Commitment to Apoptosis in CD4 T Lymphocytes ... - Journal of Virology

JOURNAL OF VIROLOGY, Oct. 2007, p. 11426–11440 0022-538X/07/$08.00⫹0 doi:10.1128/JVI.00597-07 Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Vol. 81, No. 20

Commitment to Apoptosis in CD4⫹ T Lymphocytes Productively Infected with Human Immunodeficiency Virus Type 1 Is Initiated by Lysosomal Membrane Permeabilization, Itself Induced by the Isolated Expression of the Viral Protein Nef䌤 Mireille Laforge,1 Frederic Petit,2 Je´ro ˆme Estaquier,2,3† and Anna Senik1,3†* INSERM U542, Ho ˆpital Paul Brousse, Villejuif, France1; URA CNRS 1930, Institut Pasteur, Paris, France2; and Unite´ INSERM 841, Creteil, France3 Received 21 March 2007/Accepted 21 July 2007

Primary CD4ⴙ T lymphocytes, supporting in vitro human immunodeficiency virus type 1 (HIV-1) replication, are destined to die by apoptosis. We explored the initial molecular events that act upstream from mitochondrial dysfunction in CD4ⴙ T lymphocytes exposed to the HIV-1LAI strain. We tracked by immunofluorescence the cells expressing the p24 viral antigen and used Percoll density gradients to isolate a nonapoptotic CD4ⴙ T-cell subset with a high inner mitochondrial transmembrane potential (⌬⌿m) but no outer mitochondrial membrane (OMM) rupture. In most p24ⴙ (but not bystander p24ⴚ) cells of this subset, the lysosomes were undergoing limited membrane permeabilization, allowing the lysosomal efflux of cathepsins (Cat) to the cytosol. This was also induced by HIV-1 isolates from infected patients. Using pepstatin A to inhibit Cat-D enzymatic activity and Cat-D small interfering RNA to silence the Cat-D gene, we demonstrate that once released into the cytosol, Cat-D induces the conformational change of Bax and its insertion into the OMM. Inhibition of Cat-D activity/expression also conferred a transient survival advantage upon productively HIV1-infected cells, indicating that Cat-D is an early death factor. The transfection of activated CD4ⴙ T lymphocytes with a Nef expression vector rapidly induced the permeabilization of lysosomes and the release of Cat-D, with these two events preceding OMM rupture. These results reveal a previously undocumented mechanism in which Nef acts as an internal cytopathic factor and strongly suggest that this viral protein may behave similarly in the context of productive HIV-1 infection in CD4ⴙ T lymphocytes. The progressive and massive destruction of CD4⫹ T lymphocytes by programmed cell death is a characteristic of human immunodeficiency virus type 1 (HIV-1)-related disease (35). In HIV-1-infected individuals treated with antiretroviral drugs, there is a very high turnover of virus-producing CD4⫹ T cells (26, 59). In the rhesus macaque model of acute simian immunodeficiency virus infection, memory CD4⫹ T cells of the gut are readily eliminated as a consequence of direct viral-gene expression. (36, 40). Thus, direct cytopathic effects, linked to viral replication, coexist with indirect cytopathic effects (in particular those induced by “activation-induced cell death” in bystander cells) and make a major contribution to the loss of CD4⫹ T cells during HIV-1 pathogenesis (reviewed in references 2 and 19). In this study, we were interested in examining the very early death events triggered directly by productive HIV-1 infection in primary CD4⫹ T cells. Classical apoptosis, with its main hallmarks (cell shrinkage, strong chromatin condensation, and caspase activation), has long been viewed as the major cell death mechanism affecting the cells of HIV-1-infected CD4⫹ T-cell cultures (56) However, several studies with productively

* Corresponding author. Mailing address: INSERM 841, équipe 16 “De´veloppement Lymphoide Normal et dans l’Infection par le VIH,” Faculte´ de Me´decine Henri Mondor, 8 rue du Geńe´ral Sarrail, 94010 Cre´teil, France. Phone: (33) 1 49 81 36 72. Fax: (33) 1 49 81 37 09. E-mail: [email protected]. † J.E. and A.S. contributed equally to this work. 䌤 Published ahead of print on 1 August 2007.

HIV-1-infected primary CD4⫹ T-cell cultures described a death pathway insensitive to the peptide Z-VAD.fmk [benzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethylketone] (a broad-spectrum caspase inhibitor) and to reagents that specifically inhibit the death receptors of the tumor necrosis factor receptor superfamily (16, 47). This suggests that a caspaseindependent death pathway, also operating in the cells, could substitute for the caspase-dependent one. A critical issue is the interplay between these two pathways in HIV-1-infected cells. In a number of cell death models, lysosomal destabilization and the ensuing efflux of cathepsins play early and important roles in the destruction of the cells (reviewed in references 21 and 28). Whether a cathepsin or a caspase is the prime mover of apoptosis in any given model depends on the cell type and on the nature of the stimulus. For instance cathepsin B (CatB), which is an essential mediator of tumor necrosis factor alpha-induced cell death in murine embryonic fibroblasts, depends on caspase 9-induced lysosomal-membrane permeabilization (caspase 9 cleavage is itself dependent on caspase 8 activation) (23). On the other hand, Cat-D, which is primarily involved in oxidative stress and staurosporine-induced apoptosis in human fibroblasts, acts upstream from outer mitochondrial membrane (OMM) rupture and caspase activation (3, 29). We have characterized in particular an early caspaseindependent phase of commitment to apoptosis, triggered by low concentrations of staurosporine, during which limited lysosomal destabilization and partial leakage of cathepsins into the cytosol occur. Released Cat-D is important during this

11426

VOL. 81, 2007

LYSOSOMAL PERMEABILIZATION LINKED TO HIV-1 INFECTION

stage, promoting Bax activation and insertion into the OMM, with Bax in turn promoting mitochondrial apoptosis-inducing factor (AIF) release (3). Here, we therefore examined whether lysosomal permeabilization is involved in the death of productively HIV-1-infected primary CD4⫹ T cells. We also examined whether the HIV-1 Nef protein is cytopathic for these cells. Nef is a 27-kDa myristoylated phosphoprotein essential for AIDS pathogenesis, dramatically increasing viral loads and promoting high-level viral replication (30, 43). Nef has multiple effects, including the downregulation of cell surface proteins, such as CD4, major histocompatibility complex class I, CD28 (reviewed in reference 58), and chemokine receptors (42), and an increase in the responsiveness of T cells to activation stimuli (15). Studies with cell lines stably expressing nef have shown that Nef also enhances the apoptotic responses to a number of cell death agonists (48). It is produced in large amounts shortly after viral infection (50). We therefore hypothesized that it might induce early apoptotic events in HIV-1-infected cells. Using discontinuous-density Percoll gradients, we isolated from HIV-1-infected cell cultures a CD4⫹ T-cell subset that lacked apparent signs of mitochondrial dysfunction (inner mitochondrial transmembrane potential [⌬⌿m] loss and release of apoptogenic factors). We show that the lysosomes of infected p24⫹ (but not bystander p24⫺) cells of this subset undergo limited membrane permeabilization and that released Cat-D is primarily involved in the early phase of commitment to apoptosis in these cells, acting upstream from Bax translocation and insertion into the OMM. The transfection of activated CD4⫹ T lymphocytes with an expression vector encoding Nef rapidly resulted in the permeabilization of lysosomes and the release of cathepsins before OMM rupture. Our data strongly suggest that the Nef protein may induce lysosomalmembrane permeabilization in HIV-1-infected lymphocytes, thereby playing a direct role in their death by apoptosis. MATERIALS AND METHODS Lymphocytes and culture conditions. Peripheral blood leukocytes were isolated from healthy volunteers (Etablissement Français du Sang). CD4⫹ T cells were negatively selected using a CD4 T-cell isolation kit (Miltenyi Biotech). The purity of CD4⫹ T cells was ⱖ98%. Monocytes (5%) were added back to the purified cells to allow full T-cell activation. The cells were incubated with HIV1LAI for 12 h at a multiplicity of infection (MOI) of 0.01 and then activated with 1 ␮g/ml concanavalin A (ConA) (Sigma-Aldrich) and 100 units/ml recombinant human interleukin 2 (IL-2) (Roussel-Uclaf, France). After 5 days of activation, shrunken apoptotic cells were eliminated by running the cell populations through discontinuous-density Percoll gradients according to a methodology previously validated (11). Virus preparation. High-titer stocks of the laboratory strain HIV-1LAI (106 50% tissue culture infective doses/ml) were prepared as described previously (47). Clinical HIV-1 isolates interacting with CXCR4 (isolate I) or CCR5 (isolate II) came from patients who developed AIDS and were kindly given by F. Mammano (Inserm U552, Ho ˆpital Bichat-C. Bernard, Paris, France). Flow cytometric analysis of ⌬⌿m and plasma membrane permeability. To evaluate changes in the ⌬⌿m, cells were stained for 15 min at 37°C with 40 nM of the potential-sensitive fluorescent dye DiOC6 (3.3⬘-diethyloxacarbocyanine) from Molecular Probes. Dead cells were detected by using 100 nM of the TOTO dye (Molecular Probes). Apoptotic/dead cells were also enumerated under a light microscope on the basis of abnormal cell morphology and/or trypan blue uptake. Synthetic inhibitors, enzymatic substrates, and other chemicals. The Cat-B and -L inhibitor Z-FA.fmk (benzyloxycarbonyl-Phe-Ala-fluoromethylketone), the Cat-B substrate Z-Arg-Arg-␤NA, the Cat-D substrate Bz-Arg-Gly-Phe-PhePro-4MeO␤NA, and the pancaspase inhibitors Z-VAD.fmk and Boc-D.fmk

11427

[Boc-Asp(OMe)-fluoromethylketone] were from Calbiochem. The Cat-D inhibitor pepstatin A, the ␤-hexosaminidase substrate p-nitrophenyl N-acetyl-␤-Dglucosaminide, and dideoxyinosine (ddI) were from Sigma. Enzyme assays. Cat-B and ␤-hexosaminidase activities were assayed by hydrolysis of their respective substrates as described previously (3). Cat-D activity was determined as described previously (39), and the release of 2-naphtylamine was measured in a spectrofluorometer and expressed as relative fluorescence units/ second. Immunofluorescence studies. The reagents used for immunofluorescence studies were rabbit polyclonal antibodies recognizing the cleaved form of caspase 3 (Asp175; Cell Signaling Technology), Bax-NT raised against amino acids 1 to 21 of human Bax (46), Bak-NT (Upstate Biotechnology), and Cat-D (Zymed Laboratories), a rabbit anti-human Cat-L antiserum (Athens Research and Technology), a sheep anti-cytochrome c antiserum (Sigma), anti-Cat-B monoclonal antibody (MAb) (Ab-1; Oncogene Research Products), fluorescein isothiocyanate (FITC) or RD1-labeled MAb anti-p24 antigen (KC-57; Beckman Coulter), allophycocyanin-labeled anti-CD4 (L 200; Pharmingen), and anti-Nef MAb provided by O. Schwartz (Institut Pasteur, Paris, France). Intracellular p24 antigen was assessed by flow cytometry after fixation and permeabilization of the cells with Intraprep permeabilization reagent (Coulter). Otherwise, the cells were fixed with 3% paraformaldehyde, spun on glass slides, washed with PBS, and permeabilized with 0.05% Triton X-100. After the washings, the cells were incubated with the indicated antibodies in phosphate-buffered saline supplemented with 0.5% bovine serum albumin and 2% fetal calf serum. They were then stained with FITC- and RD1-conjugated secondary antibodies (Immunotech) or with Alexa Fluor 350- and 546-conjugated secondary antibody (Molecular Probes). Nuclei were counterstained for 5 min with 5 ␮M DAPI (4⬘,6⬘diamidino-2-phenylindole) (Molecular Probes). The cells were examined by conventional or confocal fluorescence microscopy (Leica Microsyste`mes). For the detection of the active forms of Bax and Bak by conformation-specific antibodies, the cells were permeabilized with 0.0125% CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate} in PBS (27). To monitor lysosomal destabilization, the cells were incubated for 2 h with 5 mg/ml FITCdextrans of various molecular weights (Sigma) as previously described (3). After a 2-h chase period, the cells were stained with anti-p24 MAb and examined by laser scanning confocal microscopy. Subcellular fractionation. After the elimination of dead and apoptotic cells by centrifugation on Percoll gradients, cytosolic fractions and heavy membrane fractions (mitochondrial pellets) were generated from 107 cells, using a selective digitonin-based permeabilization and subcellular-fractionation technique (13). The digitonin concentration was 35 ␮g/ml. For alkali extraction, the mitochondrial pellets were treated as described previously (3). Briefly, the heavy membrane fractions were incubated on ice for 30 min in 25 ␮l of 0.1 M Na2CO3, pH 11.5, and subjected to ultracentrifugation at 100,000 ⫻ g for 10 min, yielding pellets and S100 supernatants. Twenty micrograms of the solubilized pellets and all of the S100 supernatants were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Consistent loading of the mitochondrial pellets was verified by immunoblotting, using antibodies against Bcl-2 and VDAC. Immunoblotting. The primary antibodies used for Western blotting were rabbit antisera against Bax (N-20; Santa Cruz) and actin (Sigma); goat polyclonal antibodies against Cat-B (Santa Cruz); and mouse MAbs against Cat-D (MAb 49), Lamp-1 (MAb 25 from Transduction Laboratory), VDAC (clone 31HL; Calbiochem), and Bcl-2 (Bcl-2-100; Santa Cruz). Twenty micrograms of cytosolic and mitochondrial extracts were boiled for 5 min in Laemmli buffer containing 2% SDS and 10% 2-␤-mercaptoethanol and then run on 4 to 20% polyacrylamide gels (Invitrogen). After transfer of the proteins onto polyvinylidene difluoride membranes (Bio-Rad), the immunoblots were sequentially incubated with primary antibodies and horseradish peroxidase-coupled secondary reagents (Amersham Biosciences) and developed by enhanced chemiluminescence (ECL from Amersham or West Femto from Pierce) using a charge-coupled device camera (Fuji LAS-1000plus). siRNAs and CD4ⴙ T-lymphocyte transfection. Cat-D gene expression was silenced by the small interfering RNA (siRNA) technique (41) using duplexes of 21-nucleotide siRNA with two 3⬘ overhanging TT residues (Proligo). The sense strand of the siRNA used to silence the Cat-D gene (Cat-D siRNA) was GCU GGUGGACCAGAACAUC. An inefficient CD9 oligonucleotide served as a negative control siRNA (GCCAUCCACUAUGCGUUGAAC) and was provided by C. Boucheix (Inserm U268, Villejuif, France). Purified resting CD4⫹ T cells were transfected by electroporation of siRNAs (0.75 ␮M/4 ⫻ 106 cells) using the Nucleofection system (Amaxa). Following a 16-h rest, the cells were exposed to HIV-1, and after an additional 12-h culture period, stimulated for 5 days with ConA and IL-2.

11428

LAFORGE ET AL.

J. VIROL.

VOL. 81, 2007


Transfection of CD4ⴙ T lymphocytes with Nef expression vectors. Nef and Nef-mock plasmids carrying the Nef gene (from the LAI isolate) in a sense and antisense orientation, respectively, and the plasmid carrying the Nef-G2A mutant (all under the control of the cytomegalovirus promoter) were kind gifts from O. Shwartz (55). The green fluorescent protein (GFP) vector pmaxGFP was from Amaxa. CD4⫹ T cells (activated with ConA and IL-2 for 3 days) were transfected with the Nef-WT, Nef-G2A, or mock plasmid, together with the GFP vector (0.5 ␮g of each vector/1 ⫻ 106 cells), using the Nucleofection system (Amaxa). After 3 to 24 h, the cells were washed and fixed for flow cytometry and confocalmicroscopy analyses.

RESULTS Isolation from HIV-1-infected T-lymphocyte cultures of a CD4ⴙ T-cell subset enriched for ⌬⌿mhigh p24ⴙ cells with no sign of OMM rupture. The numerous shrunken cells seen by flow cytometry in HIV-infected CD4⫹ T-cell cultures was evidence of a high cell death rate (not shown). After labeling them with the potential-sensitive dye DIOC6, a loss of ⌬⌿m was detected in 42% ⫾ 15% (n ⫽ 20) of the lymphocytes (versus 14% ⫾ 4% in control uninfected cultures), indicating that such cells had taken an irreversible step in the process of cell death. As HIV-1-infected CD4⫹ T cells are destined to die by apoptosis, we sought to separate large p24⫹ cells, presumably in the commitment phase to apoptosis, from shrunken apoptotic p24⫹ cells that had entered the execution phase of apoptosis. The cells from HIV-1-infected cultures were therefore run through discontinuous-density Percoll gradients (Fig. 1): the cell preparations recovered in the low-buoyant-density fraction of these gradients then contained ⱖ90% ⌬⌿mhigh cells (hereafter called ⌬⌿mhigh CD4⫹ T cells). Thirty percent ⫾ 5% (n ⫽ 10) of them were positive for the p24 viral antigen. Using an antibody specific for the processed form of caspase 3, we found that 40% of the whole cell populations exposed to HIV-1 displayed the active form of caspase 3 but that only 4% of isolated ⌬⌿mhigh CD4⫹ T cells presented this characteristic (not shown). Accordingly, 37% of p24⫹ cells in the whole HIV-1-infected cell populations, but only ⬃10% of p24⫹ cells in the purified ⌬⌿mhigh CD4⫹ T-cell subset, exhibited condensed nuclei and/or apoptotic bodies (Fig. 1A, b). We additionally performed double staining of the cells with anti-p24 and anti-cytochrome c antibodies to monitor by fluorescence microscopy the subcellular localization of cytochrome c, a proapoptotic factor normally residing in the intermembrane space of mitochondria that is released into the cytosol during the execution phase of apoptosis (Fig. 1B). In the whole

11429

HIV-1-infected CD4⫹ T-cell population, ⬃55% of p24⫹ cells exhibited a diffuse staining pattern for cytochrome c. In those cells, yellow staining in the overlay indicated that cytochrome c was now colocalized in the cytosol with the viral p24 antigen (versus ⬃14% of bystander p24⫺ CD4⫹ T cells and ⬃11% of control uninfected CD4⫹ T cells) (Fig. 1B, a). In contrast, in ⬃90 of ⌬⌿mhigh p24⫹ cells, punctate staining of cytochrome c was detected, consistent with exclusive mitochondrial localization of this protein and precluding OMM rupture. Similar scores were obtained when the subcellular distribution of the other mitochondrial proapoptotic factors, namely, Smac/ DIABLO, Omi/Htra2, and AIF, was tested by immunofluorescence (data not shown). Using cytosolic extracts prepared from ⌬⌿mhigh CD4⫹ T-cell populations, we also failed to detect by immunoprecipitation the presence of cytochrome c and AIF in these preparations (data not shown). Thus, p24⫹ CD4⫹ T cells in which the level of ⌬⌿m was high, the OMM intact, and caspase 3, the major executioner of apoptosis, not activated were most likely in the early phase of commitment to apoptosis. Early and limited lysosomal-membrane permeabilization in ⌬⌿mhigh CD4ⴙ T cells supporting HIV-1 replication. Because lysosomal-membrane permeabilization is an early event in a number of cell death processes, we tested whether it also affected the lysosomes of ⌬⌿mhigh productively HIV-1-infected cells and, if so, to what extent. ⌬⌿mhigh CD4⫹ T cells isolated from HIV-1-infected cultures were loaded with FITC-conjugated dextrans of various molecular masses, and after a 2-h chase allowing complete colocalization with Cat-D, the samples were stained for the p24 antigen. Laser scanning confocal microscopy (Fig. 2A) revealed a diffuse staining pattern for the 10-kDa FITC-dextran molecules in 80% of the p24⫹ cells (ⱖ200 cells were examined), which is indicative of lysosomal efflux. In bystander p24⫺ cells and in uninfected CD4⫹ cells, the 10-kDa molecules were almost totally confined to punctate structures, consistent with exclusive lysosomal localization. Redistribution of the 40-kDa FITC-dextran molecules was observed in 50% of p24⫹ cells, but not in p24⫺ cells. The 70-kDa and 250-kDa dextrans were not released at all from lysosomes. These data suggest that limited, but not massive, lysosomal destabilization occurred in p24⫹ T cells, allowing the nonspecific release of lysosomal constituents with molecular masses below 70 kDa. Lysosomal cathepsins are ⬃30- to 40-kDa monomeric pro-

FIG. 1. Productively HIV-1-infected (p24⫹) cells, contained in the ⌬⌿mhigh CD4⫹ subset, do not display OMM permeabilization. CD4⫹ T cells, highly purified by negative selection over magnetic columns, were stimulated for 5 days with 1 ␮g/ml ConA and 100 U/ml IL-2 in the presence or absence of HIV-1LAI at a MOI of 0.01. They were then run on discontinuous-density Percoll gradients to eliminate apoptotic cells that fell at the bottom of the gradients and to recover the large cells that remained in the low-buoyant-density fraction. (A) (a) ⌬⌿m loss determined by flow cytometry in the lymphocyte gate, using the green fluorescent dye DiOC6 (representative of 18 experiments). (b) cells stained with a MAb directed at the HIV-1 p24 antigen. Proportions of condensed nuclei and/or apoptotic bodies (arrowheads) are indicated. In the course of 10 experiments, ⱖ2,000 cells were examined for each double p24/DAPI staining. (B) (a) Cells were doubly stained with a MAb recognizing cytochrome c (Cyt.c) (red) and another specific for HIV-1 p24 antigen (green) and examined by fluorescence microscopy. Nuclei were counterstained with DAPI. Negative controls consisted of uninfected CD4⫹ T cells exposed to isotypic control antibodies and secondary FITC- or tetramethyl rhodamine isothiocyanate-labeled antibodies (400 cells were examined for each double staining). The yellow overlays indicate that mitochondrial cytochrome c was diffusing into the cytosol of 55% ⫾ 3% of p24⫹ cells. In ⱖ86% of bystander p24⫺ and uninfected cells, the pattern of cytochrome c distribution was punctate. (b) Double cytochrome c/p24 staining of ⌬⌿mhigh CD4⫹ cells isolated by passage on discontinuous-density Percoll gradients (ⱖ600 cells were examined for each double staining). Ninety percent ⫾ 3% of the cells displayed punctate immunofluorescence staining for cytochrome c. Negative controls were as in a. The histograms are means ⫾ standard deviations of the cell counts, made in parallel, of whole cell populations and isolated ⌬⌿mhigh cells.

11430

LAFORGE ET AL.

J. VIROL.

VOL. 81, 2007


teases, so it is likely that they were released into the cytosol. The subcellular distribution of Cat-B, -D, and -L, the most abundant proteases in lysosomes, was therefore examined by immunofluorescence in p24⫹ and p24⫺ cells after they were double labeled with an anti-p24 MAb and with antibodies directed against each cathepsin (Fig. 2B). In control uninfected cells, immunostaining of Cat-B, -D, and -L displayed a punctate distribution. In a substantial proportion of p24⫹ (but not p24⫺ bystander) cells, in contrast, all three cathepsins exhibited a diffuse distribution pattern, indicative of extralysosomal location. Incubation of the cells with 5 ␮M ddI, an inhibitor of the HIV-1 reverse transcriptase (10), prevented both the appearance of p24⫹ cells and the lysosomal efflux of cathepsins (only the inhibition of Cat-D release is shown). Lysosomalmembrane permeabilization was therefore directly linked to HIV-1 replication. Lysosomal efflux of Cat-B, -D, and -L occurred in ⬃50 to 60% of p24⫹ cells versus ⬃10% of p24⫺ bystander cells and ⬃12% of uninfected control cells (Fig. 2C). Cells infected with the laboratory-adapted strain HIV-LAI or with either of two natural HIV-1 isolates from patients who developed AIDS gave similar scores. Released Cat-D, but not Cat-B, is catalytically active in the cytosol of HIV-1-infected ⌬⌿mhigh p24ⴙ cells. Digitonin-based subcellular fractionation and immunoblot analysis revealed that small amounts of both Cat-B (representative of cysteine cathepsins) and Cat-D were present in the cytolytic extracts prepared from infected ⌬⌿mhigh CD4⫹ T-cell populations, but not in those from control uninfected cells (Fig. 3A, a and b). No ␤-hexosaminidase activity was detected in these cytosolic preparations, indicating that this 250-kDa enzyme was almost totally retained within the lysosomes (Fig. 3A, c). Lamp-1, a molecule anchored in the lysosomal membrane, was not detected either. These data therefore indicated that digitonin treatment selectively permeabilized the plasma membranes of the cells, leaving lysosomes intact. As Cat-B and -D showed diffuse staining in p24⫹ (but not p24⫺) cells under the fluorescence microscope, it was likely that the faint cathepsin bands detected by immunoblotting corresponded to the lysosomal efflux of these cathepsins in ⌬⌿mhigh p24⫹ cells. We prepared cytosolic extracts from cells that had been exposed throughout the 5-day culture period either to Z-FA.fmk, a pharmacological inhibitor of Cat-B and Cat-L, or to pepstatin A, a pharmacological inhibitor of Cat-D (at a cumulative dose of 40 ␮M). The released Cat-B and Cat-D were then assayed for hydrolysis of their respective substrates (Fig. 3B). Substantial Cat-D enzymatic activity was detected in the cytosol of infected ⌬⌿mhigh CD4⫹ T-cell pop-

11431

ulations unless they had been cultured in the presence of pepstatin A. Cat-B activity, in contrast, was hardly detected in the cytosol of these cells, despite some Cat-B protein being present. This was probably due to the inhibitory effect of cytosolic cystatins that act as “emergency inhibitors,” ready to inhibit the small amounts of cystein cathepsins (including Cat-B and Cat-L) possibly leaking from lysosomes (57). No endogenous inhibitors of Cat-D are known, which explains why Cat-D activity was readily detected. In no instance did the inhibition of Cat-B and Cat-D activities by Z-FA.fmk and pepstatin A, respectively, prevent the lysosomal release of cathepsins (Fig. 3C). The chronic presence of Boc-D.fmk, a broad-spectrum caspase inhibitor, did not prevent the permeabilization of lysosomes. This suggests that the activity of initiator caspases was not involved in this event. Bax (but not Bak) acquires a membrane-inserting conformation in ⌬⌿mhigh p24ⴙ cells, a Cat-D-dependent event. Bax, a proapoptotic Bcl-2 family protein, is found as a soluble monomeric protein in the cytosol of healthy cells but rapidly moves to the mitochondria and inserts into the OMM at the initiation of apoptosis (60). The conformation of Bax changes concurrently, so that the N terminus of the molecule, normally cryptic in nonapoptotic cells, becomes exposed (44). Because Cat-D, once translocated to the cytosol, becomes a death factor acting upstream from Bax activation and OMM permeabilization (3, 14), we assumed that Bax underwent a Cat-Dinduced conformational change in ⌬⌿mhigh p24⫹ CD4⫹ T cells. Resting CD4⫹ T cells were therefore transfected with siRNA targeting the human Cat-D mRNA. The cells were then subjected to HIV-1 infection and mitogenic stimulation. siRNA-mediated silencing of the Cat-D gene became fully operative 5 days posttransfection. At that time, Cat-D protein production was strongly decreased, whereas in mock-transfected cells and in cells transfected with control siRNA, Cat-D production was unaffected (Fig. 4A, a). In contrast, Cat-D siRNA transfection did not influence Cat-B protein production. It also did not change the proportions of p24⫹ cells that were generated in CD4⫹ T-cell populations exposed to HIV-1 (Fig. 5). Under these conditions, conventional immunofluorescence microscopy indicated that ⬃60% of mock- and control siRNA-transfected ⌬⌿mhigh p24⫹ cells (but not p24⫺ bystanders) were stained by Bax-NT, an antibody that specifically recognizes the “activated” configuration of Bax (Fig. 4A, b and c). Bax-NT immunoreactivity was, in contrast, no longer detected in ⌬⌿mhigh p24⫹ cells transfected with Cat-D siRNA. Bax-NT immunoreactivity was exclusively found among the

FIG. 2. Limited lysosomal permeabilization in ⌬⌿mhigh p24⫹ (but not p24⫺) CD4⫹ T cells. (A) After ingestion of FITC-dextran molecules of various molecular masses, ⌬⌿mhigh CD4⫹ T cells were stained with anti-p24 MAb and examined by laser scanning confocal microscopy. The insets correspond to cells doubly stained with anti-Cat-D (red) and FITC-dextrans and show the complete colocalization of these molecules in the lysosomes (yellow overlay). The white numbers within the pictures of HIV-1-infected cultures represent the percentages of p24⫹ cells exhibiting diffuse staining for FITC-dextran molecules. In uninfected cells, the background percentage of cells with diffuse staining for the 10-kDa FITC-dextran molecules was 10%. (B) Lysosomal efflux of Cat-B, Cat-D, and Cat-L preferentially occurs in productively infected p24⫹ cells. The images, obtained by fluorescence microscopy, show punctate (lysosomal) staining for all cathepsins in uninfected cells and in bystander p24⫺ cells and diffuse (cytosolic) staining in some p24⫹ cells. Incubation of the cells with 5 ␮M ddI, an inhibitor of the HIV-1 reverse transcriptase, prevented both the appearance of p24⫹ cells and the lysosomal efflux of cathepsins. (C) Similar numbers of p24⫹ cells exhibiting diffuse staining for Cat-B, -D, and -L were scored, whether the cells had been infected with the laboratory-adapted strain HIV-LAI or with either of two natural HIV-1 isolates from infected patients. The error bars indicate standard deviations.

11432

LAFORGE ET AL.

J. VIROL.

FIG. 3. Cat-D enzymatic activity is found in the cytosol of ⌬⌿mhigh CD4⫹ T-cell populations isolated from HIV-1-infected cultures. (A) Cytosolic extracts (50 ␮g of proteins) derived from ⌬⌿mhigh CD4⫹ T cells at day 5 of the cultures were analyzed by Western blotting for the presence of Cat-D (a) and Cat-B (b). Anti-Lamp-1 was used to verify the absence of lysosomal contamination and anti-actin to confirm loading of equal amounts of protein. Controls were cell lysates (20 ␮g) obtained with 1% Triton X-100. (c) Lack of ␤-hexosaminidase activity in the cytosolic extracts. OD, optical density. The results in a are representative of three separate experiments and those in b of two separate experiments. (B) Tests of Cat-B and Cat-D enzymatic activities in the cytolytic extracts and lysates (100 ␮g) of cells incubated for 5 days in the presence of Z-FA.fmk (an inhibitor of Cat-B and Cat-L) or pepstatin A (PA) (an inhibitor of Cat-D) at cumulative doses of 40 ␮M. The means plus standard deviations correspond to triplicate samples (one of two independent experiments is presented). (C) The chronic presence of Z-FA.fmk, pepstatin A, and Boc-D.fmk does not prevent the lysosomal efflux of cathepsins in p24⫹ cells (600 cells were counted for each double staining in the course of three independent experiments). The inset shows that Boc-D.fmk almost completely inhibits the processing of caspase 3 in whole HIV-1-infected CD4⫹ T-cell populations.

VOL. 81, 2007

m

a

Cat D Cat B Actin

Bax-NT

b

p24

11433

B 80

% Bax-NT positive cells

oc k

A

co nt ro l-s iR Ca NA tD - si RN A


+Z-FA.fmk + Boc-D.fmk

60 40 +PA

20 0

Dapi i

un

+ - + - + - + ted p24 p24 p24 p24 p24 p24 p24 p24

ec nf

HIV-1-infected cultures

mock controlsiRNA

mitochondrial pellet

supernatant

HI

HI V+ PA uni nfe cte d HI V HI V+ PA uni nfe cte d HI V HI V+ PA

(S100)

uni

cte d

C

-

-

-

V

nfe

Cat D-

60

Bcl-2

40

VDAC 20

Na2CO3: p

24

p2

4+

p

24

p2

4+

p

+

+

+

+

+

+

24

C siR atD NA -

p2

4+

co siR ntr NA ol-

0

m oc k

c

% Bax-NT positive cells

Bax

D

Bak-NT

p24

Dapi

7% HIV-1-infected cultures 80 % positive control (STS)

FIG. 4. siRNA-mediated Cat-D gene silencing and inhibition of Cat-D activity by pepstatin A prevent the occurrence of the membrane-inserting conformation of Bax in ⌬⌿mhigh p24⫹ cells. (A) (a) Resting CD4 T lymphocytes were either mock transfected or transfected with 0.75 ␮M control siRNA and Cat-D siRNA. After infection and activation, ⌬⌿mhigh cells were isolated on discontinuous-density Percoll gradients and subjected to lysis, SDS-PAGE, and immunoblotting. In cells transfected with Cat-D siRNA, Cat-D protein production was inhibited, whereas for the same protein loading (Actin), the abundance of Cat-B protein was unaffected. (b) The activation status of Bax was examined by fluorescence microscopy in p24⫹ and p24⫺ cells of ⌬⌿mhigh CD4⫹ T-cell populations, using an antibody raised against the amino acid 1 to 21 region of Bax (Bax-NT), together with an anti-p24 MAb. Cat-D siRNA transfection results in almost total extinction of the Bax-NT-associated immunofluorescence detected in ⌬⌿mhigh p24⫹ cells. (c) The values shown are means plus standard deviations (SD) of cell counts performed in three independent experiments (ⱖ600 cells were examined for each double staining). (B) The active conformation adopted by Bax in ⌬⌿mhigh p24⫹ cells is selectively prevented by pepstatin A. HIV-1-infected cultures were incubated with pepstatin A (PA), Z-FA.fmk (cumulative doses, 40 ␮M), or Boc-D.fmk (75 ␮M). The percentages of cells exhibiting the Bax NH2 terminus, among p24⫹ and p24⫺ cells, were determined by fluorescence microscopy. The values are means plus SD of three independent experiments (ⱖ500 cells were examined for each double staining). (C) Bax translocated to mitochondria becomes alkali resistant. The heavy membrane fractions from uninfected and HIV-1-infected ⌬⌿mhigh CD4⫹ T-cell populations were incubated in 0.1 M Na2CO3, pH 11.5, and centrifuged at 100,000 ⫻ g to yield the mitochondrial pellet and the S100 supernatant. The fractions were analyzed by Western blotting for Bax, Bcl-2, and VDAC proteins. For immunoblot analysis of VDAC, duplicate samples were run in parallel on SDS-PAGE, and the proteins were transferred onto a polyvinylidene membrane. The membrane was then probed with an anti-VDAC antibody. The data show equal loading of Bcl-2 and VDAC proteins in the pellets, indicating that the supernatants were derived from the same quantities of mitochondrial material. (D) Bak does not undergo a conformational change in ⌬⌿mhigh p24⫹ cells, as determined by using an antibody recognizing the N terminus of Bak (Bak-NT). Positive controls were activated CD4⫹ T cells exposed for 2 h to 500 nM staurosporine. The numbers in the two left panels are the percentages of Bak-NT-positive cells.

11434

LAFORGE ET AL.

J. VIROL.

VOL. 81, 2007


cells displaying diffuse Cat-D staining (data not shown). Immunofluorescence studies (Fig. 4B) confirmed that the early generation of Bax-NT-positive cells in ⌬⌿mhigh p24⫹ cells was inhibited by pepstatin A, but not by Z-FA.fmk. Neither was it inhibited by Boc-D.fmk. Thus, Bax activation was induced by Cat-D in a caspase-independent manner. As Bax can be loosely attached to the mitochondria of nonapoptotic cells but becomes fully inserted in the OMM of apoptotic cells (44), we tested which of these two situations applied to ⌬⌿mhigh p24⫹ cells. We prepared the heavy membrane (mitochondrial) fraction from uninfected and HIV-1infected ⌬⌿mhigh CD4⫹ T cells by subcellular fractionation and subjected this fraction to alkaline extraction of proteins. Bax was present in the mitochondrial preparation from uninfected cells but proved to be sensitive to alkaline extraction, indicating that it was not integrated into the OMM (Fig. 4C). Bax from HIV-1-infected ⌬⌿mhigh CD4⫹ T cells was partially resistant to extraction at alkaline pH, implying that it was partly integrated. Control Bcl-2 and VDAC, two mitochondrial transmembrane proteins that are constitutively anchored in the OMM, were completely resistant to alkaline extraction. The chronic presence of pepstatin A during the culture period inhibited the insertion of Bax into the OMM. These results demonstrate that the conformational change of Bax in ⌬⌿mhigh p24⫹ cells and its insertion into the OMM were Cat-D induced. Interestingly, Bax insertion itself was insufficient to permeabilize the mitochondria of ⌬⌿mhigh p24⫹ cells, as cytochrome c was still confined to these organelles. Bak is another proapoptotic member of the Bcl-2 family that is associated with mitochondria and whose conformational change in its N terminus is also a prelude to apoptosis (20). Using an epitope-specific antibody recognizing the N terminus of Bak, normally cryptic in healthy cells, we were unable to detect any N-terminus-associated immunofluorescence (Fig. 4D). Thus, the lysosomal perturbation in p24⫹ cells had no impact on Bak. Cat-D siRNA and pepstatin A confer a transient survival advantage on p24ⴙ CD4ⴙ T cells. Five days after transfection with Cat-D siRNA, viable cells remaining in HIV-1-infected cultures were enumerated on the basis of trypan blue exclusion and normal cell morphology. Absolute numbers of viable p24⫹ cells were deduced from the proportions of p24⫹ cells in whole transfected CD4⫹ T-cell populations as determined by flow cytometry (Fig. 5A). It appears that transfection of CD4⫹ T cells with Cat-D siRNA promoted p24⫹ survival: the number of viable cells was 131% higher following transfection with Cat-D siRNA than with control siRNA (Fig. 5B). No significant survival advantage was conferred on p24⫺ bystander cells by Cat-D siRNA transfection.

11435

After 5 days posttransfection, Cat-D siRNA-mediated gene silencing was much less efficient (not shown). We therefore performed kinetic studies, using pepstatin A, to estimate the duration of the survival advantage conferred upon p24⫹ cells by the inhibition of Cat-D activity. ⌬⌿mhigh CD4⫹ T cells were isolated from HIV-infected cultures on day 5, thoroughly washed, and replaced in IL-2-containing medium for 48 h in the presence or absence of pepstatin A (20 ␮M was added each day). Twenty-one percent of the cells harvested on day 5 were p24 positive as determined by flow cytometry (Fig. 5C). HIV-1 infection rapidly spread, resulting in ⬃45% p24⫹ cells after 24 h and ⬃78% after 48 h, whether or not pepstatin A was present, indicating that the drug did not halt virus propagation in the CD4⫹ T-cell population. We noticed that the percentage of viable cells was significantly increased in whole HIV-1-infected CD4⫹ T-cell cultures when pepstatin A was present, as estimated by trypan blue exclusion. For instance, at the 24-h time point, the percentage of viable cells was ⬃74% with pepstatin A versus ⬃53% without pepstatin A. In contrast, in uninfected cell cultures, the percentage of viable cells was not influenced by pepstatin A (⬃88% versus ⬃85%). As only p24⫹ cells presented a continuous apoptotic cell death rate, the survival advantage afforded by pepstatin A most probably concerned p24⫹ and not p24⫺ cells. The absolute numbers of viable p24⫹ cells were therefore inferred as described above from the proportions of p24⫹ cells in whole CD4⫹ T-cell populations (Fig. 5D). After 24 h of continuous presence of pepstatin A, the number of viable p24⫹ cells in HIV-1-infected cultures was 90% higher than that in controls. In contrast, pepstatin A had no effect on the number of p24⫺ bystander cells. By 48 h, most of the cells had become p24 positive, and pepstatin A no longer protected them from destruction. Presumably, death triggers other than Cat-D came into play. Transfection of activated CD4ⴙ T lymphocytes with an expression plasmid encoding Nef results in early lysosomalmembrane permeabilization. We investigated whether Nef expression was sufficient to induce the permeabilization of lysosomes by transfecting CD4⫹ T cells (activated for 3 days) with expression vectors encoding Nef-WT, the Nef-G2A mutant that has a nonmyristoylated N-terminal domain (unable to anchor to the plasma membrane), or the antisense Nef (Nefmock). Cells were cotransfected with one of these plasmids plus the pmaxGFP construct encoding GFP. We verified by fluorescence microscopy that almost all green GFP⫹ cells transfected with the Nef-WT or the Nef-G2A plasmid were stained with a Nef-specific antibody (ⱖ100 GFP⫹ cells were examined in each case), whereas the cells negative for GFP expression were not stained, making it possible to identify transfected cells. Only in cells expressing the Nef-WT con-

FIG. 5. Effects of transfection with Cat-D siRNA and exposure to pepstatin A on the survival of p24⫹ cells. (A) Flow cytometric analysis of p24 antigen expression in whole populations of siRNA-transfected CD4⫹ cells after a 5-day exposure to HIV-1. Ten thousand events were acquired in the lymphocyte gate. Uninfected cells were used to determine the position of p24-negative cells along the FL2 axis. (B) Absolute numbers of p24⫹ and p24⫺ cells were inferred from the proportions of corresponding cells as determined by flow cytometry. The bars are means plus standard deviations (SD) of triplicate samples. The experiment was repeated three times with similar results. (C and D) ⌬⌿mhigh CD4⫹ T cells were obtained by centrifugation over discontinuous-density Percoll gradients 5 days post-HIV-1 infection. The cells were washed and resuspended in IL-2-containing medium for 24 or 48 h. Pepstatin A (20 ␮M) was added each day to the cultures. (C) The proportions of p24⫹ cells were determined by flow cytometry. (D) The absolute numbers of viable p24⫹ and p24⫺ cells were calculated as before, referring to the flow cytometry analysis in panel C. The bars are means plus SD of triplicate values of an experiment repeated twice with similar results.

11436

LAFORGE ET AL.

J. VIROL.

FIG. 6. Transfection of activated CD4⫹ T lymphocytes with an expression plasmid carrying Nef results in early lysosomal membrane permeabilization. (A) Flow cytometric analysis of 10,000 live transfected cells (selected on the basis of forward scatter analysis) showing the downmodulation of cell surface CD4 antigen (FL4) 24 h after cotransfection of CD4⫹ T lymphocytes with the pmaxGFP construct and with plasmids carrying the antisense Nef (Nef-mock), Nef-WT, or the Nef-G2A mutant. Most (⬃80%) GFP⫹ cells were ⌬⌿mhigh at this time. (B) (a) Transfected CD4⫹ T lymphocytes were harvested 3 h posttransfection and analyzed by fluorescence microscopy for the subcellular distribution, in GFPexpressing cells, of cytochrome c and Cat-D, according to the images presented in b; the violet overlay in b shows the colocalization in the cytosol of cytochrome c (red) and Cat-D (blue). At least 100 GFP⫹ cells from triplicate cultures were counted for each triple staining, and the values are means plus standard deviations (SD); another experiment was performed in the presence of 10 ␮M Z-VAD.fmk and gave similar results. (C) (a) Flow cytometry analysis of the percentages of Nef-mock-, Nef-WT-, and Nef-G2A-transfected cells (detected as GFP⫹ green cells) remaining in the cultures at the 3- and 6-h posttransfection time points. The bars are means plus SD of the values obtained in the course of three separate experiments performed in the absence of Z-VAD.fmk. (b) The bars are means plus SD of quadruplicate values obtained in the presence or absence of 10 ␮M Z-VAD.fmk. The 3-h posttransfection time is presented. *, P ⬍ 0.001; **, P ⬍ 0.003.

VOL. 81, 2007


struct did we observe the expected downregulation of CD4 expression on the cell surface (Fig. 6A) (the results for the 24 h posttransfection time are presented). Nef has been reported to promote Fas ligand transcription in T cells (61). We therefore carried out some transfection experiments in the presence of the pancaspase inhibitor Z-VAD.fmk to avoid possible interference of the Fas-L/Fas death pathway. We routinely harvested the cells 3 h after transfection to monitor, by fluorescence microscopy, the integrity of lysosomal and mitochondrial external membranes shortly after Nef protein production. Preliminary experiments indicated that transfection efficiency was substantial at this time point (26% ⫾ 5%; n ⫽ 3) as determined from GFP expression in mock-cotransfected cells. The transfection efficiency was much lower at earlier time points (not shown). Immunofluorescence studies showed that most (⬃80%) of the cells transfected with the Nef-mock vector displayed punctate immunostaining for cytochrome c (red) and Cat-D (blue), consistent with these molecules being sequestered in their respective organelles (Fig. 6B). In contrast, about 80% of Nef-WT-expressing cells displayed diffuse staining for Cat-D, reflecting lysosomal-membrane permeabilization. In half of these cells, cytochrome c diffusion into the cytosol, resulting in colocalization with Cat-D (see the violet overlay), was also observed. Cytochrome c release is a characteristic of cells that have irreversibly entered the apoptotic death process. In the remaining Nef-WT-transfected cells, cytochrome c was retained in the mitochondria. This situation (punctate cytochrome c/diffuse Cat-D) is very similar to that of HIV-1LAIinfected CD4⫹ T cells in the early phase of commitment to apoptosis, in which lysosomes are permeabilized before mitochondria. Interestingly, the loss of the myristoylation signal in the G2A mutant did not prevent a small proportion of NefG2A line-expressing cells from entering the early phase of commitment to apoptosis (⬃15% above background). GFP⫹ cells progressively disappeared from Nef-WT-transfected cell populations (Fig. 6C) and to a lesser degree from Nef-G2Atransfected populations. In contrast, the percentages of GFP⫹ cells in mock-transfected cells remained stable, at least during the time window for our observations (from 3 to 6 h posttransfection). It has been shown that there is a general inhibition of protein synthesis during late apoptosis, but not during necrosis (52). Thus, the loss of GFP expression may be used as a surrogate marker of apoptotic cells. This loss may also be used as a marker of cells undergoing caspase-independent death, since it occurred to the same extent whether or nor Z-VAD.fmk was present. (Fig. 6C, b). These results demonstrate the existence of a causal relationship between isolated Nef production and early lysosomal-membrane permeabilization. DISCUSSION We report that lysosomes are rapidly permeabilized in CD4⫹ T lymphocytes productively infected with HIV-1, resulting in the release of cathepsins into the cytosol, and that the permeabilization of lysosomes precedes that of mitochondria. During the early phase of commitment to apoptosis, represented by ⌬⌿mhigh p24⫹ cells, released Cat-D acts upstream from Bax conformational change and subsequent Bax insertion into the OMM. Inhibition of Cat-D activity confers a transient survival advantage upon infected cells, indicating that Cat-D

11437

behaves as an early death factor. Isolated production of the HIV-1 Nef protein in activated CD4⫹ T lymphocytes is sufficient to trigger lysosomal-membrane permeabilization and its effects. Cell death may proceed in HIV-1-infected cultures in which caspases are inhibited by Z-VAD.fmk (47). There is therefore a caspase-independent death pathway operating in these cells, in concert with the caspase-dependent pathway. The cells subjected to caspase-independent death continued to shrink and suffered from mitochondrial deregulation (as evidenced by ⌬⌿m loss) but did not display the strong chromatin condensation typical of classical apoptosis. In view of our findings, this pathway appears to be largely lysosomal. It has been proposed that extensive lysosomal rupture with a lysosomotropic detergent triggers a necrotic form of cell death, while limited permeabilization allows the apoptotic phenotype to predominate (37). The death phenotype most probably results from the balance between cathepsins and caspases. The magnitude of the death insult may also be relevant. Thus, using CD4⫹ T cells infected with different strains of HIV-1, Lenardo et al. found that the virus-induced cytopathic effect was necrotic rather than apoptotic (34). In that study, the MOI was high, whereas we used a low MOI. Under our low-MOI conditions, limited lysosomal-membrane permeabilization occurred, in a size-selective manner, allowing the apoptotic form of cell death to predominate. As stated above, endogenous cytosolic inhibitors, called cystatins, can restrain the cysteine cathepsins, including Cat-B, that leak from lysosomes (57). Probably, the destructive potential of these cathepsins in our system was de facto minimized and the necrotic phenotype was avoided. The cytosol does not contain inhibitors of Cat-D. This explains the prominent role of Cat-D during the early phase of commitment to apoptosis, evidenced by Bax activation, and the survival advantage conferred by pepstatin A and Cat-D siRNA transfection upon HIV-1-infected cells. Note that the HIV-1 protease is also an aspartyl protease. However, Cat-D is inhibited by pepstatin A with a 50% inhibitory concentration of 10 nM (5), whereas the concentrations of pepstatin A required for inhibition of the proteolytic activity of the HIV-1 protease toward its specific substrate, the Gag precursor, are several orders of magnitude higher, with a 50% inhibitory concentration of 0.5 mM (31). The doses of pepstatin A used in our study were in fact insufficient to alter the occurrence of p24⫹ cells. Bax⫺/⫺ Bak⫺/⫺ double-knockout mouse embryonic cells fail to undergo mitochondrial-membrane permeabilization and cell death in response to lysosomotropic agents that disrupt the integrity of the lysosomal membrane (4). In this situation, Bax and Bak are each sufficient to induce cell death. This indicates that lysosomal cathepsins, once released into the cytosol, are not able to provoke mitochondrial dysfunction directly but require the activation of Bax or Bak. In HIV-1-infected cells, Bax changed its shape and integrated into the OMM, whereas Bak remained in its native conformation, indicating that of these two proapoptotic factors, Bax was the more important. As in activated T lymphocytes exposed to staurosporine (3), Bax activation was strictly dependent, during the phase of commitment to apoptosis, on Cat-D activity. By comparing Bax⫺/⫺ tumor cells and their normal counterparts, it was also found that certain anticancer drugs induce Cat-D translocation upstream from the activation of Bax, thereby provoking an

11438

LAFORGE ET AL.

apoptotic response (14). One possibility is that in all these situations, including in HIV-1-infected cells, a chaperone protein sequesters Bax in the cytosol of healthy cells and that the protein becomes the target of Cat-D proteolytic activity during apoptosis. Several cytosolic proteins, namely, Ku70 (53), 14-3 to 3␪ (45), and humanin (22), interact with Bax, stabilize its inactive conformation, and suppress its translocation to mitochondria. However, the specific target of Cat-D, allowing Bax activation in T lymphocytes, still awaits identification. During early commitment to apoptosis in ⌬⌿mhigh p24⫹ T cells, Cat-D-induced Bax conformational change and insertion at the OMM occurred, but Bax was not permeabilized. It has been reported that during commitment to apoptosis, Bax conformational change and translocation from the cytosol to the mitochondria may be reversible (18). Work with taxol-resistant tumor cells (38) and Bid-null embryo fibroblasts (51) has also demonstrated that Bax dimerization/complex formation at the mitochondrial membrane is insufficient in itself to permeabilize the OMM. A further Bax-activating signal is required. Truncated Bid (tBid), which enables oligomerized Bax to form pores in liposomes (33), is one such signal. However, Bid remained in its full-length form in ⌬⌿mhigh CD4⫹ T cells from HIV-1-infected cultures (not shown). Other death agonists, able to bind and activate Bax directly, were presumably also inactive, at least temporarily, in particular Bim, another BH3only protein of the Bcl-2 family (32), and p53, which accumulates in the cytosol during apoptosis and functions like a BH3only protein (6). It is conceivable that the proapoptotic effect of Cat-D was seconded by other lysosomal cathepsins accumulating in the cytosol and by proteins of viral origin until a point of no return represented by permeabilization of the OMM and by subsequent entry into the execution phase of apoptosis. Some Cat-B enzymatic activity was indeed detectable in the cytosol of whole HIV-1-infected CD4⫹ T-cell populations that comprise substantial percentages (⬃40%) of ⌬⌿mlow apoptotic cells (not shown). This suggests that Cat-B accumulated in the cytosol of p24⫹ cells, finally overcoming the inhibitory effect of endogenous cystatins. Note that Bid is susceptible at neutral pH to proteolytic activation by cysteine cathepsins, including Cat-B and -L (7). As expected, Bid was proteolytically activated to truncated tBid at late times (not shown). Presumably, released cysteine cathepsins may be involved, together with caspases, in the generation of tBid, which in turn interacts with Bax to permeabilize the OMM. We reasoned that Nef, an early and abundant HIV-1-encoded protein, would be a good molecular candidate for a protein inducing early lysosomal rupture in HIV-1-infected cells. In mice, nef transgene expression alone induces the development of a severe AIDS-like disease, whereas vpu, vpr, or tat transgene expression is dispensable for the emergence of this disease phenotype (24). In severe combined immunodeficient (human thymus/liver) mice, a Nef-defective virus has been shown to replicate well in the human thymic tissue without causing apparent cellular damage. Restoration of the nef gene restored the pathogenicity in this virus, as evidenced by thymocyte depletion, without significantly altering its replication kinetics (12). Thus, Nef is in itself an important pathogenic factor in HIV-1 infection, independently of viral spread and replication. Nef has also been reported to enhance (but not to induce) the apoptotic death of cell lines exposed to a

J. VIROL.

number of death agonists (48). The mechanisms by which Nef mediates pathogenicity are incompletely understood. Nef can form a signaling complex with the ␨ chain of the T-cell receptor complex, bypassing the requirement for antigen to initiate the induction of FasL expression and the subsequent killing of Fas-bearing uninfected bystander cells (61, 62). However, cellautonomous apoptosis of HIV-1-infected (tumor) cells, via the cis ligation of Fas, can be prevented by the association of Nef with Ask1 (apoptosis signal-regulating kinase 1) (17). Thus, Nef may actually protect HIV-infected tumor cells from apoptosis. In particular, Nef-induced downregulation of the cell surface density of major histocompatibility complex class 1 molecules (54) decreases the recognition of HIV-infected cells by HIV-1-specific cytotoxic T lymphocytes and prolongs the survival of infected target cells (9). We show here that Nef can drive primary activated CD4⫹ T lymphocytes into apoptosis by inducing limited lysosomal permeabilization. This event occurs very early, within 3 h of transfection, before the cell surface CD4 receptor is significantly downregulated (at 16 h posttransfection). Primary CD4⫹ T lymphocytes are probably particularly susceptible to the cytopathic effects of Nef, as tumor cell lines are not affected by the production of Nef alone (48). It remains unclear how Nef induces lysosome membrane permeabilization. The multiple interactions between Nef and cellular transduction proteins (reviewed in reference 49) illustrate the complexity of this issue. Most of the effects of Nef are attributed to its association with cellular transduction proteins, owing to the myristoylation of its N terminus. Interestingly, we observed that the transfection of CD4⫹ T cells with the nonmyristoylated Nef-G2A construct nonetheless led to some lysosomal destabilization. This may be related to the ability of the G2A mutant, which is restricted to the cytosol, to function as a phosphorylation substrate for protein kinase C, albeit with a lower efficiency than for the wild type (8), thereby activating putative downstream transduction proteins. This is also consistent with the ability of the G2A mutation to strongly attenuate, but not completely abolish, Nef-induced CD4⫹ T-cell depletion in the peripheral lymphoid organs of G2A transgenic mice, although it completely abrogates the development of an AIDS-like disease (25). The Nef-defective NL4-3HSA HIV-1 has been found to be highly cytopathic in vitro for CD4⫹ T lymphocytes, independently of the presence of Env (34), indicating that other HIV-1 accessory proteins are active in inducing the cytopathic effect of HIV-1. In particular, the viral regulatory protein (Vpr) is an important endogenous proapoptotic factor in productively infected cells, which act by provoking G2 arrest downstream from the activation of the DNA-damage signaling protein (ATR) and by inducing Bax-induced mitochondrial permeabilization (1). Our approach, consisting of transfecting CD4⫹ T lymphocytes with an expression vector encoding only the HIV-1 Nef protein, allowed us to evaluate the proper cytopathic effect of this protein. Thus, our study suggests a new role for the HIV-1 Nef protein in primary CD4⫹ T lymphocytes in which Nef induces the early permeabilization of lysosomes, leading to Cat-D activity in the cytosol, Bax activation, and the initiation of the apoptotic death machinery.

VOL. 81, 2007

LYSOSOMAL PERMEABILIZATION LINKED TO HIV-1 INFECTION ACKNOWLEDGMENTS

This work was supported by the Institut National de la Recherche Me´dicale and the Institut Pasteur. M.L. was supported by fellowships from the Ministe`re de la Recherche et de la Technologie, from the Association pour la Recherche contre le Cancer, and from SIDACTION. J.E. was supported by grants from ANRS and from the Fondation pour la Recherche Me´dicale. We warmly thank O. Schwartz for providing the Nef expression vectors. REFERENCES 1. Andersen, J. L., J. L. Dehart, E. S. Zimmerman, O. Ardon, B. Kim, G. Jacquot, S. Benichou, and V. Planelles. 2006. HIV-1 Vpr-induced apoptosis is cell cycle dependent and requires Bax but not ANT. PLoS Pathog. 2:e127. 2. Arnoult, D., F. Petit, J. D. Lelievre, and J. Estaquier. 2003. Mitochondria in HIV-1-induced apoptosis. Biochem. Biophys. Res. Commun. 304:561–574. 3. Bide`re, N., H. K. Lorenzo, S. Carmona, M. Laforge, F. Harper, C. Dumont, and A. Senik. 2003. Cathepsin D triggers Bax activation, resulting in selective apoptosis-induced factor (AIF) relocation in T lymphocytes entering the early commitment phase to apoptosis. J. Biol. Chem. 278:31401–31411. 4. Boya, P., K. Andreau, D. Poncet, N. Zamzami, J. L. Perfettini, D. Metivier, D. M. Ojcius, M. Jaattela, and G. Kroemer. 2003. Lysosomal membrane permeabilization induces cell death in a mitochondrion-dependent fashion. J. Exp. Med. 197:1323–1334. 5. Chen, C. S., W. N. Chen, M. Zhou, S. Arttamangkul, and R. P. Haugland. 2000. Probing the cathepsin D using a BODIPY FL-pepstatin A: applications in fluorescence polarization and microscopy. J. Biochem. Biophys. Methods 42:137–151. 6. Chipuk, J. E., and D. R. Green. 2006. Dissecting p53-dependent apoptosis. Cell Death Differ. 13:994–1002. 7. Cirman, T., K. Oresic, G. D. Mazovec, V. Turk, J. C. Reed, R. M. Myers, G. S. Salvesen, and B. Turk. 2004. Selective disruption of lysosomes in HeLa cells triggers apoptosis mediated by cleavage of Bid by multiple papain-like lysosomal cathepsins. J. Biol. Chem. 279:3578–3587. 8. Coates, K., S. J. Cooke, D. A. Mann, and M. P. Harris. 1997. Protein kinase C-mediated phosphorylation of HIV-I nef in human cell lines. J. Biol. Chem. 272:12289–12294. 9. Collins, K. L., B. K. Chen, S. A. Kalams, B. D. Walker, and D. Baltimore. 1998. HIV-1 Nef protein protects infected primary cells against killing by cytotoxic T lymphocytes. Nature 391:397–401. 10. De Clercq, E. 2001. New developments in anti-HIV chemotherapy. Curr. Med. Chem. 8:1543–1572. 11. Dumont, C., A. Durrbach, N. Bidere, M. Rouleau, G. Kroemer, G. Bernard, F. Hirsch, B. Charpentier, S. A. Susin, and A. Senik. 2000. Caspase-independent commitment phase to apoptosis in activated blood T lymphocytes: reversibility at low apoptotic insult. Blood 96:1030–1038. 12. Duus, K. M., E. D. Miller, J. A. Smith, G. I. Kovalev, and L. Su. 2001. Separation of human immunodeficiency virus type 1 replication from Nefmediated pathogenesis in the human thymus. J. Virol. 75:3916–3924. 13. Ekert, P. G., J. Silke, C. J. Hawkins, A. M. Verhagen, and D. L. Vaux. 2001. DIABLO promotes apoptosis by removing MIHA/XIAP from processed caspase 9. J. Cell Biol. 152:483–490. 14. Erdal, H., M. Berndtsson, J. Castro, U. Brunk, M. C. Shoshan, and S. Linder. 2005. Induction of lysosomal membrane permeabilization by compounds that activate p53-independent apoptosis. Proc. Natl. Acad. Sci. USA 102:192–197. 15. Fenard, D., W. Yonemoto, C. de Noronha, M. Cavrois, S. A. Williams, and W. C. Greene. 2005. Nef is physically recruited into the immunological synapse and potentiates T cell activation early after TCR engagement. J. Immunol. 175:6050–6057. 16. Gandhi, R. T., B. K. Chen, S. E. Straus, J. K. Dale, M. J. Lenardo, and D. Baltimore. 1998. HIV-1 directly kills CD4⫹ T cells by a Fas-independent mechanism. J. Exp. Med. 187:1113–1122. 17. Geleziunas, R., W. Xu, K. Takeda, H. Ichijo, and W. C. Greene. 2001. HIV-1 Nef inhibits ASK1-dependent death signalling providing a potential mechanism for protecting the infected host cell. Nature 410:834–838. 18. Gilmore, A. P., A. D. Metcalfe, L. H. Romer, and C. H. Streuli. 2000. Integrin-mediated survival signals regulate the apoptotic function of Bax through its conformation and subcellular localization. J. Cell Biol. 149:431– 446. 19. Gougeon, M. L. 2003. Apoptosis as an HIV strategy to escape immune attack. Nat. Rev. Immunol. 3:392–404. 20. Griffiths, G. J., L. Dubrez, C. P. Morgan, N. A. Jones, J. Whitehouse, B. M. Corfe, C. Dive, and J. A. Hickman. 1999. Cell damage-induced conformational changes of the pro-apoptotic protein Bak in vivo precede the onset of apoptosis. J. Cell Biol. 144:903–914. 21. Guicciardi, M. E., M. Leist, and G. J. Gores. 2004. Lysosomes in cell death. Oncogene 23:2881–2890. 22. Guo, B., D. Zhai, E. Cabezas, K. Welsh, S. Nouraini, A. C. Satterthwait, and

23.

24. 25.

26. 27. 28. 29. 30. 31.

32.

33.

34.

35. 36.

37. 38.

39. 40. 41. 42.

43.

44. 45.

11439

J. C. Reed. 2003. Humanin peptide suppresses apoptosis by interfering with Bax activation. Nature 423:456–461. Gyrd-Hansen, M., T. Farkas, N. Fehrenbacher, L. Bastholm, M. HoyerHansen, F. Elling, D. Wallach, R. Flavell, G. Kroemer, J. Nylandsted, and M. Jaattela. 2006. Apoptosome-independent activation of the lysosomal cell death pathway by caspase 9. Mol. Cell. Biol. 26:7880–7891. Hanna, Z., D. G. Kay, N. Rebai, A. Guimond, S. Jothy, and P. Jolicoeur. 1998. Nef harbors a major determinant of pathogenicity for an AIDS-like disease induced by HIV-1 in transgenic mice. Cell 95:163–175. Hanna, Z., E. Priceputu, D. G. Kay, J. Poudrier, P. Chrobak, and P. Jolicoeur. 2004. In vivo mutational analysis of the N-terminal region of HIV-1 Nef reveals critical motifs for the development of an AIDS-like disease in CD4C/HIV transgenic mice. Virology 327:273–286. Ho, D. D., A. U. Neumann, A. S. Perelson, W. Chen, J. M. Leonard, and M. Markowitz. 1995. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 373:123–126. Hsu, Y. T., and R. J. Youle. 1998. Bax in murine thymus is a soluble monomeric protein that displays differential detergent-induced conformations. J. Biol. Chem. 273:10777–10783. Jaattela, M. 2004. Multiple cell death pathways as regulators of tumour initiation and progression. Oncogene 23:2746–2756. Johansson, A. C., H. Steen, K. Ollinger, and K. Roberg. 2003. Cathepsin D mediates cytochrome c release and caspase activation in human fibroblast apoptosis induced by staurosporine. Cell Death Differ. 10:1253–1259. Kestler, H. W., III, D. J. Ringler, K. Mori, D. L. Panicali, P. K. Sehgal, M. D. Daniel, and R. C. Desrosiers. 1991. Importance of the nef gene for maintenance of high virus loads and for development of AIDS. Cell 65:651–662. Krausslich, H. G., H. Schneider, G. Zybarth, C. A. Carter, and E. Wimmer. 1988. Processing of in vitro-synthesized Gag precursor proteins of human immunodeficiency virus (HIV) type 1 by HIV proteinase generated in Escherichia coli. J. Virol. 62:4393–4397. Kuwana, T., L. Bouchier-Hayes, J. E. Chipuk, C. Bonzon, B. A. Sullivan, D. R. Green, and D. D. Newmeyer. 2005. BH3 domains of BH3-only proteins differentially regulate Bax-mediated mitochondrial membrane permeabilization both directly and indirectly. Mol. Cell 17:525–535. Kuwana, T., M. R. Mackey, G. Perkins, M. H. Ellisman, M. Latterich, R. Schneiter, D. R. Green, and D. D. Newmeyer. 2002. Bid, Bax, and lipids cooperate to form supramolecular openings in the outer mitochondrial membrane. Cell 111:331–342. Lenardo, M. J., S. B. Angleman, V. Bounkeua, J. Dimas, M. G. Duvall, M. B. Graubard, F. Hornung, M. C. Selkirk, C. K. Speirs, C. Trageser, J. O. Orenstein, and D. L. Bolton. 2002. Cytopathic killing of peripheral blood CD4⫹ T lymphocytes by human immunodeficiency virus type 1 appears necrotic rather than apoptotic and does not require Env. J. Virol. 76:5082– 5093. Levy, J. A. 1993. Pathogenesis of human immunodeficiency virus infection. Microbiol. Rev. 57:183–289. Li, Q., L. Duan, J. D. Estes, Z. M. Ma, T. Rourke, Y. Wang, C. Reilly, J. Carlis, C. J. Miller, and A. T. Haase. 2005. Peak SIV replication in resting memory CD4⫹ T cells depletes gut lamina propria CD4⫹ T cells. Nature 434:1148–1152. Li, W., X. Yuan, G. Nordgren, H. Dalen, G. M. Dubowchik, R. A. Firestone, and U. T. Brunk. 2000. Induction of cell death by the lysosomotropic detergent MSDH. FEBS Lett. 470:35–39. Makin, G. W., B. M. Corfe, G. J. Griffiths, A. Thistlethwaite, J. A. Hickman, and C. Dive. 2001. Damage-induced Bax N-terminal change, translocation to mitochondria and formation of Bax dimers/complexes occur regardless of cell fate. EMBO J. 20:6306–6315. Matsumoto, N., O. D. Jo, R. N. Shih, and N. Yanagawa. 2005. Altered cathepsin D metabolism in PHEX antisense human osteoblast cells. Biochem. Biophys. Res. Commun. 332:248–253. Mattapallil, J. J., D. C. Douek, B. Hill, Y. Nishimura, M. Martin, and M. Roederer. 2005. Massive infection and loss of memory CD4⫹ T cells in multiple tissues during acute SIV infection. Nature 434:1093–1097. McManus, M. T., B. B. Haines, C. P. Dillon, C. E. Whitehurst, L. van Parijs, J. Chen, and P. A. Sharp. 2002. Small interfering RNA-mediated gene silencing in T lymphocytes. J. Immunol. 169:5754–5760. Michel, N., K. Ganter, S. Venzke, J. Bitzegeio, O. T. Fackler, and O. T. Keppler. 2006. The Nef protein of human immunodeficiency virus is a broadspectrum modulator of chemokine receptor cell surface levels that acts independently of classical motifs for receptor endocytosis and Galphai signaling. Mol. Biol. Cell 17:3578–3590. Miller, M. D., M. T. Warmerdam, I. Gaston, W. C. Greene, and M. B. Feinberg. 1994. The human immunodeficiency virus-1 nef gene product: a positive factor for viral infection and replication in primary lymphocytes and macrophages. J. Exp. Med. 179:101–113. Nechushtan, A., C. L. Smith, Y. T. Hsu, and R. J. Youle. 1999. Conformation of the Bax C-terminus regulates subcellular location and cell death. EMBO J. 18:2330–2341. Nomura, M., S. Shimizu, T. Sugiyama, M. Narita, T. Ito, H. Matsuda, and Y. Tsujimoto. 2003. 14-3-3 interacts directly with and negatively regulates pro-apoptotic Bax. J. Biol. Chem. 278:2058–2065.

11440

LAFORGE ET AL.

46. Otter, I., S. Conus, U. Ravn, M. Rager, R. Olivier, L. Monney, D. Fabbro, and C. Borner. 1998. The binding properties and biological activities of Bcl-2 and Bax in cells exposed to apoptotic stimuli. J. Biol. Chem. 273:6110–6120. 47. Petit, F., D. Arnoult, J. D. Lelievre, L. Moutouh-de Parseval, A. J. Hance, P. Schneider, J. Corbeil, J. C. Ameisen, and J. Estaquier. 2002. Productive HIV-1 infection of primary CD4⫹ T cells induces mitochondrial membrane permeabilization leading to a caspase-independent cell death. J. Biol. Chem. 277:1477–1487. 48. Rasola, A., D. Gramaglia, C. Boccaccio, and P. M. Comoglio. 2001. Apoptosis enhancement by the HIV-1 Nef protein. J. Immunol. 166:81–88. 49. Renkema, G. H., and K. Saksela. 2000. Interactions of HIV-1 NEF with cellular signal transducing proteins. Front. Biosci. 5:D268–D283. 50. Robert-Guroff, M., M. Popovic, S. Gartner, P. Markham, R. C. Gallo, and M. S. Reitz. 1990. Structure and expression of tat-, rev-, and nef-specific transcripts of human immunodeficiency virus type 1 in infected lymphocytes and macrophages. J. Virol. 64:3391–3398. 51. Ruffolo, S. C., D. G. Breckenridge, M. Nguyen, I. S. Goping, A. Gross, S. J. Korsmeyer, H. Li, J. Yuan, and G. C. Shore. 2000. BID-dependent and BID-independent pathways for BAX insertion into mitochondria. Cell Death Differ. 7:1101–1108. 52. Saelens, X., N. Festjens, E. Parthoens, I. Vanoverberghe, M. Kalai, F. van Kuppeveld, and P. Vandenabeele. 2005. Protein synthesis persists during necrotic cell death. J. Cell Biol. 168:545–551. 53. Sawada, M., W. Sun, P. Hayes, K. Leskov, D. A. Boothman, and S. Matsuyama. 2003. Ku70 suppresses the apoptotic translocation of Bax to mitochondria. Nat. Cell Biol. 5:320–329. 54. Schwartz, O., V. Marechal, S. Le Gall, F. Lemonnier, and J. M. Heard. 1996. Endocytosis of major histocompatibility complex class I molecules is induced by the HIV-1 Nef protein. Nat. Med. 2:338–342.

J. VIROL. 55. Sol-Foulon, N., C. Esnault, Y. Percherancier, F. Porrot, P. Metais-Cunha, F. Bachelerie, and O. Schwartz. 2004. The effects of HIV-1 Nef on CD4 surface expression and viral infectivity in lymphoid cells are independent of rafts. J. Biol. Chem. 279:31398–31408. 56. Terai, C., R. S. Kornbluth, C. D. Pauza, D. D. Richman, and D. A. Carson. 1991. Apoptosis as a mechanism of cell death in cultured T lymphoblasts acutely infected with HIV-1. J. Clin. Investig. 87:1710–1715. 57. Turk, B., D. Turk, and G. S. Salvesen. 2002. Regulating cysteine protease activity: essential role of protease inhibitors as guardians and regulators. Curr. Pharm. Des. 8:1623–1637. 58. Wei, B. L., V. K. Arora, J. L. Foster, D. L. Sodora, and J. V. Garcia. 2003. In vivo analysis of Nef function. Curr. HIV Res. 1:41–50. 59. Wei, X., S. K. Ghosh, M. E. Taylor, V. A. Johnson, E. A. Emini, P. Deutsch, J. D. Lifson, S. Bonhoeffer, M. A. Nowak, B. H. Hahn, et al. 1995. Viral dynamics in human immunodeficiency virus type 1 infection. Nature 373: 117–122. 60. Wolter, K. G., Y. T. Hsu, C. L. Smith, A. Nechustan, X. G. Xi, and R. J. Youle. 1997. Movement of Bax from the cytosol to mitochondria during apoptosis. J. Cell Biol. 139:1281–1292. 61. Xu, X. N., B. Laffert, G. R. Screaton, M. Kraft, D. Wolf, W. Kolanus, J. Mongkolsapay, A. J. McMichael, and A. S. Baur. 1999. Induction of Fas ligand expression by HIV involves the interaction of Nef with the T cell receptor zeta chain. J. Exp. Med. 189:1489–1496. 62. Zauli, G., D. Gibellini, P. Secchiero, H. Dutartre, D. Olive, S. Capitani, and Y. Collette. 1999. Human immunodeficiency virus type 1 Nef protein sensitizes CD4⫹ T lymphoid cells to apoptosis via functional upregulation of the CD95/CD95 ligand pathway. Blood 93:1000–1010.