|Home | About | Journals | Submit | Contact Us | Français|
Human immunodeficiency virus type 1 (HIV-1) mutates to escape immune selection pressure, but there is little evidence of selection mediated through HLA-A2, the dominant class I allele in persons infected with clade B virus. Moreover, HLA-A2-restricted responses are largely absent in the acute phase of infection as the viral load is being reduced, suggesting that circulating viruses may lack immunodominant epitopes targeted through HLA-A2. Here we demonstrate an A2-restricted epitope within Vpr (Vpr59-67) that is targeted by acute-phase HIV-1-specific CD8+ T cells, but only in a subset of persons expressing HLA-A2. Individuals in the acute stage of infection with viruses containing the most common current sequence within this epitope (consensus sequence) were unable to mount epitope-specific T-cell responses, whereas subjects infected with the less frequent I60L variant all developed these responses. The I60L variant epitope was a stronger binder to HLA-A2 and was recognized by epitope-specific T cells at lower peptide concentrations than the consensus sequence epitope. These data demonstrate that HLA-A2 is capable of contributing to the acute-phase cytotoxic T-lymphocyte response in infected subjects, but that most currently circulating viruses lack a dominant immunogenic epitope presented by this allele, and suggest that immunodominant epitopes restricted by common HLA alleles may be lost as the epidemic matures.
The human immunodeficiency virus type 1 (HIV-1) epidemic is characterized by a high genetic diversity within the viral population that results from high replication and mutation rates in the presence of immunological selection pressure (29). Viral strains from the same HIV-1 clade can differ by 25% at the amino acid level, depending on the particular HIV-1 protein under consideration (29). This substantial sequence diversity poses a major challenge to the design of vaccines capable of inducing cross-reactive CD8+ T-cell responses. As a consequence, the use of clade-specific consensus sequences has been recently proposed for vaccine design (12). Clade consensus sequences have the advantage of being most similar to currently circulating strains of interest, with each amino acid corresponding to the most commonly found amino acid at that position within the overall viral population.
A number of studies have demonstrated that HIV-1 can rapidly escape from CD8+ T-cell-mediated immune pressure by sequence variation within or flanking targeted epitopes (1, 6, 11, 14, 15, 18, 20, 23-25). The accumulation of escape variants within epitopes presented by the HLA class I alleles expressed in an infected individual can result in “HLA footprints” on the viral sequence (21, 22, 30). Recent data demonstrate that these cytotoxic T-lymphocyte (CTL) escape variants can be transmitted and impair the generation of otherwise immunodominant immune responses during primary infection in a new host (1, 14, 20). The rate at which these sequence mutations within epitopes may accumulate in the viral population largely depends on the frequency of the restricting HLA allele, the impact of the mutation on viral fitness, and the genetic stability of the mutations (5, 16). Mutations restricted by common HLA alleles that do not result in a major reduction in viral fitness and do not revert following transmission into a new host are more likely to accumulate over time. Here we show that most currently circulating clade B viruses lack an HLA-A2-restricted CD8+ T-cell epitope within HIV-1 Vpr, which is otherwise targeted in the acute phase of infection.
A total of 88 individuals followed at the Partners AIDS Research Center at Massachusetts General Hospital in Boston were enrolled in this study. All of the individuals expressed the HLA class I allele HLA-A2. Study subjects included 14 individuals identified during primary HIV-1 infection and 74 individuals identified during chronic infection. Primary HIV-1 infection was defined by documented HIV-1 seroconversion within the past 6 months (3), and baseline samples were obtained from all subjects enrolled during primary infection prior to initiation of antiretroviral therapy. Ten of these 14 subjects with primary infection expressed HLA-A*0201, while 2 expressed HLA-A*0202 (AC-09 and AC-35) and 2 expressed HLA-A*0205 (AC-34 and AC-75). Chronically infected individuals were infected for more than 2 years. This study was approved by the institutional review board and conducted in accordance with human experimentation guidelines of the Massachusetts General Hospital.
HIV-1-specific CD8+ T-cell responses were quantified by gamma interferon (IFN-γ) ELISPOT assay, with a panel of peptides corresponding to previously described optimal clade B CTL epitopes (7). Peripheral blood mononuclear cells (PBMC) were plated at 100,000 per well with peptides at a final concentration of 10−5 M in 96-well plates and processed as previously described (3). PBMC were incubated with medium alone (negative control) and phytohemagglutinin (positive control). The number of specific IFN-γ-secreting T cells was determined with an automated ELISPOT reader (AID, Strassberg, Germany), calculated by subtracting the average negative control value and expressed as the number of spot-forming cells (SFC) per 106 input cells. Negative controls were always ≤30 SFC per 106 input cells. A response was considered positive if there were ≥50 SFC per 106 input cells and the activity was at least three times as great as the mean background activity. Comparisons of recognition of epitope variants were performed by ELISPOT assay with serial dilutions of truncated peptides as previously described (2).
CTL clones were isolated by limiting dilution as previously described (28), with anti-CD3-specific monoclonal antibody 12F6 as the stimulus for T-cell proliferation. Developing clones were screened for HIV-1-specific CTL activity by 51Cr release assay (28) against autologous B-cell lines (BCL) pulsed with the peptides at the concentrations indicated. HIV-1-specific clones were maintained by stimulation every 14 to 21 days with an anti-CD3 monoclonal antibody and irradiated allogeneic PBMC.
Autologous virus was sequenced from proviral DNA by population and clonal sequencing as described previously (2). Viral DNA was isolated from PBMC (5 × 106 cells), and nested PCRs were conducted with a set of primers specific for the Vpr region of HIV-1 (2). The first-round PCR cycling conditions consisted of 94°C for 2 min, 35 to 50 cycles of 30 s at 94°C, 30 s at 56°C, and 2 min at 72°C; and a final extension of 68°C for 20 min, and nested PCRs had a shortened 1-min extension time. PCR fragments were then gel purified and sequenced directly or cloned (TOPO TA cloning kit; Invitrogen, Carlsbad, Calif.). Plasmid DNA was isolated by miniprep (QiaPrep Turbo Miniprep) and sequenced bidirectionally on an ABI 3100 PRISM automated sequencer. Sequencher (Gene Codes Corp., Ann Arbor, Mich.) and MacVector 4.1 (Oxford Molecular) were used to edit and align sequences. Neighbor-joining trees were constructed with Phylip (Phylogeny Inference Package, version 3.5c).
HLA class I molecules were purified from detergent lysates of Epstein-Barr virus-transformed homozygous cell lines and used in a competitive binding assay with synthetic HIV-1 peptides containing 8 to 11 amino acids, as described previously (27). Peptides known to bind to particular HLA antigens with high affinity were iodinated by the chloramine T method and used as standards for binding assays. To measure HIV-1 peptide binding to HLA molecules, 1 nM to 1 μM purified HLA molecules were incubated with the HIV-1 test peptides at concentrations ranging from 33 μg/ml to 0.33 ng/ml, along with the radiolabeled standard peptides at 1 to 10 nM, for 48 h in phosphate-buffered saline containing 0.05% NP-40 (27). All assays were run at pH 7 in a cocktail of protease inhibitors. Following a 2-day incubation, the percentage of major histocompatibility complex-bound radioactivity was determined by capturing major histocompatibility complex-peptide complexes on W6/32 antibody (anti-HLA class I)-coated Optiplates (Packard Instruments, Meriden, Conn.) and measuring counts per minute with a Topcount (Packard Instruments) microscintillation counter.
Statistical analysis and graphical presentation were done with SigmaPlot 5.0 (SPSS Inc., Chicago, Ill.). Results are given as the mean ± the standard deviation or the median with the range. Statistical analysis of significance (P values) was based on two-tailed t tests and Fisher's exact test.
Previous studies have demonstrated differential targeting of HIV-1 epitopes by virus-specific CD8+ T cells in primary and chronic HIV-1 infections (3, 9, 13). In particular, the HLA-A2-restricted CD8+ T-cell epitope SLYNTVATL in HIV-1 p17 Gag is rarely targeted in acute HIV-1 infection but is the immunodominant HLA-A2-restricted CD8+ T-cell epitope in chronic HIV-1 infection in persons expressing HLA-A2 (13). Similarly, this epitope is not targeted in Gag CTL-positive, HIV-1-uninfected vaccine recipients expressing the HLA-A2 allele (10).
In order to further study the relative immunodominance of HLA-A2-restricted HIV-1-specific CD8+ T-cell responses in acute infection, we characterized HLA-A2-restricted CD8+ T-cell responses in 88 HLA-A2 expressing individuals, 14 with primary infections and 74 with chronic infections, with an IFN-γ ELISPOT assay. In line with previous studies (3, 9), CD8+ T-cell responses during the acute phase of HIV-1-infection, prior to initiation of antiretroviral therapy (median, 8.5 days from presentation with clinical symptoms; range, 3 to 30 days), were narrowly directed against a limited number of epitopes. Only 6 of the 16 defined HLA-A2-restricted HIV-1 epitopes were recognized by at least one individual (range, 0 to 21% of the study subjects; Fig. Fig.1)1) in the acute infection phase, but in the chronic phase all 16 peptides were targeted by at least one person (range, 5 to 71% of the study subjects [P < 0.001]; Fig. Fig.1)1) and the frequency of recognition increased for all epitopes.
The most frequently targeted HLA-A2-restricted CD8+ T-cell epitopes in chronic infection were the p17 Gag epitope SLYNTVATL (p1777-85), the p1 Gag epitope FLGKIWPSYK (p11-10), and the RT epitope ILKEPVHGV (RT309-317), which were targeted in 62, 54, and 45% of the chronically infected individuals expressing HLA-A2, respectively. These three epitopes were significantly less frequently recognized in individuals identified during primary infection (SLYNTVATL, 1 of 14 [P < 0.001]; ILKEPVHGV, 1 of 14 [P < 0.01]; FLGKIWPSYK, 0 of 14 [P < 0.001]) (Fig. (Fig.1).1). In contrast, the HLA-A2-restricted epitope AIIRILQQL in Vpr (Vpr59-67) was recognized in 3 (21%) of 14 individuals with primary infections and 18 (24%) of 74 (P > 0.05) individuals with chronic infections, respectively (Fig. (Fig.1).1). These data demonstrate that, in contrast to most other HLA-A*0201-restricted epitopes, the Vpr59-67 epitope can represent an early target for HIV-1-specific CD8+ T cells in primary HIV-1 infection in a subset of HLA-A 2-positive individuals and is targeted with similar frequencies in the acute and chronic phases of infection.
We subsequently addressed the underlying mechanisms resulting in the recognition of the Vpr59-67 epitope by CD8+ T cells in some individuals during primary infection but not in others. Variant epitope sequences within the transmitted virus have been shown to impair the generation of epitope-specific CD8+ T-cell responses in primary infection (1, 14, 20), and we therefore sequenced the autologous virus sequence of HIV-1 Vpr in the 14 study subjects identified during primary HIV-1 infection at the earliest available time point (median, 8.5 days following presentation with symptomatic acute-phase infection). In line with sequences reported in the Los Alamos database, the encoding sequence of the Vpr59-67 epitope was highly conserved at all amino acid positions except positions 2 (Vpr60) and 5 (Vpr63) in the 14 study subjects, on the basis of population sequences (Table (Table1).1). Sequence variability was observed in the autologous virus from 4 individuals at position 2 (I60L mutation in 4 of 14) and in 5 individuals at position 5 (I63T in 3 of 14, I63L in 1 of 14, and I63M in 1 of 14), and these amino acid variations reflected the variations most frequently observed in the Los Alamos database (Table (Table1).1). All three individuals who mounted a Vpr59-67-specific CD8+ T-cell response during primary infection harbored virus containing the I60L variant at position 2 (Table (Table1).1). In contrast, none of the individuals infected with HIV-1 containing the clade B consensus sequence isoleucine at position 2 (Vpr I60; 10 of 14 subjects) mounted CD8+ T-cell responses during primary infection (P = 0.01, compared to responders). Sequencing of viral clones from the fourth subject (AC-60), who did not mount detectable Vpr59-67-specific CD8+ T-cell responses during primary infection despite the presence of the I60L variant on the basis of the population sequences, revealed the presence of a mixed population of ALIRILQQL (n = 10) and AIIRILQQL (n = 6) sequences (differences are underlined). This individual initiated antiretroviral therapy during primary infection with successful suppression of viremia during the subsequent 36 months, potentially impairing the mounting of a CD8+ T-cell response against the I60L variant. Taken together, these data suggested either that the virus in all of the individuals mounting a Vpr59-67-specific CD8+ T-cell response exhibited very early escape from the to I60 to the I60L variant prior to the time when the first sequence data were obtained or that only individuals infected with the I60L variant were able to develop epitope-specific CD8+ T-cell responses during primary infection.
To determine whether the I60L variant represents an escape variant of the I60 consensus sequence Vpr59-67 epitope, we compared the abilities of the ALIRILQQL (I60L) and AIIRILQQL (I60) epitopes to bind to HLA-A2. Interestingly, the ALIRILQQL epitope was a substantially better binder to HLA molecules from the HLA-A2 superfamily compared to the consensus sequence AIIRILQQL variant (Table (Table2).2). In line with these findings, both Vpr59-67-specific CD8+ T-cell clones and PBMC from the individuals targeting the Vpr59-67 epitope (AC-04, AC-13, and AC-75) recognized the ALIRILQQL peptide better at low peptide concentrations than the AIIRILQQL peptide, measured by both lytic activity and peptide-specific IFN-γ production (Fig. 2A and B). Overall, these data indicate that the I60L variant of the Vpr59-67 epitope represents the antigenic form of the epitope. In addition, none of the individuals studied who were infected with virus containing the clade B consensus sequence of the epitope (I60) had detectable Vpr59-67-specific memory T-cell responses following peptide-specific expansion (data not shown), although this epitope was well processed and presented when expressed with HIV-1 Vpr vaccinia virus constructs containing the Vpr59-67 I60 sequence (4). These data suggest that responses specific for the I60 variant were induced in individuals infected with viruses containing the consensus sequence of the Vpr59-67 epitope.
The above data obtained from individuals with primary HIV-1 infection demonstrate that the clade B consensus sequence of the HIV-1 Vpr59-67 CD8+ T-cell epitope restricted by the most frequent HLA class I allele in Caucasians, HLA-A*0201, is less immunogenic than the minor circulating I60L variant in the P2 anchor position of this epitope. In addition, the frequency of the I60L variant in the Los Alamos database (18%) is very similar to the frequency of individuals targeting this epitope in this study (24%), further supporting the need for the leucine at position 60 to elicit an epitope-specific CD8+ T-cell response. Interestingly, the highly immunogenic I60L variant of the Vpr59-67 epitope is significantly more frequently represented in the published HIV-1 clade C sequences (34 [37%] out of 90 clade C sequences derived from the Los Alamos database) than in the published HIV-1 clade B sequences (25 [18%] out of 137 clade B sequences [P = 0.01]). Considering that the clade B epidemic is historically older than the clade C epidemic (17, 19, 26) and that HLA-A2 is more frequently expressed in the Caucasian population (allele frequencies of 23 to 39%) than in the South African (14.9 to 16.3%) and Indian (0 to 10.7%) populations, which are mostly affected by the clade C epidemic (8; F. Gao and M. Carrington, personal communication), it is tempting to speculate that viral evolution in the presence of HLA-A*0201-mediated CD8+ T-cell immune pressure in the Caucasian HIV-1 clade B epidemic has accounted for the gradual elimination of the more immunogenic I60L Vpr59-67 variant as a major circulating population in favor of the less immunogenic I60 variant. Additional studies are needed to determine at what rate accumulation of less immunogenic epitope variants restricted by common HLA alleles is occurring in the HIV-1 clade B epidemic, and their clinical implications, as the presence of these epitope variants may result in a reduced ability of the current clade B consensus sequence to induce CD8+ T-cell responses when used in a prophylactic or therapeutic vaccine.
This study was supported by the Doris Duke Charitable Foundation and the National Institutes of Health. B.D.W. is the recipient of a Doris Duke Distinguished Clinical Scientist Award.