Human Immunodeficiency Virus Type 1 Escapes from Interleukin-2-Producing CD4⁺ T-Cell Responses without High-Frequency Fixation of Mutations

Abstract

The presence of interleukin-2 (IL-2)-producing human immunodeficiency virus type 1 (HIV-1)-specific CD4⁺ T-cell responses has been associated with the immunological control of HIV-1 replication; however, the causal relationship between these factors remains unclear. Here we show that IL-2-producing HIV-1-specific CD4⁺ T cells can be cloned from acutely HIV-1-infected individuals. Despite the early presence of these cells, each of the individuals in the present study exhibited progressive disease, with one individual showing rapid progression. In this rapid progressor, three IL-2-producing HIV-1 Gag-specific CD4⁺ T-cell responses were identified and mapped to the following optimal epitopes: HIVWASRELER, REPRGSDIAGT, and FRDYVDRFYKT. Responses to these epitopes in peripheral blood mononuclear cells were monitored longitudinally to >1 year postinfection, and contemporaneous circulating plasma viruses were sequenced. A variant of the FRDYVDRFYKT epitope sequence, FRDYVDQFYKT, was observed in 1/21 plasma viruses sequenced at 5 months postinfection and 1/10 viruses at 7 months postinfection. This variant failed to stimulate the corresponding CD4⁺ T-cell clone and thus constitutes an escape mutant. Responses to each of the three Gag epitopes were rapidly lost, and this loss was accompanied by a loss of antigen-specific cells in the periphery as measured by using an FRDYVDRFYKT-presenting major histocompatibility complex class II tetramer. Highly active antiretroviral therapy was associated with the reemergence of FRDYVDRFYKT-specific cells by tetramer. Thus, our data support that IL-2-producing HIV-1-specific CD4⁺ T-cell responses can exert immune pressure during early HIV-1 infection but that the inability of these responses to enforce enduring control of viral replication is related to the deletion and/or dysfunction of HIV-1-specific CD4⁺ T cells rather than to the fixation of escape mutations at high frequencies.

In the typical course of acute human immunodeficiency virus type 1 (HIV-1) infection an initial burst of high-level viremia is reduced by at least 100-fold to a set point level (11, 12). This precipitous drop in viral load is suggestive of a partially effective host immune response to primary HIV-1 infection. Several lines of evidence support an important role for CD8⁺ T cells in suppressing HIV-1 replication in acute infection: principally, the decline in HIV-1 viremia is temporally associated with the emergence of an HIV-1-specific CD8⁺ T-cell response, and the in vivo depletion of CD8⁺ T cells in simian immunodeficiency virus-infected macaques consistently results in elevated viral loads (7, 24, 30). Consistent with the application of effective immune pressure, it has been well established that HIV-1- and simian immunodeficiency virus-specific CD8⁺ T cells drive the emergence and fixation of escape mutations in the epitopes that they target (1, 3, 8, 18, 31, 33, 34). This evidence has contributed to the prioritization of vaccine candidates that elicit potent HIV-1-specific CD8⁺ T-cell responses.

The role of CD4⁺ T-cell responses in the response to acute HIV-1 infection is less clear. There is compelling evidence that CD4⁺ T-cell help may be critical for the establishment of a qualitatively and quantitatively robust CD8⁺ T-cell memory pool for persistent virus infections (4, 9, 17, 37, 39). Furthermore, an important role for CD4⁺ help in maintaining an effective CD8⁺ T-cell response has been established in the lymphocytic choriomeningitis virus model of chronic viral infection (28, 45). Evidence in support of a role for the CD4⁺ T-cell response to HIV-1 infection in suppressing viral replication is derived from studies which demonstrated that a CD4⁺ T-cell response characterized by vigorous proliferation and production of interleukin-2 (IL-2) is associated with control of viremia (6, 35). It has further been demonstrated that the functional defect of CD8⁺ T cells observed in chronic HIV-1 infection can be induced in vitro by the depletion of CD4⁺ T cells or the addition of IL-2-neutralizing antibodies and can be corrected in vivo by vaccine-mediated augmentation of HIV-1-specific CD4⁺ T-cell responses (26). These observations have suggested that an IL-2-producing response may be necessary for controlling viremia. However, in the majority of HIV-1-infected individuals, a qualitative impairment of the HIV-1-specific CD4⁺ T-cell response occurs early after infection, resulting in the loss of proliferative capacity as well as the ability to produce IL-2 (43). This impairment correlates well with levels of antigen and viremia (29). The relationship between viral control and the presence of IL-2-producing HIV-specific CD4⁺ T-cell responses must be interpreted with caution, however, as the causal relationship between these two factors is unclear. The maintenance of an IL-2-producing HIV-1-specific CD4⁺ T-cell proliferative response could simply be the result of control of viremia achieved through another means, rather than causal in the association. Therapeutic administration of IL-2 to chronically infected individuals failed to reveal any clinical benefit, perhaps supporting that IL-2 is a marker, rather than a driver, of immunological control (25). However, it is unclear whether the systemic administration of IL-2 effectively substitutes for the targeted production of IL-2 by HIV-1-specific CD4⁺ T cells.

The fixation of escape mutations in CD4⁺ T-cell epitopes during acute infection would provide direct evidence that CD4⁺ T cells apply immunological pressure against HIV-1. Harcourt et al. identified epitopes targeted by proliferative CD4⁺ T-cell responses in chronically infected individuals and sequenced these epitopes from proviral DNA at multiple time points (16). Variations in these epitope sequences were observed over time, and a minority of these variants failed to stimulate CD4⁺ T-cell lines raised against the index peptide. This study indicated the potential for HIV-1 virus to escape within proviral populations. However, the observation that the majority of emergent variants were still able to stimulate CD4⁺ T-cell responses argues against potent selective pressure for escape mutants (16). A second study examined gamma interferon (IFN-γ)-producing CD4⁺ T-cell responses and contemporaneous circulating virus epitopes in a cohort of chronically infected, untreated, HIV-1-infected individuals. A lack of intrapatient variability within CD4⁺ T-cell epitopes was observed in this study, and while two of four subjects exhibited epitope sequences that differed from the consensus HIV-1 sequence, there was a trend to greater sequence variability outside of epitopic regions, arguing against potent immune pressure (23). These studies support that HIV-1-specific CD4⁺ T-cell responses fail to exert potent selective pressure against cognate epitopes in chronic infection; however, it is difficult to determine whether or not the observed epitopic variations are indicative of relatively weak selective pressures. Since the overall cellular immune response to HIV-1 infection is particularly robust and effective during the acute phase of infection, we examined the kinetics of the HIV-1-specific IL-2-secreting CD4⁺ T-cell-mediated immune response during acute/early HIV-1 infection and studied the effects of this response on circulating plasma viruses.

MATERIALS AND METHODS

Subjects.

Subjects were selected from participants in the Canadian Immunodeficiency Research Collaborative Cohort, Toronto, Canada. Acute HIV-1 infection was defined using standard criteria, as previously described (21), by the presence of symptomatic disease following a high-risk exposure to HIV-1 with a detectable plasma HIV-1 RNA level and either a negative HIV-1 enzyme-linked immunosorbent assay or a positive enzyme-linked immunosorbent assay result with an evolving HIV-1 Western blotting result. This study was approved by the University of Toronto Institutional Review Board and subjects gave written informed consent. The estimated date of infection for OM214 was established based on exposure history, testing history—an indeterminant Western blotting result at the initial visit (2 months after exposure) followed by complete seroconversion on follow-up—and clinical manifestations consistent with primary HIV-1 infection at the initial visit (fever and maculopapular rash). For OM348 the estimated date of infection was based upon conversion serology from negative to indeterminant to positive, as well as exposure history. The date of infection for OM296 was estimated based on testing history as well as a weakly positive Western blotting result on the date of diagnosis. The date of infection for OM255 was estimated based on an evolving positive antibody test (Western blotting), a documented negative test 3 months prior, and a seroconversion-like illness at the date of diagnosis.

HIV-1 Gag-specific CD4⁺ T-cell cloning.

Leukopheresis material was obtained from acutely HIV-1-infected individuals within 1 week of initial presentation, and peripheral blood mononuclear cells (PBMC) were isolated by Ficoll density gradient centrifugation using standard methods. These PBMC were depleted of CD8⁺ T cells using Dynal beads (Miltenyi Biotec) following the manufacturer's instructions. CD8⁺-depleted PBMC were stimulated for 16 h with 10 μg/ml of recombinant HIV-1 p55 (Gag) (Austral Biologicals) at 37°C, 5% CO₂. PBMC were enriched for cells producing IL-2 in response to p55 stimulation by using the IL-2 secretion assay enrichment and detection kit (Miltenyi Biotec). Positively selected cells were plated at 200, 50, and 10 cells/well on 96-well plates with feeder medium: 1 × 10⁶ allogeneic PBMC/ml, 1 × 10⁵ cells/ml each of two allogeneic B-lymphoblastoid cell line (BLCL) (all irradiated at 5,000 rads) in RPMI with 10% fetal bovine serum (FBS; Gibco), 50 U/ml IL-2 (Hoffmann La-Roche), 1 μM nevirapine, 1 μM zidovudine (AZT), 1 μM lamivudine (3TC) (NIH AIDS Research and Reference Reagent Program). Cells were cultured at 37°C, 5% CO₂ for 14 days and then fed with 50 μl/well RPMI with 10% FBS, 50 U/ml IL-2, 1 μM nevirapine, 1 μM AZT, 1 μM 3TC. Plates were monitored for an additional 14 days, and wells exhibiting growth were fed as above when the medium yellowed. On day 28 postplating a plate was selected which exhibited growth in <1/10 of wells. Putative clones were screened for specificity by IFN-γ enzyme-linked immunospot (ELISPOT) assay using autologous BLCL pulsed with recombinant HIV-1 Gag p55 (Austral Biologicals). Specific clones were expanded in 1 ml of feeder medium (see above) in 24-well plates and subsequently restimulated at 3-week intervals. After one round of expansion, clones were screened against a matrix of 15-mer HIV-1 Gag peptide pools by IFN-γ ELISPOT using standard methods (2). Custom peptides representing truncated versions of 15-mer peptides were manufactured for each clone (see Fig. 2, below). Clones were tested for responsiveness to titrations of these truncated peptides in a second ELISPOT assay. Optimal T-cell epitopes were defined by the shortest peptide that elicited a response with >75% the magnitude of that elicited by the corresponding 15-mer.

FIG. 2.

Responses to optimal CD4⁺ T-cell epitopes in subject OM214 are detectable at low frequencies in acute infection but are lost within 12 months postinfection. (A) Longitudinal clinical data for HIV-1 viral load (branched DNA) and absolute CD4 count for subject OM214, spanning from an estimated 2 to 14 months postinfection. (B to D) Fine mapping of T-cell epitopes recognized by CD4⁺ T-cell clones isolated from OM214. Shown are IFN-γ ELISPOT data depicting responses to a series of truncated peptides. The optimal epitope was defined as the shortest peptide which elicited a response of >75% of the magnitude of that elicited by the corresponding 15-mer. (E) Shown are representative flow cytometry data analyzing IL-2-producing CD4⁺ T-cell responses to optimal epitopes at estimated time points 2 and 12 months postinfection. The population depicted is the CD4⁺ CD8⁻ subset of viable lymphocytes.

MHC-II tetramer analysis.

Two allophycocyanin (APC)-conjugated HLA-DQA1*01/DQB1*05 MHC-II tetramers were obtained from the NIH Tetramer Facility: one presenting the HIV-1 Gag peptide FRDYVDRFYKT and the second presenting the SQDLELSWNLNGLQADLSS peptide derived from the Fc fragment of immunoglobulin E (IgE; negative control tetramer). Staining was performed on either whole cryopreserved PBMC or CD4⁺ T cells enriched using the Easysep method (Stemcell Technologies) with fluorochrome-conjugated antibodies to CD4 and CD3 (BD) and APC-conjugated MHC-II tetramer at 2.0 μg/ml. Samples were analyzed on a FACSCalibur instrument (BD), collecting at least 1 × 10⁵ events, and data were analyzed using FlowJo software (Treestar).

PD-1 and Tim-3 analysis.

Cryopreserved PBMC from OM214 taken an estimated 2 months postinfection were thawed and stained immediately in 0.5% FBS, 2 mM EDTA-phosphate-buffered saline (PBS) with anti-CD4-peridinin chorophyll protein (BD), anti-Tim-3-biotin (R&D Systems), anti-PD-1-fluorescein isothiocyanate (Biolegend), and either anti-CCR5-APC (BD) or 2.0 μg/ml of APC-conjugated HLA-DQA1*01/DQB1*05 FRDYVDRFYKT tetramer. Cells were washed once and then stained with a 1/100 dilution of streptavidin-phycoerythrin (Molecular Probes) in 0.5% FBS, 2 mM EDTA-PBS. Cells were washed two additional times and fixed in 8% formalin. Samples were analyzed on a FACSCalibur instrument (BD), collecting at least 1 × 10⁵ events, and data were analyzed using FlowJo software (Treestar).

Mapping specificities of HIV-1 Gag-specific CD4⁺ T-cell clones. (i) ELISPOT assays.

ELISPOT assays were performed using standard procedures and the Mabtech IFN- γ ELISPOT assay. Peptides were obtained at >90% purity from Mimitope. Unless otherwise noted, peptides were used at a final concentration of 2 μg/ml and HIV-1 p55 (Gag) was used at a concentration of 10 μg/ml. Clones were tested at 500 to 20,000 cells/well with 5 × 10⁵ BLCL/well as antigen-presenting cells. Incubations were performed for 16 h at 37°C, 5% CO_2.

(ii) Intracellular cytokine staining and flow cytometry.

Cryopreserved PBMC were thawed and tested directly. CD4⁺ T-cell clones were mixed at a 4:1 ratio of autologous BLCL to clone cells. In both cases, cells were stimulated with 10 μg/ml of either peptide or whole p55 antigen for 1 h at 37°C, 5% CO₂ in RPMI with 10% FBS. For experiments studying CD107a, fluorochrome-conjugated monoclonal antibodies (MAbs) to this marker were added at initiation of stimulation (BD). Brefeldin A and monensin were then added to a final concentration of 1 μg/ml each, and stimulation was allowed to continue for an additional 15 h. Cells were surface stained in 0.5% FBS, 2 mM EDTA-PBS with fluorochrome-conjugated MAbs to CD4, CD8, and CD69 (BD). Cells were permeablized with cytofix/cytoperm (BD) following the manufacturer's instructions and then stained with fluorochrome-conjugated MAbs to IFN-γ, tumor necrosis factor alpha, and IL-2 (BD). Cells were then washed with 2% paraformaldehyde in PBS and analyzed on a FACSCalibur (BD). At least 1 × 10⁶ events were recorded for studies of PBMC, and at least 1 × 10⁴ events were recorded for studies of clones.

Viral sequencing.

Whole blood was collected by venipuncture into sodium citrate-containing vacutainers (BD). Plasma was isolated by centrifugation at 1,600 × g for 15 min and cryopreserved at −150°C. Thawed samples were diluted to 10 ml in PBS and ultracentrifuged at 81,584 × g for 1.5 h. Pelleted material was resuspended in 1 ml of Trizol (Invitrogen), and RNA was isolated following the manufacturer's instructions. RNA was treated with DNase using DNA-free Turbo (Ambion) following the manufacturer's instructions and then reverse transcribed into cDNA using cloned avian myeloblastosis virus reverse transcriptase (RT; Invitrogen) following the manufacturer's instructions. Ten microliters of RNA solution was used in each reaction mixture. Parallel reactions were set up without the addition of the RT enzyme to serve as controls for the presence of DNA contamination. cDNA was diluted over 10-fold serial dilutions and added at 1 μl/reaction mixture to 50-μl first-round PCR amplification mixtures using platinum Taq DNA polymerase (Invitrogen) following the manufacturer's instructions and the following primers: p17-1, 5′-ACTAGCGGAGGCTAG-3′; p24-5R, 5′-GGTTTCCATCTTCCTGGCAA-3′. Cycling conditions were as follows: 95°C for 5 min and 40 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 2 min. One microliter of each PCR mixture was transferred to a second-round PCR amplification performed under the same conditions using the following primers: HIV-1 Gag p17 sequencing (clone A1) p17-2, 5′-ATGGGTGCGAGAGCGT-3′, and p24-2R, 5′-TGTCAGCTGCTGCTTGCTG-3′; HIV-1 Gag p24 sequencing (clones A2 and B2), p24-2, 5′-CAGCAAGCAGCAGCTGACA-3′, and p24-10R, 5′-CTTGTGGGAAGGCCAGATCTTCC-3′. The resulting PCR products were analyzed by agarose gel electrophoresis. This allowed for a coarse determination of the limiting dilution as that falling between the lowest dilution that resulted in amplification and the next 10-fold dilution which failed to amplify. A second nested PCR assay was performed on cDNA over twofold serial dilutions, each run in triplicate, spanning the coarse limiting dilution range. The dilution which resulted in amplification of <1/5 of replicates was selected as the limiting dilution. This dilution of cDNA was used to perform 60 additional nested PCRs. In each case, amplification was observed in <1/5 of wells. PCR products from these wells were isolated using the QIAquick PCR purification kit (Qiagen) and sequenced bidirectionally using primers p17-2F and p24-2R or p24-2F and p24-10R. Forward and reverse sequences were aligned, and no discrepancies between these sequences were observed in sequences spanning relevant epitopes. Hypermutated sequences were identified using the hypermut tool (www.hiv.lanl.gov) and excluded.

Production of Gag proteins.

First-round PCR products amplified from OM214 plasma virus and corresponding to desired HIV-1 Gag sequences were amplified using the following primers: p17-2, 5′-ATGGGTGCGAGAGCGT-3′; p24-11, 5′-TTATTGTGACGAGGGGTCGT-3′. PCR products were cloned into the pET-SUMO vector and expressed in BL21(DE3) Escherichia coli cells by using the pET-SUMO Champion protein expression system (Invitrogen), following the manufacturer's instructions. Bacteria were lysed using bacterial protein extraction reagent (B-PER; Pierce) supplemented with 200 μg/ml chicken egg white lysozyme, 50 μg/ml DNase I, and 10 mM MgCl₂. Lysates were denatured by the addition of 7 M guanidine-HCl (GuHCl) and clarified by centrifugation at 13,000 × g for 10 min at 4°C. Supernatants were passed through 1-ml nickel columns (B-PER His₆ fusion protein purification kit; Pierce) that had been preequilibrated with 7 M GuHCl. Refolding of proteins was induced by washes with sequentially lower concentrations of GuHCl, followed by PBS. Additional washes and elution were performed following the manufacturer's instructions. Eluted proteins were dialyzed against PBS. Endotoxin was removed using detoxi-gel columns (Pierce), and proteins were filter sterilized prior to use in ELISPOT assays. SUMO-His-tagged C-acetyltransferase was expressed in parallel and employed as a negative control in all immunoassays.

RESULTS

IL-2-producing HIV-1 Gag-specific CD4⁺ T cells can be cloned from acutely HIV-1-infected individuals.

The presence of HIV-1-specific CD4⁺ T cells that produce IL-2 upon stimulation has been associated with control of viremia (35). Thus, to identify responses with the potential to exert immunological pressure on HIV-1 we focused on Gag-specific CD4⁺ T cells with this functionality. These responses are generally present at or below the limit of detection of immune assays. To avoid any ambiguity in our characterization of these responses, we took the approach of cloning out CD4⁺ T cells which produce IL-2 in response to Gag p55 by using a cytokine capture magnetic-activated cell sorting methodology (see Materials and Methods). A total of 10 clones were obtained from ex vivo PBMC taken from four acutely infected individuals at the time of clinical presentation. All individuals were treatment naïve at initial study. Clinical data for these subjects at the time of cloning, as well as estimated dates of infection, are presented in Table 1. Dates of infection were estimated based upon testing and exposure histories, as well as the presentation of clinical manifestations consistent with acute HIV-1 infection (see Materials and Methods). The specificities of Gag-specific CD4⁺ T-cell clones were confirmed with recombinant HIV-1 p55, mapped using a peptide matrix of overlapping 15-mers by ELISPOT, and confirmed with single 15-mer peptides by ELISPOT and intracellular cytokine staining (ICS) flow cytometry. Figure 1 shows representative flow cytometry data, confirming the specificity of a CD4⁺ T-cell clone obtained from subject OM348 for the Gag p2 p7 p1 p6-derived peptide EPIDKELYPLASLRS. The specificities of each of the 10 clones obtained in this study are presented in Table 2.

TABLE 1.

Patient clinical data at time of presentation

Subject ID	Sex	Age (yrs)	Est. time since infection	Viral load (copies/ml)	Absolute CD4 cell count (cells/mm3)
OM214	M	43	60 days	>400,000	310
OM255	M	44	46 days	18,322	370
OM296	M	37	<4 mos prior	10,691	590
OM348	M	36	37 days	79,000	1,240

FIG. 1.

HIV-1 Gag-specific IL-2-producing CD4⁺ T-cell clones can be isolated from individuals acutely infected with HIV-1. CD8-depleted PBMC from acutely infected individuals were stimulated for 16 h with HIV-1 Gag p55. CD4⁺ T cells that produced IL-2 in response to this stimulation were enriched using a magnetic-activated cell sorting bead capture methodology. A total of 10 Gag-specific clones were isolated from four individuals. Shown are representative flow cytometry data depicting IL-2 production and degranulation (CD107a) in response to the HIV-1 Gag-derived peptide sequence EPIDKELYPLASLRS by a CD4⁺ T-cell clone isolated from subject OM348. 15-Mer T-cell determinant sequences for each of the 10 clones are given in Table 1.

TABLE 2.

Epitopes targeted by IL-2-producing CD4⁺ T-cell clones

Subject ID	Epitope (Gag region[s])
OM214	HIVWASRELERFAVN (p17)
	FRDYVDRFYKTLRAE (p24)
	REPRGSDIAGTTSTL (p24)
OM255	WIILGLNKIVRMYSP (p24)
	AMQMLKETINEEAAE (p24)
	FRDYVDRFYKTLRAE (p24)
OM296	EELRSLYNTVATLYC (p17)
	AMQMKLETINEEAAE (p24)
	VDRFYKTLRAEQASQ (p24)
OM348	EPIDKELYPLASLRS (p2 p7 p1 p6)

IL-2-producing HIV-1 Gag-specific CD4⁺ T cells were detectable at low frequencies at 2 months but not at 12 months postinfection in OM214.

OM214 was first sampled an estimated 60 days postinfection (see Materials and Methods for details), at which time viral load was determined to be 400, 000 copies/ml and CD4 count was 310 cells/mm³. Longitudinal clinical data for OM214 are given in Fig. 2A. At 3 months postinfection, viral load had decreased to 100, 000 copies/ml and CD4 count had recovered to 469 cells/mm³. At 5 months postinfection, viral load had rebounded to 400, 000 copies/ml and CD4 count had declined to 350 cells/mm³. The subject's CD4 count continued to decline and at 12 months postinfection was 296 cells/mm³. At this point the subject opted to begin a regimen of highly active antiretroviral therapy (HAART). The subject's clinical course is suggestive of an early degree of limited immunological control, which was rapidly lost, and thus presented an opportunity to study the nature of an initially effective immune response to HIV-1 and the mechanism by which that response failed. For this individual, we obtained three CD4⁺ T-cell clones at presentation from ex vivo PBMC. The optimal epitopes targeted by the three Gag-specific CD4⁺ T-cell clones obtained from OM214 at the 2-month postinfection time point were fine-mapped by ELISPOT using a series of truncated peptides. Optimal epitopes, defined as the shortest peptides which elicited a response of >75% the magnitude of that elicited by the 15-mer, were the following: clone A1, HIVWASRELER; clone A2, FRDYVDRFYKT; clone B2, REPRGSDIAGT (Fig. 2B to D). In order to determine the kinetics of CD4⁺ T-cell responses to these epitopes, ex vivo PBMC from 2 and 12 months postinfection were stimulated separately with each of these peptides, as well as with recombinant Gag protein generated using sequences from circulating plasma virus at 2 months postinfection. IL-2 production was measured by ICS flow cytometry, using >10⁶ PBMC per stimulation and costaining for the activation marker CD69 to improve sensitivity. Responses (percent IL-2⁺ CD69⁺ cells) were as follows: at month 2, FRDYVDRFYKT, 0.009%, REPRGSDIAGT, 0.006%, HIVWASRELER, 0.009%, and autologous Gag, 0.012%; at month 12, FRDYVDRFYKT, 0%, REPRGSDIAGT, 0%, HIVWASRELER, 0%, autologous Gag, 0% (Fig. 2E). The extremely low frequencies of these responses illustrate the importance of employing T-cell clones to identify, map, and characterize IL-2-producing CD4⁺ T-cell responses. The lower limit of sensitivity of flow cytometry is generally considered to be around 0.05%, well above the magnitude of these responses; however, the isolation of clones by IL-2 bead capture confirms their presence at this time point. At 12 months postinfection these responses could no longer be detected in PBMC, despite our performing repeated intracellular cytokine staining assays at this time point. IFN-γ production in response to these peptides was similarly not detected by ICS flow cytometry at this later time point (data not shown). Thus, consistent with previous reports, IL-2-producing CD4⁺ T-cell responses were lost with progressive infection.

Emergence of a an escape mutation to an IL-2-producing HIV-1 Gag-specific CD4⁺ T-cell response in HIV-1-infected individual OM214.

We hypothesized that the fixation of escape mutations in these CD4⁺ T-cell epitopes contributed to the loss of responses with progressive infection. To examine this, we sequenced HIV-1 Gag regions containing these epitopes from circulating plasma virus at 2, 5, 7, and 12 months postinfection. We employed a limiting dilution sequence methodology to avoid resampling of single sequences and PCR-induced errors (see Materials and Methods). The sequences of these epitopes at each time point are given in Tables 3 to 5. No variation was observed in the A1 clone epitope HIVWASRELER. An FRDYVDQFYKT variant (change indicated in bold) of the A2 epitope FRDYVDRFYKT was observed at the following frequencies: month 2 postinfection, 0/41; month 5, 1/21; month 7, 1/10; month 12, 0/12. A REPGGSDIAGT variant of the B2 epitope REPRGSDIAGT was found at the following frequencies: month 2, 1/41; month 5, 0/23; month 7, 0/10; month 12, 0/12. A variety of mutations flanking the optimal epitope were observed for each of the three regions, as shown in Tables 3 to 5. Clone A2 failed to recognize the variant FRDYVDQFYKT, while potent recognition of the wild-type FRDYVDRFYKT was observed (Fig. 3A). This pair of peptides was subsequently used to stimulate PBMC from 2 months postinfection, and responses were measured by ICS flow cytometry. As was previously observed, FRDYVDRFYKT elicited a low-frequency IL-2-producing response, whereas FRDYVDQFYKT failed to elicit a response as measured by IL-2 production (Fig. 3B), degranulation (CD107a staining), or IFN-γ production. Thus, the FRDYVDQFYKT variant of the epitope targeted by clone A2 constitutes an escape mutant from a CD4⁺ T-cell response.

TABLE 3.

Sequences of epitopes targeted by clone A1 in circulating plasma viruses in OM214

Epitope sequencea	Ratio at mo:
	2	5	7	12
RLRPGGKKKYRLKHIVWASRELERFAVNPGLLESAS	23/25	17/18	3/10	0/15
RLRPGGKKrYRLKHIVWASRELERFAVNPGLLESAS	0/25	0/18	0/10	10/15
qLRPGGKKKYRLKHIVWASRELERFAVNPGLLESAS	0/25	1/18	3/10	1/15
RLRPGGrKKYRLKHIVWASRELERFAVNPGLLESAS	0/25	0/18	0/10	1/15
RLRPGGKKKYRiKHIVWASRELERFAVNPGLLESAS	0/25	0/18	0/10	1/15
RLRPGGKKrYRiKHIVWASRELERFAVNPGLLESAS	0/25	0/18	0/10	2/15
RLRPGGKKqYRLKHIVWASRELERFAVdPGLLESAS	1/25	0/18	0/10	0/15
RLRPGGKKKYkLKHIVWASRELERFAVNPGLLEtse	1/25	0/18	1/10	0/15
RLRPGGKKKYRLKHIVWASRELERFAVdPGLLESAS	0/25	0/18	1/10	0/15
RLRPGGqKKYRLKHIVWASRELERFAVNPGLLESAS	0/25	0/18	2/10	0/15

^aPortions shown in lowercase indicate sites of sequence variation, and the underlined portions indicate optimal epitopes. The ratios in Tables 3 to 5 are the given epitope sequence/total sequences obtained.

TABLE 4.

Sequences of epitopes targeted by clone A2 in circulating plasma viruses in OM214

Epitope sequencea	Ratio at mo:
	2	5	7	12
TSILDIRQGPKEPFRDYVDRFYKT	20/41	0/21	0/10	0/12
aSILDIRQGPKEPFRDYVDqFYKT	0/41	1/21	1/10	0/12
aSILDIRQGPKEPFRDYVDRFYKT	21/41	20/21	9/10	12/12

^aPortions shown in lowercase indicate sites of sequence variation, and the underlined portions indicate optimal epitopes.

TABLE 5.

Sequences of epitopes targeted by clone B2 in circulating plasma viruses in OM214

Epitope sequencea	Ratio at mo:
	2	5	7	12
MREPRGSDIAGT	21/41	0/23	0/10	0/12
MREPgGSDIAGT	1/41	0/23	0/10	0/12
iREPRGSDIAGT	19/41	23/23	10/10	12/12

^aPortions shown in lowercase indicate sites of sequence variation, and the underlined portions indicate optimal epitopes.

FIG. 3.

The FRDYVDRFYKT-to-FRDYVDQFYKT mutation observed in circulating virus sequences in OM214 confers escape from clone A2. Mutations flanking clone A2 and B2 epitopes do not confer escape. (A) The CD4⁺ T-cell clone A2 was combined with autologous BLCL pulsed with either FRDYVDRFYKT or FRDYVDQFYKT peptide, and responsiveness was assessed by ELISPOT and flow cytometry. Shown are flow cytometry data, gated on the CD4⁺ T-cell clone population, depicting IL-2 (x axis) and IFN-γ (y axis) production in response to peptide. (B) OM214 PBMC sampled at an estimated 2 months postinfection were analyzed for IL-2-producing responses to wild-type FRDYVDRFYKT peptide and escape mutant FRDYVDQFYKT peptide by flow cytometry. Shown are representative flow cytometry plots, gated on the CD4⁺ CD8⁻ population, depicting CD69 staining (x axis) by IL-2 staining (y axis). (C) Recombinant HIV-1 Gag p55 was produced in E. coli by using Gag sequences cloned from circulating OM214 virus. HIV-1 Gag p55 containing mutations flanking the epitope recognized by clones A2 and B2 was compared with wild-type Gag p55 for the ability to stimulate these clones in an IFN-γ ELISPOT assay. Shown are representative ELISPOT data for clone B2 tested against wild-type (wt) p55 and a flanking mutation (fm) containing p55 at 10 and 1 μg/ml.

We also observed mutations outside of the optimal epitopes for clones A1, A2, and B2 which could potentially reduce recognition by interfering with epitope processing or binding to MHC-II molecules. To test for this, we cloned full-length HIV-1 Gag p55 coding sequences from plasma viruses in OM214 at 2 and 7 months postinfection into the pET-SUMO bacterial expression construct. Two constructs were selected for expression, one of which contained the TSILIDIRQGPKEPFRDYVDRFYKT-extended A2 epitope wild-type (wt) sequence and the MREPRGSDIAGT-extended B2 epitope wt sequence, and the second of which contained the ASILDIRQGPKEPFRDYVDRFYKT-extended A2 epitope flanking mutation (fm) sequence and the IREPRGSDIAGT-extended B2 epitope fm sequence. These proteins were purified (see Materials and Methods) and tested for their ability to stimulate clones A2 and B2 in ELISPOT assays. Figure 3C shows a representative result for clone B2 tested against wt and fm versions of p55 at 10 μg/ml and 1 μg/ml. No differences were observed in the degree of recognition of these proteins by either the A2 or the B2 clones. Thus, the flanking mutations to the p24 epitopes identified in this study did not confer escape from CD4⁺ T-cell responses. Clone A1, after multiple restimulations in culture, lost the ability to respond to its cognate peptide before the fm-containing recombinant p17 could be expressed, and so it was not tested.

Mutations in epitopes targeted by CD8⁺ and CD4⁺ T cells can result in altered peptide ligands (APLs) that are able to engage the T-cell receptor but fail to stimulate the T cell (5, 10, 22, 36, 38). Through this phenomenon of T-cell antagonism, an APL can interfere with the recognition of wild-type epitopes. Thus, it was conceivable that the FRDYVDQFYKT epitope expressed by a minority of circulating viruses could interfere with T-cell recognition of the predominant FRDYVDRFYKT sequence. We tested this in an ELISPOT assay by comparing the responsiveness of clone A2 to either 1 μg/ml of whole p55 antigen or 5 μg/ml of FRDYVDRFYKT peptide in isolation, compared to that in the presence of FRDYVDQFYKT peptide added in twofold serial dilutions ranging from 10 μg/ml to 10 ng/ml. No impairment in responsiveness to either p55 or FRDYVDRFYKT was observed with addition of FRDYVDQFYKT, thus FRDYVDQFYKT does not act as an APL for clone A2 (Fig. 4).

FIG. 4.

The FRDYVDQFYKT peptide does not antagonize responses to FRDYVDRFYKT. Shown are summary data from an IFN-γ ELISPOT assay depicting total number of spots (y axis) elicited by stimulation of 500 cells/well of clone A2 with autologous BLCL pulsed with either 1 μg/ml of HIV-1 Gag p55 or 5 μg/ml of FRDYVDRFYKT peptide in the context of various amounts of FRDYVDQFYKT peptide (x axis).

HIV-1 Gag FRDYVDRFYKT-specific CD4⁺ T cells displayed markers associated with T-cell exhaustion and were lost from the peripheral blood with disease progression.

The observed loss of CD4⁺ T-cell responses to the HIV-1 Gag FRDYVDRFYKT epitope in OM214 could be due to either deletion of cells of this specificity, loss of responsiveness of cells of this specificity, or a combination of both of these factors. To distinguish between these possibilities we enumerated antigen-specific cells in peripheral blood at multiple time points using APC-conjugated MHC class II tetramers. Clone A2 was determined to be restricted by HLA-DQ alleles, as demonstrated by the ability of anti-HLA-DQ antibodies, but not anti-HLA-DR antibodies, to block presentation of antigen. The specific restriction was determined to be HLA-DQA1*01/DQB1*05 by testing a panel of allogeneic BLCL matched with OM214 on at least one DQ allele for the ability to present the FRDYVDRFYKT peptide. The HLA-DQA1*01/DQB1*05 FRDYVDRFYKT tetramer, as well as a negative control HLA-DQA1*01/DQB1*05 tetramer presenting the SQDLELSWNLNGLQADLSS peptide from the Fc fragment of IgE, were manufactured at the NIH tetramer facility. The specificity of the tetramer was tested by examining its binding to the FRDYVDRFYKT-specific clone A2 and, as a negative control, the REPRGSDIAGT-specific clone B2 at concentrations of 2.0 μg/ml and 0.2 μg/ml. We observed dose-dependent binding of the tetramer to the A2 clone, but not the A1 clone (Fig. 5A). We then tested the ability of the FRDYVDRFYKT tetramer, as well as the negative control tetramer, to bind to CD4⁺-enriched cells taken from OM214 at 2 months postinfection. We observed clear staining of 0.24% CD4⁺-enriched cells with the FRDYVDRFYKT tetramer and a background of 0.016% with the negative control tetramer (Fig. 5B). The staining observed on CD4⁺-enriched cells was substantially brighter than that observed on clone A2 cells. This was likely due to TCR downregulation in clone A2, which underwent several rounds of in vitro stimulation with autologous BLCL pulsed with HIV-1 Gag p55 prior to this experiment. We then stained whole PBMC from OM214 taken 2, 6, 11, 14, and 15 months postinfection. At 2 months postinfection 0.064% of CD4⁺ T cells stained tetramer-positive by this method, with a background of 0.019% using the tetramer negative control (Fig. 5B). The background-subtracted frequencies of tetramer-positive cells at each time point are shown in Fig. 5C. Tetramer-positive cells were undetectable in the blood at 6 and 11 months postinfection but reemerged after initiation of HAART at 12 months postinfection to levels of 0.02% and 0.01% at 14 and 15 months postinfection, respectively. These data support that HIV-1 Gag-specific CD4⁺ T cells undergo deletion with disease progression in OM214 but can reemerge upon initiation of HAART.

FIG. 5.

HIV-1 Gag FRDYVDRFYKT-specific CD4⁺ T cells were rapidly lost from blood in subject OM214. Two APC-conjugated HLA-DQA1*01/DQB1*05 MHC-II tetramers were obtained from the NIH Tetramer Facility, one presenting the HIV-1 Gag peptide FRDYVDRFYKT and the second presenting the SQDLELSWNLNGLQADLSS peptide derived from the Fc fragment of IgE (negative control tetramer). (A) Representative flow cytometry data (lowest and highest concentrations) from a titration of the FRDYVDRFYKT peptide containing tetramer against the FRDYVDRFYKT-specific CD4⁺ T-cell clone A2 and the REPRGSDIAGT-specific CD4⁺ T-cell clone A1 (negative control clone). (B) Flow cytometry data showing CD4 staining (x axis) by MHC-II tetramer staining (y axis) for either CD4⁺-enriched cells (top row) or whole PBMC (bottom row) from subject OM214. The left panel shows staining with the SQDLELSWNLNGLQADLSS negative control tetramer, while the right panels show staining with the FRDYVDRFYKT tetramer. (C) Longitudinal data indicating the background-subtracted frequency of FRDYVDRFYKT-specific CD4⁺ T cells in the PBMC of OM214 as measured by flow cytometry (analogous to the lower right plot of panel B). The y axis represents the percentage of cells staining positive for FRDYVDRFYKT minus the percentage of cells staining positive for the SQDLELSWNLNGLQADLSS negative control tetramer.

The observed frequency of tetramer-positive T cells at 2 months postinfection was substantially greater than the frequency of CD4⁺ T cells that produced IL-2 in response to stimulation with the FRDYVDRFYKT peptide, suggesting that cells of this specificity may be functionally impaired. Expression of the transmembrane glycoproteins Tim-3 and PD-1 has been associated with T cells that have become dysfunctional, or “exhausted” in the context of chronic viral infection (13, 19, 20, 32, 40). We analyzed Tim-3 and PD-1 expression on total CD4⁺ T cells, antigen-experienced CCR5⁺ CD4⁺ T cells, and HLA-DQA1*01/DQB1*05 FRDYVDRFYKT tetramer⁺ CD4⁺ T cells in PBMC from OM214 at 2 months postinfection and observed the following frequencies of Tim-3⁺, PD-1⁺/Tim-3⁺, and PD-1⁺ populations: total CD4⁺, 45.7%, 26.6%, and 18.9%; CCR5⁺ CD4⁺, 61.3%, 57.3%, and 42.9%; tetramer⁺ CD4⁺, 82.3%, 67.6%, and 58.8% (Fig. 6). These data support that the low proportion of FRDYVDRFYKT-specific T cells that produce IL-2 upon stimulation is attributable to an advanced stage of T-cell exhaustion.

FIG. 6.

HIV-1 Gag FRDYVDRFYKT-specific CD4⁺ T cells express high levels of the T-cell exhaustion markers Tim-3 and PD-1. PBMC from OM214 taken at an estimated 2 months postinfection were stained with fluorochrome-conjugated antibodies to PD-1 and CD4, as well as biotinylated polyclonal anti-Tim-3. Tim-3 antibody was subsequently labeled with streptavidin-phycoerythrin. These cells were then stained with either fluorochrome-conjugated anti-CCR5 or APC-conjugated HLA-DQA1*01/DQB1*05 FRDYVDRFYKT. Shown are flow cytometry plots depicting Tim-3 staining (y axis) by PD-1 staining (x axis) after gating on total CD4⁺ cells (first panel), CCR5⁺ CD4⁺ cells (middle panel), or tetramer⁺ CD4⁺ cells (third panel).

DISCUSSION

It has previously been observed that the presence of CD4⁺ T cells that produce IL-2 upon stimulation with HIV-1 Gag is associated with the control of HIV-1 viremia. This has lead to the suggestion that CD4⁺ T cells with this functionality may play an important role in an effective cellular immune response to HIV-1 infection. In the present study, we demonstrated that IL-2-producing HIV-1 Gag-specific CD4⁺ T cells can be cloned from individuals with primary HIV-1 infection. The presence of these cells, however, failed to protect these subjects from disease progression. This was particularly clear in the case of subject OM214, who progressed rapidly to a CD4 count of 290 cells/mm³ within 12 months of infection, despite the presence of IL-2-producing CD4⁺ T-cell responses to at least three HIV-1 Gag epitopes during primary infection.

The accumulation of escape mutations within CD4⁺ T-cell epitopes presents one possible explanation for the failure of these responses to protect from disease progression. We observed the emergence of a viral variant containing an R-to-Q mutation at Gag amino acid position 167 in the FRDYVDRFYKT epitope at a frequency of 1/21 at 5 months postinfection and was maintained at a frequency of 1/10 at 7 months postinfection. The mutant epitope FRDYVDQFYKT did not antagonize responses to its wt counterpart, and mutations flanking the epitope did not impact upon recognition. Furthermore, no mutations were observed within the REPRGSDIAGT or HIVWASRELER epitopes at either 5, 7, or 12 months postinfection. Thus, the inability of IL-2-producing CD4⁺ T-cell responses to support enduring control of HIV-1 replication cannot be primarily attributed to the fixation of escape mutations.

The emergence and maintenance of this escape mutation, albeit at a low frequency, does however provide evidence consistent with the application of some degree of immunological pressure by this IL-2-producing CD4⁺ T-cell response. This mutation falls within the major homology region of HIV-1 Gag, a 20-amino-acid region which is conserved across distantly related retroviruses (41). R167, the residue mutated in FRDYVDRFYKT to give FRDYVDQFYKT, is one of the four most highly conserved residues within the major homology region, which together form a hydrogen bond network that stabilizes the conformation for facilitating Gag-Gag interactions during particle formation (15). Conservative mutations in R167, such as R to K, result in viruses with delayed in vitro replication kinetics compared to the wild-type, while nonconservative mutations, such as R to A, completely abolish viral replication (27). The highly nonconservative mutation of the basic R residue to an acidic Q residue observed in OM214 is therefore likely to carry a substantial fitness cost to the mutant virus and provides evidence for selective pressure applied against this site. We observed a lack of cytotoxic CD8⁺ T cell responses to either of the three overlapping 15-mer peptides containing R167 at both 2 and 5 months postinfection, arguing against cytotoxic T-cell-mediated pressure (data not shown). Thus, our observations are consistent with the application of immunological pressure by an HIV-1 Gag-specific CD4⁺ T-cell response.

The failure of this mutant virus to become the predominant circulating quasispecies may be related to a high fitness cost for mutating this highly conserved R167 residue. However, it may also be related to a fundamental difference in the mode of action of CD4⁺ versus CD8⁺ T cells which imposes limitations on the ability of CD4⁺T cells to apply direct immunological pressure. CD8⁺ T cells act directly against the infected cells that present their cognate epitope in the context of MHC class I on their surface, driving either direct lysis of these cells or the suppression of viral replication through noncytopathic mechanisms. Thus, there is a direct feedback mechanism that confers a selective advantage to escape variants, as cells infected by these viruses will not be targeted by these CD8⁺ T cells. CD4⁺ T cells, in contrast, act primarily by providing help to CD8⁺ T cells or B cells after recognizing their cognate epitope presented by MHC class II on either a virus-infected cell or a cell that has endocytosed exogenous viral antigens. These signals are provided either directly or through an intermediate cell type, such as a dendritic cell. The CD8⁺ T cells or B cells subsequently perform effector functions on viruses or virus-infected cells that they recognize by a distinct, MHC class I-restricted epitope or an antibody epitope. Thus, in contrast to CD8⁺ T-cell epitope escape, the emergence of a CD4⁺ T-cell epitope escape virus would be expected to confer a general survival advantage to all local viral variants, rather than feeding back specifically as an advantage to the escaping virus variant. There is, nonetheless, conclusive evidence that CD4⁺ T cells can specifically drive the fixation of escape mutations in viral epitopes in the lymphocytic choriomeningitis virus model (10). One possible mechanism by which this occurs is through direct cytotoxic effector functions of CD4⁺ T cells. There is also precedence, however, in models of evolutionary biology for CD4⁺ T cells acting in a helper capacity to effectively exert evolutionary pressure through interdemic selection. If one considers a population of HIV-1 in a single compartment, such as a lymph node, as a deme, or semiinsular population, the random acquisition of a CD4⁺ T-cell escape mutation within this population could confer a survival advantage to this deme as a whole by lessening the total degree of CD4 help. Migration from this higher fitness deme into demes of lower mean fitness (such as other lymph nodes) could drive a net selection for CD4⁺ T-cell epitope escape mutations throughout the host (42). The degree to which such interdemic pressures can apply selective pressure on CD4⁺ T epitope in vivo is unknown. These considerations caution against extrapolating relative degrees of immunological pressure applied by CD4⁺ versus CD8⁺ T cells to the relative importance of these populations in the immune response to HIV-1 infection.

The absence of IL-2-producing HIV-1 Gag-specific CD4⁺ T-cell responses, which has previously been reported in chronic infection, was reflected in the present study by the loss of detectable responses to all three epitopes within a year of primary infection in individual OM214. We reasoned that this loss was due to either deletion of HIV-1 Gag-specific CD4⁺ T cells, to loss of functionality of these cells, or to a combination of these two factors. In order to study this we obtained an MHC-II tetramer presenting the FRDYVDRFYKT epitope recognized by OM214 clone A2. We observed high levels of expression of the T-cell exhaustion-associated markers Tim-3 and PD-1 on tetramer-positive cells at 2 months postinfection. Cells of this specificity were lost rapidly, with levels becoming undetectable by flow cytometry at 5 months postinfection and remaining undetectable until the initiation of HAART at 12 months postinfection. Previous studies which have shown that HIV-1-specific CD4⁺ T cells are preferentially infected by HIV-1 and are preferentially prone to apoptosis support deletion of HIV-1-specific CD4⁺ T cells as the etiology of this loss; however, we cannot strictly rule out migration of antigen-specific cells out of the blood and into other compartments (14, 44). Upon initiation of HAART, we observed the reemergence of this HIV-1 Gag-specific CD4⁺ T-cell population in the blood.

Taken together, our results clearly demonstrate that the presence of HIV-1-specific IL-2-producing CD4⁺ T cells in acute infection is not sufficient to maintain enduring immunological control of viral replication. Progression occurred in each of the three subjects from whom we identified IL-2-producing CD4⁺ T-cell clones. As demonstrated by longitudinal viral sequencing in OM214, the failure of the immune responses to primary HIV-1 infection to establish control cannot be solely attributed to immune escape from these responses. These data support that, in the observed association between long-term immune control of HIV-1 infection and the presence of IL-2-producing CD4⁺ T cells, the latter may be maintained as a result of low levels of viremia rather than playing the dominant role in dictating clinical course. The mechanisms by which these responses are lost in progressive infection are multifactorial and likely include a component of T-cell exhaustion as well as deletion of HIV-1-specific cells. Our confirmation, at multiple time points, of a CD4⁺ T-cell response escape mutation in a functionally critical region of HIV-1 Gag does, however, provide evidence for an important role for HIV-1-specific CD4⁺ T-cell responses in acute HIV-1 infection. These data support that vaccination and immunotherapeutic strategies that induce HIV-1-specific CD4⁺ T-cell responses may result in initial immunological pressure against HIV-1 but that these responses are likely to be very transient in the face of ongoing viral replication. The potential benefit of such therapies, either in prophylactic or therapeutic settings, must be weighed carefully against the potential disadvantage of providing a prime pool of target cells for HIV-1 infection.