|Home | About | Journals | Submit | Contact Us | Français|
Apoptosis of uninfected bystander T cells contributes to T-cell depletion during human immunodeficiency virus type 1 (HIV-1) infection. HIV-1 envelope/receptor interactions and immune activation have been implicated as contributors to bystander apoptosis. To better understand the relationship between T-cell activation and bystander apoptosis during HIV-1 pathogenesis, we investigated the effects of the highly cytopathic CXCR4-tropic HIV-1 variant ELI6 on primary CD4+ and CD8+ T cells. Infection of primary T-cell cultures with ELI6 induced CD4+ T-cell depletion by direct cell lysis and bystander apoptosis. Exposure of primary CD4+ and CD8+ T cells to nonreplicating ELI6 virions induced bystander apoptosis through a Fas-independent mechanism. Bystander apoptosis of CD4+ T cells required direct contact with virions and Env/CXCR4 binding. In contrast, the apoptosis of CD8+ T cells was triggered by a soluble factor(s) secreted by CD4+ T cells. HIV-1 virions activated CD4+ and CD8+ T cells to express CD25 and HLA-DR and preferentially induced apoptosis in CD25+HLA-DR+ T cells in a CXCR4-dependent manner. Maximal levels of binding, activation, and apoptosis were induced by virions that incorporated MHC class II and B7-2 into the viral membrane. These results suggest that nonreplicating HIV-1 virions contribute to chronic immune activation and T-cell depletion during HIV-1 pathogenesis by activating CD4+ and CD8+ T cells, which then proceed to die via apoptosis. This mechanism may represent a viral immune evasion strategy to increase viral replication by activating target cells while killing immune effector cells that are not productively infected.
Human immunodeficiency virus type 1 (HIV-1) infection causes chronic immune activation and the functional impairment and loss of CD4+ T cells (16, 39, 41, 62). In up to 50% of patients, HIV-1 infection also causes CD8+ T-cell depletion (61, 64, 78). Dysregulation of T-cell homeostasis leads to immunodeficiency and AIDS, but the mechanisms leading to T-cell depletion remain controversial. Impaired thymic production of naïve T cells over time reduces the size of the T-cell pool (22, 63). Direct lysis of infected CD4+ cells also contributes to T-cell depletion (10). However, most of the apoptotic T cells in the peripheral blood and lymph nodes of HIV-1-infected patients are uninfected (5, 12, 27, 51, 64, 67). Apoptosis of uninfected bystander T cells is caused by activation-induced cell death of mature T cells following chronic immune activation, in addition to HIV-1-mediated mechanisms (36, 62).
HIV-1 induces apoptosis in uninfected T cells by several mechanisms. HIV-1 proteins, including Tat, Vpr, and Nef, have cytotoxic effects in tissue culture (31, 60, 81, 86). However, the envelope glycoproteins (Env) have been implicated as the major cause of bystander cell death in T cells and other cell types (43-45, 48, 72, 84). Nonreplicating virions induce a proapoptotic signal in uninfected CD4+ T cells through a CXCR4- or CCR5-mediated pathway that does not require CD4 signaling or membrane fusion (24, 45). In addition, soluble, virion, or cell-associated HIV-1 envelope glycoproteins can prime uninfected T cells for activation-induced apoptosis (6, 17, 25, 44, 86). Previous studies estimate that only 0.00001 to 0.01% of HIV-1 virions are infectious in vitro and in vivo (19, 76). Thus, noninfectious virions may contribute to HIV-1 pathogenicity by inducing bystander T-cell apoptosis.
Host-cell proteins that are incorporated into the HIV-1 viral membrane can increase virion binding to target cells through interactions with their cognate ligands on target cells and thereby may influence the ability of HIV-1 to induce bystander apoptosis. Cellular membrane proteins that are incorporated into virus particles include ICAM-1 (CD54), LFA-1, LFA-2, LFA-3, CD55, major histocompatibility complex (MHC) class II isoforms, CD28, and B7-2 (CD86) (3, 8, 9, 24, 29, 34, 79). The incorporation of these cellular proteins into the viral membrane is selective, as other cell surface proteins, such as CD45, CXCR4, and CD4, are not incorporated into virions (30, 55, 69). The mechanism for the selective incorporation of cellular proteins into the viral membrane is not well understood, although a role for directed HIV-1 virion budding from glycolipid-enriched membrane microdomains, or lipid rafts, has been postulated (69). The presence of ICAM-1, CD28, and MHC class II in the viral envelope has been shown to increase HIV-1 infectivity by enhancing virus binding to target cells (8, 9, 28). Additionally, incorporation of MHC class II and B7-2 into the viral membrane enhances the ability of HIV-1 virions to induce bystander cell death (23, 24). Whether these cellular proteins act solely by enhancing virion binding to T cells or whether they induce separate signals that contribute to T-cell anergy or apoptosis is unknown.
In this study, we investigated the relationship between T-cell activation and bystander apoptosis during HIV-1 pathogenesis. We demonstrate that infection of primary T-cell cultures with the CXCR4-tropic HIV-1 variant ELI6 causes CD4+ T-cell depletion by direct cell lysis and bystander apoptosis. Nonreplicating HIV-1 virions induce the activation of CD4+ and CD8+ T cells, which then proceed to die via apoptosis. CD4+ T-cell apoptosis requires virion binding to CXCR4, whereas the apoptosis of CD8+ T cells is triggered by a soluble factor(s) secreted by CD4+ T cells. Maximal levels of activation and apoptosis are induced by virions that incorporate MHC class II and B7-2 proteins into the viral membrane. These findings suggest that nonreplicating HIV-1 virions induce activation and apoptosis of CD4+ and CD8+ T cells through distinct mechanisms that may contribute to chronic immune activation and T-cell depletion in vivo.
293T cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% (vol/vol) bovine calf serum (Sigma Chemical, St. Louis, MO), along with 100 μg/ml penicillin and streptomycin. Peripheral blood mononuclear cells (PBMC) were purified from the blood of healthy HIV-1-negative donors by Ficoll-Hypaque density gradient centrifugation and cultured in RPMI 1640 supplemented with 10% (vol/vol) fetal bovine serum (FBS) and 100 μg/ml penicillin and streptomycin. PBMC were stimulated with 2 μg/ml phytohemagglutinin (Sigma) for 3 days. The total T-cell fraction and CD4+ T-cell fraction were isolated by negative selection using a pan T-cell or CD4+ T-cell isolation kit, respectively (Miltenyi Biotech, Auburn, CA). The CD8+ T-cell fraction was isolated from the T-cell fraction by using CD8+ microbeads (Miltenyi). The purity of the T-cell fractions was 98.2% ± 1.2% CD3+ (three donors) as determined by staining with anti-CD3 monoclonal antibody (Beckman-Coulter) and flow cytometry. T cells were cultured in lymphocyte culture medium (1:1 volume of RPMI 1640 and AIM-V medium [Sigma] supplemented with 10% [vol/vol] FBS, 100 μg/ml penicillin, and streptomycin) and 10 U/ml interleukin-2. Herpesvirus saimiri-immortalized CD4+ (CD4/HVS) T cells (45) were cultured in lymphocyte culture medium. Transwell experiments used 24-well plates with filter inserts containing 0.1 μM pores (Costar).
Expression plasmids encoding the HIV-1 ELI6, NL4-3Env−Luc (NL4-3 ΔEnv), and 8x-NL4-3 proviruses have been described previously (32, 42, 45, 75). To generate virus, 293T cells were transfected by the calcium phosphate method with 20 μg of provirus plasmid DNA. Alternatively, 16 μg of provirus plasmid was cotransfected with 3 μg of plasmid expressing HLA-DRα1 and 3 μg of plasmid expressing HLA-DRβ7, 3 μg of plasmid expressing B7-2, or 3 μg of plasmid expressing ICAM-1. The total amount of transfected DNA was kept constant by using empty vector plasmid. The medium was replaced after 16 h, and supernatants were harvested after 72 h and stored at −80°C. Reverse transcriptase (RT) activity in the supernatants was measured using [3H]dTTP incorporation as described previously (77). For virus produced from CD4/HVS T cells, 293T-produced virus was used to infect CD4/HVS T cells, and cultures were monitored for virus production by RT assay. After clarification by low-speed centrifugation (200 × g for 5 min), supernatants were stored at −80°C. MN virions inactivated with aldrithiol-2 were generously provided by Jeff Lifson.
Cells (3 × 106) were incubated with 5 × 104 RT cpm of virus (corresponding to ~250 HIV-1 particles/cell or ~1.5 ng of gp120 [63, 70]) for 3 h in 1 ml at 37°C. Cells were washed once and plated in a six-well tissue culture plate in 3 ml medium. At 3-day intervals, cultures were split to keep the total cell density at 1 × 106/ml. At these time points, culture supernatant was removed for RT assays and cells were removed for staining with 7-aminoactinomycin D (7AAD; Via-Probe, PharMingen), anti-CD4-phycoerythrin (PE; PharMingen), anti-CD8-PE (Beckman Coulter, Fullerton, CA), or anti-p24-PE (KC57-RD1; Coulter), along with terminal dUTP nick end labeling-fluorescein isothiocyanate (TUNEL-FITC) (in situ cell death detection kit; Roche Molecular Biochemicals, Indianapolis, IN), followed by flow cytometric analysis as described below.
Following the selection of the T-cell fraction from PBMC, cultures were allowed to recover for 24 h and then were infected with 5 × 104 RT counts of virus, as described above. Cells were plated in 3 ml in six-well plates that had been coated with 1 μg/ml of anti-CD3. Briefly, plates were incubated with 5 μg/ml goat anti-mouse immunoglobulin G (IgG) for 90 min at room temperature in 0.05 M carbonate buffer, pH 9.8. The plates were blocked with phosphate-buffered saline (PBS) containing 1% human AB serum for 30 min at room temperature and then incubated with 1 μg/ml anti-CD3 (UCHT1; PharMingen, San Diego, CA) in PBS for 1 h. The plates were washed three times with PBS, and cells were added, along with 1 μg/ml anti-CD28 (CD28.2; PharMingen). Cells were then cultured as described above, with a freshly coated plate at each split.
Cells (1 × 106) were washed twice in fluorescence-activated cell sorter buffer (PBS supplemented with 0.2% sodium azide and 10% FBS), and stained with 7AAD and/or antibodies against cell surface proteins for 20 min at room temperature. The following antibodies were used: PE anti-IgG1, anti-IgG2, anti-CD25, anti-CD28, anti-CD38, anti-CD40L, anti-CD45RA, anti-CD45RO, anti-CD69, anti-HLA-DR, anti-CXCR4, anti-CCR7, anti-CD95, and anti-CD95L (PharMingen); FITC-IgG1 and FITC anti-CD4 (PharMingen); and FITC anti-CD8 (Beckman Coulter; Fullerton, CA), which recognizes the 8a epitope. Cells were washed once in fluorescence-activated cell sorter buffer and once in annexin V binding buffer and then were stained with annexin V-FITC (PharMingen). For intracellular p24 and TUNEL staining, cells were washed twice in PBS and fixed and permeabilized using the Cytofix/Cytoperm kit (PharMingen). Cells were resuspended in 50 μl Perm/Wash buffer and incubated with a 1:200 dilution of the KC57-RD1 α-p24 monoclonal antibody or mouse IgG1-RD1 isotype control (Coulter) for 20 min at 4°C. The cells were then washed twice with Perm/Wash buffer and stained using a TUNEL kit. Cells were then washed twice with PBS and counted by using a FACScan flow cytometer (Becton Dickinson, San Jose, California) or an Epics XL flow cytometer (Coulter). Data were analyzed by using Cell Quest (Becton Dickinson) and Expo32 (Coulter) software. For all experiments, 7AAD-positive cells were gated out of the analysis unless otherwise indicated.
HIV-1 virions were UV-inactivated for 45 min (52). Some samples were also preincubated for 1 h with 1.2 μM AMD3100 (21, 80) or 10 μg/ml control IgG1, anti-Fas (Calbiochem, San Diego, CA), or anti-CD86 (PharMingen). Virus (1 × 105 RT cpm, corresponding to ~1,500 HIV-1 particles/cell or ~3 ng of gp120 [63, 70]) or an equivalent volume of supernatant from uninfected CD4/HVS T-cell cultures was added to 1 × 106 T cells. As an additional control, some samples were incubated with virus stock supernatant that had been cleared of pelletable components, including virions, by high-speed centrifugation (1 × 105 × g for 1 h). A second dose of the blocking reagents was added to the cultures after 72 h. Cells were incubated with virions for 5 days, stained as described above, and analyzed by flow cytometry.
Primary CD4+ T cells (5 × 105) were washed twice in PBS supplemented with 10% FBS and 0.02% sodium azide. Cells were preincubated with 10 μg/ml anti-CD4 (Calbiochem) for 1 h at 4°C where indicated. Cells were then incubated with 4 × 104 RT counts of 293T-produced ELI6 virions at 4°C for 2 h, washed three times in binding buffer, and lysed in 100 μl of cell culture lysis buffer (Promega, San Luis Obispo, CA), and the level of bound p24 was quantified using a p24 enzyme-linked immunosorbent assay (Perkin-Elmer, Boston, MA).
Equivalent amounts of virions (based on RT assay) were purified by high-speed centrifugation over a sucrose gradient. Virions were solubilized in lysate buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 0.05 M Tris hydrochloride buffer [pH 7.5], 0.15 M NaCl, 1 mM EDTA, 1% aprotinin, 1 mM phenylmethylsulfonyl fluoride) and run on 10% sodium dodecyl sulfate-polyacrylamide gels. Proteins were transferred onto Immobilon-P membranes by standard techniques. Specific proteins were detected by immunoblot analysis with a mouse anti-CD86 antibody (PharMingen) and a rabbit anti-p24 antibody (ABT, Cambridge, MA). Primary antibodies were detected with horseradish peroxidase-conjugated species-specific goat secondary antibodies (Bio-Rad, Hercules, CA) and enhanced-chemiluminescence reagents (Amersham, Arlington Heights, IL).
HIV-1 infection induces CD4+ T-cell depletion by direct cell lysis and apoptosis in cultures of activated primary T cells. Infection of herpesvirus saimiri-immortalized CD4+ T cells with the highly cytopathic CXCR4-tropic HIV-1 variant ELI6 (45, 87) induces high levels of apoptosis in uninfected bystander T cells (45). CD4/HVS T cells resemble activated mature human CD4+ T cells (7, 45). We investigated whether ELI6 also induces T-cell depletion and apoptosis in activated primary T-cell cultures. Primary T cells were isolated, cultured, and infected as described in Materials and Methods. Infected and control cells cultured with or without continuous anti-CD3/anti-CD28 costimulation were examined for cell lysis and apoptosis by staining with 7AAD, anti-CD4-PE, anti-CD8-PE, or anti-p24-PE, along with TUNEL-FITC. ELI6 infection induced high levels of CD4+ T-cell depletion by day 9 (Fig. 1A and B). On days 6 and 9, infected cultures contained 2- to 4-fold more 7AAD-positive cells and 2- to 10-fold more TUNEL-positive CD4+ cells than control cultures (Fig. 1B and C). The majority of TUNEL-positive CD4+ T cells stained low for CD4. Primary T-cell cultures were >98% CD3+, suggesting that the CD4low population represents apoptotic T cells that have down-regulated CD4 expression (26, 57, 83) rather than antigen-presenting cells (APC). Similarly, apoptotic CD8+ T cells were frequently CD8low (Fig. (Fig.22 and not shown) (57). Costimulation with anti-CD3/anti-CD28 increased CD4+ T-cell depletion (31% of CD4+ T cells remained on day 6 compared to 54% for untreated ELI6-infected cultures) and apoptosis (40% of CD4+ T cells were apoptotic on day 9 compared to 15% for untreated ELIb-infected cultures) compared to stimulation with phytohemagglutinin/interleukin-2 alone, which may be due to increased levels of virus replication (Fig. 1A and C). On days 9 and 12, infected cultures costimulated with anti-CD3/anti-CD28 contained two- to threefold more apoptotic CD8+ T cells than control cultures. However, the levels of CD8+ T-cell apoptosis were minor compared to the levels of CD4+ T-cell apoptosis. The low percentage of CD4+CD8+ double-positive T cells present in these cultures (~1%) did not account for the CD8+ T-cell apoptosis (not shown). On days 9 and 12, 30 to 60% of the TUNEL-positive cells stained positive for anti-p24-PE (Fig. (Fig.1C),1C), while 40 to 70% remained p24 negative. Together, these results indicate that HIV-1 infection of primary T cells induces CD4+ T-cell depletion by direct lysis of infected cells and, at later time points, also induces bystander apoptosis.
We previously demonstrated that nonreplicating ELI6 virions induce apoptosis in CD4/HVS T cells and primary T cells (45). To investigate the relationship between T-cell activation and bystander apoptosis induced by nonreplicating HIV-1 virions, we began by examining the ability of nonreplicating ELI6 virions to induce apoptosis in primary T-cell cultures. Following incubation with 1 × 105 RT units of UV-inactivated ELI6 virions (corresponding to ~1,500 HIV-1 particles/cell or ~3 ng of gp120 [63, 70]) for 5 days, cells were stained and analyzed by flow cytometry. Annexin V-FITC was used to detect apoptotic cells. Annexin Vhigh and annexin Vintermediate populations were observed (Fig. (Fig.2A2A and and3).3). Over 80% of the 7AAD-positive cells stained high for annexin V (not shown), suggesting that the annexin Vhigh 7AAD-negative population represents late apoptotic cells. The annexin Vintermediate fraction showed apoptotic morphological changes by forward- and side-scatter analysis (not shown), indicating that this population represents early apoptotic cells (71). Because 7AAD-positive cells were gated out of the analysis, both annexin Vhigh and annexin Vintermediate populations were considered apoptotic.
Incubation with ELI6 virions induced apoptosis in three- to fivefold more T cells than did control supernatants (Fig. 2A and B). In contrast to the results observed for infected cultures, incubation with nonreplicating virions induced high levels of CD8+ T-cell apoptosis (Fig. (Fig.2A).2A). Annexin Vhigh CD8+ T cells frequently stained low for CD8, probably reflecting down-regulation of CD8 expression on apoptotic cells (57, 83). HIV-1 MN virions chemically inactivated with AT-2 had a similar capacity for inducing apoptosis in primary T cells in a manner that was dose dependent (200 pg to 5 ng p24; not shown). Thus, nonreplicating HIV-1 virions induce apoptosis in both CD4+ and CD8+ primary T cells.
HIV-1 Env/CXCR4 binding induces apoptosis in CD4+ T cells (45). However, because CXCR4 binding requires CD4-induced conformational changes in gp120, the mechanism of CD8+ T-cell apoptosis is unclear. To determine if nonreplicating HIV-1 virions directly induce apoptosis in CD8+ T cells through Env/CXCR4 binding, we incubated primary CD8+ T cells with UV-inactivated ELI6 virions. After 5 days, nonreplicating ELI6 virions did not induce apoptosis in purified primary CD8+ T-cell cultures (Fig. (Fig.2B),2B), even when preincubated with 0.01 μg to 1 μg of soluble CD4 to trigger the Env into a conformation capable of binding to CXCR4 (not shown). Additionally, the CD4-independent 8x-NL4-3 HIV-1 (45) did not induce apoptosis in primary CD8+ T cells (not shown). These results suggest that Env/CXCR4 binding on CD8+ T cells is not sufficient to induce apoptosis.
The preceding results raise the possibility that the induction of apoptosis in CD8+ T cells by HIV-1 virions requires the presence of CD4+ T cells. To determine whether CD8+ T-cell apoptosis requires direct contact with CD4+ T cells or is induced by soluble factors secreted by CD4+ T cells, we used a transwell system. Purified CD4+ and CD8+ T cells isolated by negative and positive selection, respectively, were either mixed together or separated by a transwell insert. Mixed T-cell cultures or CD4+ T cells in the lower chamber were incubated with UV-inactivated ELI6 virions for 5 days, while CD4+ T cells or CD8+ T cells were cultured in transwell inserts. High levels of apoptosis were induced in CD4+ T cells following exposure to ELI6 virions (Fig. (Fig.2D).2D). However, apoptosis was not induced when CD4+ T cells were separated from HIV-1 virions by a transwell, suggesting that interactions between virions and CD4+ T cells are required. CD8+ T-cell apoptosis was observed in mixed T-cell cultures exposed to ELI6 virions or in cultures separated by a transwell from CD4+ T cells exposed to virions. Similar results were obtained when CD4+ T cells were positively selected and CD8+ T cells were negatively selected (not shown). These results suggest that apoptosis of CD8+ T cells is triggered by a soluble factor(s) produced by CD4+ T cells following exposure to HIV-1 virions.
CD4/HVS T cells, which have a highly activated phenotype, are highly susceptible to HIV-1-induced apoptosis (45). To determine if apoptosis induced by HIV-1 virions occurs preferentially in activated T-cell subsets, we examined primary T cells exposed to nonreplicating virions for the expression of activation markers and for apoptosis. Incubation with HIV-1 virions for 5 days increased the percentages of T cells expressing CD25 and HLA-DR and decreased the percentages expressing CD28 and CD38 (Fig. (Fig.3A3A and and4).4). Smaller increases were observed in the percentages of cells expressing CD40L and CD69. Similar changes were observed in the mean fluorescence intensities (MFI) of these markers (not shown). In contrast, exposure to ELI6 virions had no significant effect on expression of CD44, CD45RA, CD45R0, and CCR7. Surprisingly, no difference was observed in the percentage of CXCR4+ T cells in cultures incubated with ELI6 compared to control cultures. CXCR4 expression decreased from 90 to 95% on day 0 to 20 to 30% on day 5 in both control cultures and cultures exposed to ELI6 virions (not shown). However, ELI6 virions induced a significant decrease in the MFI of CXCR4 on day 5 compared to control cultures (67.0% ± 0.7% of the control; P < 0.05) (not shown), suggesting that cells expressing high levels of CXCR4 may be preferentially depleted.
In primary T-cell cultures incubated with ELI6 virions, over 60% of CD25+ cells and HLA-DR+ cells were apoptotic (Fig. (Fig.3).3). Changes in the expression of these activation markers may be due to up-regulation or down-regulation of markers on apoptotic cells or may represent specific subsets of T cells that are preferentially affected by HIV-1 virions. To address the latter possibility, we examined the ratio of annexin V+ marker+ T cells to annexin V+ marker− T cells between control cultures and cultures incubated with ELI6 virions (Fig. (Fig.4B).4B). As expected, annexin V staining was increased threefold in the CD25+ and HLA-DR+ fractions in cultures incubated with ELI6 virions compared to control cultures. Apoptosis was also enhanced in the CD45RO+ subset despite no change in the overall percentage of cells expressing CD45RO. Although ELI6 virions decreased the percentage of cells expressing CD28, there was no specific reduction in the CD28+annexin V+ population, since the ratio of CD28+annexin V+ to CD28−annexin V+ cells was similar to that of control cultures. Thus, nonreplicating HIV-1 preferentially induced apoptosis in activated CD25+HLA-DR+CD45RO+ T cells.
The preceding results suggest that HIV-1 virions induce apoptosis in CD4+ T cells and CD8+ T cells by distinct mechanisms (Fig. (Fig.2).2). This finding raises the possibility that HIV-1 virions may also have differential effects on the activation of CD4+ and CD8+ T cells. To investigate this possibility, we examined the expression of activation markers on CD4+ and CD8+ T cells (Fig. (Fig.4C).4C). ELI6 virions activated both CD4+ and CD8+ T cells to express CD25 and HLA-DR and decreased the percentage of cells expressing CD28. The increase in CD25 expression was greater on CD4+ T cells than on CD8+ T cells. Additionally, virions decreased CD38 expression on CD8+ T cells but not on CD4+ T cells. Thus, HIV-1 virions activate both CD4+ and CD8+ T cells but have differential effects on the expression of specific activation markers.
Previous studies on HIV-1-induced bystander apoptosis suggested a role for Fas/FasL (CD95/CD95L)-mediated cell death (1, 5, 25, 51, 82). Therefore, we also examined the expression of Fas and FasL on CD4+ and CD8+ T cells following exposure to ELI6 virions. ELI6 virions increased the percentages of CD4+ and CD8+ T cells expressing Fas and FasL (Fig. (Fig.4C)4C) and the MFI of FasL expression on CD8+ T cells (not shown). These results suggested that the apoptosis of CD4+ and CD8+ T cells might be mediated by Fas/FasL interactions. However, an anti-Fas blocking antibody had no effect on apoptosis induced by exposure to ELI6 virions, implicating a Fas-independent mechanism (Fig. (Fig.4D4D).
To examine the relationship between T-cell activation, apoptosis, and the incorporation of cellular proteins that are expressed on activated T cells into HIV-1 virions, we produced virions from 293T cells in the presence or absence of MHC class II and/or B7-2. To distinguish between effects mediated by Env and effects mediated by host cell proteins in the viral membrane, virions were produced by using either an ELI6 provirus plasmid or an Env-deleted provirus plasmid (ΔEnv). ELI6 virions produced in CD4/HVS T cells, which express high levels of MHC class II and B7-2 (not shown), induced twofold-higher levels of apoptosis than ELI6 virions produced in 293T cells (Fig. (Fig.5A).5A). However, the incorporation of MHC class II isoform DRα1/β7 or B7-2 enhanced the ability of 293T-produced virions to induce apoptosis. The MHC class II isoform DRα1/β1 had a similar effect. In contrast, ICAM-1 slightly enhanced the ability of ELI6 virions to induce apoptosis, and MHC class I had no effect (not shown). ΔEnv virions did not induce apoptosis in primary T cells. However, ΔEnv virions that incorporated MHC class II and B7-2 induced apoptosis in primary T cells, suggesting these cell surface molecules, when incorporated into HIV-1 virions, can induce T-cell apoptosis in an Env-independent manner. Increased amounts of B7-2 were present in virions upon cotransfection of higher levels of B7-2 plasmid (Fig. (Fig.5B,5B, lower panel), with the ability of ELI6 virions to induce apoptosis dependent on the amount of cotransfected B7-2 plasmid (Fig. (Fig.5B,5B, upper panel). This result suggests that B7-2 in the viral membrane enhances the ability of virions to induce apoptosis in a dose-dependent manner.
We examined whether the incorporation of MHC class II and B7-2 into the viral membrane enhances the ability of HIV-1 virions to induce T-cell activation and apoptosis. Virions produced from 293T cells in the absence of MHC class II and B7-2 induced a 1.4-fold increase in CD25+ T cells, and a twofold increase in annexin V+CD25+ T cells (Fig. (Fig.5C,5C, upper right and lower right panels), suggesting that the incorporation of these proteins is not required to induce T-cell activation. However, incorporation of MHC class II or B7-2 enhanced the ability of ELI6 virions to activate and induce apoptosis in T cells. Maximal levels of activation and apoptosis were induced by virions produced in cells expressing both MHC class II and B7-2. These results suggest that the incorporation of MHC class II or B7-2 into the viral membrane enhances the ability of HIV-1 virions to induce T-cell activation and apoptosis.
To test whether the incorporation of MHC class II and B7-2 into the viral membrane increases virion attachment to target cells, we quantified the amount of bound p24 antigen following the incubation of T cells with ELI6 virions produced in the presence or absence of these proteins. Cultures were preincubated with anti-CD4 to prevent gp120-specific binding where indicated. (Fig. (Fig.6A).6A). MHC class II and B7-2 enhanced the ability of ELI6 to bind to T cells, with maximal binding occurring when virions were produced from cells expressing both proteins. Binding of ELI6 virions but not ΔEnv virions was inhibited by anti-CD4. B7-2 increased ΔEnv virions binding to T cells, and anti-CD4 only partially inhibited the binding of ELI6 B7-2 virions. Thus, the incorporation of B7-2 into the viral membrane increases virions binding to target cells in an Env-independent manner. These results suggest that incorporation of MHC class II and B7-2 enhances the ability of virions to induce apoptosis by increasing virion attachment.
HIV-1 Env/coreceptor binding induces apoptosis in CD4 T cells (45). To investigate whether a similar mechanism is involved in primary T-cell activation and apoptosis and whether this effect can be inhibited by disrupting B7-2/CD28 interactions, we used AMD3100 and anti-CD86. UV-inactivated virions and cultures of primary T cells were preincubated for 1 h with anti-CD86 and AMD3100, respectively, prior to mixing, and these blocking reagents were added again after 72 h. AMD3100 inhibited apoptosis induced by ELI6 virions produced in CD4/HVS T cells or B7-2-expressing 293T cells by 60% (Fig. (Fig.6B,6B, upper left panel). Moreover, AMD3100 inhibited increases in expression of HLA-DR and CD25 (Fig. (Fig.6B,6B, upper and lower right panels) and inhibited apoptosis in the HLA-DR+ fraction by >80% (Fig. (Fig.6B,6B, lower left panel), suggesting that Env/CXCR4 binding induces proactivation signals in T cells. Anti-CD86 also inhibited apoptosis induced by virions produced in CD4/HVS T cells and B7-2-expressing 293T cells by 23% and 37%, respectively, and inhibited increases in activation marker expression and HLA-DR+ T-cell apoptosis. Antibodies bound to B7-2 on virions could sterically hinder virus attachment. Therefore, we cannot determine whether anti-CD86 de-creases T-cell activation and apoptosis by blocking B7-2/CD28 or Env/CXCR4 binding. These results suggest that HIV-1 virions induce both proactivation and proapoptotic signals through CXCR4 and provide further evidence that HIV-1-induced bystander apoptosis is directly enhanced by T-cell activation.
The mechanisms of CD4+ and CD8+ T-cell depletion during HIV-1 infection remain poorly understood. In this study, we investigated mechanisms of HIV-1-induced bystander cell death in primary CD4+ and CD8+ T cells. Together with our previous study (45), our data suggest a model for bystander apoptosis induced by HIV-1 virions (Fig. (Fig.7).7). In this model, HIV-1 virions bind to CD4+ T cells and induce signals through CXCR4 or CCR5. These Env-mediated signals activate CD4+ T cells and preferentially induce apoptosis in activated CD25+HLA-DR+ T cells. Exposure to HIV-1 virions also causes CD4+ T cells to release cytotoxic soluble factors that activate and induce apoptosis in CD8+ T cells. Env/coreceptor binding and subsequent proactivation and proapoptotic signaling are enhanced by host cell proteins, including MHC class II and B7-2, which are incorporated into the viral membrane and bind to their cognate receptors on target cells. These results suggest that HIV-1 virions may induce T-cell activation to increase viral replication (18, 83), whereas activated cells that do not become productively infected proceed to die via apoptosis.
The relative impact of direct lysis of infected cells versus apoptosis of bystander cells on CD4+ T-cell depletion during HIV-1 infection has been the subject of much debate (10, 11, 24, 35, 36, 44, 45, 48, 53, 54, 65). Direct cytopathic effects are the predominant cause of cell death in CD4+ T-cell lines infected with lab-adapted strains (10, 53, 54, 56, 58), whereas apoptosis of both infected and uninfected CD4+ T cells occurs in infected lymphoid tissue explants (35, 38, 48). Direct lysis and apoptosis of infected and uninfected CD4+ T cells were induced by infection of primary T-cell cultures. In contrast to our previous study of CD4/HVS T cells (45), however, direct cell lysis was the predominant mechanism of primary CD4+ T-cell depletion. This difference could be due to the highly activated phenotype and higher levels of virus replication in CD4/HVS T cells compared to those of primary T-cell cultures (7, 45). Consistent with this possibility, CD28 costimulation increased levels of bystander apoptosis in primary T cells, and HIV-1 virions preferentially induced bystander apoptosis in activated CD25+HLA-DR+ T cells. These findings suggest a direct link between CD4+ T-cell activation and bystander apoptosis.
CD8+ T-cell apoptosis was induced by a soluble factor(s) produced by CD4+ T cells following exposure to HIV-1 virions. One potential candidate for the proapoptotic soluble factor is tumor necrosis factor alpha (TNF-α) (44). However, neutralizing antibodies to TNF-α, TNF receptor I, and TNF receptor II did not inhibit CD8+ T-cell apoptosis induced by nonreplicating HIV-1 virions (G. H. Holm and D. Gabuzda, unpublished observation), suggesting the involvement of other cytotoxic soluble factors. Other studies that demonstrated HIV-1-induced CD8+ T-cell apoptosis in vitro (24, 44, 51) used PBMC cultures and suggested a requirement for the presence of monocytes/macrophages. In contrast, we found that high levels of CD8+ T-cell apoptosis were induced by HIV-1 virions in the absence of APC. It will be of interest to determine the involvement of other soluble factors, such as a soluble TNF-related apoptosis-inducing ligand, in CD8+ T-cell apoptosis induced by nonreplicating HIV-1 virions.
Nonreplicating virions activated both CD4+ and CD8+ T cells, as indicated by increases in CD25 and HLA-DR expression. A previous study demonstrated that nonreplicating HIV-1 virions increase CD25 expression on T cells (50), whereas another study (24) reported partial activation of T cells but no increase in CD25 expression. These discrepancies may be due to differences in the cell culture conditions (PBMC versus primary T-cell cultures), virus strains (NL4-3 versus ELI6), or methods of producing virus stocks. Remarkably, SDF-1, the natural ligand of CXCR4, is a costimulator of CD4+ T-cell activation and up-regulates expression of CD25 in combination with anti-CD3 stimulation (68). Thus, CXCR4-tropic HIV-1 Envs may use the same signaling pathway as SDF-1 to activate T cells. Circulating lymphocytes of HIV-1-infected patients exhibit increased HLA-DR and CD25 expression and decreased CD28 expression, consistent with chronic immune activation (40, 52, 59, 74). Furthermore, levels of CD25+HLA-DR+CD4+ T cells correlate with disease severity (74). Together, these findings suggest that nonreplicating HIV-1 virions may contribute to chronic immune activation and activation-induced cell death of T cells in vivo.
Previous studies suggest that HIV-1 infection increases the susceptibility of activated T cells to Fas/FasL-mediated apoptosis in vitro and in vivo (1, 5, 13, 20, 25). Cross-linking of CD4 by gp120 induces FasL expression on CD4+ T cells, leading to Fas-mediated apoptosis of CD4+ and CD8+ T cells (51, 66, 82). We found that noninfectious HIV-1 virions increase Fas and FasL expression on CD4+ and CD8+ T cells, similar to results reported by other groups (24, 50). However, in contrast to Kameoka et al. (50), we found that anti-Fas did not inhibit apoptosis induced by nonreplicating virions, suggesting that T cells activated by Env/CXCR4 binding undergo apoptosis via the intrinsic pathway rather than the extrinsic Fas-mediated pathway. This discrepancy may reflect differences between experimental systems, including cell culture conditions and differences between wild-type and protease-defective HIV-1 particles, which have aberrant virion morphology and incorporate higher levels of gp120 than wild-type particles (49). CCR5-tropic HIV-1 also induces signals through the Env/CCR5 binding (15, 18, 85) and can trigger apoptosis in CCR5+ T cells (2, 84, 88). However, in contrast to CXCR4-mediated apoptosis, CCR5-mediated apoptosis may be Fas-dependent (2, 84, 88). Nonreplicating CCR5-tropic HIV-1 virions also induce apoptosis in CD4/HVS T cells, which express high levels of CCR5 (45). However, the ability of nonreplicating CCR5-tropic virions to induce apoptosis in primary T cells remains to be determined.
In contrast to Esser et al. (23), we found that incorporation of B7-2 alone enhanced the ability of HIV-1 virions to induce apoptosis in T cells. B7-2 is a costimulatory molecule expressed on antigen-presenting cells and activated T cells that provides a “second signal” through its ligand, CD28, to fully activate T cells following T-cell receptor ligation (reviewed in reference 37). Other groups have postulated that MHC class II and B7-2 incorporated into HIV-1 virions may bind to the T-cell receptor and to CD28 to induce proactivation or proapoptotic signaling pathways (8, 23, 33). Microvesicles containing B7 that are secreted from APC can stimulate naïve T cells (46), and ΔEnv virions containing B7-2 induced higher levels of T-cell apoptosis than ΔEnv virions alone. These findings raise the possibility that B7-2/CD28 interactions between virions and T cells may prime T cells for viral replication and apoptosis. Consistent with this possibility, both anti-B7-2, and CTLA-4-Ig (G. H. Holm and D. Gabuzda, unpublished observation) inhibited virion-induced activation and apoptosis. However, antibodies bound to B7-2 on virions may also sterically hinder virus binding and thereby prevent gp120/CXCR4 interactions. Thus, it remains uncertain whether B7-2 incorporated into the viral membrane induces activation and apoptosis independent of its effects on virion attachment. Expression of MHC class II and B7-2 is increased on circulating T cells of HIV-1 patients (52) and on HIV-1-infected T cells in vitro (47). These findings imply that the incorporation of MHC class II and B7-2 into HIV-1 virions produced by infected macrophages and activated T cells may enhance their capacity for inducing activation and bystander apoptosis. Additionally, by activating target cells to express these molecules, HIV-1 can increase the infectivity and cytopathicity of progeny virions produced during subsequent cycles of infection.
In summary, our studies suggest that nonreplicating HIV-1 virions induce activation and apoptosis in CD4+ and CD8+ T cells via distinct mechanisms. Exposure to nonreplicating HIV-1 virions activates T cells, an action which in turn increases viral infectivity and stimulates virus production (4, 14, 73). Activated cells that do not become productively infected proceed to die via apoptosis. These mechanisms may represent an important viral replication and immune evasion strategy. Activation and apoptosis of bystander T cells contributes to chronic immune activation and dysregulation, and impairs host immune responses. Thus, therapeutic strategies to block interactions between HIV-1 virions and CXCR4 may reduce both activation and depletion of CD4+ and CD8+ T cells in AIDS patients.
We thank J. Sodroski, G. Freeman, P. Ancuta, J. Wang, R. Desrosiers, and C. Gerard for helpful discussions. We are also grateful to G. Freeman for providing the B7-2 plasmid, D. Schols and E. De Clercq for providing AMD3100, K. Peden for providing the ELI6 plasmid, J. Sodroski for the MHC class II plasmids, and J. Lifson for AT-2-inactivated MN virions.
This work was supported by National Institutes of Health grant NS35734 to D.G. Core facilities were supported by Center for AIDS Research and Dana-Farber Cancer Institute/Harvard Center for Cancer Research grants.