Apoptosis or programmed cell death may play a critical role in AIDS pathogenesis through depletion of both CD4+ and CD8+ T lymphocytes. Using a reporter virus, a recombinant HIV infectious clone expressing the green fluorescent protein (GFP), apoptosis was measured in productively infected CD4+ T lymphocytes, in the presence and absence of autologous macrophages. The presence of macrophages in the culture increased the frequency of nonapoptotic GFP-positive productively infected CD4+ T lymphocytes. The appearance of nonapoptotic productively infected CD4+ T lymphocytes in the culture required intercellular contacts between macrophages and PBLs and the expression of the HIV Nef protein. The presence of macrophages did not reduce apoptosis when CD4+ T lymphocytes were infected with a GFP-tagged virus deleted for the nef gene. TNF-α (TNF) expressed on the surface of macrophages prevented apoptosis in nef-expressing, productively infected CD4+ T lymphocytes. Similarly, following TNF stimulation, apoptosis was diminished in Jurkat T cells transfected with a nef-expressing plasmid. TNF stimulation of nef-expressing Jurkat T cells resulted in NF-κB hyperactivation, which has been shown to deliver anti-apoptotic signals. Our results indicate that intercellular contacts with macrophages increase the rate of productively infected nonapoptotic CD4+ T lymphocytes. The survival of productively infected CD4+ T lymphocytes requires Nef expression as well as activation by TNF expressed on the surface of macrophages and might participate in the formation and maintenance of viral reservoirs in HIV-infected persons.
Human immunodeficiency virus type 1 infection is characterized by the progressive depletion of both CD4+ and CD8+ T lymphocytes (1). Several hypotheses have been advanced to account for the loss of T lymphocytes. They include: 1) direct lysis of the cells by viral infection (2); 2) syncytium formation (3, 4); 3) autoimmunity (5); 4) cellular and humoral virus-specific immune responses (6); 5) superantigen-mediated depletion of specific T cell subpopulations (7); and 6) apoptosis or programmed cell death (8). The potential role of apoptosis in CD4 depletion has been examined in several studies, and increased apoptosis in freshly isolated CD4+ and CD8+ T lymphocytes in culture grown with blood isolated from HIV-positive individuals has been reported (9, 10, 11, 12, 13). Increased apoptosis in both CD4+ mature T lymphocytes and thymocytes after HIV infection in the human SCID mouse model has also been described (14, 15, 16).
During the course of HIV infection, both uninfected and infected CD4+ T lymphocytes undergo apoptosis. In vitro, evidence of a direct killing of HIV-infected CD4+ T cells has been reported using reporter virus systems based on the expression of placental alkaline phosphatase or green fluorescent protein (GFP)4 (2, 17). In vivo, most of the CD4+ T lymphocytes that undergo apoptosis in the lymph nodes of HIV-infected individuals are bystander uninfected cells (18). APCs, such as macrophages (Mφ) have been found to trigger apoptosis in bystander uninfected CD4+ and CD8+ T lymphocytes (17, 19, 20, 21). Mφ-mediated apoptosis of CD4+ T cells involves CD4 cross-linking that results in the up-regulation of Fas on CD4+ T cells and Fas ligand on Mφ, respectively, and triggers CD4+ T cell apoptosis via Fas/Fas ligand interaction (21, 22, 23). Mφ-mediated CD8+ T lymphocyte apoptosis has been shown to involve the TNF-α (TNF) pathway (20). In addition to Mφ contact, the secretion of proapoptotic cytokines, viral proteins, and the stimulation of the chemokine receptor CXCR4 may also be involved in the apoptosis of uninfected bystander T cells (20, 24, 25, 26).
Dysregulation of cytokine production is a main feature of AIDS. Parallel to the Th1/Th2 cytokine switch (27), a chronic activation of the immune system might explain the increased levels of proinflammatory cytokines detected in plasma and tissues of HIV-infected patients (1). Among the proinflammatory cytokines detected during the progression of the disease, TNF seems to play a central role. TNF is secreted by primary Mφ infected in culture by HIV-1 or treated with envelope glycoprotein gp120, and by HIV-infected monocyte-derived Mφ isolated from patients (28). In vivo, membrane-bound TNF, which is present on the surface of CD4+ T cells, might account for the polyclonal activation of B cells (29), and high levels of TNF detected in plasma and tissues might contribute to the cachexia and fever observed in HIV-infected subjects (30). TNF enhances HIV-1 replication in chronically infected promonocytic and T-lymphoid cell lines by activation of the NF-κB, which stimulates the long terminal repeat (LTR) of the provirus (31, 32). TNF has also been reported to inhibit entry of R5 monocytotropic HIV strains in primary Mφ by inducing the production of RANTES and decreasing CCR5 expression (33, 34).
Nef is a 27-kDa HIV protein that is produced early during infection, and translated from multiply spliced viral mRNAs (35). Information is beginning to emerge that suggests that Nef may have evolved a number of different, independent functional activities to enhance the replication and survival of the virus within infected cells, and to facilitate its spread in vivo (36, 37). These activities include down-regulation of the CD4 receptor from the cell surface, down-regulation of MHC class I molecules, which may protect infected cells from killing by CTL, infectivity enhancement, and modulation of lymphocyte activation (38, 39, 40, 41). In vivo, several studies have demonstrated the importance of Nef for the efficiency of viral replication and for the maintenance of high viral loads (42, 43, 44). Recently, Nef expression within Mφ has been reported to favor the recruitment of resting T cells, via the secretion of C-C chemokines, and to subsequently favor their activation, suggesting a role for Nef in lymphocyte recruitment and activation at sites of viral replication (45).
Although Mφ have been reported to trigger apoptosis in uninfected T cells, the role of Mφ in modulating apoptosis in productively HIV-infected T lymphocytes has not yet been addressed. To determine whether Mφ change the susceptibility of HIV-infected CD4+ T lymphocytes to apoptosis and also to assess the role of Nef in the apoptotic process, we used GFP-tagged HIV viruses, expressing or not nef, to infect PBLs in the presence and absence of autologous monocyte-derived Mφ. Using flow cytometry analysis to discriminate between GFP-positive (productively infected) and GFP-negative (uninfected or latently infected) T lymphocytes, we assessed the effect of both Mφ and Nef expression on the rate of nonapoptotic productively infected GFP-positive CD4+ T lymphocytes. We observed that intercellular contacts between Mφ and PBLs resulted in the appearance of a population of nonapoptotic GFP-positive (productively infected) CD4+ T lymphocytes, in vitro. This population of nonapoptotic infected CD4+ T cells was detected in the presence of Nef expression, but not in its absence. We also observed that, in addition to Nef expression, the activation of the immune system, such as TNF stimulation provided by Mφ, was required to detect nonapoptotic infected CD4+ T lymphocytes in the culture.
Materials and Methods
CHOTNF cells are Chinese hamster ovary (CHO) cells that stably express membrane-bound TNF on their surface based on the expression of an uncleavable mutant form of TNF (46). CEM × 174 is a CD4+ T cell/B cell hybrid line generated from the polyethylene glycol-mediated fusion of 721.174 and CEM.3 cells. CHO, CHOTNF, Jurkat, and CEM × 174 cells were maintained in RPMI medium supplemented with 10% FBS. CEM × 174 cells were used to allow rapid and efficient recovery of progeny viruses derived from HIV-89.6 (47).
Isolation and culture of PBL, PBMC, and primary macrophages
Human PBMC and purified PBLs were prepared from peripheral blood of healthy donors as previously described (17). For purified PBL preparation, Ficoll-Hypaque (Pharmacia, Uppsala, Sweden)-isolated PBMC were incubated for 2 h on 2% gelatin-coated plates. Nonadherent cells, >98% which were PBL as assessed by CD45/CD14 detection by flow cytometry analysis (Simultest Leucogate; Becton Dickinson, San Jose, CA), were harvested after Ficoll-Hypaque isolation and adherence. PBL or PBMC were cultivated in RPMI with 10% (v/v) FBS supplemented with human recombinant IL-2 (20 IU/ml) following treatment with PHA (5 μg/ml) for 48 h. To block PBL-Mφ intercellular contacts, the cell populations were separated by a semipermeable membrane in six-well plates (0.4-μm pore size; Transwell; Costar, Greenwich, CT). Where specified, 5 × 106 PHA/IL-2-activated PBL were cocultivated with 5 × 106 CHO cells or CHOTNF cells in RPMI with 10% (v/v) FBS supplemented with hrIL-2 (20 IU/ml). Adherent tissue culture differentiated Mφ (>94% CD14+ by flow cytometric analysis) were cultured in RPMI medium supplemented with 10% (v/v) pooled AB human serum (Sigma) (20).
Generation of GFP-tagged infectious molecular clones derived from HIV-89.6
To create the virus designed HIV-GFPΔNef, we inserted the GFPS65T cDNA (Clontech Laboratories, Palo Alto, CA) into the deleted nef open reading frame as previously described (17). To reconstitute the nef open reading frame, a SacI-SacI fragment of p89.6-3′Eco (47) was engineered as follows. Two restriction sites corresponding to SacI and BamHI were introduced upstream of the nef open reading frame of p89.6-3′Eco by PCR using the following primers: 5′-CGC GAG CTC GGA TCC TAA GAT GGG AGG CAA GTG GTC-3′ and 5′-AGC CAG AGA GCT CCC AGG CTC AGA TC-3′. This amplified SacI-SacI fragment was cloned into the pHIV-GFPΔNef plasmid digested with SacI which contains two SacI restriction sites (located in the multiple cloning site downstream of GFPS65T cDNA and in the 3′-LTR of pHIV-GFPΔNef). We refer to this construct as pHIV-GFPNef-SacI/BamHI. An internal ribosome entry site (IRES) derived from the encephalomyocarditis virus, pCITE-4a-c+ (Novagen, Madison, WI) was amplified by PCR using the following primers: 5′-CGG GAT CCT AGG GCG AAT TAA TTC CG-3′ and 5′-CGG GAT CCA TTA TCA TCG TGT TTT TCA-3′. The amplified BamHI-BamHI fragment containing the cDNA for pCITE-4a-c+ was cloned into the plasmid pHIV-GFPNef-SacI/BamHI upstream of the nef open reading frame and downstream of the cDNA of GFPS65T. We refer to this GFP-positive IRES-positive Nef-positive construct, which expresses an intact nef open reading frame containing the ATG start codon of nef, as pHIV-GFP.
Wild-type and mutant HIV-1 infectious DNA were generated after ligation of a 5′-hemigenome (EcoRI-digested pEV114 derived from pILIC19 and pNL4-3) (17, 48), with the single-LTR-containing 3′-hemigenome constructs described above (EcoRI-digested p89.6-3′Eco, pHIV-GFPΔNef, and pHIV-GFP), as previously described (17). Ten micrograms of concatemerized proviral DNA were transfected into 107 Jurkat cells by the DEAE-dextran procedure (49). Twenty-four hours after transfection, Jurkat cells were cocultivated with 107 CEM × 174 cells to allow rapid and efficient recovery of progeny virus. Virus stocks (HIV-89.6, HIV-GFP, and HIV-GFPΔNef) were prepared from supernatants after filtration through a 0.45-μm-pore-size filter, quantified by measuring reverse transcriptase (RT) activity, and stored at −80°C, as reported previously (17).
PHA/IL-2-activated PBL and PBMC were cultivated in six-well plates at a density of 5 × 106 cells/well followed by infection with HIV-89.6, HIV-GFPΔNef, or HIV-GFP infectious clones as reported previously (17). After 2 h of exposure to virus at 37°C, cells were washed three times with PBS to remove the unadsorbed inoculum and reincubated in fresh culture medium at 37°C. Culture supernatants were collected every 2 days and assayed for RT activity.
Generation of nef-expressing plasmids and transfection
Two restriction sites corresponding to BamHI were introduced at the 5′ and 3′ extremities of the nef open reading frame of p89.6-3′Eco (47) by PCR using the following primers: 5′-CGC GGA TCC ATG GGA GGC AAG TGG TCA AAA CGT AGG GCA-3′ and 5′-CGC GGA TCC TCA GTT CTT GAA GTA CTC CGG ATG CAG GTC TC-3′. This amplified BamHI-BamHI fragment was cloned into the pREP9 plasmid (Invitrogen, San Diego, CA) digested with BamHI. We refer to this construct as pNef. One microgram of either pNef or pREP9 control plasmid (pCtl) was cotransfected with 1 μg pEGFP-C1 plasmid (Clontech) into 107 Jurkat cells using the GenePorter transfection assay (Gene Therapy Systems, San Diego, CA). Twenty-four hours after transfection, 5 × 106 Jurkat cells were cocultivated in the presence of either 5 × 106 CHOTNF cells or CHO cells. Following cocultivation for 24 h, the apoptosis was measured in the GFP-positive cells by TUNEL assay, as reported previously (17).
A total of 107 Jurkat cells were transfected with 1 μg of pNF-κBluc or 1 μg of pNF-κBmutluc (50), and 1 μg of pNef or 1 μg of pCtl plasmid using GenePorter transfection assay. Twenty-four hours later, 5 × 106 transfected cells were cocultivated in the presence of either 5 × 106 CHOTNF cells or 5 × 106 CHO cells. Following cocultivation for 24 h, luciferase activity was measured in cell lysates using a luminometer (TD-20/20; Promega, Madison, WI) as previously reported (51). Values normalized to protein concentrations were expressed in fold increase over the unstimulated control.
Flow cytometry analysis
Following infection with the GFP-expressing viruses, the specific fluorescence of GFP was measured upon excitation at 488 nm. CD4 detection was performed with peridinin chlorophyll protein (PerCP)-labeled anti-human CD4 mouse IgG1 mAb (Becton Dickinson). For TUNEL assay, cells were fixed in PBS containing 3% paraformaldehyde, and were then labeled as described below. Labeled cells were analyzed by flow cytometry with a FACScan flow cytometer (Becton Dickinson). PBL were gated on the basis of side scatter and forward scatter, and identified following CD45/CD14 labeling (Simultest Leucogate; Becton Dickinson). Data from 5 × 104 cells were collected, stored, and analyzed with CellQuest software (Becton Dickinson).
The detection of apoptosis by TUNEL assay was performed as previously described (17, 20). Briefly, after fixation with 3% paraformaldehyde in PBS for 30 min, 5 × 106 cells per sample were washed three times with 0.3% (v/v) Triton X-100 in PBS. Following one wash with terminal transferase buffer (Boehringer GmbH, Mannheim, Germany) in a 50-μl final volume, cells were incubated for 1 h at 37°C in a 50-μl final volume containing 20 U of terminal transferase, 2.5 mM CoCl2, 5 μM 16-dUTP biotin, 5 μM dUTP (all from Boehringer GmbH), and 0.3% (v/v) Triton X-100. After one wash in 0.3% (v/v) Triton X-100 in PBS for 1 h, cells were incubated for 1 h at room temperature in PE-labeled streptavidin (Boehringer GmbH) diluted at 1:200 in PBS. After one wash in 0.3% (v/v) Triton X-100 in PBS, cells were resuspended in PBS and then subjected to flow cytometric analysis. Rates of apoptotic and nonapoptotic CD4+ T cells were measured in both uninfected and infected autologous PBL and PBMC based on GFP expression (17). The percentage of nonapoptotic GFP-positive CD4+ T lymphocytes was calculated as follows: (number of nonapoptotic GFP+ CD4+ T lymphocytes)/(number of GFP+ CD4+ T lymphocytes), as previously reported (17).
Detection of nonapoptotic productively infected CD4+ T lymphocytes depends on both the presence of Mφ and the expression of HIV Nef
To discriminate between infected and uninfected T cells, we constructed two infectious molecular clones derived from HIV-89.6 (47) tagged with the GFP, expressing or not Nef, called HIV-GFP and HIV-GFPΔNef, respectively (17). To determine whether HIV replication correlates with GFP expression in CD4+ T lymphocytes, we performed a time-course analysis following infection of PBL with either HIV-GFP or HIV-GFPΔNef. Three different variables were monitored as a function of time: 1) RT activity in culture supernatants; 2) HIV gene expression by GFP measurement; and 3) CD4 expression on T lymphocytes by flow cytometric analysis. The efficiency of replication of HIV-GFPΔNef was decreased by 4- to 5-fold in PBL as compared with HIV-GFP and wild-type HIV-89.6 (Fig. 1,A). In contrast, both HIV-GFP and HIV-GFPΔNef allowed the detection of two populations of GFP-negative and GFP-positive cells in both T lymphocyte and CEM×174 cultures (Fig. 1,B and data not shown). The peak of GFP-expression coincided with the peak of viral replication following infection with either GFP-expressing viruses, usually at day 9 to 11 postinfection for a multiplicity of infection (MOI) of 0.10, as determined by measurement of RT activity in culture supernatants (data not shown) (17). The percentage of GFP-positive cells present in the PBL culture following infection with HIV-GFP or HIV-GFPΔNef correlated linearly with the level of RT activity in the culture supernatants (r2 = 0.96 and 0.95, respectively) (Fig. 1 C). Therefore, cell fluorescence as a reflection of GFP expression allows identification of productively infected cells in a heterogeneous population comprising both uninfected and infected cells.
To determine whether Mφ can block the apoptosis in HIV-infected CD4+ T cells, we measured the frequency of nonapoptotic GFP+ CD4+ T lymphocytes following infection of PBL and autologous PBMC with HIV-GFP and HIV-GFPΔNef viruses (MOI of 0.10) using three-color flow cytometric analysis (Figs. 2 and 3,A). At the peak of RT activity, a 10-fold increase of the rate of nonapoptotic GFP+ CD4+ T lymphocytes was measured in PBMC cultures vs PBL cultures following infection with HIV-GFP virus (Figs. 2 and 3, A and B). The infection of PBL and autologous PBMC with the HIV-GFPΔNef virus, that is deleted for the nef gene, resulted in barely detectable rates of nonapoptotic GFP-positive CD4+ T lymphocytes in the culture (Figs. 2 and 3, A and B). Separation of Mφ from PBL by a semipermeable membrane that blocks intercellular contact between Mφ and PBL, but still allows soluble factors to diffuse, did not result in the appearance of nonapoptotic GFP+ CD4+ T lymphocytes in the culture (Fig. 3,B). The finding of increased rates of nonapoptotic GFP+ CD4+ T lymphocytes in PBMC cultures, but not in PBL cultures, following infection with a nef-expressing GFP+ virus, is therefore consistent with a pivotal role of both Mφ and Nef protein in the regulation of apoptosis in HIV-infected cells. To further confirm the role of Mφ for the resistance to apoptosis in infected CD4+ T lymphocytes, we infected PBL with HIV-GFP or HIV-GFPΔNef (MOI of 0.10) in the presence of increasing concentrations of autologous monocyte-derived Mφ. The rate of nonapoptotic GFP+ CD4+ T lymphocytes increased with the proportion of autologous Mφ present in the culture following infection with HIV-GFP (r2 = 0.98), but not after infection with HIV-GFPΔNef (r2 = 0.01) (Fig. 3 C). Therefore, both the presence of Mφ and Nef expression within the infected cell are required to detect nonapoptotic GFP+ CD4+ T lymphocytes in the culture.
To further delineate the phenotype of nonapoptotic GFP+ CD4+ T lymphocytes, we measured the levels of cell surface CD4 on GFP+ CD4+ T lymphocytes following infection of PBMC with HIVGFP (MOI of 0.10). We observed that GFP+ CD4+ T lymphocytes could be divided into two main T cell subsets expressing low or high cell surface CD4 levels. High and low levels of CD4 expression on the T cell surface were discriminated on the basis of log fluorescence > 102 and ranging from 101 to 102, respectively. The presence of Mφ in the culture increased the amount of nonapoptotic GFP+ T lymphocytes that express low levels of CD4 on the cell surface by 30-fold (Fig. 4). In contrast, the barely detectable rate of nonapoptotic GFP+ T lymphocytes that express high levels of CD4 on the cell surface was not modified by the presence of Mφ in the culture (Fig. 4). A semipermeable membrane that blocks intercellular contacts between Mφ and PBL inhibited the appearance of the nonapoptotic GFP+ CD4low T lymphocyte subset in the culture (Fig. 4). These data suggest that intercellular contacts between Mφ and PBL result in the appearance of a subset of nonapoptotic GFP+ CD4+ T lymphocytes that express low levels of CD4 on the cell surface.
Critical role for Nef in the resistance to apoptosis in HIV-infected CD4+ T cells stimulated with TNF
HIV infection is characterized by a state of immune hyperactivation with increased levels of TNF in both serum and lymphoid tissue (1). TNF activation leads to increased activation of the NF-κB, which has been shown to deliver anti-apoptotic signals (52, 53, 54). We observed that membrane-bound TNF is expressed on the cell surface of Mφ following HIV infection (20), and that Mφ favor the appearance of a subset of nonapoptotic GFP-positive infected CD4+ T lymphocytes via intercellular contacts (Figs. 3,B and 4). To investigate the effect of TNF stimulation on the resistance to apoptosis in infected CD4+ T lymphocytes, we treated PBMC infected with HIV-GFP (MOI of 0.10) with a neutralizing anti-human TNF Ab. We observed that the rate of nonapoptotic GFP+ CD4+ T lymphocytes decreased by 75% in the culture following anti-human TNF treatment vs untreated control (Fig. 5,A). To further confirm the role of TNF in the resistance to apoptosis in infected CD4+ T cells, we used engineered CHO cells that express membrane-bound TNF (CHOTNF cells) (46). PBL were left uninfected or were infected with the wild-type HIV-89.6 strain (MOI of 0.10). Nine days later, either uninfected or infected PBL were cocultivated with either CHOTNF cells or control CHO cells at a ratio of 1:1 for 24 h, and the rate of apoptosis in CD4+ T lymphocytes was measured in the cultures. TNF stimulation decreased by 4- to 5-fold the apoptosis rate of CD4+ T lymphocytes in the HIV-infected culture, but not in the uninfected culture (Fig. 5,B). We then measured the rate of nonapoptotic productively infected CD4+ T cells following stimulation with TNF. PBL were infected with either HIV-GFP or HIV-GFPΔNef (MOI of 0.10), and 9 days later were cultivated alone or were cocultivated in the presence of either CHO cells or CHOTNF cells at a ratio of 1:1 for 24 h. Nonapoptotic GFP+ CD4+ T lymphocytes were observed only when PBL infected with the Nef-expressing virus HIV-GFP were cocultivated with CHOTNF cells (Fig. 5,C). Nonapoptotic GFP+ CD4+ T lymphocytes were not detected in PBL/CHOTNF cocultures infected with HIV-GFP and treated with neutralizing anti-human TNF Ab (data not shown). The cocultivation of either HIV-GFPΔNef-infected PBL with CHOTNF cells or of HIV-GFP-infected PBL with CHO cells did not result in a significant increase in the rate of nonapoptotic GFP+ CD4+ T cells in the culture (Fig. 5 C).
The two GFP-expressing viruses, HIV-GFP and HIV-GFPΔNef, are not isogenic due to the presence of an IRES sequence in HIV-GFP, which is absent in HIV-GFPΔNef (see Materials and Methods). Therefore, we directly tested the role of Nef for the resistance to apoptosis in TNF-stimulated CD4+ T cells using a plasmid expressing the nef open reading frame of HIV-89.6. We studied the effect of Nef on apoptosis in the CD4+ T cell line Jurkat. Jurkat cells were cotransfected either with a nef-expressing plasmid, pNef, or a control empty vector, pCtl, and a GFP-expressing plasmid (pEGFP), which was used as a marker for efficient cotransfection. Twenty-four hours later, transfected Jurkat cells were cultivated alone or were cocultivated with either CHO cells or CHOTNF cells at a ratio of 1:1 (Fig. 6). Twenty-four hours later, transfected cells were harvested and analyzed by two-color flow cytometric analysis to measure the rate of apoptosis in GFP-positive transfected cells. Cocultivation of Jurkat cells transfected with the nef-expressing plasmid pNef with CHO cells, or their cultivation alone, resulted in increased rates of apoptosis vs control (8–9% vs <0.5%; Fig. 6). In contrast, the frequency of apoptotic cells decreased by >90% when Jurkat cells transfected with the nef-expressing plasmid, pNef, were cocultivated with CHOTNF cells (Fig. 6). The blockade of apoptosis in TNF-stimulated Jurkat cells transfected with pNef was abolished when the culture was treated with a neutralizing anti-human TNF mAb (data not shown). These results show that TNF activation of Nef-expressing T cells can block Nef-mediated apoptosis in the CD4+ T cell line Jurkat.
NF-κB hyperactivation in Nef-expressing CD4+ T cells stimulated with TNF
TNF stimulation results in NF-κB activation that has been shown to deliver anti-apoptotic signals (52, 54). Cell types that show increased NF-κB activation, such as monocytes/Mφ, are usually less susceptible to HIV-induced apoptosis (53, 54). Therefore, we assessed whether increased levels of NF-κB might account for the blockade of apoptosis in Nef-expressing T cells stimulated with TNF. Jurkat cells were transiently cotransfected with pNef and a target plasmid that contains the luciferase reporter gene under the control of the NF-κB promoter, pNF-κBluc (50). Twenty-four hours later, transfected cells were either cultivated alone or cocultivated with CHO cells or CHOTNF cells at a 1:1 ratio (Fig. 7). Cells were harvested 24 h later, and luciferase activity was measured in cell lysates. TNF stimulation alone increased NF-κB activation by 5-fold in Jurkat cells vs the untreated control cells, as previously reported (55). In the absence of TNF stimulation, nef-transfected cells showed a 2-fold increase in NF-κB activation vs the untreated control. In contrast, the expression of Nef protein in TNF-stimulated Jurkat cells resulted in a 13-fold increase in NF-κB activation vs control, demonstrating a synergistic effect of TNF and Nef on NF-κB activation in Jurkat T cells. Increased NF-κB activation was not observed when a plasmid containing a mutated NF-κB site, pNF-κBmutluc, was used instead of pNF-κBluc (Fig. 7). These data show that Nef expression in TNF-stimulated T cells results in hyperactivation of the anti-apoptotic transcription factor NF-κB. This suggests that NF-κB hyperactivation might be involved in the resistance to apoptosis in Nef-expressing HIV-infected CD4+ T lymphocytes stimulated by TNF expressed on the surface of Mφ.
Our results demonstrate that Mφ can increase resistance to apoptosis in CD4+ T lymphocytes infected with HIV in vitro. Mφ hereby favor the appearance of productively infected, nonapoptotic T lymphocytes expressing low levels of cell surface CD4 through intercellular contacts. Our results also show that both stimulation with TNF and expression of the HIV protein Nef play a critical role in the generation of nonapoptotic productively infected CD4+ T cells.
Even though numerous studies have investigated the effects of HIV infection on T lymphocyte apoptosis, the resistance to apoptosis in CD4+ T lymphocytes mediated by Mφ in the context of HIV infection has not been reported so far. To discriminate between productively infected and either uninfected or latently infected cells, we used HIV reporter viruses expressing the GFP. In a previously reported study we used a Nef-defective GFP-tagged HIV-1 infectious clone to measure the rate of apoptosis in both infected and uninfected T cells (17). Because the Nef protein has been shown to be critical for the progression of HIV disease (42, 43, 44, 56, 57), in the present study, we compared two GFP-tagged infectious HIV-1 clones, expressing or not Nef, for the induction of apoptosis in infected CD4+ T cells. Using the GFP+ HIV-1 infectious clone expressing Nef, our data demonstrate that a population of productively infected CD4+ T lymphocytes does not undergo apoptosis when cultivated in the presence of Mφ. Although low viral growth in the infected cell population may leave some infected T lymphocytes GFP-negative, the detection of GFP-positive T lymphocytes certainly corresponds to productively infected cells. Therefore, we believe that GFP-positive CD4+ T lymphocytes that do not undergo apoptosis represent a population of cells, which are productively infected by HIV. Several explanations could account for the low amount of nonapoptotic productively infected CD4+ T lymphocytes detected in the presence of Mφ following infection with HIV-GFP. First, we observed that direct intercellular contacts between Mφ and PBL are required to observe the resistance to apoptosis in HIV-infected CD4+ T lymphocytes. It is possible that only a fraction of HIV-infected CD4+ T lymphocytes stay in contact with Mφ, thereby limiting protection from apoptosis. Infected CD4+ T lymphocytes that are not in direct contact with Mφ might lose the stimulation provided by membrane-bound TNF and therefore be more susceptible to apoptosis mediated by Nef. Second, the measurement of GFP expression might not be sensitive enough to detect low levels of productive infection. Therefore, infected cells producing low amount of virions might not be measured as GFP-positive using flow cytometric analysis. This could result in the underestimation of the pool of productively infected nonapoptotic cells. Also, the detection of low amounts of nonapoptotic, productively infected cells in the culture, might indicate that a majority of productively infected CD4+ T lymphocytes die via apoptosis. A high level of apoptosis in productively infected CD4+ T lymphocytes could result from the overexpression of the Nef protein following infection with HIV-GFP. We used a GFP-tagged recombinant HIV clone expressing the nef gene under the control of an IRES sequence which allows a high efficiency of translation of the Nef protein. Because Nef has been shown to sensitize infected CD4+ T cells to apoptosis (58), overexpression of Nef might result in increased apoptosis in CD4+ T lymphocytes infected with HIV-GFP.
TNF is expressed on the surface of activated Mφ following HIV infection (20), and has been reported to activate NF-κB. NF-κB, in turn, stimulates HIV-1 replication via the activation of the LTR, meanwhile delivering anti-apoptotic signals (31, 32, 52, 54). Our data show that TNF stimulation can favor resistance to apoptosis in HIV-infected CD4+ T lymphocytes. The fact that mostly uninfected T cells undergo apoptosis in the lymph nodes of HIV-infected individuals (18) suggests that viral factors could indeed protect infected T cells from apoptosis. The Vpr protein has been implicated in the protection of infected T cells from apoptosis (59). Endogenous Tat protein has been reported to protect T cells from apoptosis, whereas exogenous Tat induces apoptosis in bystander uninfected T cells (25, 26). However, when we infected PBL with HIV-GFP, which expresses all viral genes including vpr and tat, we did not observe a population of nonapoptotic infected CD4+ T lymphocytes, unless Mφ were present in the culture. This observation suggests that, apart from the expression of viral gene(s), stimulatory signals, such as TNF, that are provided by Mφ in the context of chronic immune activation, may be critical for the appearance of a population of nonapoptotic infected CD4+ T cells following HIV infection. Our data also suggest that Nef within the infected CD4+ T lymphocytes might interact with the TNF/TNF receptor pathway, thereby blocking the apoptotic process that is triggered when viral genes are expressed alone in the absence of additional cellular stimuli (60, 61). In human cells, two TNFRs with molecular masses of 55 kDa (TNFR1) and 75 kDa (TNFR2) have been identified and cloned (62, 63). The biological response to TNF is believed to be a result of the balance of multiple signals delivered via both TNFR1 and TNFR2. The resistance to apoptosis in productively infected CD4+ T lymphocytes could be explained by at least two mechanisms. Nef in the presence of TNF stimulation might favor increased activation of the anti-apoptotic transcription factor NF-κB, thereby blocking caspase-8 activation that is involved in the apoptotic process (64). Nevertheless, we cannot rule out that Nef can also increase the resistance to apoptosis by directly binding to signal transduction molecules involved in the apoptotic process, thereby prolonging the survival of the infected cell. Our data show that Mφ favor resistance to apoptosis in HIV-1-infected T lymphocytes that express low levels of cell surface CD4. The existence of a subpopulation of T cells that expresses low levels of CD4 and supports HIV-1 replication without demonstrating significant cytopathic effects, has been reported and may result in a lower incidence of CD4+ T cell superinfection and decreased CD4 cross-linking (23, 65, 66). Among the HIV-1 genes vpu, env, and nef, that have been implicated in down-regulating the levels of cell surface CD4 on infected cells, a stronger dependence on Nef function for the reduction of cell surface CD4 on primary T lymphocytes has been previously described (67).
The data presented in this study demonstrate that, in vitro, resistance to apoptosis in infected CD4+ T lymphocytes is increased by Mφ, suggesting a pivotal role for APCs in the formation of a pool of nonapoptotic productively infected T cells in vivo. The long-lasting quiescence and later progression of HIV disease may be explained by the persistence of cellular reservoirs of virions. Mφ represent a main reservoir of virions and are resistant to cell death until advanced disease (68). We describe here that, in addition to latently infected CD4+ T lymphocytes (69), resistance to apoptosis in productively infected CD4+ T lymphocytes might allow the formation of reservoirs of virions. In fact, the ability of HIV-1 and SIV to infect and replicate at low levels in minimally activated T cells has been reported recently to generate a population of long-living productively infected CD4+ T cells in vivo (70). The inhibition of apoptosis in HIV-infected CD4+ T lymphocytes by pharmacological inhibitors of cell death has been reported to result in increased viral replication (71). Together, these data suggest that resistance to apoptosis in productively infected CD4+ T lymphocytes may lead to a sustained production of virions and favor the establishment of a persistent infection in vivo.
In conclusion, our study shows that a population of productively HIV-infected CD4+ T lymphocytes is resistant to apoptosis. Resistance to apoptosis in infected CD4+ T lymphocytes depends on the presence of Mφ, and involves both the expression of the Nef protein and T cell activation by TNF. A better understanding of the mechanisms underlying the survival of infected CD4+ T lymphocytes is likely to lead to new therapeutic approaches, which could help to clear the reservoirs of virions in HIV-infected individuals.
We thank the AIDS Research and Reference Program (National Institute of Allergy and Infectious Diseases, National Institutes of Health) for reagents used in this study.
This work was supported by a grant from the American Foundation for AIDS Research (02631-26-RGI; to G.H.), by institutional funds from the University of Texas Medical Branch, by grants from the National Institutes of Health of the U.S. Public Health Service NS 38414 (to W.A.O.), AI-39400, and HL-42257 (to J.L.S. and T.C. G.), and by a grant from the German National Science Foundation MA (2057/1-1; to U.M.).
Abbreviations used in this paper: GFP, green fluorescent protein; Mφ, macrophage; TNF, TNF-α; LTR, long terminal repeat; IRES, internal ribosome entry site; RT, reverse transcriptase; PerCP, peridinin chlorophyll protein; CHO, Chinese hamster ovary; MOI, multiplicity of infection.