HIV-1 Nef is preferentially recognized by CD8 T cells in primary HIV-1 infection despite a relatively high degree of genetic diversity

Objective: To compare the magnitude, breadth and protein specificity of HIV-1-specific CD8 T-cell responses against the clade B consensus sequence during primary and chronic HIV-1 infection and to analyze the impact of viral diversity on the localization of detected responses.

Methods: HIV-1-specific CD8 T-cell responses against the clade B consensus sequence in individuals with acute (n = 10), early (n = 19) and chronic (n = 10) infection were longitudinally assessed using an interferon-γ EliSpot assay.

Results: CD8 T-cell responses against clade B consensus sequences were preferentially directed against central regions of Nef during primary HIV-1 infection, despite a relatively higher degree of genetic diversity compared with other subsequently targeted regions. In subjects with acute and early infection, Nef-specific CD8 T-cell responses against the consensus Nef sequence represented 94 and 46% of the total magnitude of HIV-1-specific CD8 T-cell responses, respectively. Subjects with untreated chronic infection exhibited broadly diversified CD8 T-cell responses against more conserved viral regions, with only 17% of virus-specific T-cell responses targeting Nef. The initial immunodominance of Nef persisted in individuals with treated acute infection, but shifted rapidly to Gag, Env and Pol in subjects with continuous antigen exposure.

Conclusion: These data show that despite relatively high sequence variability, viral regions within the clade B consensus sequence of Nef are preferentially recognized during primary HIV-1 infection. Later diversification of responses to other proteins during prolonged antigen exposure provides evidence of the initial preferential immunogenicity of Nef epitopes compared to similarly conserved regions within other viral proteins.

Introduction

The hallmark of acute HIV-1 infection is overwhelming viral replication in CD4 T lymphocytes resulting in high level HIV-1 viremia [^1,2]. This initial peak of viral replication begins to decline simultaneously with the appearance of HIV-1-specific CD8 T lymphocytes, which are vigorously expanded during primary HIV-1 infection [^3,4]. These cells have the ability to eliminate HIV-1-infected cells directly by MHC class I-restricted cytolysis or indirectly through the production of cytokines, chemokines or other soluble factors, thus curtailing the generation of new viral progeny [⁵]. The biological relevance of HIV-1-specific cytotoxic T cells (CTL) in acute HIV-1 infection was supported by in-vivo studies demonstrating a dramatic rise of viremia and an accelerated clinical disease progression in SIV-infected macaques after the artificial depletion of CD8 cells [^6,7]. Additional evidence for the antiviral pressure mediated by virus-specific CTLs during primary HIV-1 and SIV infection has been provided by the demonstration of rapid selection of mutations in CTL epitopes after only a few weeks of infection [^8-10]. Thus, HIV-1-specific CD8 T cell responses arising during primary infection appear to be effective in viral containment and may represent attractive targets for HIV-1 vaccine design. The characterization of CD8 T cell epitopes that are involved in viral suppression during acute HIV-1 infection and the analysis of their fate during the ensuing disease process may therefore provide important information for the development of HIV-1 immunization strategies [¹¹].

One characteristic of the HIV-1 epidemic is the considerable viral gene diversity that is present even between subjects infected with the same HIV-1 clade in the same geographic region [¹²]. This genetic diversity poses a challenge to both the design of broadly cross-reactive vaccines [¹³] as well as the accurate assessment of immune responses in an infected individual [¹⁴]. In particular, the use of reference strains of virus in immunology assays limits detection of responses to autologous virus, specifically in regions of high sequence diversity, and thus may underestimate the magnitude, breadth and specificity of responses in both acute and chronic infection.

Despite this limitation, one can address the relative immunogenicity of relatively conserved regions of the virus using a representative reference strain of virus. Moreover, such an approach allows one to assess the relative sequence variability (entropy) of early targeted versus later targeted regions of the virus. Therefore, to begin to address the relative kinetics of induction of those cross reactive immune responses detectable in chronic infection, we utilized a set of clade B consensus sequence peptides to characterize the regions within HIV-1 that are recognized by virus-specific CD8 T cells from the time of acute infection and compared these to chronic infection. In addition, we assessed the impact of overall entropy within these regions on their frequency of recognition. These studies demonstrate that HIV-1-specific CD8 T-cell responses in acute and chronic HIV-1 infection differ in their breadth and magnitude, that ongoing viral replication results in broadening of the initial response, and that the initial immune responses cross-reactive with the clade B consensus sequence are directed against multiple epitopes in Nef, despite the fact that this region has a higher entropy value than other regions that are either targeted later or not at all.

Materials and methods

Study population

This study characterized HIV-1-specific CD8 T-cell responses in four different groups of infected individuals: patients who were diagnosed with acute (n = 10) or early (n = 10) HIV-1 infection and subsequently treated with highly active antiretroviral therapy (HAART), individuals diagnosed during early HIV-1 infection who remained antiretroviral therapy-naive over the ensuing 12 months (n = 9), and chronically HIV-1 infected subjects (n = 10). Individuals diagnosed during acute HIV-1 infection were characterized by detectable HIV-1 viremia, a positive or intermediate HIV-1-specific antibody in a standard enzyme-linked immunosorbent assay test, and detection of no more than three bands in an HIV-1-specific western blot. Individuals with early HIV-1 infection had more than three bands in the HIV-1-specific western blot and a documented HIV-1 seroconversion within the previous 6 months. The group of chronically HIV-1 infected subjects consisted of 10 individuals with documented HIV-1 infection of more than 1 year. All these individuals were antiretroviral therapy naive at the first peripheral blood mononuclear cells (PBMC) specimen collection. Study subjects were recruited from the Massachusetts General Hospital and the Fenway Community Heath Care Center in Boston. The study was approved by the respective institutional review boards and was conducted in accordance with human experimentation guidelines of the Massachusetts General Hospital.

HLA typing

High and intermediate resolution HLA class I typing was performed at a commercial laboratory (Dynal Biotech, Oxford, UK) by sequence-specific polymerase chain reactions according to standard procedures.

Synthetic HIV-1-peptides

Four hundred and ten synthetic 17-19 amino acid long peptides, overlapping by 10 amino acids and spanning the entire HIV-1 clade B 2001 consensus sequence (http://hiv-web.lanl.gov), were synthesized at the MGH Peptide Core Facility on an automated peptide synthesizer using Fmoc technology.

Elispot assays

Elispot assays were carried out as described previously [¹⁵]. Briefly, PBMC isolated by Ficoll-Hypaque (Sigma, St. Louis, Missouri, USA) density gradient centrifugation were plated in 96-well polyvinylidene plates that had been precoated with 0.5 μg/ml of an anti-human interferon-γ monoclonal antibody (Mabtech, Stockholm, Sweden). PBMC were added at a concentration of 50 000 to 100 000 cells per well in a volume of 100 μl of RPMI 1640 medium supplemented with 10% fetal calf serum (FSC), Hepes buffer (10 mmol/l), l-glutamine (2 mmol/l) and penicillin-streptomycin (50 U/ml). The final concentration of the peptides in every single well was 14 μg/ml. Plates were incubated overnight at 37°C, 5% CO₂ and developed on the next day as described elsewhere [¹⁵]. Wells containing PBMC and medium with phytohemagglutinin or without any peptide were used as a negative or positive controls, respectively, and run in triplicate on each plate. To calculate the number of specific T cells, the number of spots in the negative control wells was subtracted from the counted number of spots in each well. Responses were considered positive if there were > 50 SFC/10⁶ PBMC and had at least three times the mean number of SFC of the three control wells. CD8 T-cell dependence of responses was determined by depletion of CD4 T cells using the Minimacs cell depletion system (Miltenyi Biotec, Bergisch-Gladbach, Germany).

Calculation of peptide variability

A Shannon entropy score was calculated as described previously [¹⁶] for each position in the 2002 alignment of clade B sequences published in the Los Alamos National Database (http://hiv-web.lanl.gov). Entropy is a measure of the amino acid variability at a given position that takes into account both the number of possible amino acids allowed and their frequency. An average entropy score for all positions in each of the 410 overlapping peptides was determined to provide a single value that characterizes the overall variation of each peptide.

Statistical analysis

Results are given as means or medians with ranges. Statistical analysis was based on Student's t tests, Wilcoxon rank sum test, or a multiparametric analysis of variance (ANOVA) test, as appropriate; a P < 0.05 was considered significant. When two adjacent peptides were recognized in Elispots, we deleted the weaker of the two responses. In the case of three adjacent peptides eliciting a response, the weakest of all three peptides was deleted and the responses were counted as two epitopic regions.

Results

Relative kinetics of induction of HIV-1-specific CD8 T-cell responses and relationship to viral sequence diversity

Previous studies have indicated that HIV-1-specific CD8 T-cell responses are weak and narrowly directed against a limited number of epitopes during acute infection [^15,17]. However, no studies have assessed virus-specific T-cell responses directed against the entire HIV-1 clade B consensus sequence during the acute phase of infection and determined the impact of sequence variability on the recognition of clade B consensus peptides. Here, we conducted a longitudinal analysis of HIV-1-specific CD8 T-cell responses directed against the entire expressed HIV-1 clade B consensus sequence in 29 patients with primary HIV-1 infection (10 patients with acute infection, 19 with early infection) and 10 patients with chronic HIV-1 infection. The demographic and clinical characteristics of the study cohort are summarized in Table 1. No significant difference in sex, age, CD4 T-cell count and HLA class 1 allele frequencies were observed among the three study groups.

Table 1

No HIV-1-specific CD8 T-cell responses directed against the HIV-1 clade B consensus sequence were detectable at baseline in four of the 10 study subjects with acute infection. Although this does not preclude the presence of type-specific responses to autologous virus, these data indicate the lack of early induction of immune responses cross-reactive with the clade B consensus sequence in a significant minority of persons. In the remaining six acutely-infected individuals, generally weak responses (median of 100 SFC/10⁶ PBMC per peptide; range, 80-1150) against a median of one epitope per person (range, 1-2) were identified. Individuals with early infection had stronger (median of 204 SFC/10⁶ PBMC per peptide, range: 50-980, P = 0.44) and more diverse (median 1; range, 0-7, P = 0.07) CD8 T-cell responses than study persons with acute infection. In patients with chronic HIV-1 infection, CD8 T-cell mediated immune responses were 12 and 10 times higher when compared with individuals with acute and early infection, respectively (P < 0.01). Moreover, the median number of epitopic regions targeted by CD8 T cells in individuals with chronic HIV-1 infection was significantly higher than in individuals with acute and early infection (19 epitopic regions in chronic infection versus one in early and acute infection, respectively, P < 0.01). In neither primary nor chronic HIV-1 infection did we observe any significant correlation between the level of HIV-1 viremia and the total quantity of such cross-reactive HIV-1-specific CD8 T-cell responses (data not shown).

As previous data suggested that HIV-1-specific CD8 T-cell responses preferentially recognize conserved regions within the HIV-1 proteome, we calculated entropy scores [¹⁶] for each of the 410 overlapping peptides used to determine the impact of antigen variability on the characterization of HIV-1-specific T-cell responses in chronic and primary infection. The median entropy score (degree of sequence variability) of the entire panel of overlapping peptides spanning the HIV-1 clade B consensus sequence used was 0.16 (range, 0-1.1826). Overall, the median sequence variability of viral peptides targeted during chronic infection was significantly lower than the median sequence variability of peptides targeted during primary infection (entropy scores of 0.14 versus 0.18, respectively, P = 0.046), suggesting a differential impact of viral entropy on the recognition of clade B consensus peptides in acute and chronic infection. In addition, and in line with previous reports, the entropy of peptides targeted by CD8 T cells in individuals with chronic infection was significantly lower (median entropy score of 0.14; range, 0.007-0.68) than the entropy of the non-targeted peptides (median entropy score of 0.19; range, 0-1.18, P = 0.00017). In contrast, the entropy of the recognized and non-recognized peptides did not differ significantly during primary infection [median entropy scores of 0.18 (range, 0.03-0.46) versus 0.16 (range, 0-1.18), P = 0.55; Fig. 1a]. These data suggest that the entropy of HIV-1 does not significantly determine the pattern of clade B consensus peptides recognized during primary infection, which is in distinct contrast to the pattern seen in chronic infection.

Fig. 1

Preferential recognition of HIV-1 Nef clade B consensus sequences in acute HIV-1 infection

Given that viral entropy did not seem to significantly determine the recognition of HIV-1 clade B consensus peptides in primary HIV-1 infection, we evaluated the protein specificity of the individual epitopic responses. Strikingly, CD8 T-cell responses during acute infection were almost entirely directed against the viral Nef protein. Seventy percent of all epitopic regions targeted during acute infection were located within the Nef protein, accounting for 94% of the total magnitude of virus-specific CD8 T-cell responses (Fig. 1b). The preferential recognition of HIV-1 Nef clade B consensus sequences during acute HIV-1 infection was further emphasized when the total magnitude of virus-specific CD8 T-cell responses targeting any given HIV-1 protein was normalized for the protein length (138 SFC/10⁶ PBMC per 100 amino acids in Nef, compared with 1.6 and 1.16 SFC/10⁶ PBMC per 100 amino acids in Gag and Env, respectively; P < 0.01, ANOVA significance test).

In individuals with early HIV-1 infection, Nef was also the most frequently recognized HIV-1 protein within the clade B consensus sequence, accounting for 28% of all epitopes targeted (Fig. 1b). The magnitude of CD8 T-cell responses directed against Nef constituted 46% of the total magnitude of HIV-1-specific CD8 T-cell responses in this group, but did not reach statistical significance when compared with the magnitude of CD8 T-cell responses directed against Gag, Pol and Env and any other HIV-1 protein (P = 0.125). Yet, when evaluating the magnitude of CD8 T-cell responses after adapting to the length of the targeted protein, the number of SFC/10⁶ PBMC per 100 amino acids in Nef significantly exceeded the corresponding number of SFC in Gag, Pol, Env or any other HIV-1 protein (P < 0.01, multiparametric ANOVA significance test). In individuals with chronic HIV-1 infection, virus-specific CD8 T-cell responses were broadly diversified over the entire clade B consensus sequence, with most of the targeted epitopes being localized in the structural proteins Gag, Pol and Env (Fig. 1b).

Interestingly, the Nef epitopes that were most frequently recognized during acute infection clustered predominantly in two different central regions of the protein, ranging from amino acid (aa) 65 to aa 105 and aa 112 to aa 148, respectively. These viral regions had the highest degree of sequence conservation within the Nef protein sequence. Yet, when grading all 410 overlapping HIV-1 peptides used according to their entropy score, these most frequently recognized Nef peptides ranked in the second and third quartile (median rank, 269.5; range, 182-355), indicating that in comparison with the entire HIV-1 proteome, these Nef regions represent areas with relatively higher sequence diversity. Taken together, these data illustrate that despite a relatively high degree of genetic diversity, HIV-1 Nef clade B consensus sequences are preferentially recognized by CD8 T cells in patients with acute and early HIV-1 infection, but not in individuals with chronic infection.

Preferential targeting of HIV-1 Nef is preserved by early initiation of antiretroviral therapy, but rapidly shifted towards Gag, Pol and Env in the presence of continuous antigen exposure

To determine if the immunodominance of Nef observed during acute and early HIV-1 infection is maintained over the ensuing months, we analyzed the HIV-1-specific CD8 T-cell responses over a period of 12 months after the putative time of infection, again using the consensus clade B sequence as a reference strain (Fig. 2). Overall, we observed a moderate three- to four-fold increase of the total magnitude of HIV-1-specific CD8 T-cell responses in patients with primary infection who subsequently started antiretroviral combination therapy. In contrast, the magnitude of HIV-1-specific CD8 T-cell responses increased by seven- to eight-fold in patients who did not initiate antiretroviral treatment following primary infection. In individuals with acute HIV-1 infection who subsequently started antiretroviral therapy, HIV-1 Nef remained the immunodominant HIV-1 protein during the entire time of observation. However, although the absolute magnitude of Nef-directed CD8 T-cell responses remained considerably stable over the 12 months of observation (Table 2), the contribution of Nef-specific responses to the total virus-specific CD8 T-cell responses declined from 94% at baseline to 50% after 12 months of follow-up (Fig. 2a), reflecting the evolution of broader CD8 T-cell responses against other HIV-1 proteins. In individuals with early HIV-1 infection, the initial immunodominance of Nef was switched to the structural HIV-1 proteins Gag, Env and Pol within 1-2 months, irrespective of antiretroviral treatment initiation (Fig. 2b and c). Interestingly, the genetic diversity of peptides recognized 1 year after infection was significantly lower than the median entropy of peptides recognized during primary infection [median entropy score of 0.14 (range, 0.007-0.407) versus 0.18 (range, 0.03-0.46); P = 0.03].

Fig. 2

Table 2

Close correlation between the relative proportion of Nef-specific CD8 T-cell responses and the duration of HIV-1 infection

The rapid shift in the contribution of clade B consensus Nef-specific CD8 T-cell responses to the total magnitude of HIV-1-specific CD8 T-cell responses following acute infection prompted us to determine if the level of Nef immunodominance during primary HIV-1 infection correlates with the time passed after viral infection. Therefore, in the study subjects who mounted Nef-directed responses during primary HIV-1 infection and/or in the subsequent treatment-naive follow-up time, we correlated the total proportion of CD8 T cells targeting Nef to the number of days passed since the putative time of HIV-1 infection (Fig. 3). We observed a strong correlation between those two parameters (R = 0.8, P < 0.01), indicating that the contribution of Nef-specific CD8 T-cell responses to the total HIV-1-specific CD8 T-cell responses may serve as an indicator of the temporal stage of primary HIV-1 infection.

Fig. 3

Discussion

Recent data demonstrate that HIV-1-specific CD8 T-cell responses in acute HIV-1 infection are important for the initial control of viral replication, whereas their relevance for viral containment during chronic infection is less apparent. A better understanding of the specificities of these responses in acute infection is important for understanding the early immune reactions at the time when viral load is diminishing from peak values. This is a difficult issue to address, as comprehensive analysis of immune responses to autologous virus is not readily obtainable in large cohort studies. However, one can begin to address the relative kinetics of induction epitopes by using a reference strain of virus, and compare the earliest targeted epitopes to those that arise later. As one can compare relative induction of responses to different proteins as well as the relative entropy of the peptides targeted, this approach allows one a relative assessment of immunogenicity of regions of similar entropy.

Here, we used a comprehensive panel of overlapping peptides derived from the clade B 2001 consensus sequence to determine virus-specific CD8 T-cell responses in a cohort of subjects with primary HIV-1 infection. We show that the CD8 T-cell-mediated recognition of clade B consensus peptides during primary and chronic HIV-1 infection does not only differ with regard to the overall breadth and magnitude, but also with reference to the HIV-1 protein specificity. Interestingly, viral epitopes within Nef were the first targets for HIV-1-specific CD8 T-cell responses within the clade B consensus sequence during primary HIV-1 infection, despite the fact that the Nef regions targeted were of higher sequence variability than other non-targeted or later targeted regions. In contrast, CD8 T-cell immune responses in chronic infection were broadly diversified, with only 17% of the total virus-specific responses targeting Nef.

To date, studies of T-cell responses in primary HIV-1 infection have mainly relied on the assessment of virus-specific responses using selected described optimal T-cell epitopes [^17-20] or peptides spanning viral sequences from laboratory HIV-1 strains [^15,21-24]. These studies have demonstrated weak and narrowly directed HIV-1-specific CD8 T-cell responses in acute infection, but differed in the description of the specificity of immunodominant HIV-1-specific CD8 T cells in primary HIV-1 infection, potentially due to screening with viral peptide sequences that were largely derived from chronically HIV-1 infected individuals [¹⁴]. Here, we demonstrated that in individuals with primary HIV-1 infection, CD8 T-cell responses against the HIV-1 clade B consensus sequence preferentially recognize the central region of Nef. Although this region has been shown previously to be highly immunogenic [^24,25], the relative kinetics of induction of responses to this region compared to others has not been determined. In individuals with acute infection, Nef was clearly the most frequently recognized HIV-1 protein, with more than 90% of the total HIV-1-specific T-cell responses directed against this relatively small protein. In study subjects with early infection, Nef also represented the immunodominant HIV-1 protein; yet, the total proportion of CD8 T cells recognizing a Nef epitope was less pronounced. Overall, the degree of Nef immunodominance was closely linked to the temporal stage of primary HIV-1 infection, with subjects being diagnosed most recently after virus infection exhibiting the strongest levels of Nef immunodominance. HAART initiation during acute HIV-1 infection preserved the dominance of HIV-1-Nef-specific CD8 T-cell responses, while the delay of therapy commencement resulted in a rapid diversification of virus-specific CD8 T-cell responses.

A number of recent reports have demonstrated that HIV-1-specific CD8 T cells in chronic HIV-1 infection preferentially target regions of the HIV-1 clade B consensus sequence with low overall sequence diversity [^14,16,26]. In line with these studies, HIV-1 clade B consensus peptides with low entropy were preferentially targeted in the 10 subjects with chronic infection in this study. In contrast, the peptide entropy did not significantly impact the localization of CD8 T-cell responses against the clade B consensus sequence in primary infection; instead, the pattern of recognized peptides seemed to be determined rather by their protein specificity. The preferential targeting of HIV-1 Nef clade B consensus sequences observed in this study might result from a variety of mechanisms. The Nef mRNA does not require additional splicing and can thus be rapidly transported to the cytoplasm in a Rev-independent pathway [²⁷]. Together with the two other unspliced HIV-1 gene products Rev and Tat, HIV-1 Nef is therefore considered an early expressed HIV-1 protein and Nef-derived T-cell epitopes might be presented earlier to CD8 T cells than epitopes derived from the structural proteins [^28,29]. Furthermore, the initial presentation of Nef peptides may not yet be hampered by Nef-mediated MHC-class I down-regulation [^30,31] and might therefore induce stronger responses than viral epitopes presented later. Thus, the early targeting of Nef during primary HIV-1 infection could reflect an early presentation of Nef antigens on MHC class I molecules.

The preferential recognition of an early expressed HIV-1 gene product in acute HIV-1 infection is in line with previous reports from the SIV macaque model, demonstrating early targeting of immunodominant CD8 T-cell epitopes within SIV Tat [¹⁰] and Nef [³²], as well as recent studies in primary HIV-1 infection demonstrating early targeting of Rev and Tat [²²]. Interestingly, in SIV-infected macaques, these immune responses against the early expressed gene products Tat and Nef apparently exerted a dramatic immune pressure, leading to rapid selection of viral escape variants [^9,32]. Our data suggest that the HIV-1 Nef clade B consensus sequence can be preferentially recognized during acute HIV-1 infection in a variety of individuals from the clade B-infected population, despite divergent HLA class I backgrounds. In contrast, the clade B consensus sequences of HIV-1 Tat (median entropy score: 0.25) and Rev (median entropy score: 0.22) were less frequently recognized during acute HIV-1 infection than the central region of Nef (aa 65-148, median entropy score of 0.2), despite a similar genetic diversity (P = 0.3, ANOVA significance test) [¹⁴]. Interestingly, in five of the seven individuals in this study who did eventually develop Tat and Rev-specific CD8 T-cell responses, Nef-specific responses were induced earlier. Thus, among the early expressed HIV-1 proteins, the central region of Nef represented the most frequently recognized protein sequence during primary infection. The induction of immune responses against early-expressed HIV-1 proteins by vaccines has recently been proposed as a potential approach for HIV-1-specific immunization, arguing that the potential early recognition of infected cells during the viral life cycle can optimize the inhibition of virion production [³³]. Our data suggest that immune responses generated by vaccination with Nef-1 clade B consensus sequences could recognize a large variety of different viral sequences circulating in a clade B-infected population, whereas responses directed against the other early expressed proteins Rev and Tat may exhibit a lesser degree of cross-recognition of circulating strains [¹⁴].

Taken together, our data demonstrate that HIV-1-specific CD8 T-cell responses from individuals with acute and early HIV-1 infection preferentially recognize epitopes located within the central regions of HIV-1 Nef, despite a relatively high degree of overall entropy of these regions. Thus, these data suggest that the pattern of recognition of clade B consensus peptides in primary HIV-1 infection is not significantly determined by their genetic variability, but is rather impacted by their protein specificity.

Acknowledgements

Sponsorship: This study was supported by grants from the National Institutes of Health (M.A., T.M.A., E.S.R., B.K. and B.D.W.), the Doris Duke Charitable Foundation (M.A., E.S.R. and B.D.W.) the Foundation for AIDS and Immunology Research (X.G.Y. and M.A.) and the Deutsche Forschungsgemeinschaft (M.L. and M.M.A.).