Portal de Periódicos da CAPES

Effects of MHC Class I on HIV/SIV Disease in Primates

2002; Lippincott Williams & Wilkins; Volume: 16; Linguagem: Inglês

10.1097/00002030-200216004-00015

1473-5571

Mary Carrington, Ronald E. Bontrop,

T-cell and B-cell Immunology

Introduction The major histocompatibility complex (MHC) in primates contains the extraordinarily polymorphic class I and class II loci [1], known as HLA in humans (human leukocyte antigen), MhcPatr in chimpanzees, and MhcMamu in rhesus monkeys. Class I gene products are fundamental to acquired immune responses. The classical class I loci, HLA-A, -B, and -C in humans encode molecules that bind antigenic epitopes usually derived from intracellular pathogens and present them to CD8+ T cells, thereby initiating a cytotoxic T cell response. Orthologues of these loci, displaying a high degree of polymorphism, have been found in chimpanzees [2,3] and rhesus macaques [4]. The classical class II loci, HLA-DR, -DQ, and -DP, specify molecules that primarily bind peptides of extracellular origin and present them to CD4+ T cells, resulting in cytokine production and T cell help in antibody production. Again, evolutionary equivalents of these loci are present in most nonhuman primate species, and some lineages encoded by these loci appear to be stable over long evolutionary time spans [5]. The extensive allelic polymorphism in class I/II molecules is concentrated primarily among amino acid positions that determine specificity for foreign peptide (for a listing of HLA alleles and their sequences, see www.ebi.ac.uk/imgt/hla/allele.html) [6,7]. Diversity at these peptide binding residues is believed to be maintained through natural selection by infectious disease morbidity and mortality [8,9], and ensures that a given species is capable of resisting a wide variety of pathogenic organisms. A number of models have been proposed to explain maintenance of HLA polymorphism, each of which has likely been operative under different historic circumstances: balancing selection, where alleles conferring resistance to one disease confer susceptibility to another; heterozygote advantage (overdominant selection), where an increasing number of unique HLA types increases the breadth of peptide recognition and immune defense against infectious organisms [10–12]; and frequency-dependent selection [13], where a pathogen has evolved to escape an efficient immune response mediated by common alleles in the population, but remains susceptible to responses mediated by low frequency alleles. Functional, genetic, evolutionary and epidemiological data have indicated highly influential effects of class I genes on controlling HIV-1/SIV after exposure to the virus. Considering the large number of functionally related loci mapping within the MHC and the strength of linkage disequilibrium among these loci [14–17], it is often difficult to conclusively identify a specific disease locus within the MHC using genetic association studies in the absence of supporting functional data. Likewise, genetic association data lend credence to functional studies, such as characterization of cytotoxic T lymphocyte (CTL) recognition of specific HIV epitopes. The chimpanzee provides a crucial model to understanding mechanisms of viral resistance because, like humans, chimpanzees can be infected with HIV, but are relatively resistant to AIDS. Alternatively, SIV induced disease in rhesus monkeys represents an invaluable model system for studying the pathogenesis of human AIDS [18], because like humans, most rhesus monkeys develop AIDS after infection. These animals have been particularly useful in studying and understanding the initial phases of infection, and detailed studies of changes in viral sequences from infected animals over time as a function of CTL reactivity has enhanced our understanding of the means by which the virus escapes immune control [19]. Thus, comprehensive consideration of data generated from multiple experimental approaches is key to determining of the role of MHC class I in controlling HIV-1. Here we review recent data that implicate class I molecules in regulating progression to AIDS. Protection againstHIV Disease Progression Associated with HLA-B*57 and -B*27 The most lucid genetic effects of HLA on delayed AIDS progression include those conferred by HLA-B*27 and B*57 [20–23], and protection associated with these alleles continues to be substantiated in recent reports [24,25]. Novel effects of HLA class I have also been identified over the past year or so, some of which implicate certain class I alleles in mediating innate responses against HIV-1 [26,27]. Table 1 summarizes the significant HLA class I associations identified in our AIDS cohorts.Table 1: Consistent HLA Effects on AIDS ProgressionHLA-B*57 The strong protective effect of HLA-B*57 has been observed in many studies [20,21,24,28,29] and a recent analysis of 735 HIV-1 Caucasian seroconverters has indicated the profound influence of B*57 on delayed progression to AIDS [30] (relative hazard = 0.46, p = 0.0008 for progression to AIDS 1993]. B*57 has also been shown to be significantly associated with lower viral loads among a group of 259 Zambians who were primarily infected with clade C HIV-1 viruses [25], the only highly consistent class I associations observed between the Zambian cohort and Caucasian cohorts infected with clade B viruses. Individuals with B*57 (and B*27] also appear to generate more pronounced CTL responses against an HIV recombinant canarypox vaccine [31]. Patients with B*57 do, however, tend to exhibit increased hypersensitivity to the therapeutic HIV-1 reverse-transcriptase inhibitor, abacavir, relative to patients with other HLA Type [32,33]. The B*57 studies underscore the need to determine HLA type-specific outcomes to both vaccination and drug therapy against HIV in order to treat appropriately, and perhaps differentially, HIV positive individuals and those at high risk for infection. Although there is presently no obvious mechanism explaining the B*57 genetic association with long term nonprogression to AIDS, several intriguing characteristics of CTL responses in B*57+ long term nonprogressors (LTNP) have indicated the plasticity of B57 responses against HIV-1. CTL responses restricted by B57 molecules target multiple HIV-1 peptides, with predominant responses observed against gag and reverse transcriptase motifs [34,35] and strong B57-restricted CTL responses can occur against overlapping peptides of different lenght [36]. Cross-reactivity against HIV peptide variants has also been shown for other HLA types [37–39], however, so it is unlikely that this property of B57 responses completely accounts for long term nonprogression to disease in B57+ patients. Studies have shown that differential epitope recognition occurs in the context of primary versus chronic HIV-1 infection [40,41], and even more strikingly in HIV-1-exposed, persistently seronegative subjects versus infected individuals [42]. In this regard, it is interesting that dominant HIV-specific CTL responses among long term non-progressors that carry B*57 tend to involve HLA-B57-restricted CTL epitopes as compared to epitopes restricted by other HLA types in those same individuals [35]. Similarly, among individuals with B*57, LTNP exhibited CD8+ T cell responses that were directed primarily towards B57-restricted gag-peptides, while patients who progressed to AIDS rapidly displayed a considerably broader response to gag peptides in terms of allotype restriction [21]. Thus, the types of epitopes recognized and presented by B57 molecules may lead to a particularly effective immune response against the virus. HLA-B*27 The protective genetic effect of HLA-B*27 on AIDS progression correlates nicely with an immunodominant response to a conserved epitope in p24 gag among patients who carry HLA-B*27 [43,44]. The HLA-peptide complex is stabilized by the interaction of an arginine residue at position 2 of the HIV-1 peptide with the B pocket of the B27 peptide binding groove [44–46]. An amino acid substitution from arginine (R) to lysine (K) or glycine (G) at this position (R264K and R264G, respectively) in the gag epitope results in a peptide that binds poorly to B27 [43,47] and correlates with development of AIDS in B27+ individuals [43,48]. This mutation may therefore enable viral escape from B27 restricted CTLs, but additional, compensatory mutations in the gag protein may be required in order for the virus to maintain replication fitness [48]. However, the additional mutations may not compensate completely for the cost of R264K or R264G substitutions on viral fitness, since reversion to wild-type has been observed after the loss of the B27-restricted CTL response and in the presence of high viral load [48]. A mother-infant transmission study has also revealed the strong correlation between mutational viral escape from B27-restricted CTL responses and disease progression [49]. Transmission of HIV escape mutants involving the gag epitope from B27+ mothers to their infants perinatally resulted in the failure of the infected infants to contain HIV replicatio. Thus, the protective genetic effect observed for B27 may be due to its recognition of an epitope that is under structural restraint to remain intact and reinforces the belief that HLA molecules can exert strong selection pressure on HIV in development of mutations facilitating immune escape. Susceptibility to AIDS Progression in Patients with Specific HLA-B*35 Subtypes HLA-B*35, which is almost always found in haplotypic association with Cw*04, has been the most consistently implicated allotype in susceptibility to AIDS progression, particularly in Caucasian individuals [24,28,50,51,52,53] but not African Americans, Zambian, not Kenyan patients [25,51,54]. The difference observed across the two ethnic groups raised the question of whether B*35, Cw*04, or some other locus in linkage disequilibrium with these alleles accounts for accelerated AIDS progression. Recent data from our lab suggests that the operative susceptibility locus is indeed HLA-B [24]. B*35 subtypes among 559 Caucasian and 210 African American seroconverters were divided into two groups according to peptide-binding specificity: (i) the HLA-B*35-PY group, composed of two closely related HLA-B*35 alleles including the most common, B*3501, which primarily bind epitopes with proline in position 2 (P2) and tyrosine in position 9 [55,56] and (ii) the more broadly reactive B*35-Px group, which also prefers epitopes with proline in position 2 but accepts several amino acids different from tyrosine in position 9. Although alleles in both the B*35-Px and -PY groups are in stong linkage disequilibrium with Cw*04, accelerated AIDS progression was only observed amongst individuals with B*35-Px alleles and not with B*35-PY. The demonstration that the susceptibility effect of B*35 is attributable to only a subset of this HLA type offers compelling evidence that the locus responsible for the observed rapid AIDS progression is indeed HLA-B [24]. The dichotomy of B*35-Px vs. B*35-PY in AIDS survival outcome most likely explains the absence of a B*35 association with rapid progression in individuals of African descent. B*3501, by far the most common B*35 allele present in this ethnic group, belongs to the B*35-PY group, which had no effect on AIDS progression. B*3501 was the only B*35 subtype observed in a study of HIV infected Zambian patients [25] and in concordance with our data, this allele showed no association with high viral loads in these individuals. On the other hand, B*53, which was associated with rapid progression to AIDS in African Americans, fits the criteria used to define the B*35-Px group, and was included in the B*35-Px group in our previous study [24], showed no association with high viral load in the Zambian cohort [25], an inconsistency that may be explained by differences in HIV clades between the two studies. Indeed, clade-specific CTL have been detected previously [57] and CTL specific for a B53 epitope that maps to a highly immunogenic region of p24 showed no cross-reactivity to other HIV-1 clades [58]. The B*3501 (PY) and B*3503 (Px) molecules differ by a single amino acid at position 116, which forms the floor of the peptide-anchoring F pocket, determines the size of the peptide C-terminal residue, and directly interacts with residue P9 of the bound peptide [59,60,61]. Substitutions at position 116 have been associated with an increased risk for transplant-related death [62]. Variation at this position also affects the level of interaction between HLA class I and the peptide loading machinery in the endoplasmic reticulum for optimizing the peptide repertoire [63–65], which could account for the difference in AIDS progression between the B*35-Px and -PY groups. We don't beleive that this is the case based on two observations: (i) B*5301 (Px) is associated with rapid progression to AIDS in both Caucasians and African Americans, but like B*3501 (PY), it has a serine residue at position 116 and (ii) preliminary analyses do not suggest any differential effects of HLA alleles grouped on the basis of amino acid identity at position 116 (Fig. 1).Fig. 1.: Association between variants at position 116 of the HLA-B locus and progression to AIDS-1993 (left) and AIDS-1987 (right) in Caucasians. Seroconverters were used in Kaplan Meier and Cox proportional hazards model. Relative hazards (RH) values calculated by the Cox model are given for HLA-B alleles grouped on the basis of amino acid type at position 116. No significant differences were observed.HLA-B and the Killer Immunoglobulin-like Receptor allele, KIR3DS1, interact synergistically to protect against HIV-1. All HLA-B molecules express one of two mutually exclusive serological epitopes termed Bw4 and Bw6, which are distinguished by variable amino acids within the motif spanning amino acid positions 77–83 [66]. A recent report has indicated that homozygosity for any two copies of Bw4 delays AIDS progression [26], raising the intriguing possibility that natural killer cells are involved in regulating AIDS progression since HLA-B molecules with the Bw4 motif serve as ligands for one of the natural killer cell receptors, KIR3DL1 [67]. Natural killer (NK) cells defend against viral infections by producing cytokines and killing virally infected cells [68]. Killer immunoglobulin-like receptors (KIRs) expressed on NK cells (and a subset of T cells) regulate inhibition and activation of NK-cell responses through recognition of HLA class I molecules on target cells [69]. The KIR loci are unlinked to the MHC, mapping to chromosome 19q13.4, and they are exteremely diverse in terms of the number and types of genes present on a given haplotype [70]. Recently KIR3DS1 in combination with HLA-B alleles that encode molecules with isoleucine at position 80 (Bw4–80Ile) was shown to be associated with delayed progression to AIDS in HIV-1 infected seroconverters [27]. The data strongly supported a model involving a protective epistatic interaction between KIR3DS1 and Bw4–80Ile, and could not be attributed to additive effects of these genes based on the observations that: (i) HLA-B Bw4Ile80 alleles (as a group) in the absence of KIR3DS1 were not associated with any AIDS outcomes measured and (ii) KIR3DS1 showed a weak, but significant recessive association with rapid progression to AIDS amongst individuals who did not have Bw4-Ile80. The synergistic effect of KIR3DS1-Bw4–80Ile was most pronounced on early outcomes after HIV infection involving CD4+ T cell depletion, concurring with the defensive role of NK cells in the innate immune response soon after infection. KIR loci have also been identified in chimpanzees and rhesus monkeys [71,72] and their characterization in the context of SIV infection will be of particular interest. Functional evidence for the role of NK cells in defense against HIV/SIV may be particularly illuminating in providing new therapeutic approaches to controlling AIDS. Influence of HLA on Resistance to HIV Infection An increasing amount of genetic data has consistently indicated a protective role for HLA in HIV transmission and infection [54,73–75]. An increase in the risk of HIV-1 transmission has been observed for concordance at HLA class I loci in studies of mother-infant pairs [73] and heterosexual couples [74]. Both studies suggest that allogeneic immune responses may be protective against HIV transmission in a dose-dependent manner, a phenomenon that may be mediated by the high expression of HLA molecules present on the envelope of HIV-1 [76]. Groupings of HLA molecules, known as HLA supertypes, have been proposed based on similarity in structure, peptide-binding motif, epitope presentation and evolutionary affinity of class I molecules [77] [78]. The A2/6802 supertype (A*0202, A*0205, A*0214, and A*6802) was shown to be significantly associated with decreased frequency of HIV-1 seroconversion among a cohort of highly HIV-exposed prostitutes in Nairobi [54] and similar associations were observed in a comparison of HIV+ and HIV-individuals from Botswana [79]. HIV-specific CTL, which persisted for several years over the course of follow-up, have been identified in a group of seronegative prostitutes from The Gambia [80], potentially explaining the genetic associations observed between HLA class I and resistance to HIV-1 infection. CTL responses to multiple conserved HIV epitopes have also been observed in highly exposed seronegative prostitutes from Nairobi, some of which are restricted by HLA molecules associated with resistance to infection in this group of women [81]. A striking association between resistance to HIV-1 infection and recognition of CTL epitopes distinct from those targeted in HIV+ individuals with the same HLA types has been reported [42] and the significant differences in epitope specificity involved only those HLA class I types that have been associated with resistance to infection in this cohort [54]. The seronegative women with the longest duration of HIV-1 exposure had the most frequent and strongest HIV-1-specific CD8+ T cell responses [42], strongly supporting a model in which resistance to HIV infection is mediated by CD8+ T cell responses. Nevertheless, consistent antigenic exposure appears to be necessary in order to maintain an effective HIV-specific CTL response [82]. The Chimpanzee Story At the DNA level, humans and chimpanzees share more than 95% sequence similarity. This high degree of genetic likeness has a number of biological consequences. For example, apart from humans, the chimpanzee is the only species susceptible to infection with HIV-1. In the quest for an AIDS vaccine, more than 150 chimpanzees have been infected with various HIV-1 isolates in different research institutes over the past few decades. In contrast to most human individuals, chimpanzees are able to maintain relatively low virus titers and as such appear to control viral escape. Based on these observations, it is thought that chimpanzees have a natural resistance to AIDS-induced pathogenesis. Thus far only one animal, which was infected multiple times with a combination of two virus strains, has developed AIDS-like pathogenesis [83,84]. A recombinant strain, that probably caused the loss of CD4 positive T cells and increased lymphocyte apoptosis, was isolated from this animal at the time of onset of disease. When this strain was used to infect three other chimpanzees, signs of immune activation were observed. These studies suggest that natural resistance in chimpanzees can be by-passed by (recombined) viruses that have been partially edited in the human population. Some chimpanzees in captivity as well as a few free ranging animals were shown to have natural infections with SIVcpz, a lentivirus closely related to HIV-1. In some animals SIVcpz may show relatively high copy numbers, but evidence for SIVcpz induced disease has not been documented. SIVcpz is the most likely ancestor of all known HIV-1 isolates. It is believed that this chimpanzee virus crossed a species barrier approximately 50 years ago [85], and this zoonosis is held responsible for the contemporary HIV-1 pandemic in the human population [86]. The question arises as to which factors in chimpanzees may contribute to their natural resistance. In chimpanzees, the CCR5 HIV-1 coreceptor is intact and can not be held responsible for the resistance trait [87]. MHC class I molecules are known to play an important role in the defense against viruses and on first sight, chimpanzees appear to display a high degree of allelic polymorphism at the Patr-A, -B and -C [2,3,88] and at the Patr-DR, -DQ and -DP loci [5,89,90]. The Patr class I alleles themselves, however, appear to be rather homogeneous. Interestingly, all described chimpanzee Patr-A alleles group into the HLA-A1/A3/A11 family, representing only one of the five HLA-A families [3,4,91]. Although a negative selection event may explain this reduced repertoire, convergent evolution could provide an alternative explanation. A subsequent study on MHC class I gene associated intron 2 variation illustrated that chimpanzees display far less variation than the human population [92]. This is an unexpected observation, since all other polymorphic systems studied thus far have shown more variation in the contemporary chimpanzee population. The explanation for this extensive diversity in chimpanzees is that, as a species, they are older and as such, have had more time to accumulate variation. Furthermore, the human population has experienced a few bottlenecks. In conclusion, the low intron 2 variation in chimpanzees (Fig. 2) provides evidence that chimpanzees experienced a selective sweep targeting the MHC class I loci. We have hypothesized that this selective sweep was caused by an ancient AIDS-like pandemic [92]. If SIVcpz or a closely related lentivirus was responsible for this repertoire reduction, one would expect that the contemporary Patr class I molecules would be capable of binding and presenting HIV/SIV epitopes to CTL in an efficient manner. The peptide binding profiles of several Patr-class I molecules have been defined [93–96] and it appears that Patr-B molecules may target conserved HIV-1 epitopes similar to those bound by HLA-B27 and -B57 molecules present in some human long term nonprogressor cohorts [97]. Specifically, Patr-B02 molecules bind the HQAISPRTL gag peptide, which is also recognized by HLA-B5701 molecules, whereas Patr-B03 molecules select another gag epitope, KRWIIGLN, that is also bound by HLA-B2705 positive cells. It is now known that these human and chimpanzee molecules group into different evolutionary clusters, as evidenced by dissimilar anchor residues, thus illustrating that the quality of MHC class I molecules may be important in protection from disease.Fig. 2.: Polymorphic nucleotide positions in HLA and Patr intron 2 sequences showing how conserved the chimpanzee sequences are relative to humans. Identity to the consensus sequence is depicted by dashes. Substitutions and inserts have been depicted by the conventional code. The brackets show the division of the alleles into trans-species lineages, which was made based on phylogenetic analyses. For details see [92].In conclusion, chimpanzees may have a selected set of MHC class I molecules that efficiently bind conserved epitopes of SIVcpz/HIV. The efficient control of CTL may play an important role in immune surveillance and as such, limit viral replication and subsequent escape by variants. Future studies are needed to develop an inventory of the peptide binding profiles of many other Patr class I molecules. AIDS-like disease in Rhesus Macaques and Impact of MHC class I Factors. Several African Old World monkey species such as African green monkeys, mandrills, Sykes’ monkeys and Sootey mangabeys are naturally infected with particular SIV strains. These animals remain asymptomatic despite chronic infections. In these cases, the natural host species and virus have co-evolved in such a way that disease is absent in the contemporary population. In most of these species, little is known about the MHC repertoire of the host and the extent to which MHC factors may contribute to protection. However, when such a natural SIV strain or a derivative is used to infect another Old World primate species, such as rhesus macaques, AIDS-like pathogenesis may develop. Such experimental protocols are very similar to the human situation, since the AIDS pandemic may have been started by infection of the human population with a chimpanzee virus. The rhesus macaque model has provided fundamental knowledge, because it allows researchers to use molecularly cloned viruses and study immune responses in cohorts of animals directly after infection. Further, the MHC class I and II repertoires of rhesus monkeys have been studied rather thoroughly [4,5]. Most recent research avenues have focused on the role of CTL and their interaction with Mamu class I molecules containing viral peptides. Like HIV in humans, SIV infections are followed by early peaks and subsequent declines in viremia as strong cellular immune responses develop. CTL responses have been correlated with protection against infection observed in a challenge study with SIV-infected peripheral blood cells. Only monkeys that were treated with SIV infected cells that shared a particular serologically defined MHC class I allele (Mamu-A26] with the donor were protected from infection [98]. Also, SIV strains are involved in the destruction of CD4 positive cells and the elimination of these cells impairs the function of CTL [99]. Other studies documented that SIV replication was not controlled in monkeys depleted of CD8 positive T cells [100,101] whereas administration of anti-CD8 monoclonal antibody was shown to interfere with the clearance of SHIV during primary infections in rhesus macaques [102]. SIV infected rhesus macaques are used to study candidate AIDS vaccines and the current concept is that these vaccines should elicit strong CTL responses. Many of those studies have been undertaken in Mamu-A*01 positive animals because the corresponding peptide motifs have been identified [103–107]. These studies illustrated that CTL from SIV infected rhesus monkeys may recognize at least 14 different epitopes in the context of Mamu-A*01 molecules. Peptide binding motifs have also been defined for Mamu-A*02, -A*11, -B*03, B*04 and B*17 molecules [108–111]. In the first phases of infection, many of the CTL appear to target infected cells by recognizing the viral Tat protein. These Mamu-A*01 restricted, Tat specific CTL could be excellent candidates to eliminate viral infection. Allen and coworkers, however, showed dramatic viral escape from the Tat 28–35 epitope (SL8) early in the acute phase of the infection [112]. The Mamu-A*01 restricted Tat specific CTL responses peaked between 3 and 4 weeks after infection but declined after the acute phase of the infection. In five out of the ten Mamu-A*01 positive animals, all sequenced SIV clones contained mutations in the Mamu-A*01-restricted SL8 epitope. In contrast, little or no amino acid variation was observed outside this SL8 epitope in the Mamu-A*01 positive animals. Acute phase CTL escape appears to represent a hallmark of SIV infection [113]. One could argue that pre-existing Tat-specific immune responses at the time of virus infection could be protective. Unfortunately Tat vaccinated macaques do not control SIV replication [113]. At present, the Mamu-A*01 molecule is considered to represent one of the factors that controls slow progression to AIDS-like disease in rhesus macaques. Interestingly, animals that are positive for the combination of Mamu-A*01 and particular other Mamu class I specificities (heterozygous advantage) may be resistant to disease (O'Connor et al. Submitted for publication). This indicates that not only the quality of MHC class I molecules but also the quantity of epitopes recognized may play an important role in disease. Summary Data indicate that resistance to HIV-1 disease involves an array of contrasting HLA genotypic effects that are subtle, but significant, particularly when these genetic effects are considered as a whole. Numerous reports attributing a role for HLA genotype in AIDS outcomes have been reported, and a few of these have been affirmed in multiple studies. Functional studies of immune cell recognition have provided clues to the underlying mechanisms behind some of the strongest HLA associations, suggesting the means by which relative resistance or susceptibility to the virus may occur. SIV infection in non-human primates has served as an invaluable model for understanding AIDS pathogenesis (in rhesus monkeys) and viral resistance (in chimpanzee). The effect of rhesus MHC class I molecules on the evolution of SIV has been convincingly described [19], and a recent study in humans has suggested that selection pressure conferred by HLA molecules is responsible for specific genetic variation in HIV-1 [114]. HIV-1 may eventually have conspicuous evolutionary effects on HLA and other AIDS restriction genes, a prolonged process that could have occurred in chimpanzee [92]. To prevent such an outcome, it will be necessary to approach the disease from many perspectives, and apply comprehensively the knowledge gained to the successful control of the virus.