T-Cell Receptor-Mediated Anergy of a Human Immunodeficiency Virus (HIV) gp120-Specific CD4⁺ Cytotoxic T-Cell Clone, Induced by a Natural HIV Type 1 Variant Peptide

Abstract

Human immunodeficiency virus type 1 (HIV-1) infection triggers a cytotoxic T-lymphocyte (CTL) response mediated by CD8⁺ and perhaps CD4⁺ CTLs. The mechanisms by which HIV-1 escapes from this CTL response are only beginning to be understood. However, it is already clear that the extreme genetic variability of the virus is a major contributing factor. Because of the well-known ability of altered peptide ligands (APL) to induce a T-cell receptor (TCR)-mediated anergic state in CD4⁺ helper T cells, we investigated the effects of HIV-1 sequence variations on the proliferation and cytotoxic activation of a human CD4⁺ CTL clone (Een217) specific for an epitope composed of amino acids 410 to 429 of HIV-1 gp120. We report that a natural variant of this epitope induced a functional anergic state rendering the T cells unable to respond to their antigenic ligand and preventing the proliferation and cytotoxic activation normally induced by the original antigenic peptide. Furthermore, the stimulation of Een217 cells with this APL generated altered TCR-proximal signaling events that have been associated with the induction of T-cell anergy in CD4⁺ T cells. Importantly, the APL-induced anergic state of the Een217 T cells could be prevented by the addition of interleukin 2, which restored their ability to respond to their nominal antigen. Our data therefore suggest that HIV-1 variants can induce a state of anergy in HIV-specific CD4⁺ CTLs. Such a mechanism may allow a viral variant to not only escape the CTL response but also facilitate the persistence of other viral strains that may otherwise be recognized and eliminated by HIV-specific CTLs.

Despite a functional dysfunction of CD4⁺ T helper lymphocytes, individuals infected with the human immunodeficiency virus type 1 (HIV-1) respond very rapidly to the presence of the retrovirus by a strong increase in HIV-1-specific cytotoxic T lymphocytes (CTLs). The appearance of this CTL response is associated with a sharp decline in the initial viremia (2). Although vigorous during the early stages of infection, this cell-mediated cytotoxic response appears unable to contain the virus, which reemerges at the end of a long asymptomatic period, this time unchallenged by the immune system (39). The mechanism(s) by which the virus escapes the immune system is now the focus of intense investigation.

From the onset of the CTL response, a war in adaptability is waged between the patient's immune system and the virus. Rapidly, CTL clones responding best to certain HIV epitopes are selected over other clones responding less vigorously to the same epitopes or responding to functionally less important epitopes. Conversely, the virus responds to this immune pressure by mutating or deleting the epitope(s) targeted by these CTLs (28, 53). It is estimated that at least 100 million mutant viruses can be produced every day, as a result of a high replication rate and the poor accuracy of the viral reverse transcriptase (18, 52, 53), exposing the immune system of the HIV-infected patient to every possible point mutation on a daily basis (39). The emergence of mutated CTL epitopes may therefore facilitate the persistence, and ultimate escape, of a broad range of HIV-1 variants (28, 30). For example, variations in key CTL epitopes may prevent either major histocompatibility complex (MHC) binding and presentation of viral peptides (44) or the recognition of MHC-bound variant peptides by the T-cell receptor (TCR) of HIV-specific CTLs. The complete deletion of a CTL epitope has also been described (26). Finally, TCR antagonism may also play an important role in the persistence of HIV-1. It was shown recently that naturally occurring variations in HIV-1 Gag CTL epitopes generated viral altered peptide ligands (APLs) that antagonized CTL activation by the original peptide epitope (1, 25, 49, 56) when both variant and native epitopes were presented to the CTL.

Virus-specific CTLs are frequently of the CD8⁺ surface phenotype, although HIV-specific CTLs of the CD4⁺ phenotype have also been described. CD4⁺, MHC class II-restricted CTL clones have been derived by in vitro stimulation of peripheral blood T cells from a normal individual with purified recombinant envelope glycoproteins (57) and have also been isolated from the blood of individuals vaccinated with HIV-1 envelope proteins (43). HIV-1 Gag- and gp120-specific CD4⁺ CTLs have also been recovered from HIV-1-infected patients (21, 22) as well as healthy individuals immunized with a recombinant gp160-derived vaccine (10, 16, 35). These observations indicate that HIV-derived peptide epitopes can be presented by class II MHC gene products and induce a class II MHC-restricted CTL response. It is, however, unclear whether these CD4⁺ cytotoxic T cells truly participate in the antiviral cytotoxic response.

T lymphocytes recognize short peptides, presented by dedicated antigen-presenting cells (APCs), in association with the products of the MHC. The TCR is composed of an antigen-specific αβ heterodimer that interacts with the peptide antigen-MHC complex. The TCR is expressed at the cell surface in noncovalent association with the CD3 and ζ chains, which are responsible for transducing the extracellular signal that results in T-cell activation (reviewed in reference 66). The CD3 γ, δ, and ɛ molecules each contain a single signaling module, called an immunoreceptor tyrosine-based activation motif (ITAM), characterized by the consensus sequence Yxx(L/I)x_6–8Yxx(L/I), whereas the ζ chain contains three such ITAMs (51). Upon ligand binding, the tyrosine residues of the ITAMs become phosphorylated, possibly by the two members of the Src family of nonreceptor protein tyrosine kinases (PTKs), Lck and Fyn, that are physically associated with the CD4/CD8 surface glycoproteins and the TCR-CD3 complex, respectively. The phosphorylation of the two tyrosine residues of each ITAM allows the modules to serve as docking sites for the tandem Src homology 2 (SH2) domains of the PTK Zap-70, which binds predominantly to the phosphorylated ITAMs of the ζ and CD3ɛ chains (17). This interaction of Zap-70 with phosphorylated ITAMs allows Lck to interact with Zap-70 and phosphorylate regulatory tyrosine residues, resulting in the catalytic activation of Zap-70 (6, 7, 27). Once these TCR-proximal events have taken place, the activated Zap-70 will recruit and phosphorylate diverse cellular substrates, including the adapter proteins LAT (69) and SLP-76 (19, 50, 65), and initiate a signal transduction cascade culminating in T-cell activation and proliferation.

Stimulation of T lymphocytes by antigen peptides (or agonist peptide ligands) specifically recognized by the TCR induces interleukin-2 (IL-2) secretion and proliferation of the T cells (66). Conversely, T-cell stimulation with antigenic peptide variants (or APLs), which may differ from the native antigenic peptide by as little as a single amino acid, can result in the induction of a state of unresponsiveness termed anergy. T-cell anergy can be operationally defined as the state in which a lymphocyte no longer responds to its nominal agonist ligand by IL-2 secretion, proliferation, or functional activation. The molecular events triggered by anergy-inducing APLs and the mechanisms causing sustained T-cell unresponsiveness remain to be defined. However, initial studies have shown that the stimulation of T lymphocytes with anergy-inducing APLs generates modified early signaling events characterized by an incomplete phosphorylation of the ζ chain and the recruitment of inactive and unphosphorylated Zap-70 to the ζ chain (30, 38, 59, 60). This finding suggested that the initial events normally triggered in response to TCR stimulation (e.g., Lck-mediated phosphorylation of the ITAM modules and the subsequent activation of Zap-70) are somehow deficient.

Because of the extreme genetic variability of the viral genome during HIV infection and the knowledge that certain APLs can trigger TCR-mediated T-cell anergy in a variety of systems, we explored the possibility that HIV-specific CTLs could be rendered anergic by TCR stimulation with natural variants of HIV-1. To address this question, we used a human, nontransformed CD4⁺ cytotoxic T-cell clone, Een217, which was developed by in vitro stimulation of nonadhering peripheral blood mononuclear cells (PBMCs), isolated from a healthy donor, with recombinant gp120 and cloning of proliferating clones in soft agar (57). We report that the stimulation of these cells with a naturally occurring variant peptide, which differs from the original peptide antigen by only two amino acids, induced altered TCR-mediated signaling events and a state of unresponsiveness characteristic of anergic cells.

MATERIALS AND METHODS

Cells and peptides.

The Een217 T-cell clone was generated by in vitro stimulation of PBMCs isolated from the blood of a normal, seronegative individual with autologous monocytes pulsed with recombinant gp120 (57). This CD4⁺, HLA-DR4-restricted human cytotoxic T-cell clone is specific for gp120 residues 410 to 429 of HIV-1 strain PV22, presented in association with the class II MHC gene product HLA-DR4. The generation of HLA-DR4-transfected murine L cells, used as antigen-presenting cells in our assays, has also been described (57). Een217 T cells were maintained in culture (RPMI 1640 supplemented with 10% fetal bovine serum, penicillin, and streptomycin [each at 50 μg/ml]) by periodic stimulation with 1 to 2 μg of phytohemagglutinin/ml in the presence of recombinant human IL-2 (rhIL-2; 50 U/ml; Gibco-BRL) and irradiated (5,000 rads) allogeneic PBMCs (2 × 10⁶/well). The cells were rested for 2.5 to 3 weeks prior to experimentation. The peptides representing gp120 residues 410 to 429 of HIV-1 strain PV22 (GSDTITLPCRIKQFINMWQE) and four variant strains; HXB2 (GSDTITLPCRIKQIINMWQK), CDC42 (TGDIITLPCRIKQII-NRWQV), EL1 (TNTNITLQCRIKQIIKMVAG), and Z3 (CTGNITLPCRIKQIINMWQE), were synthesized at the Sheldon Biotechnology Center (McGill University, Montréal, Québec, Canada), using standard solid-phase methods. Crude peptides were purified by reversed-phase high-pressure liquid chromatography. Peptides were analyzed for homogeneity by thin-layer chromatography, and their composition was assessed by amino acid analysis of acid-hydrolyzed peptides. The molecular weight of each peptide was determined by mass spectrometry analysis. All peptides were >90% pure.

Proliferation and cytotoxicity assays.

CD4⁺ APCs (murine L cells stably expressing the human class II MHC molecule HLA-DR4) were irradiated (5,000 rads) and plated in 96 flat-bottomed wells at 3 × 10⁴ cells/well in the presence of culture medium alone or with various concentrations of the indicated peptides at a final volume of 100 μl/well. After overnight incubation at 37°C, the excess peptide was removed and the Een217 T cells were added at 5 × 10⁴ cells/well in a final volume of 200 μl of culture medium without rhIL-2. After 24 h, 1 μCi of [³H]TdR (Mendel Scientific Co., Ltd., St. Laurent, Québec, Canada) was added to each well. Forty-eight hours later, cells were harvested and [³H]TdR incorporation was measured by scintillation counting. All determinations were performed in triplicate. The cytolytic activity was measured in standard ⁵¹Cr release assays. Briefly, target cells (HLA-DR4⁺ L cells; 2 × 10⁶ cells/ml) were pulsed with the indicated concentrations of peptides for 16 h at 37°C, washed, and labeled for 1.5 h at 37°C with 100 μCi of sodium [⁵¹Cr]chromate (Mendel Scientific). Labeled target cells were washed three times and added to v-bottomed 96-well plates at 10⁴ cells/well in a volume of 100 μl of medium without rhIL-2. Resting Een217 T cells (resuspended in 100 μl of the same medium) were subsequently added to give an effector-to-target ratio of 20:1. The plates were centrifuged at 1,800 rpm for 1 min to promote cell contact and were incubated at 37°C for 6.5 h. The plates were then centrifuged at 4°C, supernatants were collected, and gamma radioactivity release was measured and converted to percent specific lysis, according to the formula [(E − C)/(M − C)] × 100, with E as the experimental value in counts per minute, C as the control release value, and M as the maximal release value. C was determined as the average release in control wells from which effector cells were omitted; M was determined as the average release in wells where a 0.5% Triton X-100 solution was added in place of effector cells. All determinations were made in triplicate.

Anergy induction assay.

Irradiated (5,000 rads) HLA-DR4⁺ L cells were seeded in 24-well tissue culture plates at 5 × 10⁵ cells/well and incubated alone or with 10 nM peptide overnight at 37°C in a final volume of 700 μl. The cells were washed, and resting Een217 T cells were added at 5 × 10⁵ cells/well in a final volume of 1 ml of culture medium without rhIL-2. Forty-eight hours later, the T cells were removed from the APCs by vigorous washing and centrifugation over Ficoll Paque Plus gradients (Pharmacia) at 1,800 rpm for 25 min. This technique allows the separation of the T lymphocytes from the adherent fibroblasts used as APCs, as microscopic examination and flow cytometric analysis of the cells fail to indicate the presence of CD4⁻ cells (not shown). The recovered Een217 cells were washed three times and rested for 2 days in fresh culture medium without rhIL-2. The Een217 cells were finally challenged with 10 nM PV22 in a standard proliferation assay as described above. All determinations were made in triplicate.

Determination of TCR-CD3 and CD4 surface expression.

Een217 cells were cocultured in 24-well plates with irradiated HLA-DR4⁺ L cells alone or L cells pulsed with 10⁻⁹ to 10⁻⁴ M PV22, HXB2, or CDC42 peptide for 48 h at 37°C. The cells were washed and stained with 10 μl of fluorescein isothiocyanate-conjugated anti-CD4 monoclonal antibody (MAb; clone SFCI12T4D11; Beckman Coulter) and 10 μl of rhodamine-conjugated anti-CD3 MAb (clone UCHT1; Beckman Coulter) for 1 h at room temperature. The cells were washed and assessed for CD3/CD4 surface expression using an EPICS flow cytometer (Coulter, Miami, Fla.). In these experiments, negative cells were gated to fall within the 0.1 to 1.0 window.

Stimulation of Een217 T cells, immunoprecipitations, and immunoblotting.

For the stimulation of Een217 cells, HLA-DR4⁺ L cells were seeded in 24-well tissue culture plates at 10⁶ cells/well and incubated at 37°C overnight in medium alone or medium containing indicated concentrations of the PV22 or HXB2 peptide. The peptide-pulsed APCs were washed, and resting Een217 T cells were added at 12 × 10⁶ cells/well in a final volume of 500 μl of medium without rhIL-2. The plates were centrifuged for 1 min at 1,800 rpm to promote cell contact and were incubated at 37°C for 5 min. The stimulation was terminated by adding 500 μl of ice-cold phosphate-buffered saline in each well. The plates were transferred immediately to ice, and the cells were harvested with a cell scraper. The cells were then transferred to microcentrifuge tubes and centrifuged at 1,500 rpm for 30 s. Cell pellets were lysed with 500 μl of ice-cold lysis buffer containing 1% Triton, 20 mM Tris-HCl (pH 8), 150 mM NaCl, 5 mM NaF, 5 mM EDTA, protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 10 μg of leupeptin/ml, 10 μg of aprotinin/ml), and phosphotyrosine phosphatase inhibitor (1 mM sodium orthovanadate). Lysates were incubated for 30 min on ice and cleared of insoluble material by ultracentrifugation. Zap-70 was immunoprecipitated from postnuclear lysates exactly as described elsewhere (8). Immune complexes were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 13% polyacrylamide gels, transferred to nitrocellulose, and immunoblotted with the antiphosphotyrosine MAb 4G10 or anti-Zap-70 MAb 2F3.2 (Upstate Biotechnology, Lake Placid, N.Y.), followed by peroxidase-conjugated secondary antibodies. The blots were developed by enhanced chemiluminescence (Amersham) according to the manufacturer's instructions. The data presented are representative of experiments performed three times.

GST-SH2 pull-down of tyrosine-phosphorylated TCRζ from Een217 cells.

The tandem SH2 domain region of the human Zap-70 cDNA (corresponding to amino acids 1 to 260) was amplified by PCR using the cDNA described by Chan et al. (7) as a template. The 780-nucleotide PCR fragment was subcloned into the pGEX-3T expression vector and sequenced. The resulting Zap(SH2)₂–glutathione S-transferase (GST) fusion protein was expressed in Escherichia coli and purified as previously described (8, 9). The purified fusion protein precipitated from lysates of pervanadate-treated Jurkat T cells a 21- to 23-kDa tyrosine-phosphorylated protein that reacted with the anti-TCRζ MAb 6B10 (Zymed) in immunoblotting experiments (not shown), which demonstrates that the Zap(SH2)₂-GST fusion protein has the expected binding specificity. For precipitation of the tyrosine-phosphorylated ζ chain in peptide-stimulated Een217 cells, the cell lysates were incubated with ∼5 μg of Zap(SH2)₂-GST bound to glutathione-Sepharose beads for 1 h at 4°C. The beads were washed, and the bound proteins were separated on 13% SDS–polyacrylamide gels, transferred to nitrocellulose, and immunoblotted with the antiphosphotyrosine MAb 4G10 as described above.

For the pretreatment experiments, resting Een217 cells were incubated for 48 h at 37°C with HLA-DR4⁺ L cells alone or L cells pulsed with 1, 10, or 100 nM HXB2 or CDC42 peptide in a 24-well tissue culture plate. The T cells were removed from the APCs, washed, and rested 2 days in fresh culture medium without rhIL-2. The cells were then stimulated for 5 min with 10 μM PV22 peptide and lysed, and the ζ chain was precipitated using the Zap(SH2)₂-GST fusion protein as described above.

RESULTS

Functional response of Een217 cells to stimulation by native and variant gp120 peptides.

The Een217 T-cell clone is an HLA-DR4-restricted CD4⁺ cytotoxic T cell specific for gp120, amino acid residues 410 to 429, of the PV22 strain of HIV-1 (57). To determine the effects of natural variations in HIV-1 epitopes on CTL activation and function, we first measured the activity of synthetic peptides, corresponding to amino acids 410 to 429 of gp120 molecules from different HIV-1 strains, in a standard proliferation assay using L cells stably expressing human HLA-DR4 as APCs (Fig. 1A). We found that the CDC42-, EL1-, and Z3-derived peptides did not induce the proliferation of Een217 cells. However, the peptide derived from HXB2, which differs from the PV22 sequence by only two amino acids (F23I and E29K) (Table 1), induced proliferation but was consistently found to be less potent than PV22, by a factor of ∼25 (Table 1).

FIG. 1.

Functional response of Een217 cells to HIV-1 gp120-derived peptides. (A) Proliferation of Een217 cells in response to DR4-expressing L cells pulsed with the indicated concentrations of the agonist antigenic ligand (PV22) or natural peptide variants from the CDC42, EL1, Z3, and HXB2 strains of HIV-1. Note that the HXB2-derived sequence can induce the proliferation of Een217 cells but is less potent than the PV22 peptide. (B) Cytotoxic activity of Een217 cells toward ⁵¹Cr-loaded DR4⁺ L cells pulsed with the indicated concentrations of peptides. In this experiment, the cells were cocultured for 6.5 h with an effector-to-target ratio of 20:1 as described elsewhere (57). Note that the HXB2 peptide is at least 100 times less efficient than the PV22 peptide in this assay. Each determination was made in triplicate. The results shown are representative of three independent experiments.

TABLE 1.

Amino acid sequences of peptides used in this study and their relative abilities to induce proliferation and cytotoxic activity of Een217 cells

Peptide	Sequence	Relative potencya
		Proliferation	Cytotoxicity
PV22	GSDTITLPCRIKQFINMWQE	1.000	1.000
CDC42	TG-I---------I--R--V	<0.0005	<0.0005
EL1	TNTN---Q-----I-K-VAG	<0.0005	<0.0005
Z3	CTGN---------I------	<0.0005	<0.0005
HXB2	-------------I-----K	0.043	0.015

^aThe relative potency of a given peptide is calculated as the ratio between the concentration of the PV22-derived peptide required for 50% maximal response and the concentration of each variant peptide required for 50% maximal response. In both cases, the concentration of PV22 required to induce 50% maximal activation was given an arbitrary value of 1.0. 

We next determined whether the cytolytic function of Een217 cells could be induced by stimulation with ⁵¹Cr-loaded HLA-DR4⁺ L cells pulsed with increasing concentrations of the five HIV-1 peptides. In this assay, we found that the Een217 cells responded strongly to PV22, as measured by the release of ⁵¹Cr from PV22-pulsed L cells (Fig. 1B), while the peptides that were unable to induce the proliferation of Een217 cells were also inactive in the cytotoxicity assay. The HXB2 peptide, however, could induce the cytotoxic function of Een217 T cells, but only when used at a 100-fold-higher concentration than the PV22-derived sequence. Together, these data indicate that Een217 cells can respond to the HXB2 peptide in vitro but only when elevated peptide concentrations are used.

Induction of T-cell anergy by the HXB2 natural variant peptide.

It was recently reported that the copresentation of native and altered peptides to HIV-specific CD8⁺ CTLs was required to prevent the activation and function of the CTLs (49, 56). The authors concluded that this “defect” was the result of antagonism, where the APL competes with the agonist peptide for TCR binding and prevent the proper activation of the CTL. Conceptually, T-cell anergy differs from antagonism by the ability of an APL, when presented alone, to induce a long-lasting state of unresponsiveness where the T cell is unable to respond to mitogenic concentrations of its agonist ligand.

In the initial characterization of the Een217 clone, the ability of several natural HIV-1 variant peptides to stimulate the proliferation and function of Een217 cells had been partly determined (5, 42, 46, 57). However, the possibility that these natural APLs could induce TCR-dependent anergy of the CTL clone had never been tested. To determine whether natural gp120 variants could induce T-cell anergy, Een217 cells were pretreated with 10 nM each variant peptide for 48 h, rested for another 48 h, and finally stimulated with a mitogenic concentration (10 nM) of PV22 in a standard proliferation assay. Our results indicate that at a concentration where the peptide induces little proliferation and function of the T cells, HXB2 completely blocked PV22-induced proliferation of Een217 cells (Fig. 2). Under the same conditions, the CDC42- and EL1-derived peptides had a marginal effect on PV22-stimulated proliferation of Een217 cells, whereas the Z3 peptide reduced proliferation by ∼60% upon secondary stimulation with the PV22 peptide. The ability of the HXB2 peptide to induce the anergy of Een217 T cells was found to be dose dependent and to increase when higher concentrations of peptide were used in the pretreatment of the T cells. However, HXB2 was effective in attenuating the PV22-induced proliferation of Een217 cells at concentrations as low as 1 nM, where a 62% inhibition of Een217 proliferation was observed (data not shown). Importantly, Een217 T-cell anergy induced by 1 to 10 nM HXB2 could not be reversed by restimulation of the cells with up to 10 μM PV22. These results clearly demonstrate that the pretreatment of CTLs with a gp120-derived variant peptide alone can induce anergy and prevent subsequent CTL activation by the native gp120 epitope.

FIG. 2.

Induction of Een217 T-cell anergy by pretreatment with the HXB2 natural variant peptide. Een217 cells were cocultured for 48 h with APCs alone or APCs pulsed with 10 nM peptide, rested for 48 h, and restimulated with 10 nM PV22 in a standard proliferation assay. Each determination was made in triplicate. The percentage of inhibition obtained with each peptide is indicated. The data presented are representative of three independent experiments.

Inhibition of PV22-mediated cytotoxic activation of Een217 cells by the HXB2 variant peptide.

Since the pretreatment of Een217 T cells with the HXB2 variant peptide prevented PV22-induced proliferation of the T cells, we next tested the ability of the HXB2 peptide to prevent the induction of the CTL activity observed after PV22 stimulation of Een217 cells. T cells were first pretreated with increasing concentrations of HXB2 (presented by DR4⁺ L cells) for 48 h, rested for 48 h, and challenged with 10 nM PV22 in a standard cytotoxicity assay, using PV22-pulsed, ⁵¹Cr-loaded L cells as targets. Our data (Fig. 3B) indicate that the HXB2 peptide, at concentrations ranging from 0.1 to 10 nM, completely abrogated the induction of the cytotoxic activity of Een217 T cells normally observed in response to 10 nM PV22, a peptide concentration that induces ∼50 to 60% maximal CTL activity (Fig. 1 and 3A). When higher concentrations (100 nM to 1 μM) of HXB2 were used for the pretreatment of Een217 cells, low levels of cytotoxicity, inferior to 10% cell lysis, were observed (Fig. 3B). This may be due to the fact that high concentrations of HXB2 can induce the proliferation and cytotoxic activation of Een217 cells (Fig. 1). It is therefore likely that the CTL activity detected in these samples represents a residual function triggered by high concentrations of HXB2.

FIG. 3.

Inhibition of PV22-induced Een217 T-cell cytotoxicity by pretreatment of the cells with the HXB2 peptide. (A) Cytotoxic activity of Een217 cells, pretreated with culture medium alone for 48 h, toward ⁵¹Cr-loaded DR4⁺ L cells pulsed with the indicated concentrations of the PV22 peptide. (B) Cytotoxicity of Een217 cells, pretreated with the indicated concentrations of the HXB2 peptide for 48 h, and then stimulated with ⁵¹Cr-loaded DR4⁺ L cells pulsed with 10 nM PV22. The specific lysis measurement obtained with 10 nM PV22 (A) is represented by the asterisk in panel B. The data presented in panels A and B were derived from the same experiment.

TCR-CD3 and CD4 expression by Een217 cells stimulated by agonist and variant peptides.

Exposure of T lymphocytes to mitogenic concentrations of an agonist peptide, presented in association with the proper MHC molecule by specialized APCs, can induce an internalization of the TCR-CD3 complex (30, 47) through a mechanism that involves serial binding of the receptor to low numbers of peptide-MHC complexes and downregulation of the ligated TCR (62). This phenomenon is probably mediated by the activation of the Src family kinases Lck and Fyn (31) and does not occur with weak agonists (1) or TCR antagonists (30). However, since high concentrations of HXB2 stimulated Een217 cells to proliferate and exhibit cytotoxic activity (Fig. 1), we examined the possibility that the exposure of Een217 cells to HXB2 could modulate the expression of the TCR-CD3 complex and prevent later stimulation with the PV22 peptide.

Een217 cells were stimulated with 10⁻⁹ to 10⁻⁴ M PV22, HXB2, or CDC42 peptide for 48 h and examined by flow cytometry for surface expression of the TCR-CD3 complex, using a MAb to the CD3 ɛ chain. Our analysis revealed that the treatment of Een217 cells with all concentrations of these three peptides did not significantly alter the proportion of Een217 cells expressing the TCR-CD3 surface antigens, which remained at ≥97% (not shown). However, the surface expression levels of the TCR-CD3 complex, as measured by the mean fluorescence intensity, varied greatly according to the peptide and peptide concentration used (Fig. 4A). Indeed, the PV22 peptide induced an internalization of the TCR-CD3 complex that was dose dependent, reaching up to ∼70% reduction at the highest concentration used (10⁻⁴ M). At every peptide concentration tested, the HXB2 and CDC42 peptides induced lower levels of TCR internalization than the PV22 peptide. Similar results were obtained when the surface expression of CD4 was monitored after peptide stimulation of the Een217 cells (Fig. 4B). Most important, these data indicate that the stimulation of Een217 cells with 10 nM HXB2 or CDC42-derived peptide did not cause significant reductions in the surface expression levels of the TCR-CD3 complex or CD4, which remained at ≥92%. It is therefore extremely unlikely that the 95% reduction in PV22-induced proliferation of Een217 cells, and the total inhibition of cytotoxic function observed after the pretreatment of the cells with 10 nM HXB2, was the result of a dramatic reduction in TCR-CD3 or CD4 surface expression levels prior to PV22 stimulation.

FIG. 4.

Modulation of the TCR-CD3 complex (A) and CD4 (B) surface expression levels induced by stimulation of Een217 T cells for 48 h with the indicated concentrations of the PV22-, HXB2-, and CDC42-derived peptides. The data are presented as relative expression index, representing the ratio of the mean fluorescence intensity of each surface antigen in peptide-stimulated cells to that measured in nonstimulated Een217 cells (mean fluorescence intensities for CD3 and CD4 surface expression in nontreated cells were 8.97 and 4.54, respectively). In all cases, the proportion of TCR-CD3- and CD4-positive cells remained unchanged, indicating that the surface density of these molecules decreased following peptide treatment.

Early signal transduction events induced by the PV22 and HXB2 peptides.

APLs can, when presented by the proper MHC products, bind to the TCR and initiate early signaling events that differ from those initiated by the agonist antigenic peptide (reviewed in reference 36), resulting in either partial T-cell activation or functional anergy (11, 12, 58). Some of the intracellular events triggered by anergy-inducing APLs include an incomplete phosphorylation of the TCR-associated ζ chain and recruitment of unphosphorylated and inactive Zap-70 to the TCR-CD3 complex (8, 59). Since our data suggested that the HXB2 peptide variant induces anergy, we examined the TCR-proximal signaling events triggered in Een217 cells in response to the HXB2 peptide.

Een217 cells were stimulated with the indicated concentrations of either PV22 or HXB2 for 5 min. Zap-70 was immunoprecipitated from postnuclear detergent lysates with a previously described antiserum (8) and subjected to immunoblot analysis using antiphosphotyrosine or anti-Zap-70 MAbs. We found that Zap-70 became phosphorylated on tyrosine residues in response to the PV22 peptide (Fig. 5A, left panel), but not in response to equivalent concentrations of HXB2 (right panel). Notably, tyrosine-phosphorylated TCR ζ was present in all Zap-70 immunoprecipitates (except those performed on APCs alone). However, only the native PV22 peptide could induce the complete phosphorylation of the ζ chain, which is detected as a tyrosine-phosphorylated doublet (p21 and p23) in cells stimulated with peptide concentrations of 10 μM or higher (Fig. 5A).

FIG. 5.

Induction of Zap-70 tyrosine phosphorylation and association with the TCRζ chain in response to PV22 and HXB2 stimulation of Een217 cells. (A) Immunoprecipitation of Zap-70 from APCs alone or Een217 cells stimulated for 5 min with DR4-expressing L cells pulsed overnight with the indicated concentrations of PV22 and HXB2. The immunoprecipitates were subjected to SDS-PAGE and immunoblotted with antiphosphotyrosine antibodies (αPTyr; upper panel). Note that PV22 induced the phosphorylation of Zap-70 and its association with the fully phosphorylated form of TCR ζ, which appears as a p21/p23 doublet. The membrane was stripped and reblotted with a Zap-70-specific MAb (αZap-70) to ensure that equal amounts of Zap-70 were present in all immunoprecipitates (lower panel). As expected, Zap-70 was absent from anti-Zap-70 immunoprecipitates in assays performed on APCs alone (first lane of each panel). (B) Cell lysates of APCs or Een217 cells stimulated as described above were incubated in presence of ∼5 μg of the purified Zap(SH2)₂-GST fusion protein. The Zap(SH2)₂-GST-bound material was precipitated with glutathione-Sepharose beads and analyzed by immunoblotting with antiphosphotyrosine antibodies. The results presented were reproduced in four independent experiments. IgH, immunoglobulin heavy chain.

The differential phosphorylation of TCRζ in response to agonist and APL peptides was also detected when the ζ chain was isolated from peptide-stimulated Een217 cells by using a GST fusion protein containing the tandem SH2 domain region (amino acids 1 to 260) of Zap-70. Indeed, antiphosphotyrosine immunoblot analysis of the material bound to the Zap(SH2)₂-GST fusion protein revealed the same phosphorylation pattern (p21/23 doublet) as that coprecipitating with Zap-70 (Fig. 5B). This confirms that the 21/23-kDa phosphoproteins detected in the Zap-70 immunoprecipitates were in fact derived from the ζ chain. Finally, stimulation of Een217 cells with the CDC42-, EL1-, and Z3-derived peptides did not induce the phosphorylation of Zap-70 or p23ζ on tyrosine residues, even at peptide concentrations greater than 1 mM (not shown).

Although the maximum proliferation of Een217 cells was detected after stimulation with 10 nM PV22 (Fig. 1), we were unable to detect tyrosine phosphorylation of Zap-70 or p23ζ at peptide concentrations below 10 μM (Fig. 5A). This may reflect limitations in our experimental procedures but most likely resulted from the shorter duration of Een217 peptide stimulation in these experiments than in the proliferation assays (5 min versus 48 h). The same phenomenon was observed by other investigators in similar systems (30, 59).

Reduction of PV22-induced TCRζ phosphorylation by pretreatment with HXB2.

Our demonstration that the pretreatment of Een217 cells with the HXB2-derived peptide reduced the ability of the Een217 cells to respond to their antigenic peptide prompted us to determine whether the PV22-induced phosphorylation of the ζ chain could also be downregulated by HXB2 pretreatment. Een217 cells were incubated for 48 h with DR4⁺ APCs alone or APCs pulsed with 1, 10, or 100 nM HXB2 or CDC42 peptide. The T cells were then removed from the APCs, rested for 48 h, and challenged for 5 min with APCs pulsed with 10 μM PV22. The Een217 cells were then lysed, and the TCRζ chain was precipitated with the Zap(SH2)₂-GST fusion protein as described above. The SH2-bound material was then analyzed by antiphosphotyrosine immunoblotting.

As expected, PV22 stimulation increased the phosphorylation of the TCRζ chain at least fivefold (Fig. 6, lanes 2 and 3). This 21-kDa phosphopeptide originated from the T cells, as no such band was detected when APCs were tested in absence of T cells (lane 1). For unknown reasons, PV22 stimulation of Een217 cells preincubated with APCs alone (in absence of HIV-1-derived peptide) or APCs loaded with the CDC42 or HXB2-derived peptides consistently failed to induce the appearance of the p21/23 phosphorylated doublet normally observed in agonist-stimulated T cells. This may be due to fibroblast-derived cytokines or to a negative signal generated by the interaction of the T cells with the fibroblasts. Other investigators (49), using different experimental systems, also failed to detect the TCRζ-derived doublet following peptide stimulation of HIV-specific T cells. Nonetheless, we found that the pretreatment of Een217 cells with 100 nM HXB2 greatly reduced the level of TCRζ phosphorylation induced by PV22 stimulation (lane 4). A similar treatment of the T cells with 100 nM CDC42 failed to inhibit PV22-mediated TCRζ phosphorylation, which is consistent with the inability of the CDC42 peptide to induce T-cell anergy. Because the association of the Zap(SH2)₂-GST fusion protein with the TCRζ chain depends on the double phosphorylation of the ITAM modules of TCRζ, different amounts of ζ are expected in every lane. It is therefore difficult to ascertain that the reduced phosphorylation of TCRζ observed following pretreatment with 100 nM HXB2 (lane 4) is due to a reduction in the phosphotyrosine content of ζ and not to a gel loading artifact. However, this result was reproducible, and we verified that equal amounts of GST fusion proteins were present in all precipitates (not shown).

FIG. 6.

Inhibition of PV22-induced phosphorylation of TCRζ by pretreatment of Een217 cells with the HXB2 peptide. Een217 cells were preincubated for 48 h with APCs alone or APCs pulsed with 1, 10, or 100 nM HXB2 or CDC42 peptide. The Een217 cells were recovered, rested for 48 h, and restimulated for 5 min with APCs alone (lane 2) or APCs pulsed with 10 μM of PV22 (lanes 3 to 9). The TCRζ chain was then isolated from cellular lysates by using the Zap(SH2)₂-GST fusion protein and analyzed by immunoblotting with antiphosphotyrosine antibodies. The tyrosine-phosphorylated p21/23 protein originated from the Een217 T cells, as T-cell-free APCs, used as a negative control (lane 1), did not contain this phosphopeptide. Note that HXB2 attenuated the PV22-induced ζ phosphorylation, whereas the CDC42 peptide, which does not induce Een217 T-cell anergy, did not. Identical results were obtained in two independent experiments.

Prevention and reversion of HXB2-induced Een217 T-cell anergy by exogenous IL-2.

In several experimental systems, T-cell anergy, induced by either APLs or TCR ligation in absence of costimulation, is a result of an active repression of TCR-induced transcriptional activation of the IL-2 gene (4, 71), which prevents agonist-stimulated T cells from entering the cell cycle. T-cell anergy is therefore expected to be reversible. Indeed, the addition of exogenous IL-2 either at the moment of stimulation with the APL or during the recovery period preceding the stimulation with the agonist peptide is sufficient to prevent or reverse the induction of the anergic state, respectively (37). It was therefore important to determine whether the addition of IL-2 can prevent or reverse the induction of Een217 T-cell anergy by the HXB2 variant peptide. Een217 cells were pretreated with DR4⁺ L cells pulsed with 10 nM PV22, CDC42, or HXB2, in the absence or presence of exogenous rhIL-2 for 48 h. The T cells were then recovered and rested for 48 h in the absence or presence of rhIL-2 and finally challenged with 10 nM PV22 agonist peptide, in the absence of rhIL-2, in a standard proliferation assay. In this experiment, the pretreatment of Een217 cells with the HXB2 peptide reduced the PV22-induced proliferation of Een217 cells by ∼90%, whereas the CDC42 peptide had minimal effects (Fig. 7A). The addition of rhIL-2 (50 U/ml) at the time of pretreatment with the different peptides (Fig. 7B) or during the 48-h recovery period following the pretreatment (Fig. 7C) blocked the induction of Een217 T-cell anergy by the HXB2 peptide.

FIG. 7.

Prevention and reversion of HXB2-induced Een217 T-cell anergy by exogenous IL-2. Een217 T cells were pretreated with either medium alone or 10 nM peptide (in the presence of DR4-transfected L cells) for 48 h, separated from the APCs, rested for 48 h, and finally stimulated with 10 nM PV22 peptide as before. The anergy assay (A) was performed in the absence of IL-2. The addition of IL-2 (50 U/ml) during the first peptide stimulation (B) or during the 48-h rest period (C) prevented or reverted the induction of anergy by the HXB2 peptide. All determinations were made in triplicate.

DISCUSSION

During HIV-1 infection, CTLs are believed to exert a significant selective pressure on the virus population and cause the emergence of escape variants (26, 28, 30, 44, 48). A number of mechanisms have been proposed to explain the ability of HIV-1 to escape the virus-specific CTL response during asymptomatic HIV-1 infection. These include the mutation or deletion of major viral epitopes, resulting in an inefficient MHC-dependent presentation and impaired TCR recognition (5, 22), or antagonism (1, 25, 49, 56), where the copresentation of the agonist and variant peptides prevents CTL activation.

Here, we show that a naturally occurring HIV-1-derived antigenic peptide variant has the ability to induce the anergy of the CD4⁺ CTL clone Een217 to the wild-type (PV22) viral antigen. The pretreatment of Een217 cells with nonmitogenic concentrations (1 to 10 nM) of the HXB2-derived peptide variant prevented the TCR-induced proliferation and functional cytotoxic activation of the T cells by mitogenic concentrations (≥10 nM) of the PV22 antigenic peptide. This inhibition did not require copresentation of the variant (HXB2) and antigenic peptides to the T cells, indicating that the functional inhibition observed in this system is most likely a result of TCR-mediated anergy. Despite our efforts to remove excess peptide by extensive washing of peptide-pulsed APCs, we cannot formally rule out the unlikely possibility that free HXB2 peptide, originating from the initial peptide treatment, was picked up by HLA-DR4⁺ Een217 cells and cross-presented (16) to the T cells simultaneously with APC-bound PV22 peptide, thereby antagonizing PV22-induced T-cell activation.

The dual effect of the HXB2 peptide, i.e., anergy at low concentrations and proliferation and cytotoxicity at higher concentrations, is unique. What would be the most likely outcome of in vivo stimulation of Een217 T cells with an emerging epitope variant? Since the concentration of such a variant peptide at the surface of an infected cell would increase with time, it is reasonable to assume that the concentration required for anergy induction would be reached first. Although a peptide concentration of 1 to 10 nM has been referred to as a “physiological” concentration (49), the exact correlation between the concentration of exogenously added peptides to APCs in vitro and the actual cell surface concentration of a naturally processed HIV-derived peptide is unclear. It is therefore difficult to determine whether this anergy-inducing peptide concentration can be reached in vivo.

Despite the fact that the vast majority of HIV-derived CTL epitopes identified so far are MHC class I restricted and trigger a CD8⁺ T-cell response, MHC class II-restricted viral epitopes have also been identified (21, 32, 57; reviewed in reference 39). Although there is still no direct evidence for the participation of CD4⁺ cytotoxic T cells in HIV-specific antiviral immune response, two groups have reported the isolation of HIV-1 Gag- and gp120-specific CD4⁺ CTLs from HIV-infected patients (10, 35). The rare occurrence of these CD4⁺ CTLs in HIV-infected patients contrasts with the fact that CD4⁺ CTLs can readily be isolated following in vitro stimulation of normal PBMCs with recombinant HIV-derived gp120 or following immunization of normal seronegative individuals by gp160/120-derived vaccines (16, 21, 22, 43, 57). Since virus-specific CD4⁺ CTLs can also be isolated from patients or animals infected with other viral pathogens (29, 33, 40, 64, 67, 68), these findings may indicate that CD4⁺ CTLs could indeed be generated following HIV infection, perhaps at the earliest stages, but may be primary targets of the virus and disappear early after the initial contact with the virus.

Provided that CD4⁺ CTLs and/or helper T cells are generated as a result of HIV infection, then the virus must have developed different strategies to escape both CD8⁺ and CD4⁺ effector cells. Viral epitope variations have been shown to prevent antigen presentation by APCs or productive TCR stimulation. In addition, the HIV-1 accessory gene products Vpu (23) and Nef (55) have been shown to downregulate the expression of class I MHC antigens at the surface of infected cells, hence interfering with the recognition of viral antigens by CD8⁺ CTLs. By analogy to a variety of experimental systems where CD4⁺ T cells (both human and murine) have been shown to respond to TCR epitope variants by complete functional anergy, our analysis indicates that virus-specific CD4⁺ CTLs may also be sensitive to APL-induced T-cell anergy. Since most cellular targets of HIV-1, such as macrophages, dendritic cells, and activated T lymphocytes, express high levels of MHC class II antigens, a class II MHC-restricted CTL response may have a role in eliminating these viral reservoirs throughout the infection.

The conclusion that HXB2 stimulation of Een217 cells induced T-cell anergy is supported by our biochemical analysis of the early signal transduction events induced by this peptide. Indeed, the stimulation of Een217 cells with the HXB2-derived peptide, even at elevated concentrations, induced only a partial phosphorylation of the TCR-associated ζ chain, which appeared as a single tyrosine-phosphorylated band of ∼21 kDa. Since the peptides used in this study bind to HLA-DR4 with similar affinities (46), this may reflect differences in the affinity and duration of the interactions between the different peptide-MHC complexes and the TCR, rather than the inability of the DR4⁺ APCs to present a variant peptide. Based on the data obtained by Kersh et al. (24), the p21 ζ phosphoprotein arising in response to APL stimulation is not expected to contain doubly phosphorylated ITAM modules. Yet the partly phosphorylated ζ chain from HXB2-treated cells can still interact with Zap-70 or with the Zap(SH2)₂-GST fusion protein. This suggests that at least one of the three ITAM motifs present in each TCRζ chain must be fully phosphorylated in order to generate a suitable docking site for the tandem SH2 domains of Zap-70. In addition, the fact that the TCRζ-associated Zap-70 is not phosphorylated on tyrosine, an event normally mediated by the nonreceptor PTK Lck, suggests that HXB2 failed to induce the early Lck-mediated phosphorylation and activation of Zap-70 (13). Since Lck is thought to be responsible for both the phosphorylation of the TCRζ chain and the activation of Zap-70 (63), it is unclear how one event can occur in the absence of the other. It is, however, conceivable that in response to APL stimulation, Zap-70 is recruited to the TCR-CD3 complex but maintained in an unphosphorylated and inactive form by a putative phosphotyrosine phosphatase. Candidates for such a phosphatase would include SHP-1, known to downregulate T-cell activation and Zap-70 phosphorylation (20, 45), and CD45, which has been shown to interact with the ζ chain and dephosphorylate TCRζ and Zap-70 in vitro (15, 41). Interestingly, the pretreatment of Een217 cells with 100 nM HXB2 48 h before TCR stimulation with the agonist ligand reduced the ability of the PV22 peptide to induce TCRζ phosphorylation. Since we found no significant reduction of TCR-CD3 surface expression under these conditions, this observation would be consistent with a phosphatase-mediated inhibition of the TCR proximal events resulting in TCRζ phosphorylation.

While the differential phosphorylation of TCRζ, as well as CD3_ɛ in some experimental systems, was observed in T cells anergized by TCR stimulation with APLs, the induction of T-cell anergy by TCR stimulation with agonist ligands, in the absence of costimulation, occurs with no major alterations in the early TCR-mediated signaling events, including the phosphorylation of TCRζ and Zap-70. Therefore, it is now understood that the lack of IL-2 production, rather than the altered pattern of TCR-mediated phosphorylation, is the crucial factor controlling T-cell anergy (37). This view is supported by our observation that the addition of exogenous IL-2, either at the moment of HXB2 pretreatment or during the recovery period preceding PV22 stimulation, is sufficient to prevent HXB2-induced Een217 T-cell anergy.

The molecular events responsible for the induction of T-cell anergy are unknown, but several biochemical events required for the maintenance of the anergic state have been identified. An impaired Ras activation, possibly leading to a deficiency in the activation of the mitogen-activated protein kinases Erk-2 and JNK, has been observed (14, 34, 54). Recently, the constitutive activation of the Rap1 GTPase, induced by TCR ligation in absence of CD28-derived costimulation, has been shown to be responsible for the sustained repression of IL-2 gene expression in anergized T cells (4). This elevated activity of Rap1 may result from an elevated phosphorylation and tyrosine kinase activity of the Fyn PTK and its association with a tyrosine-phosphorylated form of the proto-oncoprotein Cbl. Our attempt to determine whether similar events could be induced in HXB2-treated Een217 cells revealed no such association between Fyn and Cbl and no increase in Cbl phosphorylation. Therefore, the events responsible for inducing T-cell anergy in APL-stimulated T cells may be different from those induced by TCR ligation in absence of costimulation.

Taken together, our data suggest that T-cell anergy, induced by HIV-derived natural variants, is a plausible addition to the putative mechanisms that allow immune escape and viral persistence of HIV-1 in infected individuals. Such a mechanism may not only allow a viral variant to escape the CTL response but also facilitate the persistence of other viral strains that may otherwise be recognized and eliminated by HIV-specific CTLs. Whether CD4⁺ CTLs indeed exist in HIV-1-infected patients and participate in the cytotoxic response to the virus is still unclear. Nonetheless, our findings may have to be taken into consideration for the design and utilization of HIV-specific vaccines. Indeed, viral epitope-specific CD4⁺ CTLs, induced by vaccination, may be rendered anergic if an individual is infected with a viral population containing an epitope variant having partial agonist and anergic properties. Finally, it should be noted that it may be possible to manipulate anergized HIV-specific CTLs to restore their ability to respond to antigen stimulation. Such a strategy may consist in exposing APL-anergized T cells to heteroclitic peptides, which are essentially mutated versions of the native peptide antigen that can stimulate T cells more efficiently than the native peptide itself. It has recently been reported that treatment of in vivo-tolerized T cells, a phenomenon involving the induction of T-cell anergy, with heteroclitic peptides reverted the anergic phenotype of the T cells (70). This finding suggests that T-cell anergy may not be irreversible and that superinduction of the TCR expressed by anergic clones by heteroclitic peptides may rescue the cells and restore responsiveness to TCR ligation. Whether these phenomena also apply to CD8⁺ CTLs is now the focus of our investigations.