Immunobiology of cytotoxic T-cell escape mutants of lymphocytic choriomeningitis virus

Dynamics of cytotoxic T-lymphocyte exhaustion

We examine simple mathematical models to investigate the circumstances under which the dynamics of cytotoxic T-lymphocyte (CTL) activation and differentiation may result in the loss of virus specific CD8+ cells, a process known as CTL exhaustion. We distinguish between two general classes of viruses: (i) viruses infecting cells that are not involved in the immune response; and (ii) viruses infecting antigen presenting cells (APCs) and helper cells. The models specify host and viral properties that lead to CTL exhaustion and indicate that this phenomenon is only likely to be observed with viruses infecting APCs and helper cells. Moreover, it is found that for such viruses, a high rate of replication and a low degree of cytopathogenicity promote the exhaustion of the CTL response. In addition, a high initial virus load and a low CD4+ cell count promote the occurrence of CTL exhaustion. These conclusions are discussed with reference to empirical data on lymphocytic choriomeningitis virus and on human immunodeficiency virus.

A non-retroviral RNA virus persists in DNA form

Infection of adult mice with lymphocytic choriomeningitis virus (LCMV), a non-cytopathic segmented RNA virus, leads initially to generalized infection, followed by clearance and subsequent life-long immunity. Indirect evidence has suggested that viral antigens may persist in lymphoid tissues during the phase of immunological memory, but viral genomic RNA has not been detected in previous studies. During a search for persistent virus in the spleen, we identified LCMV-specific sequences present as DNA by polymerase chain reaction (PCR) in mice over 200 days after infection. In vivo and in vitro studies revealed that reverse transcription of viral RNA into complementary DNA occurred after acute infection of cells of its natural hosts, mouse and hamster, but not of other species and could be inhibited in vitro by azidothymidine (AZT), indicating that this was mediated by endogenous reverse transcriptase activity. These findings reveal a surprising and new pathway of interaction between exogenous RNA viruses and endogenous retroviral, and perhaps other host components, that results in the persistence of virally determined DNA. We speculate that the latter may function in vivo as a form of DNA vaccine.

Co-evolution of human immunodeficiency virus and cytotoxic T-lymphocyte responses

After more than a decade of intensive research, the precise role of human immunodeficiency virus (HIV)-specific cytotoxic T lymphocytes (CTL) in determining the course of the infection remains open to argument. It is established that HIV-specific CTL appear early in the infection and are temporally associated with the clearance of culturable virus from the blood; that CTL are generally detectable at very high levels throughout the asymptomatic phase and decline at the time of progression to AIDS; and that CTL-mediated killing is sufficiently fast to prevent production of new virions by HIV-infected cells. However, viral turnover is high throughout the course of the infection, and infected individuals progress inexorably to disease in spite of the CTL response. In order to address the question of whether CTL play an active part in influencing the course of HIV infection, one approach has been to seek evidence for CTL-mediated selection pressure on the virus. Several clear examples of CTL epitope-specific mutations selected to fixation are described. We argue that CTL escape is a common event which occurs at all stages of the infection. Detailed longitudinal studies are required to detect CTL escape and to understand the complexities contributed by factors such as a polyvalent CTL response and the presence of epitope variants which antagonise the CTL response. In conclusion, there is strong evidence of a dynamic process in which CTL impose important selection constraints upon HIV from which the virus attempts to escape; ultimately, at the time of disease progression, the tenuous control of CTL over the virus is lost.

Cytotoxic T-cell-resistant variants arise at early times after infection in C57BL/6 but not in SCID mice infected with a neurotropic coronavirus

Under certain conditions, C57BL/6 mice persistently infected with mouse hepatitis virus strain JHM (MHV-JHM) develop clinical disease and histological evidence of demyelination several weeks after inoculation with virus. In a previous report, we showed that mutations in the RNA encoding an immunodominant CD8 T-cell epitope within the surface glycoprotein (epitope S-510-518) were present in all persistently infected animals and that these mutations abrogated recognition by virus-specific cytotoxic T cells (CTLs) in direct ex vivo cytotoxicity assays. To obtain further evidence that these mutations were necessary for the development of clinical disease, the temporal course of their appearance was determined. Mutations in the epitope were identified by 10 to 12 days after inoculation, and in some mice, virus containing mutated epitope was the dominant species detected by 15 days. In addition, most mice that remain asymptomatic at 80 days after inoculation, a time after which clinical disease almost never develops, were infected with only wild-type virus. Finally, analysis of virus isolated from mice with severe combined immunodeficiency (SCID) revealed the presence only of wild-type epitope S-510-518. These results, by showing that mutations are not selected in SCID mice and occur at early times after inoculation in C57BL/6 mice, support the view that they result from immune pressure and contribute to virus persistence and demyelination in mice infected persistently with MHV-JHM.

What can we learn about human immunodeficiency virus infection from a study of lymphocytic choriomeningitis virus?

The role of cytotoxic T lymphocytes (CTL) in human immunodeficiency virus (HIV) infection remains elusive. Since the discovery 10 years ago of high levels of specific CTL in this disease, some have argued that they play an important role in virus control, others that they drive disease progression through destruction of T helper cells, and others still that they play no obvious role at all. By contrast, the central role of CTL in murine lymphocytic choriomeningitis virus (LCMV) infection has been very clearly worked out through the use of in vivo depletion and adoptive transfer experiments, as well as knockout and transgenic mice. To interpret the possible roles for CTL in HIV, we have therefore made a comparison between what is known about CTL and their interaction with virus-infected cells in these two infections. This illustrates a potential critical role for these cells in both control of HIV replication and immune-mediated pathology, but one that is highly dependent on virus dose, distribution and dynamics.

Virus load and antigenic diversity

In this paper, we analyse mathematical models for the interaction between virus replication and immune responses. We show that the immune system can provide selection pressure for or against viral diversity. The paper provides new insights into the relationship between virus load (=the abundance of virus in an infected individual) and antigenic diversity. Antigenic variation can increase virus load during infections, but the correlation between load and diversity in comparisons among different infected individuals can be positive or negative, depending on whether individuals differ in their cross-reactive or strain-specific immune responses. We derive two models: our first model applies to any replicating parasite that can escape from immune responses; our second model includes immune function impairment, and specifically describes infections with the human immunodeficiency virus (HIV).

How viruses escape from cytotoxic T lymphocytes: molecular parameters and players

Viruses that persist in infected hosts must evolve successful strategies to avoid recognition by the immune system. The primary player in antiviral immune surveillance is the CD8+ cytotoxic T lymphocyte (CTL), and the battle drawn between the CTLs and viruses is the focus of this review. In this struggle, viruses can follow multiple distinct pathways. For example, DNA viruses often adopt the strategy of encoding proteins that interfere with the immune response along routes of antigen presentation. Such interference prevents the viral peptide from binding to the major histocompatibility complex (MHC) class I glycoprotein; therefore, no virus-MHC complex forms for recognition by antiviral CTLs. RNA viruses, having fewer genes, generate swarms of quasispecies that can contain mutated viral proteins. When such mutants occur in viral peptides presented to the MHC protein or the residue recognized by the CTL receptor, CTL recognition and activation fail. If, instead, the mutation occurs in the viral peptide flanking sequence, the infected cell may not process the viral peptide from the cytosol to the endoplasmic reticulum. Viruses can also directly or indirectly attack dendritic cells and CD4+ or CD8+ T lymphocytes, other routes that interfere with immune functions. Dendritic cells are the primary professional antigen-presenting cells and are critical for the activation of CTL responses. CD4+ T lymphocytes provide help for long-term CD8+ CTL activity and are necessary for its maintenance. Consequently, interference with either dendritic or CD4+ cell types constitutes yet another way that viruses can disable the immune response and persistently infect their host.

The signal sequence of lymphocytic choriomeningitis virus contains an immunodominant cytotoxic T cell epitope that is restricted by both H-2D(b) and H-2K(b) molecules

Infection of H-2b mice with lymphocytic choriomeningitis virus (LCMV) generates three well-characterized H-2D(b)-restricted immunodominant epitopes delineated in the NP, GP1, and GP2 proteins. Here we report that the H-2D(b)-restricted GP1 epitope GP33-41/43 (KAVYNFATC/GI) located in the signal sequence of LCMV is also the immunodominant epitope recognized by CTL at the surface of the same infected cells in the context of H-2K(b) restriction. The GP1 epitope bound to H-2D(b) and H-2K(b) molecules with comparable affinities. The respective binding processes involved different sets of peptide anchoring residues and required dramatically different conformations of the peptide backbone as well as rearrangement of residue side chains. The 10-mer peptide GP34-43 (AVYNFATCGI) was the optimal H-2K(b)-binding sequence and the 8-mer peptide GP34-41 (AVYNFATC) the minimal sequence for optimal H-2K(b)-restricted CTL recognition. Comparison of lytic activities of primary splenic anti-LCMV CTL from C57BL/6 (D(b+)/K(b+)), B10A.[5R] (D(b-)/K(b+)), and B10A.[2R] (D(b+)/K(b-)) mice against LCMV-infected or peptide-coated target cells expressing either one or the two MHC alleles revealed that the H-2K(b)-restricted component of the anti-GP1 CTL response was mounted independently of but as efficiently as its H-2D(b) counterpart. Analysis of the immune response against a GP1 variant that escapes CTL recognition showed that the GP1 epitope: (i) was likely the only immunodominant LCMV epitope in the context of H-2K(b), and (ii) could efficiently evade H-2D(b) and H-2K(b)-restricted CTL mediated lysis.

Evolution of mouse hepatitis virus: detection and characterization of spike deletion variants during persistent infection

High-frequency RNA recombination has been proposed as an important mechanism for generating viral deletion variants of murine coronavirus. Indeed, a number of variants with deletions in the spike glycoprotein have been isolated from persistently infected animals. However, the significance of generating and potentially accumulating deletion variants in the persisting viral RNA population is unclear. To study this issue, we evaluated the evolution of spike variants by examining the population of spike RNA sequences detected in the brains and spinal cords of mice inoculated with coronavirus and sacrificed at 4, 42, or 100 days postinoculation. We focused on the S1 hypervariable region since previous investigators had shown that this region is subject to recombination and deletion. RNA isolated from the brains or spinal cords of infected mice was rescued by reverse transcription-PCR, and the amplified products were cloned and used in differential colony hybridizations to identify individual isolates with deletions. We found that 11 of 20 persistently infected mice harbored spike deletion variants (SDVs), indicating that deletions are common but not required for persistent infection. To determine if a specific type of SDV accumulated during persistence, we sequenced 106 of the deletion isolates. We identified 23 distinct patterns of SDVs, including 5 double-deletion variants. Furthermore, we found that each mouse harbored distinct variants in its central nervous system (CNS), suggesting that SDVs are generated during viral replication in the CNS. Interestingly, mice with the most severe and persisting neurological disease harbored the most prevalent and diverse quasispecies of SDVs. Overall, these findings illustrate the complexity of the population of persisting viral RNAs which may contribute to chronic disease.

Generation of cytotoxic T lymphocytes against immunorecessive epitopes after multiple immunizations with adenovirus vectors is dependent on haplotype

Currently, adenovirus (Ad) is being considered as a vector for the treatment of cystic fibrosis as well as other diseases. However, the cytotoxic T lymphocyte (CTL) response to Ad could limit the effectiveness of such approaches. Since the CTL response to virus infection is often focused on one or a few immunodominant epitopes, one approach to circumvent this response is to create vectors that lack these immunodominant epitopes. The effectiveness of this approach was tested by immunizing mice with human group C adenoviruses. Three mouse strains (C57BL/10SnJ [H-2b], C3HeB/FeJ [H-2k], and BALB/cByJ [H-2d]) were immunized with wild-type Ad or Ad vectors lacking the immunodominant antigen(s), and the CTL responses were measured. In C57BL/10 (B10) mice, a single inoculation intraperitoneally (i.p.) led to the recognition of an immunodominant antigen in E1A. When B10 mice were inoculated multiple times either i.p. or intranasally with wild-type Ad or an Ad vector lacking most of the E1 region, subdominant epitopes outside this region were recognized. In contrast, C3H mice inoculated with wild-type Ad recognized an epitope mapping within E1B. When inoculated twice with Ad vectors lacking both E1A and E1B, no immunorecessive epitopes were recognized. The immune response to Ad in BALB/c mice was more complex. CTLs from BALB/c mice inoculated i.p. with wild-type Ad recognized E1B in the context of the major histocompatibility complex (MHC) class I Dd allele and a region outside E1 associated with the Kd allele. When BALB/c mice were inoculated with E1-deleted Ad vectors, only the immunodominant Kd-restricted epitope was recognized, and Dd-restricted CTLs did not develop. This report indicates that the emergence of CTLs against immunorecessive epitopes following multiple administrations of Ad vectors lacking immunodominant antigens is dependent on haplotype and could present an obstacle to gene therapy in an MHC-diverse human population.

Antiviral pressure exerted by HIV-1-specific cytotoxic T lymphocytes (CTLs) during primary infection demonstrated by rapid selection of CTL escape virus

The HIV-1-specific cytotoxic T lymphocyte (CTL) response is temporally associated with the decline in viremia during primary HIV-1 infection, but definitive evidence that it is of importance in virus containment has been lacking. Here we show that in a patient whose early CTL response was focused on a highly immunodominant epitope in gp 160, there was rapid elimination of the transmitted virus strain and selection for a virus population bearing amino acid changes at a single residue within this epitope, which conferred escape from recognition by epitope-specific CTL. The magnitude (> 100-fold), kinetics (30-72 days from onset of symptoms) and genetic pathways of virus escape from CTL pressure were comparable to virus escape from antiretroviral therapy, indicating the biological significance of the CTL response in vivo. One aim of HIV-1 vaccines should thus be to elicit strong CTL responses against multiple codominant viral epitopes.

Overlapping epitopes in human immunodeficiency virus type 1 gp120 presented by HLA A, B, and C molecules: effects of viral variation on cytotoxic T-lymphocyte recognition

Human immunodeficiency virus (HIV)-specific cytotoxic T lymphocytes (CTL) are thought to exert immunologic selection pressure in infected persons, yet few data regarding the effects of this constraint on viral sequence variation in vivo, particularly in the highly variable Env protein, are available. In this study, CD8+ HIV type 1 (HIV-1) envelope-specific CTL clones specific for gp120 were isolated from peripheral blood mononuclear cells of four HIV-infected individuals, all of which recognized the same 25-amino-acid (aa) peptide (aa 371 to 395), which is partially contained in the CD4-binding domain of HIV-1 gp120. Fine mapping studies revealed that two of the clones optimally recognized the 9-aa sequence 375 to 383 (SFNCGGEFF), while the two other clones optimally recognized the epitope contained in the overlapping 9-aa sequence 376 to 384 (FNCGGEFFY). Lysis of target cells by the two clones recognizing aa 375 to 383 was restricted by HLA B15 and Cw4, respectively, whereas both clones recognizing aa 376 to 384 were restricted by HLA A29. Sequence variation, relative to the IIIB strain sequence used to identify CTL clones, was observed in autologous viruses in the epitope-containing region in all four subjects. However, poorly recognized autologous sequence variants were predominantly seen for the A29-restricted clones, whereas the clones specific for SFNCGGEFF continued to recognize the predominant autologous sequences. These results suggest that the HLA profile of an individual may not only be important in determining the specificity of CTL recognition but may also affect the ability to recognize virus variants and suppress escape from CTL recognition. These results also identify overlapping viral CTL epitopes which can be presented by HLA A, B, and C molecules.

Escape of human immunodeficiency virus from immune control

Cytotoxic T lymphocytes (CTL) play a crucial role in the attempt to control infection with human immunodeficiency virus (HIV). Variation in epitopes recognized by CTL is common and frequently offers potential escape routes for mutant virus. Proof of escape, however, requires demonstration of increased frequency of virus particles or provirus that carry the escape sequence. There are now several recorded examples of virus variants that escape from CTL and are then selected. Most dramatic are those in which the CTL response has been dominated by CTL recognizing a single epitope that has suddenly changed, resulting in escape to fixation. This has been seen both early and late in the infection, leaving no doubt that escape occurs. Such escape is likely to be favored when the antiviral CTL response is oligoclonal and focused on a small number of immunodominant epitopes. The heterogeneous CTL response seen in many HIV-infected patients may result from successive waves of virus escape followed by new CTL responses specific for subdominant epitopes. Mutant virus can escape by several different routes, including failure of the mutated peptide to bind to the presenting HLA molecule and altered interactions with T cell receptors (TCR), including antagonism.

Cytotoxic T cell-resistant variants are selected in a virus-induced demyelinating disease

C57BI/6 mice infected with mouse hepatitis virus, strain JHM (MHV-JHM) develop a chronic demyelinating encephalomyelitis. Infectious virus can be isolated only from symptomatic mice. In C57BI/6 mice, two CD8+ T cell epitopes within the MHV-JHM surface glycoprotein were previously identified. Here, we show that mutations in the RNA encoding the immunodominant of the epitopes are present in nearly all virus samples isolated from these mice. Mutations are not present in sequences flanking this epitope or in other CD8+ or CD4+ T cell epitopes. Furthermore, analysis of five peptides corresponding to variant epitopes in direct ex vivo cytotoxicity assays showed that each mutation caused a loss of epitope recognition. These results suggest that escape from CD8+ T cell recognition is necessary for enhanced virus replication and development of clinical disease in these MHV-JHM-infected mice.

Binding of viral antigens to major histocompatibility complex class I H-2Db molecules is controlled by dominant negative elements at peptide non-anchor residues. Implications for peptide selection and presentation

Binding of viral antigens to major histocompatibility complex (MHC) class I molecules is a critical step in the activation process of CD8(+) cytotoxic T lymphocytes. In this study, we investigated the impact of structural factors at non-anchor residues in peptide-MHC interaction using the model of lymphocytic choriomeningitis virus (LCMV) infection of its natural host, the mouse. Altering viral genes by making reassortants, recombinants, and using synthetic peptides, CD8(+) cytotoxic T lymphocytes were shown to recognize only three H-2Db-restricted epitopes, GP amino acids 33-41/43, GP 276-286, and NP 396-404. However, LCMV NP and GP proteins contain 31 other peptides bearing the H-2Db motif. These 34 LCMV peptides and 11 other known H2-Db-restricted peptides were synthesized and examined for MHC binding properties. Despite the presence of the H-2Db binding motif, the majority of LCMV peptides showed weak or no affinity for H-2Db. We observed that dominant negative structural elements located at non-anchor positions played a crucial role in peptide-MHC interaction. By comparative sequence analysis of strong versus non-binders and using molecular modeling, we delineated these negative elements and evaluated their impact on peptide-MHC interaction. Our findings were validated by showing that a single mutation of a favorable non-anchor residue in the sequence of known viral epitopes for a negative element resulted in dramatic reduction of antigen presentation properties, while conversely, substitution of one negative for a positive element in the sequence of a non-binder conferred to the peptide an ability to now bind to MHC molecules.

Specificity of the H-2 L(d)-restricted cytotoxic T-lymphocyte response to the mouse hepatitis virus nucleocapsid protein

Cytotoxic T lymphocytes provide protection against persistent infection of the central nervous system by the JHM strain of mouse hepatitis virus. In BALB/c (H-2d) mice, the dominant response is directed against an Ld-restricted peptide in the nucleocapsid protein (APTAGAFFF). Characterization of the fine specificity of this response revealed that the predicted anchor residues at positions 2 and 9 were the most critical for class I binding. Amino acids at positions 7 and 8 were identified as T-cell receptor contact residues. Virus-induced cytotoxic T lymphocytes to other Ld motif-containing nucleocapsid peptides were not detected, despite the identification of two epitopes with reduced Ld affinity. These data suggest that mutations within four residues of the dominant epitope could contribute to the persistence of the JHM strain of mouse hepatitis virus.

Strong cytotoxic T cell and weak neutralizing antibody responses in a subset of persons with stable nonprogressing HIV type 1 infection

Some individuals in well-defined cohorts have now been infected with HIV-1 for well over a decade and yet remain clinically asymptomatic with normal CD4 counts. To determine immunologic and virologic parameters in these individuals, we examined 10 persons from the San Francisco City Clinic with firmly documented infection of 11-15 years duration who had maintained stable CD4 counts above 500 cells/microliters. Our results indicate that long-term nonprogressors are a heterogeneous group with respect to viral load and HIV-1-specific immune responses, and that progression can occur even after 15 years of stable infection. However, in a subset of persons with the lowest viral loads and persistent nonprogressive infection, we detected strong CTL responses, whereas neutralizing antibody studies revealed weak to undetectable titers against a panel of 10 primary isolates. This study demonstrates that a vigorous in vivo activated HIV-1-specific CTL response can be part of the host immune response in stable nonprogressive HIV-1 infection, and that circulating activated CTL can be detected in the setting of an extremely low viral load. These results also indicate that long-term nonprogressing HIV-1 infection does not require the presence of broadly cross-reactive neutralizing antibodies.

Population dynamics of immune responses to persistent viruses

Mathematical models, which are based on a firm understanding of biological interactions, can provide nonintuitive insights into the dynamics of host responses to infectious agents and can suggest new avenues for experimentation. Here, a simple mathematical approach is developed to explore the relation between antiviral immune responses, virus load, and virus diversity. The model results are compared to data on cytotoxic T cell responses and viral diversity in infections with the human T cell leukemia virus (HTLV-1) and the human immunodeficiency virus (HIV-1).

Immunology taught by viruses

The survival of viruses depends on the survival of susceptible hosts. The vertebrate immune system and viruses have therefore coevolved complementary facets. Evidence from various balanced virus-host relationships illustrates that immunological specificity and memory may best be defined biologically and that the mature immune system does not discriminate between "self" and "nonself." Rather, B cells distinguish antigen patterns, whereas T cell responses depend on localization, transport, and kinetics of antigen within lymphatic organs.

CD8+ T-cell epitopes within the surface glycoprotein of a neurotropic coronavirus and correlation with pathogenicity

CD8+ T cells with cytotoxic activity against the surface glycoprotein (S) of mouse hepatitis virus, strain JHM, have been identified in the central nervous system (CNS) of both acutely and chronically infected C57BL/6 mice. In this report, two specific epitopes recognized by these CNS-derived cells were identified, using a panel of peptides chosen because they conformed to the allele-specific binding motif for MHC class I H-2Kb and H-2Db. The active peptides encompassed residues 510 to 518 (CSLWNGPHL, H-2Db) and 598 to 605 (RCQIFANI, H-2Kb). Both epitopes are located within the region of the S protein previously shown to be prone to deletion after passage in animals. These deleted strains are generally less neurovirulent than the wild-type virus but still are able to cause demyelination. Since C57BL/6 mice become persistently infected more commonly than many other strains of mice, these data are consistent with a role for CD8+ T-cell escape mutants in the pathogenesis of the demyelinating disease. This is the first report of CD8+ T-cell epitope localization within the S protein, the protein most strongly implicated thus far in pathogenesis in the host.

Emergence of virus escape mutants after immunization with epitope vaccine

BALB/c and C57BL/6J mice were immunized with recombinant vaccines consisting of lymphocytic choriomeningitis virus CD8+ T-lymphocyte epitopes and a carrier protein. During challenge infection with WE strain lymphocytic choriomeningitis virus, mutants with alterations in distinct amino acid residues of the epitopic nonapeptides appeared and multiplied. Splenocytes from WE-infected BALB/c mice lysed cells coated with the WE-type epitope; lysis was considerably less effective when the epitopic nonapeptide with which the syngeneic cells had been sensitized was the mutated form. Neither target was lysed by splenocytes from BALB/c mice infected with the variant virus. Mutants were not detected in F1 hybrid mice immunized with two viral epitopes that were restricted by class I molecules of both parents.

The effects of natural altered peptide ligands on the whole blood cytotoxic T lymphocyte response to human immunodeficiency virus

Cytotoxic T lymphocytes (CTL) directed against human immunodeficiency virus (HIV)-1 are detectable in the majority of infected individuals, and their early appearance as the initial viremia is suppressed is thought to represent a potent antiviral response. Variation which arises in CTL epitopes can affect recognition by CTL, and we have observed previously that variant epitopes in HIV-1 gag which arise in HIV-1-seropositive donors may act as T cell receptor (TCR) antagonists of their own CTL (Klenerman et al., Nature 1994, 369: 403). The most important question arising from these observations is the extent of these immune escape mechanisms in vivo. Here we show that fresh, uncultured lymphocytes taken directly from HIV-1-infected patients are susceptible to TCR antagonism by variants present within their own virus. In contrast to HLA Class II-restricted T cell responses, where anergy may be induced, we find that in vitro, natural variants may stimulate and sustain growth of CTL. These CTL lines retain lytic specificity exclusively for the original peptide. If this represents events in vivo, natural HIV altered peptide ligands (APL) have the capacity to inhibit the range of CTL directed against an epitope, not simply those clones selected in vitro. Partial activation of CTL by APL could also act to drive an ineffectual CTL response incapable of lysing infected cells bearing these natural antigenic variants. Distortion of lymphocyte populations and function by APL might represent a further mechanism of immune evasion by HIV.

Antigenic oscillations and shifting immunodominance in HIV-1 infections

A typical protein antigen contains several epitopes that can be recognized by cytotoxic T lymphocytes (CTL), but in a characteristic antiviral immune response in vivo, CTL recognize only a small number of these potential epitopes, sometimes only one, this phenomenon is known as immunodominance. Antigenic variation within CTL epitopes has been demonstrated for the human immunodeficiency virus HIV-1 (ref. 11) and other viruses and such 'antigenic escape' may be responsible for viral persistence. Here we develop a new mathematical model that deals with the interaction between CTL and multiple epitopes of a genetically variable pathogen, and show that the nonlinear competition among CTL responses against different epitopes can explain immunodominance. This model suggests that an antigenically homogeneous pathogen population tends to induce a dominant response against a single epitope, whereas a heterogeneous pathogen population can stimulate complicated fluctuating responses against multiple epitopes. Antigenic variation in the immunodominant epitope can shift responses to weaker epitopes and thereby reduce immunological control of the pathogen population. These ideas are consistent with detailed longitudinal studies of CTL responses in HIV-1 infected patients. For vaccine design, the model suggests that the major response should be directed against conserved epitopes even if they are subdominant.