|Home | About | Journals | Submit | Contact Us | Français|
Human immunodeficiency virus type 1 (HIV-1) viral protein R (Vpr) plays a crucial role in viral replication and pathogenesis by inducing cell cycle arrest, apoptosis, translocation of preintegration complex, potentiation of glucocorticoid action, impairment of dendritic cell (DC) maturation, and T-cell activation. Recent studies involving the direct effects of Vpr on DCs and T cells indicated that HIV-1 containing Vpr selectively impairs phenotypic maturation, cytokine network, and antigen presentation in DCs and dysregulates costimulatory molecules and cytokine production in T cells. Here, we have further investigated the indirect effect of HIV-1 Vpr+ virus-infected DCs on the bystander CD8+ T-cell population. Our results indicate that HIV-1 Vpr+ virus-infected DCs dysregulate CD8+ T-cell proliferation and induce apoptosis. Vpr-containing virus-infected DC-mediated CD8+ T-cell killing occurred in part through enhanced tumor necrosis factor alpha production by infected DCs and subsequent induction of death receptor signaling and activation of the caspase 8-dependent pathway in CD8+ T cells. Collectively, these results provide evidence that Vpr could be one of the important contributors to the host immune escape by HIV-1 through its ability to dysregulate both directly and indirectly the DC biology and T-cell functions.
Upon initial exposure to the mucosal surface, human immunodeficiency virus type 1 (HIV-1) targets different cell types of the immune system, including immature dendritic cells (DC), Langerhans cells, and resting T cells. These infected cells facilitate viral replication and disseminate virus via the immune synapses (23, 31, 51). Irrespective of the presence of viral antigens in infected antigen-presenting cells (APCs), the immune response eventually fails to control HIV-1 disease progression due to induction of anergy or apoptosis of naive T cells, thereby contributing to the early loss of CD4+ T cells and, eventually, the apoptosis of CD8+ T cells (27, 28, 52, 53, 57). Recent studies have also demonstrated that DC infected with HIV-1 selectively fail to mature and lack the potential to elicit a mixed lymphocyte reaction as well as a defect in interleukin-12 (IL-12) production (18, 49). HIV-1 has also developed strategies to trigger apoptotic signaling in both infected and uninfected bystander cells to disarm the host immune response (17). In draining lymph nodes of HIV-1-infected individuals, apoptosis is evident primarily in uninfected bystander cells (T and B cells), providing further important clues for the existence of an indirect mechanism of cell death (1, 15, 33). Apoptosis of uninfected T cells could be due to CD4 cross-linking (55), secretion of viral proteins, and/or proapoptotic cytokines from infected cells (DC and T cells) as well as through cell-cell contact followed by activation and tumor necrosis factor (TNF)-related death factor action (17, 27, 28, 53).
HIV-1 viral proteins have been implicated in the selective dysregulation of the host cellular immune response as part of the viral strategy to hijack host immune surveillance. DC exposed to gp120 display a functionally immature phenotype, as they retain the capacity for antigen uptake, are impaired in their ability to secrete cytokines and chemokines, and inhibit T-cell proliferation (14). Similarly, HIV-1 Tat has been shown to inhibit antigen-induced lymphocyte proliferation, whereas native Tat induces DC maturation (59). Nef is involved in the induction of chemokines in primary macrophages, which are required to facilitate lymphocyte recruitment and activation (50, 54). While protecting infected APC, Nef can trigger apoptosis of bystander cells and sensitize cytotoxic T lymphocytes to CD95-restricted apoptosis by activation-induced cell death (37, 60). Nef is also known to trigger the initiation of apoptosis by functional upregulation of FasL on the surface of infected cells and by modulation of caspase-dependent activation of death signaling events (45, 60, 62). A recent study by Quaranta et al. (42) suggested that exposure to Nef primes DC to induce apoptosis of alloreactive CD8+ T cells through functional upregulation of TNF-α and FasL.
HIV-1 Vpr, a 14-kDa pleiotropic accessory protein, is present at detectable level in the virion, thus making it one of the HIV proteins seen by the host cells during the early phase of infection (12). Direct effects of Vpr on host cells include cell cycle arrest, apoptosis, and alteration of host gene expression (9, 56, 64). We and others have recently shown that Vpr selectively impaired phenotype maturation and functions of both infected T cells and DC (30, 35, 58). Cell surface molecules and soluble factors released by DC are known to play a role in T-cell activation, anergy, and apoptosis (10, 38, 48). The present study aims to elucidate the role of Vpr-containing virus-infected DC on bystander CD8+ T-cell dysfunction and to further identify the factor(s) that is involved in this dysfunction. We have shown that DC infected with HIV-1 Vpr+ virus trigger apoptosis of CD8+ T cells, and this DC-mediated T-cell apoptosis is in part through the enhanced production of TNF-α by DC and the activation of the caspase-dependent CD8+ T-cell death receptor signaling pathway.
Monocyte-derived DC were generated from peripheral blood mononuclear cells (PBMCs). Heparinized blood samples were obtained with written consent from normal, healthy donors. PBMCs were isolated by Ficoll-Hypaque gradient centrifugation. CD14+ monocytes were purified by positive selection using anti-CD14 monoclonal antibody-coated magnetic microbeads (Miltenyi Biotech, Auburn, CA) as described previously (30). Purity of CD14+ cells was tested by flow cytometry using CD14-phycoerythrin (BD-Pharmingen, San Diego, CA) and CD1a-fluorescein isothiocyanate (CD1a-FITC; ImmunoTech, Miami, FL) antibodies. Greater than 95% of isolated cells were CD14 positive (characteristic phenotype of myeloid-derived monocytes) at day zero, and 85% were CD1a positive (characteristic of immature DC) at day 7, as described previously (30). To obtain monocyte-derived DC, CD14+ cells (0.5 × 106 cells/ml) were cultured in 60-mm culture plates in a total volume of 10 ml RPMI medium (GIBCO, Gaithersburg, MD) containing 10% fetal bovine serum (FBS), 1% l-glutamine (Cambrex, Walkersville, MD), 1% penicillin-streptomycin (GIBCO), 25 ng/ml IL-4 (R&D Systems, Minneapolis, MN), and 50 ng/ml granulocyte-macrophage colony-stimulating factor (R&D Systems). Half the volume of medium was replaced every third day with fresh medium containing IL-4 and granulocyte-macrophage colony-stimulating factor throughout the entire culture period as described elsewhere (30). HEK293T cells were maintained in Dulbecco's modified Eagle's medium containing 10% FBS, 1% l-glutamine (Cambrex), and 1% penicillin-streptomycin (GIBCO).
HIV-1 pNL43 env− vpr+ and pNL43 env− vpr− constructs (nef was deleted in these constructs through insertion of mouse heat stable antigen as described previously ) were pseudotyped with vesicular stomatitis virus (VSV-G) envelope and are denoted as HIV-1 Vpr+ virus and HIV-1 Vpr− virus, respectively. HEK293T cells (2 × 106) were cotransfected with 7.5 μg of HIV-1 proviral construct (pNL43 env− vpr+ or pNL43 env− vpr−) and 2.5 μg VSV-glycoprotein-envelope (VSV-G-Env) expression plasmids by the calcium phosphate precipitation method (5). Forty-eight hours posttransfection, supernatants were collected, spun at 3,000 rpm, and filtered through a 0.4-μm filter to remove cellular debris. All virus stocks were further concentrated by ultracentrifugation at 22,000 rpm for 1 h at 4°C. Virus pellets were dissolved in phosphate-buffered saline (PBS) and stored at −80°C for subsequent assays. The virus titer was measured by p24 enzyme-linked immunosorbent assay (ELISA), and viral infectivity was assessed by determination of the multiplicity of infection using the HIV-1 reporter cell line cMAGI (NIH AIDS Research Reference Reagent Program [NIH AIDS RRRP]). The replication-incompetent (noninfectious) forms of these viruses were prepared by the chemical inactivation method as described elsewhere (47). Briefly, virus stocks were treated with 250 μM 2′,2′-dithiodipyridine (AT-2; Sigma, St. Louis, MO) for 1 h in a rocking 37°C H2O bath. Post-AT-2 treatment, virus stocks were further concentrated by ultracentrifugation at 22,000 rpm for 1 h. Virus pellets were dissolved in PBS and quantitated by p24 ELISA for further use. HIV-1 Vpr was purified as a GST-Vpr fusion protein using a bacterial expression system as described by the manufacturer (Novagen, Madison, WI). Glutathione S-transferase (GST)-Vpr was further cleaved by thrombin protease, and Vpr was purified by high-performance liquid chromatography (Biotechnology Core Facility, University of Pittsburgh, PA). Purified Vpr protein was further characterized for lipopolysaccharide (LPS) contamination and biological activity (data not shown). Results indicated that high-performance liquid chromatography-purified recombinant Vpr is devoid of any detectable LPS contaminant and capable of inducing cell cycle arrest in Jurkat cells.
Five-day-old immature DC (5 × 106) maintained as described above were infected at a multiplicity of infection of 2 with HIV-1 Vpr+ or HIV-1 Vpr− virus. In the case of AT-2-treated virus, 500 ng of a p24 equivalent of virus was used for infection. Seventy-two hours postinfection, cells were stimulated with LPS (1 μg/ml; Sigma, St. Louis, MO) or CD40L (100 ng/ml; Alexis, San Diego, CA) for an additional 24 h. Similarly, DC were treated with recombinant Vpr or control GST protein (100 ng/ml) for 72 h and stimulated with LPS or CD40L as described above. To assess the effect of infected DC on CD8+ T cells, CD8+ T cells were isolated using a Dynal CD8 negative isolation kit (Dynal BioTech, Brown Deer, WI) following the manufacturer's instructions, using normal PBMCs from allogeneic donors. Purity of these cells was >95% as confirmed by fluorescence-activated cell sorting using anti-human CD8-PC5 antibody (Immunotech). CD8+ T cells (1 × 107 cells/ml) resuspended in medium containing 200 U/ml of IL-2 (NIH AIDS RRRP) were cocultured either with virus-infected and -exposed DC (1 × 106) that were subsequently stimulated with LPS or CD40L or, alternatively, incubated with the supernatants derived from the same DC culture.
To evaluate the effect of Vpr-infected DC on CD8+ T-cell proliferation, infected DC with or without stimulation were cocultured with alloreactive CD8+ T cells at ratios of 1:10 and 1:20 for 72 h in the presence of IL-2. DC infected with HIV-1 Vpr+ or HIV-1 Vpr− virus exhibited 35 ± 7% (mean ± standard error of the mean) p24 positivity as detected by flow cytometry (data not shown). Proliferation of the CD8+ T cells following 72 h of coculture with DC infected or exposed to HIV-1 Vpr+ or HIV-1 Vpr− virus was measured. A Cell Titer 96 nonradioactive cell proliferation assay was conducted as suggested by the manufacturer (Promega, Madison, WI) in triplicate. The optical density (OD) was measured at 490 nm, and the percent proliferation was calculated. The OD of CD8+ T cells cultured in the presence of uninfected, unstimulated DC (NT) was considered 100%. We also performed carboxyfluorescein diacetate succinimidyl ester (CFSE) labeling of CD8+ T cells to track the proliferating CD8+ T cells in a mixed lymphocyte reaction setup by flow cytometry. Briefly, freshly isolated CD8+ T cells were washed in PBS containing 0.2% bovine serum albumin (BSA) and resuspended in 400 μl 0.2% BSA containing PBS. Cell-Trace CFSE stock solution (5 mM; Molecular Probes, Carlsbad, CA) was prepared as per the manufacturer's instructions by adding 18 μl dimethyl sulfoxide in each tube. Eight μl CFSE solution was added to 2 ml PBS with 0.2% BSA. A total of 400 μl CFSE solution was immediately added to 400 μl cell suspension and mixed properly. Cells were incubated at 37°C for 10 min in the dark and further diluted by adding 10 ml cold RPMI medium containing 10% FBS (R = 10). Cells were incubated in ice and finally washed three times in cold R = 10. CD8+ T cells labeled in this way were assessed by flow cytometry on day zero to confirm staining quality. DC infected as described previously were stimulated with LPS or CD40L. DC and CSFE-labeled CD8+ T cells were cocultured at a DC:CD8+ T cell ratio of 1:10 in 24-well culture plates in a total volume of 500 μl in the presence of IL-2 (200 U/ml) for 4 days. CD8+ T cells labeled with CSFE and cultured without IL-2 were used as a negative control for proliferation. CD8+ T-cell proliferation was assessed by flow cytometry.
Analysis of apoptosis was carried out using an apoptosis detection kit (BD Biosciences, San Diego, CA) as per the manufacturer's instructions. Briefly, DC were infected with or exposed to HIV-1 Vpr+ or HIV-1 Vpr− virus, stimulated with LPS for 24 h, and incubated with CD8+ T cells for 24 and 48 h at DC:CD8+ T-cell ratios of 1:10 and 1:20. At the end of coculture, cells were washed twice with cold fluorescence-activated cell sorting buffer and stained for surface expression of CD8 using CD8-PC5 antibody. To detect apoptosis, cells were resuspended in 100 μl sterile binding buffer containing 10 mM HEPES-NaOH (pH 7.4), 140 mM NaCl, and 2.5 mM CaCl2, incubated with Annexin V-FITC for 15 min at room temperature in the dark, and analyzed by flow cytometry. The percentages of CD8-PC5- and Annexin V-FITC-positive cells were estimated using FlowJo software (Tree Star, Ashland, OR).
DC infected with HIV-1 Vpr+ or HIV-1 Vpr− virus, exposed to AT-2-inactivated HIV-1 Vpr+ or HIV-1 Vpr− virus, or treated with recombinant GST or GST-Vpr protein were further stimulated with LPS or CD40L for 24 h. To further characterize the specific involvement of different signaling molecules in Vpr-mediated TNF-α production, DC were pretreated with the phosphatidylinositol 3-kinase (PI3K) inhibitor LY294002 (1 μM and 10 μM), the p38-mitogen-activated protein kinase (p38-MAPK) inhibitor SB203580 (1 μM and 10 μM), and the histone deacetylase inhibitor Trichostatin A (TSA; 100 nM and 10 nM) (all from Calbiochem, San Diego, CA) for 4 h and were washed thoroughly prior to stimulation with LPS for an additional 12 h. Following stimulation, supernatants were collected and analyzed for the presence of TNF-α using an Opti-EIA ELISA kit (BD Biosciences) according to the manufacturer's protocol.
Real time reverse transcription-PCR (RT-PCR) was used to further assess quantitatively the transcriptional regulation of TNF-α. DC (5 × 106) were cultured and infected with HIV-1 Vpr+ or HIV-1 Vpr− virus as described previously (30). Additionally, AT-2-treated HIV-1 Vpr− or HIV-1 Vpr+ virus-exposed DC and GST- or Vpr-treated DC were used to demonstrate the ability of Vpr, from different sources, to alter TNF-α mRNA expression. Four hours poststimulation with LPS (the optimal time for peak TNF-α RNA expression determined previously), cells were lysed and the total RNA was isolated. Briefly, cells were washed once with cold PBS, and total cellular RNA was extracted using the RNeasy mini kit (QIAGEN, CA) according to the manufacturer's protocol, with additional on-column DNase I digestion (RNase-free DNase kit; QIAGEN). RNA concentration was determined by spectrophotometry. The integrity of RNA was assessed by the A260/A280 ratio and analyzed by agarose gel electrophoresis. Two-step RT-PCR was performed as follows: RNA (0.2 to 0.5 μg) was reverse transcribed using TaqMan reverse transcription reagents (Applied Biosystems, Foster City, CA). Real-time PCR was carried out in triplicate using commercially available primer/probe sets specific for TNF-α and ribosomal large protein (RPLPO; Applied Biosystems). The comparative CT method was used to determine the relative level of TNF-α transcript by normalizing to the RPLPO control transcript. Internal RNA controls of uninfected and LPS-stimulated DC were included to validate DC-mediated upregulation of TNF-α mRNA expression.
A total of 5 × 106 CD8+ T cells cocultured for 48 h with supernatants derived from uninfected or HIV-1 Vpr+ or HIV-1 Vpr− virus-infected DC stimulated with LPS were washed twice with PBS and lysed in RIPA buffer containing 50 mM Tris (pH 7.5), 150 mM NaCl, 1% Triton X-100, 1 mM sodium orthovanadate, 10 mM sodium fluoride, 1.0 mM phenylmethylsulfonyl fluoride, 0.05% deoxycholate, 10% sodium dodecyl sulfate (SDS), 0.07 U/ml trypsin inhibitor aprotinin, and protease inhibitors leupeptin, chymostatin, and pepstatin (1 μg/ml; Sigma). As a positive control, 10 ng/ml recombinant human TNF-α (Sigma, St. Louis, MO) was used to induce CD8+ T-cell apoptosis. Cell lysates were clarified by centrifugation, and total cell lysates (50 μg) were separated on an SDS-polyacrylamide gel and transferred, and the proteins involved in the CD8+ T-cell apoptosis pathway were detected with anti-caspase 3 (total and active; Cell Signaling Technology, Beverly, CA) and anti-active caspase 8 antibodies (Upstate Biotechnology, Charlottesville, VA). Anti-tubulin-α (NeoMarkers, Fremont, CA) was used as a loading control. Similarly, cell lysates prepared from infected and uninfected DC at 0 min, 30 min, and 1 and 4 h post-LPS stimulation were immunoblotted with anti-p38-MAPK and phospho-p38-MAPK (Santa Cruz Biotechnology, Santa Cruz, CA), anti-STAT-1 and anti-phospho-STAT-1 (Upstate Biotechnology), and anti-HIV-1 p24 Gag (NIH ARRRP) and α-tubulin (NeoMarkers) antibodies. Blots were developed using an enhanced chemiluminescence kit (Pierce, Rockford, IL).
Results were expressed as means ± standard errors of the means. The data were analyzed using the Student t test for normally distributed data with equal variances, and a P value of <0.05 was considered significant.
Previously we had shown that HIV-1 Vpr+ virus-infected DC exhibited an immature phenotype in concurrence with impaired cytokine production and defective antigen-specific CD8+ T-cell activation upon LPS or CD40L stimulation (30). Here, we further evaluated the ability of HIV-1 Vpr+ virus-infected DC to induce a proliferative response of allogeneic CD8+ T cells. Since mature DC have robust priming capacity, we compared the effect of HIV-1 Vpr− virus to HIV-1 Vpr+ virus in the context of both LPS and CD40L stimuli. In an allogeneic mixed lymphocyte reaction, DC infected with HIV-1 Vpr− or HIV-1 Vpr+ virus and stimulated with LPS were used for priming alloreactive CD8+ T cells (Fig. (Fig.1A).1A). Results indicated that there was a 2.5-fold inhibition of CD8+ T-cell proliferation when cells were primed with HIV-1 Vpr+ virus-infected DC compared to HIV-1 Vpr− virus-infected DC at a DC/T-cell ratio of 1:10 (Fig. (Fig.1A,1A, left panel), whereas at a 1:20 DC/T-cell ratio this reduction was about 1.5-fold (data not shown). These results indicate that DC infected with HIV-1 Vpr+ virus are defective in their ability to prime alloreactive CD8+ T cells, whereas no marked augmentation of CD8+ T-cell viability was observed when these cells were primed with LPS and HIV-1 Vpr− virus-infected DC, suggesting a specific regulatory role for Vpr in the CD8+ T-cell priming event.
Different forms of Vpr present within infected host are capable of inducing similar biological functions (56). Therefore, we next investigated whether virion-associated Vpr (using AT-2-inactivated virus) or free Vpr (using recombinant Vpr protein) was capable of mediating similar inhibitory effects through exposed DC on CD8+ T-cell proliferation and viability. Results indicated that the AT-2-inactivated HIV-1 Vpr+ virus also retained its ability to reduce CD8+ T-cell proliferation similar to replication-competent HIV-1 Vpr+ virus infection (Fig. (Fig.1A,1A, middle panel). Following 72 h of coculture of CD8+ T cells with LPS-stimulated DC preexposed to AT-2-treated HIV-1 Vpr+ virus, we observed a 45% reduction in CD8+ T-cell proliferation compared to that cocultured with HIV-1 Vpr−-treated DC. Similar inhibition (~40%) was also seen when the CD8+ T cells were cocultured with supernatant derived from AT-2-treated HIV-1 Vpr+-exposed and LPS-stimulated DC for 72 h (Fig. (Fig.1A).1A). Recombinant Vpr also demonstrated a similar effect on CD8+ T-cell proliferation when DC were exposed to Vpr compared to GST alone (Fig. (Fig.1A,1A, right panel). In order to further confirm that the observed effect of DC on T-cell proliferation can be achieved with DC matured in the presence of their natural stimuli, CD40L, we infected or exposed DC with all three forms of Vpr, as mentioned earlier, and subsequently stimulated them with soluble CD40L (100 ng/ml) for 24 h. CD8+ T cells were similarly cocultured with infected DC for 3 days, and proliferation was measured. Similar to LPS stimulation, DC stimulated with CD40L also resulted in suppressed CD8+ T-cell proliferation (Fig. (Fig.1B).1B). These results suggest that Vpr from multiple sources, upon various DC maturation conditions, induces similar impairing effects on bystander allogeneic CD8+ T cells.
To further confirm these results, additional experiments using CSFE-labeled CD8+ T cells were performed in a mixed lymphocyte reaction (Fig. (Fig.1C).1C). Results demonstrated that compared to the no-treatment control, DC stimulated with LPS resulted in ~2.5-fold increase in CD8+ T-cell proliferation. There was a twofold reduction in the number of proliferating CD8+ T cells when cocultured with HIV-1 Vpr+-infected DC stimulated with LPS. However, when AT-2-treated virus was used to infect DC, this reduction was approximately 1.5-fold (Fig. (Fig.1C).1C). These results further confirm that the reduction of CD8+ T-cell proliferation observed in cell titer assay is very similar to CFSE analysis in multiple donors (n = 3).
HIV-1 Vpr, in the context of both infection and exposure, is known to inhibit cell proliferation and to induce apoptosis (9, 56, 64). To elucidate whether the observed reduction in CD8+ T-cell proliferation in DC-T-cell coculture experiments (Fig. 1A to C) is due to de novo-synthesized Vpr released by infected DC, we performed additional experiments using HIV-1 Vpr− and HIV-1 Vpr+ virus-infected DC, without LPS stimulation, and cocultured with CD8+ T cells; proliferation was measured as described above. In the absence of maturation stimulus, HIV-1 Vpr+ virus-infected DC failed to inhibit proliferation of CD8+ T cells, further confirming that the observed differences in Fig. 1A, B, and C are not due to a direct effect of Vpr on allogeneic T cells (Fig. (Fig.1D1D).
One of the mechanisms of HIV-1-mediated host immune dysfunction is the selective depletion of uninfected bystander CD8+ cytotoxic effector cells (41). This targeted removal of functionally effective cytotoxic T lymphocytes within the infected host is one of the key strategies implemented by HIV-1 to evade the protective immune response (28, 41, 42). Therefore, we next investigated apoptosis induced by DC infected with HIV-1 Vpr+ virus. This effect could be due to either direct contact between these two cell types or through the release of an apoptosis-inducing factor(s) from infected DC. To delineate this, we performed coculture experiments using LPS-stimulated, infected DC and DC supernatant with CD8+ T cells and measured apoptosis by Annexin V staining in CD8+ T cells gated as shown in Fig. Fig.2A.2A. In an effort to further understand whether DC maturation is required for this effect, we performed coculture experiments using infected DC prior to LPS stimulation. Infected (unstimulated) DC were cocultured with CD8+ T cells, and Annexin V staining was performed (Fig. (Fig.2B).2B). The number of CD8+ T cells positive for Annexin V did not show any change under the different treatment conditions (18 to 22%), indicating that infection alone in the absence of maturation stimulation did not trigger a significant level of apoptosis, suggesting that DC maturation-associated factors might be involved.
Upon LPS stimulation, DC infected with HIV-1 Vpr+ virus efficiently triggered apoptosis of alloreactive CD8+ T cells as early as 24 h following DC-T-cell coculture as well as of T cells treated with DC supernatants (Fig. (Fig.2C).2C). In the case of uninfected or HIV-1 Vpr− virus-infected DC, the percentage of CD8+ T cells positive for Annexin V was around 21%, whereas in HIV-1 Vpr+ virus-infected DC, the percentage of apoptotic CD8+ T cells was 40%. Direct contact of CD8+ T cells with DC infected with HIV-1 Vpr+ virus induced a twofold increase in apoptosis (Fig. (Fig.2C,2C, top panel). A similar but lesser (1.5-fold) increase in apoptosis was noticed when CD8+ T cells were incubated with DC supernatants from the same culture derived from HIV-1 Vpr+-infected DC stimulated with LPS (Fig. (Fig.2C,2C, bottom panel). The effect of HIV-1 Vpr+ virus-infected DC was more pronounced (2.4-fold increase in the case of direct coculture with DC and 2.1-fold increase in the case of DC supernatant) following 48 h of coincubation (Fig. (Fig.2D),2D), suggesting either the requirement for a longer interaction through direct contact or prolonged exposure of CD8+ T cells to soluble apoptotic mediators and/or both, as the percentage of the late apoptotic CD8+ T-cell population increased at the 48-h time point. No marked difference of apoptosis in HIV-1 Vpr− or HIV-1 Vpr+-infected DC stimulated with LPS in the coculture was observed during the indicated time periods (data not shown). The increase (n-fold) in apoptotic CD8+ T cells cocultured with HIV-1 Vpr+ virus-infected DC and DC supernatants at the 24-h time point ranged between 2 to 3 and 1.75 to 2.5, respectively, in multiple (n = 5) donors (Fig. (Fig.2E).2E). The same increase at 48 h was found, with a range from 2 to 3.2 in the case of direct culture and 1.75 to 2.5 in the case of cocultured CD8+ T cells with HIV-1 Vpr+ supernatants (Fig. (Fig.2F).2F). A similar effect of Vpr was observed when DC were stimulated with CD40L (data not shown). Together, these results indicate that both membrane-bound and soluble factors released from HIV-1 Vpr+ virus-infected DC are capable of inducing apoptosis. Moreover, the extent of induction of apoptosis seen in the CD8+ T cells by LPS-activated HIV-1 Vpr+ virus-infected DC was higher than that observed with DC supernatant, further suggesting that cell-cell contact along with the continued production of a soluble factor(s) could have been responsible for the observed additive effect. Collectively, these results indicate that HIV-1 Vpr+ virus-infected DC induce T-cell apoptosis via both soluble and membrane-bound factors.
The observation that supernatants from DC infected with HIV-1 Vpr+ virus triggered CD8+ T-cell apoptosis prompted us to identify the potential soluble factor(s) involved in this event. TNF-α is one of the key cytokines produced by DC and is involved in both activation of APC and induction of apoptosis through death receptor signaling (2, 29). Supernatants from the HIV-1 Vpr+ and HIV-1 Vpr− virus-infected DC were assessed for the presence of TNF-α by ELISA. As shown in Fig. Fig.3A,3A, a 2.9-fold increase of TNF-α production was observed in HIV-1 Vpr+-infected DC compared to HIV-1 Vpr−-infected DC upon LPS stimulation in multiple donors (n = 4), whereas HIV-1 Vpr−-infected DC did not produce a significant amount of TNF-α compared to the LPS control. It is also interesting that virus infection alone without LPS stimulation did not induce any significant increase in TNF-α production (data not shown). To determine whether virion-associated Vpr could induce similar effects, DC were treated with AT-2-inactivated virus, stimulated with LPS, and measured for TNF-α production using the same donor DC. A similar upregulation of TNF-α by AT-2-treated HIV-1 Vpr+ virus-exposed DC upon LPS stimulation was observed (Fig. (Fig.3B).3B). However, the difference was a 2.5-fold increase above the LPS control, compared to the 2.9-fold increase observed in panel A. The greater increase seen in the infection model could be due to the de novo synthesis of Vpr by the infected DC compared to the noninfectious nature of AT-2-treated virus. Additionally, we also treated DC with recombinant Vpr protein or control GST protein (100 ng/ml) and observed a 1.6-fold increase in TNF-α production in the presence of recombinant Vpr (Fig. (Fig.3C).3C). This difference is similar to what is demonstrated with virion-associated Vpr in the absence of de novo synthesis, further supporting the specific role of Vpr in the absence of other viral proteins on TNF-α production. A similar upregulation of TNF-α production by different Vpr-treated DC compared to the control was also evident when DC maturation was promoted by CD40L stimulation for 24 h, although the overall production of TNF-α was relatively less compared to that of LPS in the same donor DC (Fig. 3A to C). Collectively, these results indicate that DC treated with different forms of Vpr induce TNF-α production comparable to that observed with HIV-1 Vpr+ virus-infected DC (Fig. (Fig.3A3A).
To delineate whether the observed Vpr-induced upregulation of TNF-α also occurred at the transcriptional level, we quantitated the amount of TNF-α mRNA. Infected DC were stimulated with LPS for 4 h. RNA isolated from the DC was assessed by real-time RT-PCR using a TNF-α-specific primer and probe set. Our data showed a 3.6-fold-higher TNF-α mRNA level in HIV-1 Vpr+ virus-infected DC compared to HIV-1 Vpr− virus-infected DC at 4 h post-LPS stimulation (Fig. (Fig.3D).3D). Further evaluation using AT-2-treated HIV-1 Vpr+ virus or HIV-1 Vpr− virus-exposed DC, and also GST- or GST-Vpr-treated DC from the same donor upon LPS stimulation, revealed that both virion-associated Vpr and free Vpr protein markedly elevated the TNF-α mRNA expression level (Fig. (Fig.3D).3D). Together, these results indicate that there is a significant increase in the level of TNF-α production by the DC infected with HIV-1 Vpr+ virus compared to the DC infected with HIV-1 Vpr− virus in response to LPS as well as CD40L stimulation, indicating that TNF-α could be one of the soluble factors present in the DC supernatant that induces apoptosis in CD8+ T cells.
Results stated above clearly indicate that TNF-α might be one of the soluble factors present in a significant amount in the supernatants derived from DC infected with HIV-1 Vpr+ virus; thus, it is capable of inducing apoptosis in alloreactive CD8+ T cells. To further confirm the specificity of TNF-α-mediated apoptosis of CD8+ T cells, DC or DC supernatant was pretreated with or without anti-TNF-α-neutralizing antibody (10 μg/ml; R&D Systems) or a corresponding immunoglobulin G (IgG) control (Sigma) antibody for 2 h prior to coincubation with purified CD8+ T cells. Cells were analyzed for apoptosis 48 h postcoculture (Fig. (Fig.4A).4A). Results indicated that the addition of anti-TNF-α neutralization antibody markedly inhibits CD8+ T-cell apoptosis by 63% in DC coculture compared to the isotype control. Furthermore, it is interesting that the inhibition is higher in the case of DC supernatant than DC-CD8+ T-cell coculture, indicating that continuous release of TNF-α by DC could not be completely neutralized by a one-time treatment with anti-TNF-α antibody. Furthermore, the range of inhibition of Vpr-induced DC-mediated apoptotic CD8+ T cells following neutralization that varied in different donors may be due to different levels of TNF-α produced by DC isolated from different donors (Fig. (Fig.4B).4B). These results indicate that TNF-α released by HIV-1 Vpr+ virus-infected DC is in part responsible for the observed T-cell apoptosis.
Two major pathways that mediate cellular apoptosis are signaling through surface death receptors, such as Fas and the TNF receptor, and signaling that involves members of the Bcl-2 family. In both the pathways, caspase 8 (initiator) and caspase 3 (executioner) are central mediators (2). In order to delineate the components essentially involved in Vpr-mediated DC-induced CD8+ T-cell apoptosis, we measured the endogenous regulator caspase 8 and the downstream effector caspase 3. Our results revealed that when CD8+ T cells were cultured with HIV-1 Vpr+ virus-infected DC supernatant, there was a marked increase of the expression of active caspase 8 coupled with cleaved caspase 3 (Fig. (Fig.5)5) and a concomitant decrease in total caspase 3 compared to the same incubated with supernatants derived from HIV-1 Vpr− virus-infected and uninfected DC stimulated with LPS. The observed activation of the caspase 8 pathway in CD8+ T cells following coincubation with HIV-1 Vpr+ infected-LPS stimulated supernatants was comparable with that treated with recombinant TNF-α (10 ng/ml) alone. Collectively, these results indicate that Vpr-infected DC efficiently triggered the caspase 8-dependent extrinsic pathway via TNF-α-mediated death receptor signaling.
The profound change in CD8+ T cells in conjunction with upregulation of TNF-α expression as observed in the presence of HIV-1 Vpr+ virus-infected DC or DC supernatants further motivated us to determine the signaling events leading to the production of these death-inducing molecules. Recently, Herbeuval et al. (22) showed that monocytes upon infection with chemically inactivated HIV-1 resulted in an increase of TNF-α-related apoptosis-inducing ligand (TRAIL) production due to activation of the alpha/beta interferon (IFN-α/β)-mediated STAT-1 signaling cascade. On the other hand, Nef-induced killing of bystander CD8+ T cells by monocytes has been shown to be through the activation of p38-MAPK (34). It is interesting that p38 is upstream of STAT-1 signaling, as demonstrated recently in primary cells and a transformed macrophage cell line in the context of LPS stimulation (25), thereby suggesting its role in regulating common downstream events. Lysates from infected DC cultures were tested by Western blotting for various signaling molecules. A significant increase in p38-MAPK phosphorylation was noted as early as 30 min post-LPS stimulation, followed by the subsequent upregulation of STAT-1 expression and phosphorylation at 4 h in DC infected with HIV-1 Vpr+ virus compared to DC infected with HIV-1 Vpr− virus (Fig. (Fig.6).6). We further delineated the signaling molecules involved in Vpr-mediated enhanced TNF-α production using specific inhibitors. As shown in Fig. Fig.6B,6B, pretreatment of HIV-1 Vpr+ virus-infected DC with different doses of PI3K and p38-MAPK inhibitors significantly suppressed TNF-α production following LPS treatment. In contrast, no enhancement or repression of TNF-α production was observed when infected DC were pretreated with TSA. The observed inhibition of LPS-induced DC-mediated TNF-α production following infection and treatment with different doses of inhibitors was not due to reduced viability of these cells, as there was no difference in the number of viable DC compared to the uninfected and non-LPS-stimulated control at the point of detection of TNF-α (data not shown). These results indicate that Vpr-induced upregulation of TNF-α production in DC in response to LPS uses multiple signaling pathways, which may act synergistically to enhance LPS-dependent TNF-α production by these cells.
Recent studies have shown that Vpr selectively impairs the surface expression of DC maturation/costimulatory molecules and IL-12 production and leads to a defective antigen-specific T-cell activation (30, 35). In the present study, we highlighted a new role for HIV-1 Vpr in host immune evasion. We observed that either direct contact of HIV-1 Vpr+ virus-infected DC or the soluble factor(s) present in the supernatant released by the infected DC efficiently induced CD8+ T-cell apoptosis in the absence of Nef. One of the major components of soluble cytotoxic mediators responsible for bystander CD8+ T-cell death was TNF-α, and its production was upregulated in the HIV-1 Vpr+ virus-infected DC upon both LPS and CD40L stimulation compared to their Vpr− counterpart. This finding was further confirmed by ELISA and TNF-α neutralization assays using anti-TNF-α-blocking antibodies.
CD8+ T cells are known to play an important role in HIV-1 antiviral immunity. These antiviral effects could be mediated through direct cytolysis of HIV-1 infected cells as well as through the secretion of soluble factors that suppress viral replication (41). Apoptosis of both CD8+ and CD4+ T cells has been considered one of the strategies used by HIV-1 to evade the effective host antiviral immune response. Studies using a large cohort of HIV-1 patients at various stages of disease showed that a significantly higher degree of apoptosis of T cells correlated with rapid progression of the disease (17). Apoptosis occurs primarily in uninfected bystander cells, mainly in CD4+ T cells and CD8+ T cells in tonsillar tissues (46) and T cells, B cells, and DC in the lymph nodes (33) in vivo. These apoptotic cells have been shown to exceed the numbers of productively infected cells, further suggesting the occurrence of bystander cell death (11). In contrast, long-term nonprogressors are characterized by a low frequency of CD8+ and CD4+ apoptotic T cells (32).
Several viral and virus-mediated cellular factors are associated with the loss of bystander T cells (17, 37, 40, 41). HIV-1 viral proteins, such as Nef, Tat, gp120, and Vpr, are prominently associated with the continuous induction of proapoptotic factors (6, 10, 13, 21, 38, 42). Additionally, the expression of viral proteins facilitates the destruction of uninfected T cells through modulation of DC. Here, using an in vitro infection model, we further elucidated the regulatory role of HIV-1 Vpr on DC function and its subsequent implications on alloreactive CD8+ T-cell priming. We delineated the effect of HIV-1 Vpr from different sources on bystander CD8+ T cells and observed that infection of DC with HIV-1 Vpr+ virus upon ligand-induced activation reduces the proliferation of alloreactive CD8+ T cells and impairs their priming. The functional competence of CD8+ T cells readily primed with antigen-pulsed Vpr-infected DC was severely dysregulated, as evidenced by diminished IFN-γ production (30), further supporting the regulatory role of Vpr in immune dysfunction. That the presence of LPS could influence systemic immune effector function, such as DC-induced T-cell dysregulation, in synergy with other host and viral factors is supported by a recent study which demonstrated that high serum LPS levels, as a result of microbial translocation from the gastrointestinal tract in infected individuals, are correlated with viral load and HIV-1 disease progression (7).
Though multiple factors are involved in CD8+ T-cell apoptosis during HIV-1 infection, FasL, TNF-α, TRAIL, and TWEAK are the major players in this network (2, 17). TNF-α is a proinflammatory cytokine produced primarily by macrophages, DC-activated T cells, and NK cells. HIV-1 not only favors persistence of infection by activating macrophage and DC through TNF-α production (4, 19) but also interacts independently and/or synergistically with other ligands to induce apoptosis in both CD4+ and CD8+ lymphocytes (19, 42). An elevated level of TNF-α has been reported in HIV-1-infected individuals, and it has been associated with disease progression (4, 8, 24). Similar findings were also reported in other viral infections, such as cytomegalovirus (44) and Ebola virus (61). Importantly, when the latently infected promonocytic cell line U1 was exposed to recombinant Vpr, virus replication was activated through increased production of TNF-α (36). These studies are in agreement with our results indicating that Vpr could be one of the viral proteins that have a role in TNF-α production.
HIV-1 Vpr has also been known to both induce apoptosis in primary and established cell lines as well as protect infected cells from apoptosis (3, 13). Our present study ruled out the possibility that Vpr released from infected DC was directly involved in bystander CD8+ T cell killing, as DC infected with HIV-1 Vpr+ virus or their supernatant without LPS stimulation did not induce a significant reduction of CD8+ T cell numbers. This was further supported by the fact that anti-TNF-α antibody markedly reversed (>65%) the observed Vpr-mediated apoptosis of these cells, implicating the specific contribution of TNF-α produced by HIV-1 Vpr+ virus-infected DC upon LPS stimulation. However, this does not rule out the possibility that viral and other cellular factors, including Vpr, could be present, and this was supported by the fact that the anti-TNF-α-neutralizing antibody even at the highest concentration (10 μg/ml) could not fully restore the apoptosis in CD8+ T cells. Although the exact mechanism of induction of death receptor signaling is not known, earlier reports demonstrated the roles of cFLIP, a negative regulator of caspase 8-mediated apoptosis in infected APC, and involvement of other signaling molecules, such as apoptosis signal regulating kinase 1 (16) and p38-MAPK, has also been implicated (39).
Despite evidence ascribing the role of TNF-α to apoptosis induction via multiple pathways and complex cross talk during HIV-1 infection, the underlying mechanism is still not clearly understood. Several viral and cellular factors are involved in activating the death receptor pathway and its downstream molecular events via direct as well as indirect mechanisms. For example, HIV-1 infection promotes monocyte differentiation indirectly by inducing IFN-α (63) and the secretion of proapoptotic cytokines, such as TNF-α, CD30L, and FasL (37). Similarly, chemically inactivated HIV-1 exposure of monocytes enhanced IFN-α-mediated STAT-1 activation for inducing the release of soluble apoptosis mediators (22). Although Vpr-augmented production of TNF-α and other proapoptotic cytokines is not fully understood, our studies using specific inhibitors for signaling molecules involved in TNF-α production suggest that in LPS-activated DC, the presence of Vpr enhanced the production of TNF-α by activating PI3K and its downstream p38-MAPK and eventually STAT-1. Similar upregulation of TNF-α by other viral proteins using PI3K and p38-MAPK was also reported by other investigators (26), further supporting the role of Vpr in the interaction of different signaling pathways in order to trigger LPS-mediated TNF-α production by DC.
The use of isogenic virus, with and without the vpr gene, further supports that Vpr might work in an additive manner. Together, these data provide a new insight into the role of Vpr in HIV-1 pathogenesis, in which this protein, while protecting the infected cells from apoptosis, successfully promotes killing of bystander CD8+ T cells through the upregulation of TNF-α and activation of the caspase 8-dependent death receptor signaling pathway. This study, therefore, provides important information in our efforts to develop an effective antiviral strategy to control HIV-1 infection.
HIV-1 pNL43 R+E− and pNL43 R−E− and pVSV-G-Env plasmids were obtained from the NIH AIDS Research and Reference Reagent Program. We thank Charles R. Rinaldo, Jr., and Paolo Piazza for critical reading of the manuscript.
This work was supported by the grant AI-50463 to V.A. from the NIAID, National Institutes of Health.
Published ahead of print on 2 May 2007.