuropharmacology, The Scripps Research Institute, La Jolla, CA 92037 Our laboratory is interested in the effects of SIV infection in the central nervous system of rhesus macaques. To enrich for neuroinvasive and neurovirulent viruses, microglia were isolated from infected monkeys and used t o infect new, uninfected monkeys. Such microglia-mediated infection resulted in the production of neuropathological changes, including giant cells, macrophage infiltrates and microglial nodules in recipient animals within 4 months.

Microglial cells isolated from SIV-infected monkeys produced virus in vitro as measured by reverse transcription (RT) and p27 production. Treatment of microglia with recombinant human interferon alpha (rHulFN-a) resulted in a sharp decrease in viral activity (both RT and p27 production) suggesting that rHulFN-a is able t o modulate viral activity in infected microglia. We have analyzed SlVenv sequences by PCR amplification directly from microglia DNA preparations from monkeys. Nucleotide sequence analysis results in an enrichment of unique sequences in the V1 region of the SIV env gene. The majority (>95%) of nucleotide changes encoded amino acid changes, indicating that these envelope sequences evolved as a result of selection. Moreover, sequential passage of SIVassociated microglia resulted in an increase in potential N-linked glycosylation sites within the V1 region of the env gene when compared with the parental virus. These data suggest that sequential passage of microgliaassociated SIV may select for neuroinvasive, neurovirulent variants. The adoptive transfer of CTL specific for an Ld-restricted epitope within the nucleocapsid protein of the JHMV strain of mouse hepatitis virus both protect from acute infection and reduce virus replication in the MHC class 1 positive cells within the CNS. The source of these CTL and the route of their delivery is critical in the outcome of this protection. For example, 10 fold less spleen cells activated in vitro with the pN peptide are required for protection via the direct i.c. route than the i.v. route. In addition, CTL clones are unable to protect via the i.v. route and are very efficient via the i.c. route. These data suggested the possibility that the CD4+ T cells within the polyclonal activated spleen cell population derived from in vitro culture on the pN peptide were facilitating access to the CNS. To examine this question, polyclonal pN-specific T cells were either depleted of CD4+ T cells prior to transfer to infected recipients or untreated cells were transferred to recipients depleted of CD4+ T cells with monoclonal antibody GK1.5. Both of these treatments eliminated the ability of the CTL to reduce virus replication within the CNS, suggesting that CD4+ T cells in the peripheral compartment are required for the entry of CTL into the parenchyma of the CNS during acute CNS encephalomyelitis.

Division of Retrovirology, Walter Reed Army Institute of Research and Henry M. Jackson Foundation, Rockville, MD 20850; Department of Retrovirology, Armed Forces Research Institute of Medical Sciences, Bangkok, Thailand Background The HN-1 epidemic in Thailand is largely due to two highly divergent subtypes of virus, B and E. Dual infection with distinct HN-1 subtypes, which has not been reported previously, would suggest that antiviral immunity evoked by one subtype can be incompletely protective against a second. Merhoak: PCR typing and serologic typing were used to screen a panel of non-random convenience specimens from HIV-1 infected subjects in Thailand. Specimens that showed dual subtype reactivity in these assays were subjected to differential pmbe hybridization and nucleotide sequence analysis of multiple molecular clones to c o n f m the presence of dual infection. Results. Two individuals were shown to simultaneously harbor HIV-1 of env subtypes B and E (Table) . Additionally, both subtypes were identified in co-cultured PBMC from one individual. Conclusions. These data provide the fmt evidence of dual HIV-I infection in humans and reinforce the need for polyvalent vaccines. Infection by herpes simplex virus WSV) induces in man and in mice cytolytic T lymphocytes (CTL) which recognize the immediateearly protein ICP27. Because of its early expression during the HSV replication cycle, lCP27 represents a prime target for specific T cell responses susceptible of controlling virus replication. We have expressed in E. coli a remmbinant construct coding for a fusion protein consisting of a fragment of influenza virus non-structural protein4 (NSI) and the lCP27 sequence of HSV-2. The NSI-ICP27 protein was purified by preparative eleclmplmresis and formulated in oil-in-water emulsions with monophosphoryl lipid A (MPL) and QSZl adjuvants. BalWc mice were immunized by two intrafootpad injections of formulations containing 5 pg of NSI-ICP27. Responder cells obtained from draining lymphnodes were re-stimulated in vitro with P815 cells lransfected with ICP27 and then lesled for cytolytic activity on ICP27-P815 and control P815. The induction of ICP27 specific CTL by different formulations was observed and will be discussed. The induction of heterologous cytotoxic T lymphocytes (CTL) using cassettes of multiple conserved T cell epitopes derived from different proteins and/or virus strains is envisioned as a promising vaccine approach. To study the effects of antigen processing on peptide presentation from chimeric epitope precursors we are using a model system comprising two distinct viral epitopes which are immunodominant in the H-2d haplotype: a Dd restricted epitope from the gp160 protein of HIV-1 and an Ld restricted epitope from the murine hepatitis virus nucleocapsid protein (MHV N).

The influence of proximity and flanking sequences of epitopes on antigen presentation was analyzed using vaccinia virus (vv) recombinants in which the epitopes were expressed as chimeras containing the individual epitopes in reverse order or separated by different spacer residues. Whereas individually expressed epitopes were efficiently recognized by protein-specific CTL, recognition of peptides derived from tandem constructs varied significantly with closer epitope proximity and sequential order. Following immunization with the recombinant viruses, the chimeras were all able to induce antiviral CTL specific for the native proteins. However, Cn, frequency analysis indicated that the number of responder cells to the same epitope dramatically depends on its context within the chimera and correlates with antigen recognition in vitro. The profound effect of flanking regions on CTL induction suggests that the context of an epitope will require careful evaluation in the design of recombinant multivalent minigene vaccines to induce an optimal T cell mediated immune response. and Robert E. Johnston', Departments of 'Microbiology & Immunology and %iochemistxy, Univ. of North Carolina, Chapel Hill, NC 27599 A hll-length cDNA clone of Venezuelan equine encephalitis virus W E ) has been altered to contain two strongly attenuating mutations and a second subgenomic RNA promoter immediately downstream of the structural gene region. Expression ofthe influenza HA protein from this second promoter in baby hamster kidney (BHK) cells was approximately 50?? of the level in influenza virus-infected cells, as measured by immunoprecipitation. Fourweek-old CD-1 mice were inoculated subcutaneously with 2 x 10' p h of the HA vector, vector alone or diluent. Expression of HA mRNA was detected in the draining lymph node of HA vector-inoculated mice by in situ hybridization, consistent with the organ tropism of VEE. Mice were challenged three weeks after imnmization by intranasal administration of lo5 ED, of influenza h s . AU 24 corn1 mice suffered severe disease and 50% died. Only one of 12 HA vector-inoculated mice died, and another exhibited signs of disease for one day and recovered. The geometric mean ELISA titer of anti-HA serum IgG in the HA-vector inoculated mice was 246, while only three control mice had measurable serum reactivity, and that was at the lowest dilution tested, 150. In a parallel experiment, no influenza infectivity was detected in the lungs of 12 HA vector immunized mice at 4 days postchallenge. In contrast, 8/12 PBS-inoculated mice and 5/12 inoculated with vector alone were positive for influenza infectivity and had geometric mean titers of 3.04 and 1.93 x lo6 pWgm, respectively. This vector also has been used to express the H N W C A protein in a form recognized by patient sera and a specific antibody on Western blots. These experiments demonstrate the feasibility of using vectors based on attenuated VEE cDNA clones for protective immunization against heterologous human and animal pathogens. Dose/response curves have been used to compare different routes of immunization with plasmid DNA encoding the H1 hemagglutinin glycoprotein of influenza virus. Routes of inoculation included intramuscular, intradermal and gene gun delivery of DNA. From 100 to 0.1 ug of DNA was inoculated by intramuscular and intradermal routes. From 0.4 ug to 0.0004 ug of DNA was inoculated by gene gun. Each route was evaluated for single and boosted immunizations. Antibody titers were followed Over a 20 week period, following which animals were evaluated for protection against a lethal challenge. Each of the routes raised both antibody and protective responses. Gene gun-delivery of DNA required 250 to 2,500 times less DNA to raise responses than the intramuscular and intradermal inoculations. Boosts did not have much of an effect on antibody titer or protection except at low dose inoculations (4 ng and lower for the gene gun). For each of the routes, antibody responses showed good persistence over the 20 weeks of the experiment. Inoculation of mice with plasmid vectors carrying a microbial gene under the control of an appropriate promoter results in a full spectrum of immune responses to the vectorencoded antigen. Using a murine rabies model a plasmid termed pSGSrab.gp expressing the full-length rabies virus glycoprotein regulated by an SV40 promoter was shown to induce upon inmuscular inoculation a rabies virus specific T helper cell response. of the Thl type, cytolytic T cells and virus neutralizing antibodies resulting in protection against a subsequent challenge with live rabies virus given either peripherally or directly into the cenaal nervous system. A response comparable in magnitude was also induce upon inoculation of a vector expressing a secreted form of the rabies virus glycoprotein. The immune response to the DNA vaccine could be modulated by co-injection of the rabies virus glycoprotein-expressing vector with plasmids expressing mouse cytokines. Inoculation of mice with the pSG5rab.gp vector and a vector expressing granulocyte/macrophage colony stimulating factor (GM-CSF) enhanced both the T helper and the B cell response to rabies virus thus improving vaccine efficacy. Co-inoculation with vectors expressing interferon-g failed to improve the response. Co-inoculation of the antigen-expressing vector with a plasmid encoding mouse I L 4 caused a reduction of both the T helper cell response and the B cell response to rabies VlNS.

HPVl6 E7 HPV associated cervical cancer cells express HPV16E7 protein and antibody to HPV16 E7 can be detected in the blood of cancer patients, yet the twnours are. not rejected. A mouse transgenic for the E7 protein of HPV16, and expressing E7 protein in the skin, has recently been described (1) and these mice develop spontaneous humoral immunity to E7 protein similar to patients with cervical cancer(2). To determine whether immunisation could induce immunity to E7 sufficient to allow tumour rejection, we firstly demonstrated that immunisation of H-Zb mice with HPV16E7 protein with Quil A as adjuvant could induce cytotoxic T cells able to kill HPV16 E7 expressing tumour cells in nrro. We then used similar immunisation with E7IQuil A to induce E7 specific immunity in FVB (H-29) mice. H-2qskin @s expressing E7 were not rejected by E7 immunised H-Zq mice, though immunisation induced antibody to E7, and similar grafts were rejected, as expected, across an doantigen mismatch in H-Zb mice. We conclude either that HPVl6 E7 lack a Tc epitope in the context of H-Zq, or that expression ofE7 in the skin from the E7 transgenic mice is insufficient for recognition by primed effector cells, and further experiments will address this distinction. Cervical carcinoma Is strongly associated with infection by human papillomavirus (HPV) types 16 or 18, and continued expression of the E6 and E7 gene products. This provides an opportunity for an immunotherapeutic approach to the treatment of cervical carcinoma by activation of immune reponses directed against these virally encoded tumour specific antigens. We have constructed a recornbinant vaccinia virus expressing E6 and E7 from HPV16 and 18 with the aim of inducing E6 and E7 specific HLA class I restricted cytotoxic T lymphocytes (CTL). The sequences have been inserted into the Wyeth vaccine strain of vaccinia virus at a single locus in the form of two separate fused E6/E7 reading frames, each under the control of an early v a d n i a promoter, and each modified to inactivate the Rb binding site. The virus has been characterised with respect to its ability to synthesise the expected HPV proteins, its genetic stability, and growth and virulence in a mouse model prior to use in human clinical trials.

Analysis of HPV16 €7 specific CTL from C57BU6 mice immunised with this recombinant virus show the response to be equivalent to that generated by a control vaccinia recombinant expressing non-modified HPVI 6 E7 alone, with similar recognition of the defined immunodominant H-2Db restricted epitope, E7 residues 49-57.

ABILITY OF MICE TO RESIST INFLUENZA CHALLENGE, Arthur Friedman, Douglas Martinez, John J. Donnelly and Margaret A. Liu, Department of Virus and Cell Biology Research, Merck Research Laboratories, West Point, PA 19486 Mice infected with the laboratory strains of A/PR/8/34 (HlN1) or the mouse adapted A/HK/68 (H3N2) show complete protection against challenge with a different strain of Influenza A. Humans, however, undergo multiple influenza infections as previous infections appear to provide weak or short-lived protection against the continual antigenic change of strains. We have previously shown that immunization of naive mice with DNA encoding the conserved internal antigen nucleoprotein (NP) provides protection against both H1 and H3 strains of A/influenza. Although such mice became infected they were resistant to weight loss and death this differed substantially from A/PR8 and A/HK recovered mice which were resistant to subsequent infection. To produce a more representative model of human infection, we infected the lungs of mice with currently circulating strains of human influenza. Mice that had been given lung infections with A/Beijing/92 were susceptible to subsequent infection with the A/HK/68 strain although they were resistant to weight loss and death. Other strains such as A/Beijing/89 or A/Georgia/93 provided only marginal protection against weight loss and death against A/HK challenge. Mice that were immunized with NP DNA had greater resistance to weight loss and death after A/HK/68 challenge than mice previously infected with A/Bei/89 and A/Ga/93, and were similar to mice that had been previously infected with A/Bei/92. Thus, infection with different virus strains provide various levels of cross strain protection and the level of protection provided by immunization with DNA can exceed that induced by live influenza infection. The development of Sendai virus-specific cytotoxic T lymphocyte (CTL) effectors and precursors (p) has been compared for mice that are homozygous (-/-) for a disruption of the H-21-Ab class II major histocompatibility complex (MHC) glycoprotein, and for normal (+I+) controls. The generation of CD8+ CTLp was not diminished in the (-/-) mice, although they failed to make virusspecific IgG class antibodies. While the cellularity of the regional lymph nodes was decreased, the inflammatory process assayed by bronchoalveolar lavage (BAL) of the infected lung was not modified and potent CTL effectors were present in BAL populations recovered from both groups at day 10 after infection. There was little effect on virus clearance. As found previously with CD4-depleted H-2b mice, the absence of a concurrent class Il-MHC-restricted response does not compromise the development of Sendai virus -specific CD8+ T cell-mediated immunity. The importance of cytoxic T lymphocytes in defense against acute and chronic viral infections is gaining increasing recognition. Our approach to investigating the structure-function relationship between immunogens and their in vivo ability to elicit cytotoxic T lymphocyte responses has been to formulate simple, well-defined structures that vary in their ability to introduce associated antigens directly into the cytoplasm of antigen presenting cells.

We have introduced methods for the preparation of unique, lipid-matrix based immunogens, which are highly effective in mice and monkeys for stimulating strong CD8+ cytotoxic T cell responses, (CTL). Antigens used have been proteins or peptides derived from influenza, parainfluenza, and HIV viruses, and whole formalin-fixed SIV. CTL can be induced by parenteral as well as oral administration.

Comparing the physical and chemical nature of our formulations with those from other laboratories which have reported the use of subunit preparations to induce CD8+ CTL, leads us to propose that a minimal immunogenic formulation capable of eliciting CD8+, MHC Class I restricted cytotoxic T lymphocytes includes: I) a peptide that represents a MHC Class I epitope; ii) a component that enhances the aftinity of the immunogen for MHC Class I positive antigen presenting cells ; iii) properties that can compromise the integrity of a lipid bilayer, facilitating delivery of the antigen directly into the cytoplasm for class I presentation.

CD8+ responses to peptides, glycoproteins, and even whole fixed viruses, makes them attractive candidates for diseases where clearance of infected cells is important in protection and recovery. Cani ne rabies is uncontrolled. Rabies also is epizootifally active in several species in most areas of the wor!d. Thm, vaccination of animals, both wild and domestic, as well as postexposure treatment of humans remains a global concern. Unfortunately, in those countries in which,people most need postexposure prophylaxis, the best vaccines are expenswe and in limited supply, whereas available vaccines are of questionable immunogenic efficiency, are otten contaminated and may produce neurological complications. The goal of this study was to determine whether a rabies vaccine for global use is complete one round of replication have the potential to be used as vaccines. We have previously reported the abiliity of a gHdeleted Herpes Simplex Virus Type 1 (HSV-1) to protect mice and guinea-pigs from subsequent challenge with wild-type HSV.

This virus, which we have called DISC (Disabled Infectious Single Cycle) virus, can infect normal cells but the absence of gH in the progeny virus prevents further rounds of infection.

As DISC HSV clearly has potential as a vaccine, it is important to determine the durability of the immune response elicited by this virus. We have investigated the ability of DISC HSV-1 to protect mice from a wild-type virus challenge six months post vaccination using the ear model of HSV infection. Two immunisations on day 0 and day 21 resulted in a considerable reduction in virus titres in the challenged ears, and an almost complete absence of virus In the dorsal root ganglia.

HSV-specific antibody titres as determined by neutralisation and ELlSA were maintained for the six months period. It was possible to demonstrate an HSVspecific cytotoxic T-cell response in the DISC HSV-1 vaccinated mice following challenge; this CTL activity was similar to that observed in mice vaccinated with wild-type virus and challenged after the same time period. Animals vaccinated with inactivated virus or control mock-vaccinated mice showed a low level of CTL activity typical of a primary CTL response following challenge.

These results indicate that an effective cell-mediated and humoral anti-HSV immune response can be maintained for at least six months following vaccination with DISC HSV-1.

Viruses which lack an essential gene and thus can only 

The lmmunogenicity of two CTL @topes. influenza NPl47-158 and plasmodium berghei CS protein 252-260 were studled in BALBlc mice.

Paptides were formulated as a) a IipopepUdepeplkle conlugated to Irlpalmltoyl-Sgly~~l cysteine (Pam3Cyj) and dissolved in a 1% DMSOlglycerol solution. b) micmparticles prepared with poly @.L ladide-coglywlide) using a solvent evaporation technique. The micropaltides were administered as a suspension in phosphate buffered saline or c) an emulsion prepared wilh egg lecithin and 10% soya oil in water. 1 Wpg of peptide or controls (the welght equivalent of blanks) were administered to groups of 3 mice intra-peritoneally or subcutaneously at 1.10 and 20 days. 7 days following the last immunization splenocytes were cukured in vlro in the presence of appropriate pepWe or wntml M h rat con A supematant as a source of omwth fadors. CTL adivity was measured in a standard 4 hour chromium release assay and results expressed as % specific lysis.

CTL could be elicited in vivo with all three formulations. At an ewedoctarget ratio of 1W:l the plasmodium berghei peptide encapsulated in micmpartides gave 47% iysls on peptide pulsed target calls. Levels of lysis were similar for the peptide in emulslons. The iipopeplide P3CCS252-260 gave a level of lysis of 82% at an E:T raliOOf1W1.

These results demonstrate that peplides edminldered in a variety of formulations can induce a systemic CTL response in vivo. Peplide vaccines using such formulations wuld be used to stimulate CTL responses as part of a prophyladic vaccines or as immunotherapeulics.

ATTENUATED The attenuated Sabin strains of poliovirus have been used for many years to elicit protective immunity to poliovirus. Oral vaccination with replicating polioviruses generates both mucosal and systemic immunity. therefore, use of recombinant polioviruses expressing heterologous antigens as vaccine delivery vectors should provide a system for generating protective immunity to those antigens. cDNA copies of the poliovirus genome has been used to construct vectors containing a multiple cloning site for insertion of heterologous genes. A pilot enhanced-potency inactivated poliovirus vaccine (EJPV) with assumably improved immunogenicity containing win-treated type 3 poliovirus (shah Sauketf) together with the regular type 1 and type 2 canponena was subjected to s*mdard safety and potency tesls in the labmatory and laken through Wase I and I1 clinical aials. In Balb/c mice, the lrypin-mcdit%d E-IF' V CfrYIPV) was found to induce antibodies targeted ouaide the uypsin-sensitive BC-loop of capsid protein VP1. as previously shown for hypsin-mated type 3 poliovirus @VM alone. Trypsin used to modify the type 3 component at the bulk phase was removed by the vaccine manufacturer (RIVM) in the regular purification process. Absence of uypsin in the final product was further confumed by immunizing mice and rabbits with 10-fold concentrated type 3 component of TryIPV. Assays for lrypin antibodies using EIA and

Westem blot techniques were newve. In the clinical phase 1 aial six adult volunteers with existing immunity to poliovirus were given increasing doses of TryIPV. Already one tenth of the regular dose induced a booster effect in neuarlizing antibodies to both intact and mypsin-treated type 3 poliovirus. No unexpected sideeffects were recorded phase I1 trials comprised 50 adult volunteers with at least 5 years since the last dose of poliovirus vaccine and 50 children who were due to receive the third dose of the regular immunization schedule at about 2 years. In both groups. 25 individuals received TryIPV and 25 were injected with the regular enhanced potency IPV (E-IPV). Serum specimens drawn before injection and one month after were tested for neunalizing antibodies using standard microneuwlization assays (all mutypes) and the RACINA test (intact and uypsin-mated type 3 poliovirus). In all volunteers TryIPV was at least as immunogenic t w the regular E-IPV according to all assays. No statistically significant differences in side effects were reported. 

A murine/influenza virus model has been used to evaluate the longevity of antibody and protective responses raised by gene gun delivery of a hemagglutinin-expressing DNA. Mice were immunized and boosted at one month with 0.4 ug of an H1 expressing plasmid DNA (pCMVRI1). Antibody responses and protection against a lethal challenge were followed over the next year. Antibody responses had good longevity exhibiting comparable titers at one year post boost as at 10 days post boost Protection against the lethal challenge was complete at 10 days, 1 month and four months post boost, but only partial at one year.

A TRANSGENIC MOUSE MODEL FOR IDENTIFING HTLV-1 T-CELL EPITOPES: GENERATION OF HLA-B*3501-RESTRICTED CTL DIRECTED AGAINST SYNTHETIC PEPTIDES AND NATURALLY PROCESSED VIRAL ANTIGENS, Christian Schiinbach*, Ai Kariyone*, Kiyoshi NokiharaA+, Karl-Heinz Wiesmulle6 and Masafumi Takiguchil, Departments of Tumor Biology* and Immunology#, Institute of Medical Science, University of Tokyo, Tokyo "Tokyo University of Agriculture and Technology, Tokyo +Biotechnology Instruments Department, Shimadzu Corp., Kyoto, Japan $Natural and Medical Science Institute at the University of Tubingen, Reutlingen, Germany The majority of human T-cell leukemia virus type-I (HTLV-l), HLA class I-resmcted T-cell epitopes have been identified by cloning HTLV-1 patient-derived T cells. Here we describe for the frst time a rapid method (reverse immunogenetics) for identifing T-cell epitopes, together with a transgenic mouse model as a guide for testing the cellular immune response to a mixture of the lipohexapeptide immunoadjuvant PamgCys-Ser-(Lys)4 and synthetic HTLV-1 peptides which seem suitable for vaccine design. HTLV-1 amino acid sequences were searched for eight to 14mer patterns carrying the anchor residues of the HLA-B*3501 peptide motif at positions two and eight to fourteen. 65 candidate peptides were synthesized according to the matched sequence patterns. Their HLA-B*3501 affinity was quantitatively analyzed in an indirect immunofluorescence peptide binding assay using RMA-S-B*3501 cells. The fourth group (controls) were inoculated with H3N2 (IN) thereby providing heterotypic CTL immunity in the context of a natural infection without the confounding effects of humoral immunity against surface antigens. All four types of inoculations have been shown to protect normal (class I expressing) mice from a lethal challenge with influenza, presumably mediated by class I restricted cytotoxic T cells. The two groups inoculated via the intranasal route may gain additional protection by activating the mucosal immune system (IgA). None of these types of inoculations has been evaluated in the context of class I1 restricted cytotoxic T cells, the only CTLs found in class I deficient mice. For all four types of inoculations, MHC class I deficient mice lost significantly more weight than the class I expressing control groups (seven mice per group) indicating the importance of class I restricted T-cells in protection. Within the class I expressing groups, there was no significant difference between the four types of inoculations; within the class I deficient groups the VAC-NP IM immunized mice lost significantly more weight than the H3N2 group;the other two groups, VAC-NP IN and genetically immunized groups had intermediary results. These data lend support for a protective role for mucosal immunity. Results on both class I and class I1 CTL activity for the four types of inoculations will also be presented. We tested the PBMCs of patients participating in two vaccine therapy trials for their ability to recognize overlapping peptides of the gp120 LA1 sequence. Seventeen patients participating in a Phase I gp160 protocol and 13 patients participating in a Phase I gp120 protocol had their PBMCs isolated by Ficoll separation of heparinized venous blood. The fresh PBMCs were plated, in triplicate, into 96 well plates containing peptides overlapping the LA1 sequence of gp120, pulsed on day 7 with tritiated thymidine and harvested and counted on day 8. Results: The percentage of patient's PBMCs from each trial with an LSI 2 5 to each peptide are depicted below.

Conclusions: The PBMCs of HIV-infected volunteers who have been multiply immunized with either gp160 or gp120 proliferate to multiple peptides within the gp120 molecule. Reactivity from the end of C1 through early C2 (LAI #I 12-21 1) is particularly prominent and contains previously undescribed Th epitopes (asterisks). Conspicuously missing is reactivity to the V3 loop peptide (LA1 #300). Although the percent reactivity to the entire gp120 molecule is similar between the immunization groups, there is differential recognition of some of the individual peptides, particularly peptides in early C3 (LAI #319-348). The intracytoplasmic lifecycle of Listeriu mmcytogenes (LM) enables it to be a convenient vaccine vehicle for the introduction of foreign proteins into the MHC class I pathway of antigen presentation. Taking advantage of these properties, we have inserted the nucleoprotein (NP) gene from lymphocytic choriorneningitis virus (LCMV) into the LM chromosome by site specific homologous recombination. Infection of mice with recombinant LM expressing LCMV-NP elicited a virus-specific CTL response. We were able to recover LCMV-NP specific CTL precursers from recombinant LM vaccinated mice as shown by vigorous secondary CTL responses after in vitro stimulation. In contrast to mice immunized with wild type LM, mice vaccinated with NP-recombinant LM were protected against challenge with immunosuppressive LCMV variants. Protection was demonstrated by reduced viral titers or complete clearance of LCMV from serum and various organs including, spleen, liver, lung, kidney, and brain. The kinetics of the LCMV challenge indicate that mice vaccinated with recombinant LM were able to arrest viral growth early in the infection due to a strong CTL response and did not exhibit the immunopathology associated with infection of naive mice. Since LM not only delivers antigens into the MHC class I pathway but also induces IL-12 production, it has the potential to function simultaneously as a vehicle for expressing foreign antigens and as an adjuvant promoting cell mediated immunity. , latent membrane protein [LMP] 1 and 2A) in chromium release assays. We were fortunate in identifying one child from whom cryopresetved PBMC samples were available before. and during EBV seroconversion. EBV-specific CTL activity was demonstrated concurrent with initial detection of virus in the peripheral blood by EBV-DNA PCR. in the absence of detectable serum antibody. CTL lines from all nine children recognized one or more EBV latent gene prcduct(s). All children demonstrated CTL responses against one or more EBNA 3 proteins (3A, 3B, 3C). and EBNA 3C was recognized most frequently. No CTL responses were detected against the EBV latent proteins EBNA 1, 2, LP or LMP 1. The EBV-specific CTL lines expressed CD3/CD8 and mAb blocking experiments demonstrated that the majority of target cell lysis was inhibited by antibody against MHC class-I but not antibody against MHC class-11. These results represent one of the first reports characterizing EBV-specific CTL responses in young children. The striking similarity between EBV-specific CTL responses described here in young children and those reported for adults suggests that the EBNA 3 family of proteins and LMP 2A should be considered for inclusion in candidate EBV vaccines.

EVALUATION OF CELLULAR IMMUNE RESPONSES TO ADENOVIRUS VECTORS IN THE COTTON RAT. Soonpin Yei,' Gary Kikuchi,' Ke Tang' and Bruce C. Trapnell.' Departments of Virology' and Immunology,2 Genetic Therapy, Inc., Gaithersburg, Maryland 20878 Replication deficient recombinant adenovirus (Av) vectors are efficient gene delivery vehicles currently being developed for a variety of in vivo gene therapy strategies such a s for the fatal pulmonary component of cystic fibrosis. The cotton rat (Sigmodon hispidus) is one of the most widely accepted animal models for studying these Av vectors because wild type human adenovirus replicates in cotton rats and because of the histopathologic similarities of infected respiratory epithelial tissues from humans and cotton rats. Despite this, methods for studying immunologic responses in the cotton rat have not been developed. Importantly, recent studies in the cotton rat (Gene Tber. 1 :192-200; 1994) in our laboratory suggest that a dose-dependent specific immune response to Av vectors can limit expression of the transgene. In this context, w e have established methods to evaluate cytotoxic lymphocyte (CTL) responses to Av vectors in the cotton rat. To accomplish this, a CTL target cell line was established consisting of primary cotton rat lung fibroblasts (CRLF). Splenocytes from cotton rats exposed previously to an Av vector were harvested, cultured in virro with irradiated, AddB274nfected CRLF. Cultured (effector) splenocytes were then incubated with S'Cr-labelled CRLF (target) cells a t effectoctarget (E:T) ratios of 100, 50 and 10. In parallel, splenocytes from naive cotton rats served as negative controls. Results demonstrated vector-specific CTL lysis of target cells significantly greater than controls: 80.3 f 1.3% vs 6.2* 0.5%. 49.6*1.6% v s 5.7*0.4%. and 22.8*3.5% vs4.8*0.5% (MeanrtS.E.M., n=3; p<O.Ol, all comparisons) a t E:T ratios of 100, 50 and 10, respectively. Thus, a specific CTL response does develop in cotton rats following in vivo Av vector administration. This observation will be useful in guiding vector modifications in the rational development of improved vectors for clinical use. The identification of CTL epitopes is essential for subunit vaccine design against AIDS. A major portion of the anti-viral CTL responses after infection with HTV in humans or SIV in rhesus macaques is made up by anti-gag specificities. Here, effector cell populations were established from HIV seropositive individuals by specific stimulation of PBL with pfa fixed autologous B-LCL infected with TW gag. Expanded cultures were then tested for CTL activity against autologous B-LCL pulsed with overlapping 20-mer peptides derived from p24. Two individuals (008 and 617) had CTL against p24.14 (a(mino) a(cids) 263-282, KRWIILGLNKNRMYSFTSI), two others (038 and 157) recognized p24.17 (aa293-312, FRDWDRFYKTLRAEQASQD).

Similarly we found two SIV infected rhesus macaques reacting against the analogous aa sequences of SIV; it. p26.14 (8653) and p26.17 (IVY). This led us to further fine map the human and rhesus CTL reactivities using sets of 9 aa overlapping 10-mers and to test for HIV-1ISIV crossreactivity. Both two p24.14 epitopes (possibly restricted by HLA-A2) and the p26.14 epitope were non-overlapping, and all three epitopes were non-crossreactive. Finemapping of the p24.17 epitopes is in progress. The SN p26.17 epitope was also non-crossreactive, despite only one aa difference (position 6, S+T) between the SIV and HIV sequence. Others have also mapped Cl'L epitopes to the same regions of p24 in humans (Johnson JI 147,1512 ,1991 , Buseyne JV 67,694,1993 and to p26.17 in cynomolgus macaques (Gotch JMPrim 22.1 19,1993) . In conclusion, multiple non-cross reactive CTL epitopes are clustered within corresponding regions of HIV and SIV gag. CTL hotspots may be important for inclusion into vaccines.

forming virus) with CD-4 and CD-8 1-cells, B-cells and adherent cells, while an average of 5% of virus was associated with LPS-stimulated 8-cells and adherent cells.

In the second experiment the combined function of antigen presenting cells, 1-helper cells and 8-cells (humoral immune response) to a third party antigen (sRBC) during the acute phase of enterovirus induced disease was increased at day 0, and decreased at days 2, 3 and 4 post-CVB3, Infection relative to unlnlected animals.

Conclusions: These data indicate that CVB3, associates with and replicates in rnurine splenic immune cells, and that infection of susceptible mice with CVB3, modifies the immune response to a third party antigen during the acute stages of disease. An understanding of such CVB3,-i mmune cell interactions is criitical for gaining insight into the pathogenesis and persistence of enteroviral infection both within and outside of the immune system. Understanding the pathogenic mechanisms of virus infections depends upon identification of viral gene products involved in host-virus interactions in vivo. Many viral gene products which affect pathogenesis in vivo are nonessential for virus replication in vitro. Therefore, we constructed insertion or deletion mutants in the Hind I11 J and I regions of murine cytomegalovirus (MCMV) to assess the importance of these genes in in vivo replication of virus as well as interaction with immune regulatory and effector cells. Three mutants replicated in NIH 3T3 fibroblasts as efficiently as wild-type virus, thus confirming the nonessential nature of the genes. Deletion of 2.8 Kb in Hind I11 J and of 7.7 Kb in Hind I11 J and I resulted in mutants with reduced replication in target organs in vivo. An insertion in Hind I11 J had no significant effect on replication in mouse salivary gland, but curtailed growth of the virus in footpad fibroblasts and subsequent priming of MCMVspecific CPL precursors in the draining lymph node. In addition, the 7.7 Kb deletion mutant was unique in its failure to replicate in a permissive, differentiated macrophage cell line. Current studies aim to further define the role of this gene region in HCMV pathogenesis. that in VV5-4 progression o f V V life cycle is blocked at the the uncoating II or replication stage. Furthermore, host protein synthesis in V V 5 -4 cells is not shut-off after V V infection suggesting uncoatingll/replication is important for determining cell fate. Since this mutant clone VV5-4 was isolated b y retroviral insertional mutagenesis. A cellular gene that was integrated by single retrovirus was isolate and codes for a 5Kb transcript in Northern blot analysis. A 2 Kb cDNA was isolated and codes f o r a novel polypeptide o f 800 amino acids that show no homology with known proteins.

Experiments are in progress to understand t h e role of this cellular gene in VV infection. In addition, a two-hybrid system will be employed to identify a n y viral gene product interacting with this cellular target U s i n g o u r laboratory's s t a n d a r d s t r a i n o f A/Japan/305/57 in in uitro assays of viral infectivity of the mouse B-lymphoma A20-1.11. we have found that although infection sensitizes the A20 cells for recognition by both Class I and Class I1 restricted CTL.'s and leads t o high surface expression levels o f hemaglutinin (HA), the infected A20 cells grow through t h e infection, HA expression declines, and few of the infected cells die. In contrast, using a second s t r a b of A/Japan/305/57 from a different passage history, we find that n o t only are the infected A20 cells sensitized for lysis by both Class I and Class II restricted CTL's, exhibiting even higher levels o f HA surface expression, but in addition, most cells in the infected population die with characteristics of apoptosis.

The two virus strabs appear to be highly related because in various assays we have found t h e m antigenically indistinguishable in both their T-and B-cell epitopes.

Using these two virus strains in a n in uiuo dose response assay, we have found that o u r laboratory's standard strain will kill all mice infected with a 10-2 The therapeutic and prophylactic antiviral efficacy of interleukin-12 (IL-12) was studied using murine models of herpes simplex virus (HSV) and murine cytomegalovirus (mCMV) infection. Therapeutic intraperitoneal administration of IL-12 commenced 6 hours after mice were infected w i t h HSV, and was continued daily for a total of 5 days. IL-12 therapy improved the survival rates of mice w i t h systemic HSV infection, compared to that of placebo treated infected mice. Since neutrophilia and thrombocytopenia were common in dengue patients,we are investigating whether BM may be accompanied with abnormal hematopoiesis. Using Ficoll-Hypaque separation, mononuclear cells prepared from the total BM cells. 1 .0x106we~ells/well of both adherent and nonadherent cell types were infected with two different DEN-2 virus strains : (1) Thiland strain isolated from the dengue hemorrhagic fever(DHF1 patient and (2) Taiwan strain isolated from the dengue fever(DF) patient at MOI=l and collected the suprnatants at various time points for assay. The preliminary data showed that : (a) adherent cells were infected with DEN2 and DHF virus strain replicated much more efficiently (2.3x104PFU/ml on day 3 p.i.1 than DF strain; (b) nonadherent cells were 10 times less efficient to produce viral yields than adherent cells (DHF and2 DF strains peaked at 6.3~10' PFU/ml and 1.5~10 PFU/ml on day 3 p.i., respectively);

(c) addition of GM-CSF to nonadherent cells increased virus infectivity and productivity. In conclusion, DEN viruses were able to infect BM cells and the more virulent virus strain isolated from DHF patients had better replication capability in human bone marrow cells. Future studies on cytokines and immunologic evaluation are in progress. number of gene products whose function is to modify host responses. Recent evidence suggests that a subclass of poxvirus-encoded host response modifiers interferes with host defense responses, and thus can be termed immune or inflammatory response modifiers (irms). Several poxvirusencoded irms possess homology to mammalian proteins which are associated with host defenses, while other poxvirus irms have little sequence homology to immune-effector molecules or their receptors. The Ultrecht strain of rabbitpox (RPV) virus induces a large reddish pock on the chorioallantoic membrane (CAM) of the chicken embryo. RPV mutants lacking the SPI-2/crmN38K gene trigger an inflammatory response by 36-48 hr after inoculation, similarly to that found for cowpox virus (Brighton red strain) mutants. RPV mutant viruses with a deletion of the PS/HR gene, a gene which has recently been associated with virus host range and virion maturation, elicited an inflammatory response on the chicken embryo CAM at 48-72 hr after virus inoculation. Thus, another poxvirus-encoded gene product has been identified that has irm functional activity which would not have been predicted by sequence analyses.

Poxviruses have recently been found to possess a

Herbert W. Virgin IV, Center for Immunology and Division of Infectious Diseases, Washington University School of Medicine, St. Louis, MO 63110

Murine cytomegalovirus (MCMV) follows three stages of infection in the immunowmpetent host. Following the initial acute infection, a persistent infection remains in which lytic virus resides in the acinar cells of the salivary glands. Finally a third stage of infection exists which is characterized by a lack of detectable infectious virus. It is not known whether MCMV remains persistent at some low level during this thiid stage or establishes molecular latency. To test this, spleens, kidneys and Salivary glands from mice 2-8 months post infection were evaluated by two sensitive assays. Sensitivity was tested using virus co-cultured on murine embryo fibroblasts (MEFs) for 14 days. By this method, 1 pfu MCMV was reproducibly detected in the presence of sonicated spleen, kidney and salivary glands diluted 1:lO. SCID mice were used as a second detection assay for MCMV. Using 46 SCID mice and different doses of virus, the LDs, of MCMV in SCID mice was 2-3 pfu. MCMV infections were detected when 10-25 pfu were added to sonicated spleens, kidneys, and salivary glands prior to injection into SCID mice. Using both co-culture with MEFs and injection of tissue into SCID mice, we have tested a total of 34 spleens, 34 kidneys, and 37 salivary glands and have detected no persistent virus. While no preformed virus existed in these tissues, reactivation from explanted spleen, kidney and peritoneal exudate cells occurs in culture after 10-40 days. In addition, lytic MCMV was detected in SCID mice 28 days after injection with 2(10)' kidney cells from latently infected mice. Following identical transfers of latent spleen cells, MCMV was not detected in SCID mice although explants from the same organs reactivated in vitro. Using FACS sorting, panning, magnetic bead separation, and reactivation assays, individual cell types have been analyzed for the presence of MCMV genome and the ability to reactivate virus. Heat shock proteins (HSP) are expressed in all organisms in response to a variety of stresses, including virus infection. However, since viruses are not known to encode HSP genes, this represents expression of cellular HSPs. We have found a dramatic induction of the 72kDa HSP in the ovaries of W-infected mice, and using primary ovarian fibroblasts, demonstrated HSP72 mRNA within virus-infected cells. The physiological role of HSPs expressed during virus infection is unknown, although HSPs act as molecular chaperones to facilitate protein folding and assembly in unstressed cells, and a number of bacterial HSPs are essential for bacteriophage replication and morphogenesis. In order to determine whether HSP72 expression was beneficial to VV replication, we constructed a recombinant VV which encoded murine HSP72, but found that this virus grew to similar titres as control viruses, both in vitro and in vivo. In particular, kinetic studies of virus growth in an HSP72-negative cell line showed no difference from control viruses, and the presence of HSP72 did not influence the virulence of infection in immunocompromised nude mice. These results are interesting given that HSP70 proteins act as molecular chaperones and that HSP72 can be found in association with VV proteins. We suggest that this association may simply reflect the inherent affinity of HSP70 for non-native proteins, and the close physiological proximity of HSP and nascent viral proteins within the virus infected cell. Our results indicate that the functional significance of these interactions are minimal with respect to the outcome of virus infection. Furthermore, although the expression of HSW2 in VV-infected mice coincides with the peak anti-viral cellular immune response, HSP72 does not appear to be immunologically significant, as we cannot demonstrate any immunological responses to HSP72 during the course of VV infection. The effect of in vivo neutralisation of IFN-y on cytokine production during influenza infection was compared in CS7/BL6 and BALB/c mice. Similar cytokine profiles were obtained on in v i m restimulation of MLN or spleen cells from virus-infected mice of each strain. At days 7 and 10 postinfection, MLN or spleen cells from both strains of mice produced IL-2, IL-6, IL-10 and IFN-y, but little or no IL-4 or IL-5. However, following in vivo treatment with anti-IFN-y antibody, production of 1L-4 and IL-S was induced in MLN and spleen cells of BALB/c mice whereas those from CS7/BL6 mice showed little change in cytokine profile. An increase in IL-6 and 1L-10 production was also observed on in virro restimulation of splenocytes from BALB/c mice. However, significant amounts of 1L-2 and IFN-y were still secreted in these cultures. In v i m neutralisation of IFN-y in the restimulated cultures had little effect on the production of other cytokines, regardless of whether this cytokine had also been neutralised in vivo .Despite the change in cytokine profile, anti-IFN-y treatment had no effect on the kinetics of viral clearance in BALB/c mice. A similar situation was seen for C57/BL6 mice. Congenic strains of mice are currently being used to determine whether the effect on cytokine profiles is associated with the MHC haplotype.

Wunner, and Steven L. Spitalnik, Department of Pathology and Laboratory Medicine, University of Pennsylvania and The Wistar Institute, Philadelphia, PA 19104 Rabies virus glycoprotein (RGP) is the only glycoprotein on the viral surface, is the target of viral neutralizing antibodies, and is the viral protein that interacts with the host cell receptor. RGP of the ERA strain is a 505 amino acid Type I membrane glycoprotein with 3 potential N-glycosylation sites at Asn37, Asn247, and Asn319. Due to RGPs central role in the immunology and biology of rabies, it is important to determine its three-dimensional structure. Towards this end we produced a recombinant form of RGP that may be more amenable to analysis by x-ray crystallography. To avoid the use of detergents, we constructed a soluble 434 amino acid form of RGP lacking a transmembrane domain (RGPT434). RGPT434 was secreted by transfected cells and was appropriately glycosylated, assembled, antigenic, and immunogenic. Since N-glycosylation was required for RGPT434 secretion, we minimized the number of N-glycans by site-directed mutagenesis. Interestingly, RGPT434 was secreted only when Asn3l 9 was N-glycosylated. In contrast, full-length RGP was expressed at the cell surface a any of the three potential sites was N-glycosylated. Soluble inhibitors of N-glycosylation were used to simplify the structures of the N-glycans on RGPT434. Although inhibiting alpha-glucosidases with castanospermine blocked secretion, inhibiting any further processing with deoxymannojirimycin had no effect on secretion. Finally, to simplify the purification process, a tail of 6 His residues was added to the COOH-terminus of RGPT434. This modification did not affect secretion, antigenicity, of assembly of the recombinant glycoprotein, but did allow purification using Ni+2-agarose chromatography. In summary, we constructed a soluble form of RGP with a minimal number of homogeneous N-glycans and an "epitope" tag. This form of RGP is structurally similar to the extracellular domain of the full-length protein and may be amenable to analysis by x-ray crystallography. Endogenous retroviral genes (ev) are well characterized in chickens. Our laboratory has established chicken lines with single and multiple ev loci and, compared to ev negative birds, these lines exhibit a delayed and suppressed titer of neutralizing antibody in responses to avian leukosis virus (ALV) challenge. To address the effects of ev genes on the cell mediated immune response we have developed an in-vitro assay to monitor the in-vivo induction of MHC restricted cytotoxic T cells (CTL) induced by ALV. Using this assay, we find the CTL response in birds expressing the ev-21 locus is reduced 85 to 95% compared to ev negative birds having nearly identical (eighth backcross) genetic backgrounds. Other ev loci (1,6,10 and 11 ) also reduce the CTL response to ALV in MHC matched chickens with different genetic backgrounds. ' National Public Health Institute. Helsinki. Finland

Universities of Tampere, Qulu and 'Helsinki, Finland

Coxsackie B virus infections have been associated with clinical IDDM in seveml previous studies but their initiating role in the slowly progressing beta-cell damage was for the fmt time directly implied in our recent prospective. nationwide DiMe study. Siblings of the index IDDM cases who progressed to clinical IDDM experienced enterovim infections more frequently than nonprogressing siblings during prospective follow-up period. Enterovim infections were initially demonstrated by detecting significant increases in serum class-specific cross-reactive enteroviral antibodies by using radioimmunoassay (RIA) with the heavy-chain capture principle. Causal association benveen these infections and the pathogenesis of IDDM was suggested by temporal association of sercconvenions of viral antibodies with increases in islet cell antibody levels which are considered to be markers OK ficell damage. The possibility of the existmce of specific diabetogenic s & s of enternviruses was then considered. To identify the serotypzs responsible for the infections, successive sera from subjects exhibiting seroconvenion of autoantibodies coinciding with an increase in enteroviral antibodies were anlyzed by plaque n e u n a t i o n assay using a set of different enterovhes. The results indicate that several enternvirus serotypes may be associated with the progressive prediabetic p~ocess. Not only different coxsackie B but also coxsackievirus A9 -infections were found to coincide with seroconvenions of islet-cell and insulin auwantibodies. It is known that an immunogenic region of GAD 65. one of the major isletcell antigens, shows a high degree of homology with the 2C protein of CBV-4. Studies on potential immunological cross-reactions between enteroviruses and two significant autoantigens, HSP60 and GAD65, are in progress. This is carried out by using immunohistochemistry in infected cell lines and in frown sections of human pancreas with antibodies induced by synthetic peptides and natural and expressed viral proteins, The role of p53 alterations in human malignant pleural mesotheliomas (MPM) is unclear. Immunostaining (IPOX) of snap frozen MPM specimens revealed p53 expression in 31 of 52 (60%). p53 detection by IPOX is usually associated with mutated p53. SSCP for exons 5-9 revealed p53 alterations in 2 of 25 specimens, and LOH at exon 4 was seen in only 2 of 12 evaluable specimens. Thus, most of these specimens appeared to contain wild-type (wt) p53. We recently reported that SV40 induces mesotheliomas in rodents, and that SV40-like sequences are present and expressed in MPM. Of note, SV40 large T antigen (Tag) binds and inactivates wt p53 i n ! , abolishing its tumor suppressor function. We theorized that the immunohistochemical persistence of p53 may result from its physicdl binding with Tag. Tag was detected by IPOX in 32 (62%) of these MPM specimens. Expression of Tag was significantly associated with coexpression of wt p53 (p2 = 0.008, Fisher's exact text), and preliminary experiments suggest co-precipitation of p53 and Tag. Finally, we demonstrate that WAF-1 is not detectable in MPM expressing wt p53 suggesting that p53 is inactive. We speculate that Tag expression in MPM binds to and inactivates p53 contributing to the malignant phenotype. Human Papilloma Virus (HPV) infection is involved in 90% of cervical carcinomas and possibly 95% of cervical intraepithelial neoplasia (CIN). At present it is unclear whether patients infected with the transforming HPV types can mount efficient T cell responses. To address this issue we examined the CTL response of peripheral blood lymphocytes from patients attending a cervical neoplasia outpatient clinic. All were diagnosed HPV positive with various stages of C M and nine patients were selected for HLA-A2 expression by FACS analysis. PBMC were stimulated with CaSki, a cervical carcinoma cell line demonstrated to be HF' V type 16' and HLA-A2'. Culture media consisted of RPMI-10% human serum supplemented in three cases with 15 U/ml IL-7, in another three cases with 3 U/ml IL-2 and 15 U/ml IL-7, and in the final three cases with 10 U/ml IL-2 and 15 U/ml IL-7. The CTL were screened for reactivity to CaSki and in addition the cervical carcinoma cell line C33A (HPV, HLA-A2') transfected with either the HPV type 16 E6 or E7 genes. One patient's cells maintained with 10 Uiml IL-2 and 15 Ulml IL-7 showed HPV E7 protein specific cytotoxicity in a HLA-A2 restricted fashion. These CTL lysed the CaSki stimulator cell line but not the NK sensitive target K-562. The CTL efFIciently lysed a C33A-E7 clone but in contrast a vector only transfectant was not lysed. The lysis of C33A-E7 was blocked with a-CD3 and a-CD8 antibodies at 66% and 41% respectively. FACS analysis determined that this CTL population was 92.8 % CD3'CDS' and 7.3% CD3'CD4'. Limited E6 recognition was also observed but has not been hlly characterized To our knowledge this report represents the first demonstration of HPV E7 responsive CTL isolated from the peripheral blood of HF' V infected patients. We have previously shown that proviral variants found in the peripheral blood mononuclear cells (PBMC) of HIV-1 infected individuals can interfere with recognition by autologous cytotoxic T-lymphocytes (CTL). Abolition of recognition could be caused by interference with t h e processing of viral peptides, by a failure of the variant peptides to bind to MHC class-I molecules or by a failure of the MHC class-Vpeptide complex to efficiently activate the TCR (T-cell receptor). These mechanisms could allow the virus to escape the immune surveillance by CTL.

Most studies so far have addressed the question of viral variation in T-cell epitopes by analysing proviral DNA, extracted from the PBMC of HIV-1 infected individuals. However, the analysis of virion-associated RNA offers several advantages. Sequences found in viral RNA are likely to be functional, a t least u p to the point of packaging into virus particles. Furthermore, RNA sequences represent virus which is currently actively produced.

Here, we present the sequence variation i n two HLA-B8 restricted epitopes: p17-3 gag and amino acids 18-26 of the pol gene. Both proviral DNA from PBMC and viral RNA sequences from plasma are discussed. The ability of viral variants to hind to MHC class-I molecules was assessed and the recognition of variant peptides by CTL was tested. Variants were also tested for their ability to antagonize recognition of the index sequences. We show that viral variants with the ability to interfere with recognition by CTL are not only found in proviral DNA but also in virion associated RNA. Objective: Clinical course of HIV-1 infection may be influenced by an effective host immune response against HIV-1 and/or by viral properties. To investigate the cellular immune response to HN-1 we analyzed CTL activity against different HIV-1 proteins in long-term asymptomatic HIV-1 infection as well as during rapid progression to AIDS.

Methods: 10 Longterm asymptomatic individuals with CD4 counts >500 celllpl after more than 8 years of infection were selected from The Amsterdam Cohort Study on AIDS versus 10 subjects who progressed to AIDS < 5 years.

CTL activity was measured on "Cr labelled HLA matched or autologous B-LCL, infected with rVV expressing HIV-1 Ag. Both bulk and limiting dilution CTL assays were performed longitudinally with PBMC after Ag-specific stimulation. Sequences of CTL epitopes were determined in homologous virus isolates Resulrs: Different kinetics of anti-Gag CTL responses were observed in rapid progressors. In any case CTL responses disappeared during progression to AIDS. In long-term asymptomatic subjects persistent CTL responses were observed together with low viral load.

Conclusions: Sustained, broad anti-HIV cellular immunity may correlate with maintenance of the asymptomatic state in long-term survival by controlling viral replication. Enteroviruses are a large group of positive stranded RNA viruses known to be responsible for a number of distinct disease entities. Recombination is thought to be capable of generating new enterovirus strains that cause significant morbidity. For example, enterovirus 70 which was responsible for a pandemic of haemorrhagic conjunctivitis and poliomyelitis is thought to have originated by recombination between a coxsackie B like virus and another unidentified enterovirus. We are studying a group of echovirus 11 isolates from an outbreak of disease in southem India. Sequence analysis within the 5' untranslated region reveals that these isolates fall into two groups that differ by -20% (equivalent diversity to that seen between between published sequences of poliovirus 1 and coxsackie A 9 virus).

These two groups of viruses also differ in their cell tropism. Isolates defined as group 1 by their 5'UTR sequence grow equally well on HT29 cells (a human colon carcinoma cell line) and Vero cells. Isolates of group 2, with one exception, grow only on HT29 cells. Analysis of the structural proteins of these isolates revealed differences in migration that correlated with their cellular tropism. Thus, significant genotypic and biological diversity exists amongst these virus isolates. One virus isolate had the 5' untranslated region sequence of a group 1 virus but the protein profile and cellular tropism of a group 2 virus. The best explanation of these findings is that this anornolous isolate is a natural recombinant between the parented strains. Both the ease with which viable recombinants are generated and the diversity present within this one enterovirus serotype increase the potential for the production of novel pathogenic enterovirus strains.

DOMINANT SUSCEPTIBILITY TO POLYOMA TUMORS IN INBRED WILD MICE, Sharon R. Nahill, Yupo Ma, John Carroll and Thomas L. Benjamin, Department of Pathology, Harvard Medical School, Boston, MA 021 15 Polyoma virus (F' y) is a mouse DNA tumor virus which, under appropriate conditions, causes tumors in a wide variety of cell types. Generation of tumors is a function of both the viral and host genomes. Lukacher et al. have recently described a dominant gene, Pyv', carried by the C3-i mouse strain, which confers susceptibility to PY-induced tumors Mapping and immunological analyses indicate that Py4 is the mouse mammary tumor virus 7 superantigen (Mtv 7 sad gene, which deletes T cells required for Py tumor immunosurveillance in H-2' mice. To determine the generality of endogenous superantigens as determinants of susceptibility and to reveal potentially novel mechanisms of susceptibility, we have looked for dominant Susceptibility @ S ) gene(s) id newly established and genetically diverse inbred wild mouse strains, Czech I1 and Pedatteck (Peru). Both strains are susceptible to PY as 100% of infected animals develop a full profile of tumors. Crosses between CS7BR, whose resistance is contributed by the major histocompatibility (MHC) locus, and susceptible Peru or Czech 11, yield F1 progeny which are fully susceptible, indicating a dominant inheritance pattern of susceptibility The incidence of tumor-bearing backcross animals [((Peru x CS7BR) x C57BR) and ((Czech I1 x CS7BR) x C57BR)I suggests that DS is due to at least one, but not more than two genes. Amplification of genomic DNA from the Czech I1 and Peru mice by PCR using primers specific for MTV 7 sag indicates that both strains are negative for proviral MTV 7 sag. Furthermore, the mechanism ofDS in these mice may be independent of all MTV sag as PCR using primers specific for the highly conserved region of MTV sag is unable to amplify MTV DNA from Peru or Czech I1 genomic DNA. These results indicate that, like the C3HiBi, the Pedatteck and Czech I1 contain gene(s) which overide the resistance to PY-induced tumors contributed by the MHC of the C57BR parent and which may cause tumors via a novel, MTV sag-independent mechanism. We have initiated efforts to map the DS in Peru and Czech I1 mice using PCR and primer pairs flanking simple sequence length polymorphisms. FIS-2 is a low leukemogenic, but relatively strong immunosuppressive variant of Friend Murine Leukemia Virus (F-MuLV). This variant was originally isolated from T-helper cells of FLC-infected adult NMRl mice. Compared to F-MuLV, FIS-2 suppresses primary antibody response more efficiently in infected mice. Some of the FTS-2 infected adult NMRI mice developed a disease resembling the acquired immunodeficiency syndrome induced by HIV. Restriction mapping and nucleotide sequence analysis of FIS-2 show a high degee of homology between this variant and the prototype F-MuLV clone 57.

In this study we have attempted to localize the genomic determinant of FIS-2 which is responsible for induction of a strong suppression of primary antibody response. Six chimeric viruses of FIS-2 and F-MuLV were constructed. The primary antibody response of the mice infected with these chimeric viruses were investigated. The results of these experiments will be presented. anti-FMDV antibodies, as measured in an ELISA capture assay, were cross reactive. b) Cellular:

Proliferative (CD4) T cell responses of peripheral blood mononuclear cells (PBMC) were low or undetectable during primary responses to vaccine or virus, and frequently low during secondary responses. For good T cell proliferation in vitro, multiple immunisation is required. This may reflect preferential stimulation of the Th2 CD4

T cell subset.

Interestingly, when CD4 responses were observed, CD8 Tcell responses were also detectable. 2 . RECOGNITION OF INDIVIDUAL VIRAL PROTEINS a) Expression Cloning: Structural and non-structural protein pseudogenes were cloned from cDNA by PCR. expressed in pGEX-3XUC. and purified by SDS-PAGE. b) Humoral: Structural and non-structural proteins were recognised by infected animals. A good anamneetic antinon-structural response was only observed when the boosting serotype differed from the serotype stimulating the primary response. c) Cellular: Both structural and non-structural proteins were recognised and some were cross reactive. Interestingly, VP1 was strain specific, and the polymerase (3D) was the most immunogenic and cross reactive. d) A construct comprising 3D and the immunodominant VP1 epitopes was prepared and tested. In common with other herpesviruses, the envelope glycoproteins of equine herpesvirus 1 (EHV-1; equine abortion virus) are major determinants of the infectious process and pathogenicity, and are inducers of humoral and cell-mediated immune responses. As such, they are candidates for components of subunit vaccines against EHV-1. To generate useful amounts of individual EHV-1 glycoproteins, we have constructed recombinant badoviruses capable of expressing glycoproteins C, D, H (gC, gD, gH ) in insect cells, and have evaluated the recombinant products as innnunogens in a murine model of EHV-1 infection. AU three glycoproteins induced serum (ELISA) antihodies to EHV-1, and EHV-1 gC and gD also induced neutralizing antibody responses. Following intranad challenge with infectious EHV-1, protective immunity, as demonstrated by acelerated clearance of virus fiom respiratory tissues to below detectable levels, was evident in mice immunized with either recombinant gC or gD. In contrast, gH-immmkd mice did not develop detectable neutralizing antibody, and did not clear challenge virus more rapidly than controls. Delayed type hypersensitivity and lymphoproliferation responses to EHV-1 antigen were observed for each of the EHV-1 glycoproteins, and in experiments with gDimmunized mice, a role for cell mediated immunity in protection was confirmed by adoptive transfer and T-cell depletion experiments. The data provide support for the potential of glycoproteins C and D as a subunit vaccine against EHV-1.

Molecular Pathogenesis of Ural Infeetiom 52-331 ENTEROVIRUS-IMMUNE CELL INTERACTIONS: IMPLICATIONS IN ENTEROVIRUS-INDUCED DISEASES

We have also evaluated the effect of virus infection on the humoral immune response to CVB3, infection in adolescent C3H/HeSnJ mice. Antigen presenting cell, 1-helper cell and 8-cell function were evaluated utllizlng a sheep red blood cell (sRBC) plaque assay. Mice were injected intrapentoneally (ip) with lo5 plaque forming units of CVB3, at day 0 and with lo7 sRBC's at days 0, 2, 3 and 4 post-CVBB, Infection. Splenocytes were harvested 4 days post-sRBC injection, mixed with target sRBC's and guinea pig complement and incubated. Plaques were then quantitated. Results: CVB3, was associated with 12.9% to 17.4% of CD-8 positive T-cells and w a 11 % to 26% of adherent splenocytes. After mitogen (LPS and Con A) stimulation, B-cells and adherent cells were demonstrated to be permissive for viral replication. A 248% and 738%

Under non-stimulated conditions. an average of 1 % of virus is cell-associated (plaque

NORTH AMERICA. Bruce Anderson, Teny Yates, Norah Torrez-Martinez, Wanmin Song, Brian Hjelle. University of New Mexico, Albuquerque, N.M.We recently identified a new species of hantavirus (HMV) associated with the harvest mouse Reithrodontomys megalotis (Hjelle B et al, J. ViroJ. 1994, in press ). An Arizona woodrat (Neotoma mexicana) was found to he infected with HMV, presumably through "spillover". HMV is most closely related to the Four Comers hantavirus (FCV) of deer mice (genus Peromyscus). The nucleocapsid gene and protein of HMV differ from those of FCV by 24% and 15% of residues, and the 1896 nt S genome is shorter by 163 nt. We surveyed 174 Reithrodontomys animals captured in the U.S. and Mexico for hantavirus antibodies; 27 (15.6%) were positive. S segment cDNAs were amplified and sequenced from seropositive animals captured in California (4), Arizona (3), New Mexico (l), and Mexico (2). A monophyletic clade of HMV-like agents was identified at all sites, although an R. megalotis infected with an FCV-like virus was also identified in the state of Zacatecas, Mexico. Nucleotide sequence distances among members of the HMV clade were up to 15.5%. but amino acid distances were less than 2%. HMV is enzootic in harvest mice throughout much of North America, and can also infect wood rats. HTLV I-associated myelopathy/tropical spastic paraparesis (HAMnSP) is a slowly pro ressive neurological disease characterized by perivascuaar mononuclear infiltrates in the CNS. HTLV I-specific CD8+ CTL are found in PBL and CSF of infected patients with HTLV I-associated neurological disease but not in HTLV I seropositive individuals without neurological involvement. Previous studies have shown that in HLA-A2+ patients, HTLV I-specific CD8+ CTL restricted by HLA-A2 recognize a peptide derived from the HTLV I Tax protein (Tax 11-19 LLFGYPVW). In the present study, we have analyzed the potential of these Tax-specific CTL to recognize addtional peptides. Our results demonstrate that a subpopulation of high affinity CD8' Tax 11-19 specific CTL clones cross-react on a self peptide derived from the se uence of myelin-associated glycoprotein (MAG 556-564 VL&SDFRI) presented by HLA-A2. These ObSeNatlOnS suggest that the demyelination process in HAMlTSP may be,due, in part, to virus-specific CTL recognition of a self myelin component that is independent of HTLV I infection. Development of pathology varies widely between different strains of mice after intracerebral inoculation with the so-called 'docile' isolate of Lymphocytic Choriomeningitis (LCM) virus. The C3HeBFeJ and BlO.BR/SgSnJ mouse strains have been of special interest because they display autoimmune hemolytic anaemia with varying degrees of apparent immunological involvement. In this study, we examined the role of CD4+ T helper cells in this autoimmune response by treating mice with the CW-specific GK1.5 monoclonal antibody. We also determined if polyclonal activation of B lymphocytes, induced either by LCM virus or by lactate dehydrogenase-elevating virus, another well known B cell activator, correlated with the development of anaemia in these mice. Our results strengthened the central role of the immune system in the anaemia in C3H mice by showing that depletion of CD4+ cells largely, if not completely, abrogated this anti-erythrocyte autoimmune reaction. As reported by others, we found that the anaemia was more mild in B 1O.BR mice than in C3H mice. However, we could not confirm the difference in the degree of B lymphocyte polyclonal activation between these mice. Furthermore, lactate dehydrogenase-elevating virus had no apparent effect on erythrocytes, even though this virus also induced a sharp increase in plasma IgG levels. One of the two class I MHC (H-PKd)-restricted immunogenic sites identified on the influenza strain AIJapanl57 (H2N2) hemagglutinin (HA) encompasses two distinct partially overlapping epitopes, mapping to residues 204-212 and 210.219. When we investigated the magnitude of the CTL responses of BALWc mice to the two overlapping epitopes, we found that while the NHrterminal nonamer epitope is immunodominant, eliciting vigorous CTL responses in NJapanl57-immunized BALB/c mice, the CTL responses to the COOH-terminal decamer epitope are weak and variable. The C-terminal epitope subdominance seems to be due to factors other than inefficient processing of the epitope in vivo because CTLs generated by priming mice with recombinant Sindbis viruses expressing only one of the HA 204-219 subsites displayed patterns of responsiveness similar to that of influenza virus primed CTLs. Limiting dilution CTL assays showed that the CTL precursor frequency (pCTL) of the Nterminal epitope is at least ten fold higher than the pCTL of the Cterminal epitope, implying that the low and variable pattern of Cterminal specific responsiveness was due to the limited T cell precursors in the C-terminal specific CTL repertoire. This was further confirmed by the limited heterogeneity in the cross reactivity patterns displayed by the C-terminal specific CTL for an Ig VH fragment and the HA 210-219 epitope of influenza strain A I M 5 7 in short term bulk cultures, and the FACS analysis of TCR Vg chain usage. Taking these together with our previous observation that some JHA 210-219 specific CTLs can also crossrecognize an Ig VH fragment. these studies had provided a strong evidence that Ig gene products may influence T lymphocyte function and repertoire development. We have previously described the identification of homologous regions in the C-terminus of HIV-1 gp41 and in the N-terminus of HLA class I1 beta chains. Forty percent of patients infected with HIV-I virus were shown to have antibodies which bind to the homologous sequences, as well as to native HLA class I1 molecules. Affinity purified crossreactive antibodies (CrAb) were shown to have direct blocking effects on normal T cell responses to recall antigens, and could mediate ADCC of HLA class II+ cell lines.In order to determine the contribution of such antibodies to disease progression, we obtained longitudinal plasma samples from patients in the MACS study. In a first study, it was found that the presence of high titers CrAbs correlated with a more rapid disease progression (P = 0.027 by Fisher two tail analysis)In a second, 7 year-longitudinal study of 12 progre.ssors and 12 stable patients we found: (1) The production of CrAb was seen in 70 -80% of rapid progresson, while the true stables produce only infrequent low-titers CrAb. (2) In rapid pmgressors, production of CrAb preceded by 2-3 years the marked drop in CD4 counts. (3) CrAb production did not correlate with the degree of hyperglobulinemia in these patients. (4) The presence of CrAb during the asymptomatic stage correlated with early loss of T-helper responses to recall antigens.We are currently establishing whether periodic measurements of CrAb in patients sera could be valuable in predicting a drop in CD4 counts and disease progression. The lymphokine IFN-y is I pleiotropic insnunomodulator and possesses intrinsic antiviral activity. We studied its significance in the development of antiviral immune responses using IFN-7 receptor deficient (IFN-yR-'.) mice. After inoculation with live attenuated pseudorabies virus (PRV) the mutant mice showed no infectivity titers in various tissues and transient viral Ag expression only in the spleen similar as in wild-type mice. However, the absence of the IFN-yR resulted in increased proliferative splenocyte responses. The PRV-immune animals showed a normal IFN-1 and 11-2 production, without detectable 11-4, and with decreased 11-10 secretion in response to viral Ag or con A. Immunohistochemically, an increased ratio of IFNy/I1-4 producing spleen cells was found. After immunization with either live attenuated or inactivated PRV, IFN-yR"' mice produced significantly less antiviral antibody (Ab), and more succumbed to challenge infection than the intact control animals. The reduction in Ah titers in the mutant mice correlated with lower protection by their sera in transfer experiments. Thes? findings are in line with the strong enhancing effect of exogenous IFN-y on rabies virusand PRV-specific IgG responses. Our data demonstrate that a physiological IFN-y system is surprisingly not critical for the generation of antiviral Th-I-type and the suppression of Th-2-type cytokine responses. The lymphokine, however, is an important mediator in the generation of protective antiviral Ab.