Vpu Increases Susceptibility of Human Immunodeficiency Virus Type 1-Infected Cells to Fas Killing

Abstract

The importance of the Fas death pathway in human immunodeficiency virus (HIV) infection has been the subject of many studies. Missing from these studies is direct measurement of infected cell susceptibility to Fas-induced death. To address this question, we investigated whether T cells infected with HIV are more susceptible to Fas-induced death. We found that Fas cross-linking caused a decrease in the number of HIV-infected Jurkat T cells and CD4⁺ peripheral blood leukocytes (PBLs). We confirmed this finding by demonstrating that there were more apoptotic infected than uninfected cells after Fas ligation. The increase in sensitivity of HIV-infected cells to Fas killing mapped to vpu, while nef, vif, vpr, and second exon of tat did not appear to contribute. Furthermore, expression of Vpu in Jurkat T cells rendered them more susceptible to Fas-induced death. These results show that HIV-infected cells are more sensitive to Fas-induced death and that the Vpu protein of HIV contributes to this sensitivity. The increased sensitivity of HIV-infected cells to Fas-induced death might help explain why these cells have such a short in vivo half-life.

The in vivo half-life of CD4⁺ T cells infected with human immunodeficiency virus (HIV) is between 1 and 2 days (55). Why do infected cells have such a short in vivo half-life? Does the virus directly cause the death of infected cells, or is the immune system very efficient at clearing infected cells from the body? One mechanism used by the immune system to eliminate unwanted T cells is the Fas/Fas ligand (FasL) pathway (47). Fas is expressed on T cells, and its ligation by FasL can lead to apoptosis of the cell. The FasL used to cause this death can be expressed either on the same cell or on a neighboring cell (10, 14, 29). The cellular cascade of events from Fas ligation to apoptosis has been extensively studied (11). The cross-linking of Fas leads to the recruitment of FADD (FLICE-associated death domain) to the receptor complex. FADD recruits the zymogen FLICE (caspase 8) to the Fas receptor complex through interactions with its death effector domain. FLICE then cleaves itself (46) to form an active caspase which in turn activates other caspases. The nuclease, caspase-activated DNase, is activated by caspase 3 (57), and the cellular DNA is cleaved, killing the cell.

Controversy surrounds the issue of whether the Fas death pathway is a significant mechanism of infected cell death. Many studies have tried to determine the mechanism by which HIV kills cells in vitro. One report showed a small contribution of Fas/FasL to the death of infected cells in vitro (37). However, most in vitro systems have demonstrated that HIV causes death of cultured T cells in a Fas-independent manner (19, 21, 50). Apoptosis of peripheral blood lymphocytes (PBLs) from HIV-infected people can be detected immediately ex vivo (13) and following in vitro culture (17, 23, 24). The apoptosis seen upon in vitro culture cannot be blocked by blocking the Fas/FasL pathway (19, 50) or completely blocked by using caspase inhibitors (31). This may indicate that either the cells are primed to undergo apoptosis in vivo and are already past the point where blocking the Fas pathway or caspase inhibitors can work, or they are dying in a caspase-independent manner. Although culturing PBLs from HIV-positive people may help elucidate why, in general, CD4⁺ cells die, <1% of those cells are productively infected (12) and these studies do not elucidate the mechanism by which the infected cells die in vivo.

The question remains, why do infected cells die in vivo? Others have addressed this question by attempting to inhibit HIV-induced death in vitro with agents that block Fas/FasL signaling. We have chosen to test directly whether HIV-infected cells exposed to Fas cross-linking are more susceptible than uninfected cells to death. Addressing the question in this manner in vitro may give insight into whether infected cells are susceptible to this pathway of death in vivo, where there are many more cell types that might express FasL than in cell culture systems.

One possible source of FasL in infected individuals is cytotoxic T cells (CTLs) (4, 30, 42). The contribution of HIV-specific CTL response to decreasing viral load is disputed (25–27, 39, 56, 72). A recent study that used more sensitive techniques to measure antigen-specific CTLs (3) found that the number of CD8⁺ HIV-specific CTLs is inversely correlated to the number of infected cells (51). Thus, HIV-specific CTLs might play an important role in killing infected cells. Furthermore, there have been reports that macrophages express FasL (9, 32) and that infection of macrophages by HIV increases FasL expression (9). Also, macrophages from HIV-infected people are more able to kill CD4⁺ cells than macrophages from uninfected individuals. This death can be partially blocked by blocking the Fas death pathway (52). HIV-specific CTLs and macrophages which are low to absent in in vitro cultures of infected cells might be a source of FasL in vivo.

Here, we confirmed the results of Kobayshi et al. (35) in finding that infected T-cell lines are more susceptible to Fas-induced death and further show that HIV-infected primary CD4⁺ lymphocytes are also more susceptible to Fas-induced death. We also demonstrate that at least part of HIV’s increased susceptibility to Fas-induced death maps to Vpu and that Vpu can function alone in its ability to augment Fas-induced death when tested in Jurkat cells.

MATERIALS AND METHODS

Cell culture.

293 cells and CD4⁺ β-galactosidase (β-Gal) indicator cells were maintained in Dulbecco modified Eagle medium supplemented with 7% fetal calf serum (FCS; Gemini Bioproducts, Calabasas, Calif.). CD4⁺ β-Gal indicator cells were also cultured in the presence of G418 (200 μg/ml; Mediatech, Inc., Herndon, Va.) and hygromycin B (100 μg/ml; Sigma, St. Louis, Mo.). Jurkat T cells, CEM T cells, and CD4⁺ human PBLs were cultured in RPMI supplemented with 10% FCS, and penicillin G (100 U/ml), and streptomycin (100 μg/ml). PBLs were isolated by Ficoll-Hypaque (Pharmacia, Piscataway, N.J.) density gradient centrifugation of heparinized blood obtained from healthy donors. CD4⁺ cells were purified by negative selection. PBLs were incubated with 20 μg of anti-CD8 (OKT8; American Type Culture Collection) per ml and applied to a column of goat anti-mouse immunoglobulin G-coated Immulon beads (Biotecx Laboratories, Houston, Tex.). The purified cell population, 80 to 95% CD4⁺, was cultured in medium containing phytohemagglutinin (2.5 μg/ml; Murex, Research Triangle Park, N.C.) for 3 days before infection.

Preparation of viral stocks.

Viral constructs p83-2 (5′ half of NL4-3), p83-10 (3′ half of NL4-3), p1971-1 (5′ ΔVif), p210-19 (5′ ΔVpr), p210-13 (3′ ΔVpu), and p210-5 (3′ ΔNef) were obtained from Ronald Desrosiers from the AIDS Research and Reference Reagent Program (20). Each half of the proviral DNA (3 μg) was cut with EcoRI, ligated, and used to transfect 293 cells by the 2-bromoethanesulfonic acid transfection method (58). Culture supernatants were collected 24 to 48 h after transfection and used to infect CEM T cells in the presence of DEAE-dextran (20 μg/ml). At 7 to 10 days after infection, culture supernatants were clarified by centrifugation at 250 × g, collected, and frozen at −75°C. Viral titers were measured on CD4⁺ β-Gal indicator cells as described previously (33). The mutant lacking the second exon of tat (TatSE) was made by changing the sequence of the first amino acid of the second exon of TatSE from CAG to TAG by PCR cloning in the vector p83-10. Sequence analysis was performed to ensure that only the desired nucleotide was changed. Virus was made as described above. To ensure that the NL4-3ΔTatSE virus had not reverted its genotype, viral stocks were used to infect Jurkat cells, cell lysates were collected, and PCR was used to amplify the Tat region of the cellular proviral DNA. The Tat region was cloned into a sequencing vector, and five bacterial colonies of each viral stock were checked for reversion to wild type by sequence analysis. All colonies tested contained the stop codon mutation and had not reverted.

Fas killing assays.

Jurkat cells were infected in duplicate with viral supernatants collected from CEM T cells in the presence of DEAE-dextran (20 μg/ml), using a range of multiplicities of infection (0.012 to 0.075). Two days after infection, the cells were plated at 5 × 10⁵ to 10 × 10⁵/ml in medium containing anti-Fas antibody CH-11 (0, 6, 25, or 100 ng/ml; Oncor, Gaithersburg, Md., or Upstate Biotechnology, Inc., Lake Placid, N.Y.) or anti-Fas antibody M33 (100 ng/ml; Immunex, Seattle, Wash.) (1); 22 to 24 h later, the cells were harvested, fixed in 1% paraformaldehyde, washed twice in phosphate-buffered saline (PBS), and permeabilized with 0.1% saponin in the presence of 10% FCS in PBS. Then 2.5 μg of anti-Gag antibody K57-RD1 or K57-fluorescein isothiocyanate (FITC) (Coulter, Hialeah, Fla.) was added, and the cells were incubated at room temperature for 30 min. The cells were washed twice in staining buffer (balanced salt solution, 2% FCS, 0.1% sodium azide) and kept at 4°C until fluorescence-activated cell sorting analysis. Cells were analyzed with either a FACSCalibur or FACScan flow cytometer (Becton Dickinson, San Jose, Calif.). CD4⁺ PBLs were infected 3 days after phytohemagglutinin activation at a multiplicity of infection of 0.06 to 0.2. Cells were plated in medium in the presence of recombinant human interleukin-2 (20 IU/ml; R&D Systems, Minneapolis, Minn.) at a concentration of 1.5 × 10⁶ to 2 × 10⁶ cells/ml. Three to four days after infection, the cells were counted and plated at a concentration of 10⁶ cells/ml with or without anti-Fas antibody CH-11 or M33 (400 ng/ml) in medium supplemented with recombinant human interleukin-2 (5 IU/ml); 48 h later, the cells were harvested and analyzed for Gag expression as described above. Vpu Jurkat cell clones, control Jurkat cell clones, and parental Jurkat cells were plated at 0.5 × 10⁶ to 1 × 10⁶ cells/ml in medium that contained either CH-11 or M33 anti-Fas antibody (25 ng/ml) for 22 to 24 h. The cells were harvested and stained for annexin V binding as described below.

Annexin V binding.

NL4-3, NL4-3ΔVpu, and mock-infected Jurkat T cells were induced to undergo Fas death as described above. The cells were harvested 20 to 24 h after Fas antibody incubation, washed once in PBS, and incubated with 5 μl of annexin V-FITC (Caltag Laboratories, Burlingame, Calif.) in 100 μl of binding buffer (BB; 2.5 mM CaCl₂, 10 mM HEPES-NaOH [pH 7.4], 140 mM NaCl) for 15 min in the dark. The cells were then washed twice in BB, resuspended in 150 μl of BB and 50 μl of 4% paraformaldehyde, and incubated for 30 min. After fixing, the cells were washed twice in BB and then stained for Gag antigen as described above except that 2.5 mM CaCl₂ was included in the mixture. Following Gag staining, the cells were washed twice in BB, resuspended in BB, and kept at 4°C until analysis on a FACScan (Becton Dickinson). In the Fas killing experiment with the Jurkat cell clones described above, annexin V binding was tested as follows. The cells were washed once in PBS and then resuspended in 100 μl of BB plus 4 to 5 μl of annexin V-biotin (Caltag) for 15 min at room temperature. The cells were washed with 1 ml of BB and incubated with 4 μl of streptavidin-peridinin chlorophyl protein (PerCP) (Becton Dickinson) in 100 μl of BB for 15 min in the dark at room temperature. The cells were placed in 400 μl of cold BB and analyzed on a FACScan flow cytometer within 1 h of staining.

Jurkat cell clones.

pCLIRES-GFP was made by inserting an internal ribosome entry site (IRES)-green fluorescent protein (GFP) cassette, kindly provided by Brian Schaefer, into the BamHI site in the vector pCLXSN, described previously (49). pCLVPU-GFP was derived from pCLRIES-GPF by the addition of the coding sequence of Vpu into the EcoRI site in front of the IRES-GFP cassette. The coding sequence of Vpu was obtained from NL4-3 by performing PCR on p83-10, a vector containing the 3′ half of NL4-3 (see above). The addition of an unaltered Vpu to the vector was confirmed by sequence analysis. The vectors pCLIRES-GPF and pCLVPU-GFP were transfected either into the Phoenix amphotropic packaging cell line (kind gift of Gary P. Nolan) or into 293 cells in the presence of pCLAmpho (49) by the Lipofectamine (Gibco BRL, Grand Island, N.Y.) or 2-bromoethanesulfonic acid (58) transfection method, respectively. Virus was collected 24 to 48 h after transfection and used to infect Jurkat T-cell clones in the presence of DEAE-dextran (20 μg/ml). Two days later, the cells were analyzed on a FACScan or FACSCalibur flow cytometer for the number of GFP-positive cells; 0.3 to 0.5 GFP-positive Jurkat cells were cultured per well in a 96-well plate in media containing G418 (1 mg/ml). Neomycin-resistant colonies were grown and checked for the presence of GFP by flow cytometric analysis. GFP-positive clones from pCLVPU-GFP were further screened by Western blotting for the presence of Vpu.

Western Blotting.

Cell lysates were collected by lysing 3 × 10⁶ cells in 50 μl of TENC {50 mM Tris-HCl (pH 7.4), 5 mM EDTA, 150 mM NaCl, 0.5% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS)}, followed by the addition of 5 μl TENC–2% sodium deoxycholate. The cells were incubated on ice for 15 min and then subjected to a 5-min centrifugation at 700 × g. The amount of cell protein was determined by using the DC protein assay (Bio-Rad Laboratories, Hercules, Calif.). Equal amounts of cell lysate were run on a sodium dodecyl sulfate–15% polyacrylamide gel and transferred to nitrocellulose in 10 mM CAPS–10% methanol at 1 mA for 1 h at 4°C. The blots were blocked with 5% dry milk in PBS overnight at 4°C. The blots were then probed with either rabbit anti-Vpu anti-serum (1:100 dilution; AIDS Research and Reference Reagent Program) (43) or mouse anti-Gag antibody K57-FITC (0.5 μg/ml Coulter) in PBS–5% dry milk for 1 to 2 h at room temperature. The blot was washed five times in PBS-T (PBS, 0.2% Tween 20), incubated with a 1:2,000 dilution of horseradish peroxidase-linked protein A (Amersham, Arlington Heights, Ill.) in PBS–5% dry milk for 1 h at room temperature, washed five times in PBS-T, exposed to Super Signal chemiluminescent substrate (Pierce, Rockford, Ill.) for 1 min, exposed to film, and developed. Blots were stripped (100 mM 2-mercaptoethanol, 2% sodium dodecyl sulfate, 62.5 mM Tris-HCl [pH 6.7]) for 30 min at 50°C, washed in PBS-T, and probed as described above.

Analysis of data.

Flow cytometric analysis was performed with either the Cellquest or Cell Lysis analysis software (Becton Dickinson). Statistical analysis was performed with JMP software (SAS Institute Inc., Cary, N.C.). Two-tailed independent t test or analysis of variance models, which account for the effects of virus type, dose, and experiment, were used. In the analysis of variance models, tests for differences between virus type were performed by using individual linear contrasts in which the effects of dose and experimental variation have been removed.

RESULTS

HIV infection renders T cells more susceptible to Fas killing.

We wanted to ascertain if infected cells were more or less resistant to Fas-induced death than their uninfected counterparts. To determine the susceptibility of HIV-infected cells to Fas-induced apoptosis, we used the Jurkat T-cell line because it is responsive to Fas-induced death upon the addition of certain anti-Fas antibodies. We used the HIV type 1 (HIV-1) isolate NL4-3 because of the readily available mutants that could be used to map any effect of HIV infection on Fas killing. NL4-3-infected Jurkat cells were cultured for 20 to 24 h in the presence of either a Fas antibody that induced death (CH-11) or an antibody that bound Fas but did not cause death (M33) (1). The number of infected cells was then determined by staining cells for the presence of the intracellular HIV antigen Gag and analyzing fluorescence patterns with flow cytometry (Fig. 1B). We found that the number of infected cells in the live population was reduced in the cultures treated with CH-11 but not M33 (Fig. 1A). We determined the percent survival by dividing the number of live infected cells in the CH-11 treatment group by the number of live infected cells in the M33 treatment group (Fig. 1A and B). The killing induced by antibody CH-11 was dose responsive, with the lowest dose (6 ng/ml) resulting in more infected cell survival (77%) than the 25-ng/ml dose (68%) (Fig. 1A). However, the effect plateaued at higher doses, with 65% survival seen at the 100-ng/ml dose of CH-11. This effect was reproducible over a large range of infected cell percentages (26 to 85%). We conclude that infected cells are more susceptible to Fas-induced death than uninfected cells in the same culture.

FIG. 1.

HIV-infected Jurkat T cells are more susceptible to Fas killing. This effect maps to Vpu. (A) Jurkat T cells infected with NL4-3 were incubated for 22 h in the presence of the indicated amounts of anti-Fas antibody CH-11. Survival was calculated by the following formula: % Gag⁺ cells following CH-11 antibody treatment/% Gag⁺ cells following M33 antibody treatment. Percent survival was calculated for each concentration of CH-11 used (0, 6, 25, and 100 ng/ml). The results are shown as mean ± standard error of 12 independent experiments with duplicate samples in each experiment. (B) Measurement of viral infection and enhanced survival of NL4-3ΔVpu-infected cells. Histograms show infected Jurkat T cells treated with antibodies M33 and CH-11 (100 ng/ml) stained for the presence of intracellular Gag. The cells represented in the histogram were gated through the live cell population. Numbers over markers represent proportions of Gag⁺ cells in the population; numbers at the right represent percent survival of Gag⁺ cells treated with 100 ng of antibody CH-11 per ml as calculated for panel A. The apparent increase in Gag⁺ cells seen with the NL4-3ΔVpr compared to NL4-3 is not reproducible and is most likely due to experimental variation. (C) Fas killing experiments of NL4-3 and mutant viruses were performed as described in Materials and Methods. NL4-3 is represented in all graphs with square boxes; individual mutant viruses, indicated in the upper right of each graph, are depicted by diamonds. The graphs represent the means ± standard errors of eight, three, four, two, or three independent experiments, each performed in duplicate for NL4-3ΔVpu, NL4-3ΔNef, NL4-3ΔVpr, NL4-3ΔVif, or NL4-3ΔTatSE compared to NL4-3 in those experiments, respectively. Percent survival is represented as in panel A. (D and E) Western blots of infected cell lysates. (D) Blot probed with rabbit anti-Vpu antiserum; (E) the blot in panel D, stripped and reprobed with a monoclonal mouse anti-Gag antibody.

The increased sensitivity to Fas killing maps to Vpu.

To determine which viral gene(s) contributed to the increased sensitivity of HIV-infected cells to Fas killing, we infected Jurkat T cells with replication-competent proviral genomes (NL4-3) deleted singly in either Vpr, Vif, Nef, TatSE, or Vpu. We determined the percent survival of the mutant viruses in comparison to wild-type NL4-3 in the Fas killing assay described above (Fig. 1C). We found no large difference in the ability of viruses deleted in Vpr, Vif, Nef, or TatSE compared to wild-type NL4-3 to induce sensitivity to Fas killing (Fig. 1B and C). However, there was on average a 20% increase in survival of infected cells deleted in Vpu over all tested concentrations of antibody CH-11 (Fig. 1B and C). To confirm that the virus used was deficient for Vpu, we performed a Western blot analysis of cell lysates from cells infected with either NL4-3 or NL4-3ΔVpu (Fig. 1D). When the cell lysates were probed with the anti-Vpu antiserum, Vpu was seen only in infections with wild-type virus. As a control, equal amounts of Gag were detected in both the NL4-3- and the NL4-3ΔVpu-infected cell lysates (Fig. 1E). Hence, expression of Vpu in HIV-infected cells increases their susceptibility to Fas killing.

Primary cells infected with HIV are more sensitive to Fas-induced death; this effect is diminished after infection with Vpu-deleted viruses.

To determine whether the results obtained with Jurkat T cells were reproducible in primary cells, we performed the Fas killing experiments with primary human PBLs enriched for CD4⁺ cells. We infected CD4⁺ PBLs with NL4-3 or NL4-3ΔVpu virus and then cultured these cells in the presence or absence of the anti-Fas antibody CH-11 or M33 for 48 h. The percentage of infected cells remaining in the Fas-treated cultures was determined by staining the infected cells intracellularly with an anti-Gag antibody and subsequent flow cytometric analysis (Fig. 2A). Figure 2B shows the results of three independent experiments with CD4⁺ PBLs from two different blood donors. Not only were the NL4-3-infected cells more susceptible to Fas killing (60% survival), but the effect was partially reversed by using a virus lacking Vpu (83% survival) (Fig. 2B). This result was statistically significant, using the two-tailed independent t test at P < 0.025. Thus, primary cells show the same effect seen in Jurkat cells: HIV-infected cells are more susceptible to Fas killing, and deletion of Vpu from the virus diminishes this effect.

FIG. 2.

HIV-infected CD4⁺-enriched PBLs are more susceptible to CH-11-induced killing. Deletion of Vpu diminishes this effect. CD4⁺-enriched PBLs were infected with either mock, NL4-3, or NL4-3ΔVpu clarified supernatants; 3 to 4 days after infection, the cells were incubated with anti-Fas antibody M33 or CH-11 (400 ng/ml) for 2 days. The cells were then harvested and stained for the presence of the HIV Gag antigen in order to visualize infected cells. (A) Histograms from a representative experiment. Numbers over markers represent proportions of Gag⁺ cells in the population; numbers at the right represent the percent survival of infected cells after CH-11 treatment, as described in the legend to Fig. 1A. (B) Mean survival ± standard error of three independent experiments with two different blood donors. The asterisk indicates a significant difference between viruses as determined by a two-tailed independent t test at P < 0.025.

HIV-infected cells are more apoptotic in response to Fas killing.

To determine whether the decrease in infected cells in response to antibody CH-11 treatment can be attributed to a higher level of infected cell apoptosis, we stained Jurkat T cells after Fas antibody treatment for apoptosis and infection (Fig. 3). We saw no consistent difference between apoptosis in infected cells and apoptosis in mock-infected cultures when all cells (live, apoptotic, and dead) in the scatter plot were used to assess apoptosis and infection (data not shown). This was due either to an inability to detect late-stage apoptotic infected cells or to the loss of these cells from the culture. This interpretation was substantiated by our observation that the intensity of the Gag stain decreased in apoptotic cells and that there was an overall loss of cells from Fas CH-11-treated cultures (data not shown). To circumvent these problems, we examined apoptosis only in the live population of cells (R1; Fig. 3A to D). Annexin V staining appears to be an early step in apoptosis, occurring before cell shrinkage and changing of the scatter properties of the cell (36). Furthermore, cells in the live R1 region excluded compounds like 7-aminoactinomycin D that stain necrotic cells (data not shown). By gating through the live population of cells, we detect only cells in the early phase of apoptosis, not annexin-positive, necrotic cells. With this live gate, we detected a significant increase in the apoptosis in NL4-3-infected cells over mock infected cells (P < 3⁻¹⁰) (Fig. 3E). Also, consistent with our findings that infected cell loss is lower in cells infected with NL4-3ΔVpu, there was significantly less apoptosis in NL4-3ΔVpu-infected cells than in NL4-3 infected cells (P < 6⁻⁵) (Fig. 3E).

FIG. 3.

CH-11 treatment results in more apoptosis in HIV-infected cells than in uninfected cells. Jurkat T cells infected with mock, NL4-3, or NL4-3ΔVpr clarified supernatants were treated with anti-Fas antibody CH-11 (0, 6, 25, or 100 ng/ml) or M33 (100 ng/ml). After 22 h of treatment, the cells were stained with annexin V-FITC, fixed, and stained for Gag antigen; 60,000 to 120,000 events were collected per sample on a FACScan (Becton Dickinson). (A to C) SSC, side scatter; FSC, forward scatter. A representative experiment of three independent experiments is shown. Live cells are shown gated in the R1 region of the scatter plots. Dot plots of annexin V-versus Gag-stained cells gated through the R1 region are shown immediately to the right of the corresponding scatter plot. Numbers in the lower and upper right corners denote the proportions of annexin V-positive cells in the lower and upper portions of the grid, respectively. (D) Representative graph from one of three independent experiments described above. Squares denote NL4-3 live, Gag⁺, and annexin V⁺ cells; diamonds denote NL4-3ΔVpu live, Gag⁺, and annexin V⁺ cells; circles denote mock-treated, live, Gag⁻, annexin V⁺ cells. (E) Statistical analysis of three independent experiments, performed by using the analysis of variance models which account for the effects of virus type, dose, and experiment. Tests for differences between virus type were performed by using individual linear contrasts in which the effects of dose and experimental variation have been removed.

We also detected slightly higher levels of apoptosis in the CH-11-treated Gag⁻ population in infected cultures (Fig. 3A to C). This higher level of apoptosis could be due to an increase in death of uninfected cells in infected cultures or to death of newly infected cells that have yet to express sufficient levels of Gag to be detected by our antibody.

Vpu-expressing Jurkat T-cell clones are more susceptible to Fas killing.

To determine if Vpu alone can result in an increased sensitivity to Fas-induced death or if other HIV gene products are also required, we expressed Vpu in Jurkat cells and tested their susceptibility to death induced by Fas cross-linking. We created Jurkat T-cell clones expressing either Vpu linked by an IRES to GFP or GFP expressed alone (control clones) (Fig. 4A). We were able to isolate several Vpu and control Jurkat cell clones that expressed GFP. However, we were not able to obtain clones that expressed levels of Vpu as high as those seen in infected cells (Fig. 4B). It is possible that there are differences in the strength of the HIV promoter and the murine long terminal repeat promoter used in our clones. Alternatively, high levels of stable Vpu expression might be toxic to the cell. We stained each clone for surface Fas expression and isolated four control and three Vpu-expressing clones in which surface expression of Fas was equivalent to that in the parental Jurkat T-cell line (data not shown). The parental Jurkat cell line and the clones were incubated with anti-Fas antibodies CH-11 and M33, and the amount of cell death was determined by annexin V staining. Figure 4C shows the average amount of cell death over the parental Jurkat cell line for the control clones (0.4%) and the average amount of death for the Vpu-expressing clones (18%). The difference in Fas killing between the control and Vpu-expressing clones was statistically significant at P < 0.0002.

FIG. 4.

Jurkat cells expressing Vpu are more sensitive to CH-11-induced death. (A) The construct pCLIRES-GFP was derived from the vector pCLXSN described previously (49). pCLXSN was modified by the addition of an IRES-GFP cassette, a kind gift of Brian Schaefer. This vector was used as the control vector. pCLVPU-GFP was obtained by adding the NL4-3 coding sequence of Vpu in front of the IRES-GFP cassette in pCLIRES-GFP. The addition of an unaltered Vpu to the vector was confirmed by sequence analysis. CMV, cytomegalovirus; SV, simian virus 40; LTR, long terminal repeat. (B) Western blot of NL4-3-infected cells, one control clone, and the three Vpu-expressing clones probed for Vpu by using rabbit anti-Vpu antiserum. Equal amounts of protein were loaded per lane. Vpu is indicated by the arrow. The NL4-3-infected cells were 62% Gag⁺ as detected by flow cytometry. (C) The parental Jurkat cell line, four control, and three Vpu-expressing Jurkat cell clones were incubated with anti-Fas antibody CH-11 or M33 (25 ng/ml) for 22 to 24 h. The cells were collected and stained for annexin V-biotin followed by incubation with streptavidin-PerCP; 20,000 to 30,000 cells were collected on a Becton Dickinson FACScan flow cytometer. CH-11-induced death was determined by subtracting the number of annexin V-positive cells in the M33 treatment group from the number in the CH-11 treatment group. M33-treated cells ranged from 3 to 8% annexin V positive. The amount of CH-11-induced death of the clones was subtracted from the amount of CH-11-induced death of the parental Jurkat cells in each experiment. CH-11-induced death of parental Jurkat cells ranged from 52 to 65%. The bar graph represents the mean of three independent experiments ± standard error. The asterisk represents the significant difference between the control clones and the Vpu-expressing clones as determined by analysis of variance at P < 0.0002.

DISCUSSION

In this study, we assessed the susceptibility of HIV-infected cells to Fas-induced death. We showed that Fas receptor cross-linking by the anti-Fas antibody CH-11 decreased the number of live infected cells in the culture and resulted in increased levels of infected cell apoptosis. Our results are in agreement with those of Kobayshi et al., who found a decrease in cell viability in HIV-infected cultures treated with an anti-Fas antibody (35). We have expanded on the results of Kobayshi et al. by testing primary PBLs. Also, we increased the sensitivity of the assay by using flow cytometry to quantify the number of infected cells lost. This assay allowed us to perform mapping studies to determine which viral gene is responsible for the increased Fas killing of infected cells.

We found that deleting Nef, Vif, Vpr, or TatSE from HIV-1 had little or no effect on the susceptibility of infected cells to Fas-induced death. Of the gene products tested, only Vpu seemed to have a large effect on susceptibility to Fas-induced apoptosis. Deletion of Vpu from HIV increased survival in response to Fas cross-linking by an average of 20% in Jurkat T cells and by 23% in CD4⁺-enriched PBLs. We also detected enhanced death in Jurkat cell clones expressing Vpu compared to control clones. It is interesting that even the low levels of Vpu expressed in our Jurkat cell clones were correlated with increase Fas-induced death. It appears that levels of viral RNA are about 4- to 15-fold lower in PBLs in vivo than in tissue culture-infected PBLs (28). Thus, while levels of Vpu in the Jurkat cell clones might not approximate in vitro infection levels, they could be closer to levels seen in an in vivo infection.

Deleting Vpu from HIV did not completely abolish the sensitivity of HIV-1-infected cells to Fas killing. It has been reported that CD4 cross-linking by Env and anti-CD4 antibodies can cause an increase in Fas expression levels (2, 53, 74) and that Tat can increase FasL expression (75). Also, addition of Tat and Env gp120 exogenously to cells increases their sensitivity to Fas-induced death (75). Furthermore, expression of Tat alone in cells has been shown to increase their susceptibility to Fas-induced death (37). Thus, it is possible that in addition to Vpu, the first exon of Tat and gp120 also act to increase the susceptibility of HIV-infected cells to Fas killing.

How might Vpu make cells more sensitive to Fas killing? Vpu is a 16-kDa type I integral membrane protein (43) that has been detected in the Golgi and endoplasmic reticular membranes (60). Since there are no antibodies that can sensitively detect the presence of small amounts of Vpu, we do not know if small amounts of Vpu are found in other cellular membranes or in the virion. Vpu has been ascribed two separate functions in the life cycle of HIV. One is its ability to enhance virus release from infected cells (34, 68, 71, 79). This property is not restricted to HIV, as Vpu can also enhance the release of virus particles from retroviruses distantly related to HIV (22). The other known role of Vpu in the virus life cycle is the disruption of envelope glycoprotein-CD4 interactions in the endoplasmic reticulum (76, 77). Vpu appears to interact with the cytoplasmic tail of CD4, targeting it to a proteosome where CD4 is degraded (8, 18, 40, 44, 61, 73, 77). The two functions of Vpu can be separated into distinct regions of the protein. The transmembrane region of Vpu appears to be responsible for the increase in virus release, while its cytoplasmic tail is needed for CD4 degradation (62, 64).

Vpu is structurally similar to the ion channel protein M2 of influenza virus (67). This similarity has led some workers to investigate the potential ion channel properties of Vpu. Vpu seems able to form cation-selective ion channels in planar lipid bilayers, in Escherichia coli, and in Xenopus oocytes (16, 63). The ion channel activity of Vpu appears to map to the transmembrane region of the protein. Scrambling the amino acids in the transmembrane portion of Vpu abolishes its ion channel-forming ability while still allowing Vpu to insert into membranes (63). Because the ion channel and the increased virus release function of Vpu both map to the transmembrane region, it has been postulated that Vpu’s formation of ion channels augments virus release. However, the means by which Vpu’s ion channel activity might increase virus release is not clear. Although the idea that Vpu forms ion channels is gaining wider acceptance, more work needs to be done to clearly demonstrate that it occurs in infected cells (38).

It is tempting to speculate that the ion channel properties of Vpu cause the cells to be more responsive to apoptotic stimuli. Perturbations of K levels in neurons and decreased K levels in T cells have been shown to cause apoptosis (6, 15, 48, 65, 70). Also, Bcl-2, Bcl-XL, and BAX have all been shown to have ion channel properties (5, 45, 58, 59, 61). These proteins play an important role in regulating apoptosis in a wide variety of cells (54, 80). BAX and Bcl-2 also seem to regulate mitochondrial membrane potential (ΔΨ) (69, 78). Perturbations of ΔΨ appear to be one of the earliest events of apoptosis. Because Vpu might also have ion channel properties, we tried to ascertain if Vpu might also perturb ΔΨ. These preliminary studies did not show any correlation between Vpu’s expression and changes in ΔΨ (data not shown). Other potential mechanisms by which Vpu increases sensitivity to Fas killing are currently being studied in our lab.

Why would HIV encode a gene that causes it to be more susceptible to Fas-induced death? Since there is no homolog of Vpu in simian immunodeficiency virus (SIV), it has been difficult to determine whether this gene contributes to the pathogenesis of the virus. However, studies have been performed with a chimeric virus, SHIV, that contains the 5′ half of SIV and 3′ half of HIV HXB-2. Since the Vpu start codon of HXB-2 has been altered to ACG, it is also altered in the chimera, SHIV. When SHIV was recovered from a pig-tailed macaque that developed AIDS-like symptoms, the ACG of Vpu had reverted to ATG (66). Also, in a study (41) that used SHIVs that either did or did not contain a functional Vpu to infect cynomolgus monkeys, monkeys infected with the Vpu containing viruses showed higher levels of viral RNA in plasma and greater Env variation than monkeys infected with a Vpu-nonexpressing SHIV. The greater Env variation seen in Vpu⁺ SHIV infection is more characteristic of human infection with HIV (41). Furthermore, HIV-2, which also does not contain Vpu, has retained the ability to enhance virion release. This function performed by Vpu in HIV-1 is performed by Env in HIV-2 (7). From these studies, it appears that Vpu has a function that is beneficial to the viral life cycle. Thus, any deleterious effects of Vpu, such as an increase in Fas-induced death, might be overshadowed by the ability of the virus to augment virus replication.