Human Immunodeficiency Virus 1 Envelope Glycoprotein Complex-Induced Apoptosis Involves Mammalian Target of Rapamycin/Fkbp12-Rapamycin–Associated Protein–Mediated P53 Phosphorylation

HIV-1 infection can cause apoptosis via a variety of mechanisms, some of which rely on the intricate virus host cell interaction (in vitro) and some of which involve activation of the host's inflammatory and immune systems (in vivo; references 1 and 2). The envelope glycoprotein complex (Env)(gp120/gp41) appears to be one of the dominant apoptosis-inducing molecules encoded by the HIV-1 genome. Env expressed on the plasma membrane of infected cells can interact with CD4 and a suitable coreceptor (e.g., CXCR4) to trigger cell-to-cell fusion; the resulting syncytia subsequently undergo apoptosis. This applies to primary CD4⁺ T lymphocytes inoculated by HIV-1 (references 3 and 4), as well as to cell lines stably transfected with human CD4 cocultured with cells expressing a lymphotropic HIV-1 Env gene 5,6,7. A positive correlation between CD4⁺ T cell decline and infection by syncytium-inducing HIV-1 or SIV-1 variants has been established in vitro 3,4,8,9 and, more importantly, in vivo, in humans with AIDS 10,11, humanized severe combined immunodeficient mice 12, and monkeys 13. This suggests that fusion-induced apoptosis is relevant to AIDS pathogenesis, although syncytia are relatively infrequent in lymphoid tissues of HIV-1⁺ donors 14. In lymph nodes from HIV-1–infected individuals, syncytia express markers of early apoptosis such as tissue transglutaminase (tTGase; reference 15). The scarcity of syncytia in vivo may thus be attributed to their rapid apoptotic degeneration and phagocytic removal, yet does not argue against the pathophysiological importance of fusion-induced apoptosis.

The apoptosis-associated activation of caspases can be triggered by two pathways 16,17. In the “extrinsic” pathway, the ligation of plasma membrane death receptors (e.g., CD95, TNF-R) leads to the recruitment of caspase-8 to the receptor complex, culminating in its proteolytic autoactivation 16. In the “intrinsic” pathway, mitochondrial membrane permeabilization (MMP) results in the release of cytochrome c (Cyt. c) from the mitochondrial intermembrane space. Once in the cytosol, Cyt. c triggers the oligomerization of Apaf-1, which in turn recruits procaspase-9 and procaspase-3 into the apoptosome, the caspase activation multiprotein complex 18. MMP may be elicited by a variety of different proapoptotic second messengers (e.g., Ca²⁺, ganglioside GD3, reactive oxygen species), as well as proapoptotic members of the Bcl-2 family such as Bax, Bak, Bid, Bim, etc. 17. Nuclear DNA damage is one of the best studied initiators of the intrinsic pathway of apoptosis 19,20. DNA double-strand breaks are sensed by two enzymes from a protein family termed the phosphatidylinositol kinase-related kinases (or PIKKs), in particular ataxia teleangiectasia mutated (ATM) and DNA-dependent protein kinase (DNA-PK). UV-induced damage is mainly sensed by ataxia teleangiectasia Rad3 related (ATR), yet another member of the PIKK family. DNA-PK, ATM, ATR, as well as the downstream DNA-damage-induced kinases Chk1 (activated by UV) and Chk2 (activated by ionizing radiation), phosphorylate serine residues within the NH₂ terminus of p53, a modification that by itself may increase its transactivating function and/or inhibits the interaction of p53 with MDM2 and so prevents degradation of the p53 protein, which accumulates and can act as a transcription factor, stimulating the expression of several MMP-inducing proteins including Bax 21.

Syncytial apoptosis induced by the interaction between Env and CD4/CXCR4, in the absence of viral infection, involves a mitochondrial pathway characterized by the following sequence of events: (i) translocation of Bax from the cytosol to mitochondria; (ii) Bax-mediated MMP with loss of the mitochondrial transmembrane potential (ΔΨ_m), release of apoptogenic intermembrane proteins, in particular apoptosis-inducing factor and Cyt. c, (iii) caspase activation, and (iv) nuclear chromatin condensation 6. This sequence of events has been confirmed, in principle, for HIV-1–infected HeLa cells transfected with CD4, as well as for primary CD4⁺ lymphoblasts infected with lymphotropic HIV-1 6,22. No information was available on the premitochondrial proapoptotic signal transduction pathways involved in Env-induced syncytial apoptosis. Therefore, we have employed a systematic approach, based on proteomics and inhibitor screening assays, to define such upstream signals accounting for the induction and mitochondrial translocation of Bax. Here we show that phosphorylation of p53, via a pathway not related to DNA damage, is involved in Env-induced syncytial apoptosis. Indeed, it appears that p53 is phosphorylated on serine 15 by mammalian target of rapamycin (mTOR), also called FKBP12-rapamycin-associated protein (FRAP), a protein from the PIKK family that previously has been suggested to serve as a master regulator of the balance between protein synthesis and degradation 23, thereby participating in the control of cell size 24,25 and perhaps in oncogenic transformation 26.

HeLa cells transfected with the Env gene of HIV-1 LAI (HeLa Env, reference 27) and HeLa cells transfected with CD4 (HeLa CD4, reference 5) were cultured alone or together (1:1 ratio) in medium supplemented with 10% FCS, as described previously 5,6,27. Mouse embryonic fibroblasts (MEFs) with different genotypes (p53^+/+, p53^−/+, p53^−/−) were a gift from T. Jacks, MIT, Boston, MA. MEFs were cultured on coverslips and were fused with polyethylene glycol (PEG) (45 s of incubation with prewarmed [37°C] 50% wt/vol PEG from Sigma-Aldrich, 1,450, in Ca²⁺-free PBS, pH 7.2; preceded by two washings with FCS-free medium), followed by four washings with complete medium. N-benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (Z-VAD. fmk, used at 100 μM) was added each 24 h, whereas other inhibitors were added only once, from the beginning of the culture. Transfection with pcDNA3.1 vector only, pBR322 plasmid constructs containing wild-type p53 or mutant (H175, H273) p53 28 (gift of T. Soussi, Institut Curie, Paris, France), expressed under the control of the cytomegalovirus promoter, luciferase, terminal oligopyrimidine (TOP)-luciferase constructs 29 (gift of J. Chen, Illinois University, Urbana, IL), or a p53-responsive enhanced green fluorescent protein (GFP) plasmid (gift from K. Wiman, Karolinska Cancer Center, Stockholm, Sweden) was performed by electroporation 30 24 h before coculture of HeLa CD4 and HeLa Env cells. The p53-responsive GFP plasmid was generated by replacing the luciferase gene in PG13PY Luc (a luciferase construct containing 13 repeats of the p53-binding oligonucleotide 5′-CCTGCCCTGGGACTTGCCTGG-3′, gift from Bert Vogelstein, Howard Hughes Medical Institute, Baltimore, MD) with an EcoA7III-MluI fragment from pEGFP-C1. In some experiments, all syncytia growing on a premarked V-shaped area of a coverslip (>200 per experiment) were microinjected into the cytoplasm 31, with PBS only, pH 7.2, recombinant human Bcl-2 (amino acids 1–218; 500 ng/μl; reference 32), a mAb specific for mTOR/FRAP (BD Transduction Laboratories) or an isotype-(IgG2a) matched LAMP-1 mAb (Oncogene Research Products) (both at 50 ng/μl). Comet assays were performed using a kit from Trevigen, on HeLa Env cells (untreated or treated with 100 μM H₂O₂ for 30 min) or 24-h-old HeLa Env/CD4 syncytia (nonapoptotic adherent or apoptotic cells).

Peripheral blood samples were obtained from 49 HIV-infected individuals from the National Institute for Infectious Diseases (IRCCS Lazzaro Spallanzani, Rome, Italy). Patients (mean age 34 ± 11 y, n = 49) were included in this study according to the following criteria: asymptomatic HIV-1 infection (A, CDC 1993), under at least 2 mo of antiretroviral therapy interruption, CD4⁺ T lymphocytes in peripheral blood in the range of 300–600/μl, plasma HIV-RNA levels >80 copies per milliliter, no treatment with IFNs or corticosteroids, absence of infection with hepatitis C virus, hepatitis B virus, and autoimmune disease. All patients provided informed written consent according to the Ethical Committee. PBMCs were isolated by Ficoll/Hypaque (Amersham Pharmacia Biotech) centrifugation of heparinized blood from either healthy donors and HIV-seropositive individuals and fixed with 4% formaldehyde in PBS, pH 7.2. Plasma HIV-1 RNA levels were determined by the Nucleic Acid Sequence Based Amplification (NASBA) procedure (HIV-1 RNA QT assay; Organon Teknika). Plasma levels of HIV-1 RNA were expressed as HIV-RNA copies per milliliter with a threshold level of 80 copies per milliliter. Biopsies of axillary lymph nodes, obtained from healthy donors and HIV-1⁺ asymptomatic patients, naive for antiretroviral therapy, were immediately fixed with 10% formalin neutral buffered, dehydrated, and paraffin embedded.

5 × 10⁶ HeLa CD4 cells were cultured in the presence of 2.5 × 10⁶ chronically HIV-1–infected H9/IIIB cells (obtained from R.C. Gallo, National Institutes of Health, Bethesda, MD) in 5 ml of RPMI 1640 supplemented with 10% heat inactivated FCS and 2 mM glutamine. Alternatively, CEM cells containing a plasmid encoding GFP driven by the HIV-1 long terminal repeat 33; clone 8D6; obtained from J. Corbel, University of California at San Diego, San Diego, CA) were cocultured (1:1 ratio) with HIV-1–infected H9/IIIB cells. Cell-to-cell fusion was assessed by determining the GFP-dependent fluorescence (excitation 458 nm, emission 515 nm) in a Fluoroscan plate fluorometer (Labsystems). Alternatively, cells were resuspended in a solution containing 3.4 mM sodium citrate, 0.05 mg/ml propidium iodide, 0.1 mM EDTA, 1 mM Tris, pH 8, and 0.1% Triton X-100 34, incubated for 60 min at reverse transcription, and analyzed for DNA content in a FACScalibur™ (Becton Dickinson). This method allows for the quantitation of all apoptotic nuclei, including those of syncytia, which are disrupted by the hypotonic buffer. CD4⁺ T cells were purified from freshly isolated PBMCs by immunomagnetic negative selection (Stem Cell Technologies). Cells were cultured for 2 d in RPMI 1640 medium supplemented with 10% FCS (GIBCO BRL) containing 3 μg/ml PHA (Sigma-Aldrich) and 6 U/ml IL-2 (IL-2; Boehringer Mannheim). For HIV-1 infection, 10 × 10⁶ cells were incubated with the T cell line–adapted HIV-1 strain IIIB (500 ng of p24) for 4 h at 37°C. After washing out unabsorbed virus, cells were cultured at a density of 10⁶ cells per milliliter in RPMI 1640 medium containing 20% FCS and 10 UI/ml IL-2 in 24-well plates. 2 d after infection, when the first syncytia appeared in the infected cultures, infected and uninfected cultures were treated with 5 μg/ml AMD3100, 100 μM ZVAD.fmk, or the indicated dose of rapamycin. Infection was allowed to proceed for 2 more d. At this time mitochondrial function and nuclear morphology were evaluated after JC1/Hoechst staining as described below. Viral replication was evaluated by measuring the level of HIV-1 core protein p24 in the supernatant (Innogenetics ELISA kit). Promonocytic U937 cell line were infected at a multiplicity of infection of 0.5 TCID₅₀ per cell, for 2 h at 37°C washed three times in PBS.

Protein samples were simultaneously prepared from HeLa Env and HeLa CD4 single cells (SCs) mixed at a 1:1 ratio in lysis buffer (0 control) or from HeLa Env/CD4 syncytia obtained after 18 or 36 h of coculture. These samples were then processed by the PowerBlot facility (Becton Dickinson), which determined the expression level of ∼800 different signal-transducing proteins using a combination of SDS-PAGE (5–15% gradient), immunoblotting with specific monoclonal antibodies revealed by a secondary goat anti–mouse horseradish peroxidase, capture of chemiluminesence data by a CDD camera, and computerized processing of densitometric data (for details consult www.translab. com/TOC.shtml). Data were normalized by dividing the signal obtained for each protein by the sum of signals obtained for all 800 proteins for one given sample. For proteins revealing a variation in the expression of at least 20%, determined in two initial runs, expression levels were verified in a third independent experiment.

Aliquots of total protein extracts (40 μg), were run on SDS-PAGE and electroblotted overnight at 4°C onto nitrocellulose membrane. Immunodetection involved antibodies specific for p53, Bax, Bcl-2, mTOR/FRAP (BD Transduction Laboratories), phosphorylation of p53 on serine 15 (p53S15P) (Cell Signaling Technology), appropriate secondary antibodies (goat anti–rabbit or goat anti–mouse; Bio-Rad Laboratories) conjugated to horseradish peroxidase, and the enhanced ECL chemiluminescence detection system (Amersham Pharmacia Biotech). Equal loading and transfer was monitored by Ponceau red staining of nitrocellulose membranes. For immunoprecipitation, lysates from sonicated, Triton X-100-solublized cells (60 μg protein in 100 μl PBS with protease inhibitors) were incubated for 90 min at 37°C with 500 ng affinity-purified rabbit polyclonal antibodies specific for p53, p53S15P, or Bcl-2, followed by addition of 10 μl-packed protein A/G-agarose beads (30 min, 37°C; Santa Cruz Biotechnology, Inc.), vigorous washing of the pellet (10 min at 10,000 g, 3 ×) in PBS, 5% SDS PAGE, and immunodetection with an mTOR/FRAP-specific mAb.

For the simultaneous assessment of mitochondrial and nuclear features of apoptosis, live cells were stained with the potentiometric dye 5,5′,6,6′-tetrachloro-1,1′, 3,3′-tetraethylbenzimidazolylcarbocyanine iodide (2 μM JC-1; Molecular Probes), as well as Hoechst 33342 (2 μM; Sigma-Aldrich) for 30 min at 37°C in complete culture medium 6. A rabbit antiserum specific for p53S15P was used on paraformaldehyde (4% wt:vol) and picric acid-fixed (0.19% vol:vol) cells and revealed with a goat anti–rabbit IgG conjugated to PE (Southern Biotechnology Associates, Inc.). Cells were also stained for the detection of Cyt-c (mAb 6H2.B4 from BD PharMingen), Hsp60 (mAb H4149 from Sigma-Aldrich), Bax (mAb 6A7; BD PharMingen), p21 (mAb EA10; Oncogene Research Products), MDM2 (mAb IF2; Oncogene Research Products), mTOR/FRAP (mAb 3O; BD Transduction Laboratories), all revealed by a goat anti–mouse IgG ALEXA488 conjugate; Molecular Probes), and/or chromatin (Hoechst 33342; reference 35). Immunocytochemical staining was performed using polyclonal anti-p53S15P or monoclonal tTG (Neomarkers) antibodies. A biotinylated goat anti–mouse or anti–rabbit IgG, as a secondary antibody, was used, followed by a preformed horseradish peroxidase-conjugated streptavidin (Biogenex). The reaction was developed using aminoethylcarbazole and 3-3′ diaminobenzidine (Biogenex) as chromogenic substrates and 0.01% H₂O₂. Cells were counterstained in Mayer's acid hemalum. Endogenous peroxidase activity was blocked by preincubation with 3% H₂O₂. Results were quantified by two independent investigators on an average of 500 cells.

Syncytia formed by coculture of two different HeLa cell lines expressing HIV-1 Env or CD4/CXCR4 spontaneously undergo apoptosis 6,7. Among a panel of different apoptosis-regulatory proteins, we found only one major alteration, namely an upregulation of Bax that is well detectable 36 h after syncytium formation (Fig. 1 A). Significant alterations were also found in the expression level of the proapoptotic proteins Bid (which is downregulated) and caspase-7 (which is upregulated by ∼50%). Whereas total p53 levels did not change after formation of syncytia, an increased phosphorylation of serine 15 (but not serines 6, 9, 20, 37, and 397) of p53 (p53S15) was detected, using a panel of phosphoserine epitope-specific antibodies (Fig. 1 B). Phosphorylated p53S15 (p53S15P) was confined to the nuclei of syncytia (Fig. 1 C), and was not detected in SCs (Fig. 1 D). p53S15P could be detected early (6 h) after syncytium formation, well before Bax was overexpressed (Fig. 1 A and B) and apoptotic chromatin condensation became detectable (Fig. 1 D). Among p53S15P⁺ syncytia, only a fraction exhibited a conformational change in Bax linked to mitochondrial translocation (Fig. 1 E) as well as the release of Cyt. c from mitochondria (Fig. 1 F). Conversely, the translocation of Bax to mitochondria and of Cyt. c from mitochondria was only found among p53S15P⁺ (not p53S15P⁻) syncytia (Fig. 1 G), indicating that these changes occur after p53S15 phosphorylation. Accordingly, inhibition of Cyt. c-mediated caspase activation by Z-VAD.fmk (Fig. 1 H) or suppression of Bax translocation by microinjection of recombinant Bcl-2 protein 36 into syncytia (Fig. 1 I) failed to affect p53S15 phosphorylation. Both Z-VAD.fmk and Bcl-2 did, however, prevent nuclear apoptosis as an internal control of their efficacy (Fig. 1 H and I). Altogether, these data place p53S15 phosphorylation upstream of Bax upregulation/translocation, MMP, and subsequent caspase activation.

p53S15 phosphorylation is known to increase the transactivating function of p53 37 and thus may contribute to transcriptional activation of the bax gene 38 and apoptosis induction 21. p53 is not mutated in HeLa cells 39, yet has been reported to be functionally inactive due to the presence of E6 from papilloma virus 40. However, evidence that the low amount of p53 contained in HeLa cells is functional has been reported 41,42. It appears that p53S15 phosphorylation of HeLa syncytia elicited by the Env–CD4 interaction does result in transcriptional activation of p53 target genes, based on three lines of evidence. First, expression of the Bax protein increases (Fig. 1 B), as does that of Bax mRNA (data not shown). Second, the expression levels of two other prominent p53-activated gene products, namely MDM2 (Fig. 2 A and C) and p21 (Fig. 2 B and C), increase upon syncytium formation, with a similar delay, and this enhanced expression is restricted to the p53S15P⁺ subpopulation of syncytia (Fig. 2 A and B). Third, we employed a p53-inducible GFP construct that, when transiently transfected into HeLa Env cells, cultured in the absence of HeLa CD4 cells, fails to be expressed, unless it is cotransfected with p53 (but not with a dominant negative p53 mutant, p53H175) (Fig. 2 D). Upon coculture with HeLa CD4 cells, this p53-inducible GFP construct was strongly expressed in a subpopulation of p53S15P⁺ syncytia (but not in p53S15P⁻ syncytia nor in individual cells) (Fig. 2 E).

To validate our initial observations, obtained in HeLa syncytia, we investigated whether p53S15P is induced by HIV infection in vitro and in vivo. p53S15P was induced in heterokaryons generated by coculturing HeLa CD4 cells with a lymphoid cell line chronically infected with a syncytium-inducing HIV-1 isolate (Fig. 3 A), as well as in HIV-1–infected U937 myelomonocytary cells (Fig. 3 B), thus confirming that HIV-1 infection in vitro suffices to induce p53S15P 22. Importantly, p53S15P was also found among syncytia localized in the T cell area of lymph nodes from HIV-1⁺ patients (Fig. 3 C). Such cells also expressed the preapoptotic marker tTgase (Fig. 3 C). p53S15P was detected in PBMCs of HIV-1⁺ donors but not in PBMCs from HIV-1⁻ controls (<0.1% positive cells) (Fig. 3 D). The frequency of p53Ser15P⁺ PBMCs correlated with the percentage of tTgase⁺ cells (Fig. 3 E), as well as with viral load (Fig. 3 F). In conclusion, p53S15P is associated with HIV-1 infection, both upon in vitro infection of human cells as and vivo, in patient-derived samples. p53S15P expression correlates with HIV-1 viremia and preapoptosis.

p53S15 is known to be phosphorylated by stress kinases, in particular ERK and p38 kinase 43, as well as by several members of the PIKK family, in particular DNA-PK 44, ATM 45, ATR 46, and mTOR/FRAP 47. Among a large panel of serine kinases, phosphatases, and serine kinase activators, the only two proteins whose expression level changed in Env-induced syncytia were mTOR/FRAP, which is upregulated, and the protein phosphatase PP2A, which is downregulated (Fig. 4 A). PP2A is known to be inhibited by mTOR/FRAP 48, suggesting that its late (36 h) downregulation may be secondary to the upregulation/activation of mTOR/FRAP detectable at 18 h. Rapamycin (a specific inhibitor of mTOR/FRAP) and LY294002 (a general inhibitor of PI-3 kinases including mTOR/FRAP) significantly inhibited p53S15 phosphorylation (Fig. 4 B). In contrast, the ERK inhibitor PD98059, the p38 inhibitor SB203580, the ATM/DNA-PK inhibitor wortmannin 49,50 and the ATM/ATR inhibitor caffeine 46,51 failed to prevent the syncytium-specific p53S15 phosphorylation (Fig. 4 B). Accordingly, no phosphorylation of ERK (motif: pTEpY) or p38 kinase (motif: pTGpY) was detectable in syncytia (data not shown), indicating that these kinases are not activated. Moreover, comet assays did not detect double- strand breaks among nonapoptotic p53S15P⁺ syncytia (Fig. 4 C and D), further arguing against the implication of DNA-PK, ATM, and ATR (all of which are, in principle, activated by double-strand breaks; references 44–46) in p53S15 phosphorylation. As an alternative to rapamycin, microinjection of a mAb specific for mTOR/FRAP into newly formed syncytia caused a significant inhibition of p53S15 phosphorylation (Fig. 4 E). In conclusion, p53S15 phosphorylation is mediated via mTOR/FRAP.

mTOR/FRAP mediates translational control via phosphorylation of the p70^s6k protein kinase and the 4E-binding protein (4EBP) 1 23,24,52. In a control experiment performed on SCs, serum withdrawal (which inactivates mTOR/FRAP) was found to cause the dephosphorylation of p70^s6k and 4EBP1 (lane 4 in Fig. 5 A), which were phosphorylated again upon readdition of serum (lane 5). This serum-stimulated phosphorylation was fully inhibited by rapamycin (lane 6) as an internal control of its efficacy. In contrast, both SCs (lane 1) and syncytia (lane 2) cultured in complete, serum-containing medium exhibited constitutive phosphorylation of p70^s6k and 4EBP (Fig. 5 A), as well as constitutively high expression of a mTOR/FRAP-controlled 5′-TOP-luciferase reporter construct (Fig. 5 B). Accordingly, the expression of several mTOR/FRAP-controlled gene products including c-Myc, HIF-1, Rb, Rb2, and signal transducer and activator of transcription 3 was not increased upon syncytium formation (data not shown). Neither the constitutive phosphorylation of p70^s6k and 4EBP nor the translation of TOP-luciferase were inhibited by a 24-h exposure to rapamycin (lane 3 in Fig. 5 A and B), and rapamycin did not block the cell cycle of HeLa cells (data not shown). In contrast, within the same time frame, the fusion-dependent p53Ser15 phosphorylation was inhibited by rapamycin (Fig. 5 A). mTOR/FRAP coimmunoprecipitated with p53 and p53S15P in syncytia, not in SCs (Fig. 5 C). Moreover, we found that in syncytia mTOR/FRAP was enriched in the nucleus, as compared with SCs. This alteration was observed both in p53S15P⁻ and p53S15P⁺ syncytia (Fig. 5 D), and kinetic analyses revealed that the nuclear accumulation of mTOR/FRAP precedes the phosphorylation of p53S15P (Fig. 5 E). Altogether, these data suggest that mTOR/FRAP is activated in syncytia, where it directly phosphorylates p53 within the nucleus. In contrast, other cytoplasmic targets of mTOR/FRAP such as p70^s6k and 4EBP1 are unlikely to explain the rapamycin effects on syncytia.

When added to HeLa Env/HeLa CD4 cocultures, rapamycin and LY294002 inhibited all signs of syncytial apoptosis, including the loss of the ΔΨ_m (indicated by a red→green shift of the fluorescence emitted by the ΔΨ_m-sensitive fluorochrome JC-1), chromatin condensation (detected with the blue fluorochrome Hoechst 33342) (Fig. 6 A and B), the translocation of Bax to mitochondria, and the Bax-mediated release of Cyt. c from mitochondria (Fig. 6 B). Concomitantly, rapamycin and LY294002 allowed for the formation of larger syncytia (Fig. 6 C), an effect quantitatively similar to what could be seen with the pancaspase inhibitor Z-VAD.fmk (Fig. 6 C and reference 6). The apoptosis-inhibitory effect of rapamycin was durable, with a strong inhibitory effect on apoptosis of 6-d-old syncytia (81 ± 4% of viable, ΔΨ_m^high syncytia with 1 μM rapamycin versus only 25 ± 3% ΔΨ_m^high syncytia in untreated controls, n = 3). The antiapoptotic effects of rapamycin and LY294002 were observed at doses similar to those required for inhibition of p53S15 phosphorylation (Fig. 4 B). In contrast, wortmannin, caffeine, PD98059, SB203580, CsA, or FK506, which do not affect p53S15 phosphorylation (Fig. 4 B), also failed to prevent apoptosis (Fig. 6 D). Thus, a large panel of agents which inhibit apoptosis in other experimental systems failed to prevent syncytial apoptosis, with the notable exception of pyrrolidine dithiocarbamate (PDTC, Fig. 6 G), a pleiotropic agent known to interfere with p53 function 53,54. Of note, PDTC also reduced the phosphorylation of p53S15 (Fig. 4 B), underlining the intimate relationship between p53S15P and syncytial apoptosis.

To directly assess the role of p53 in syncytial apoptosis, Env and CD4/CXCR4-expressing cells were transiently transfected with two dominant negative (DN) p53 constructs (p53H175 and p53H273) 24 h before fusion. As compared with vector-only control transfections, DN-p53 conferred protection against all the mitochondrial and nuclear hallmarks of syncytial apoptosis including the translocation of Bax to mitochondria (Fig. 7). This protection was partial (∼50%), in agreement with the transfection efficiency (∼50%). These data indicate that the mTOR/FRAP→p53S15P pathway delineated here dictates the fate of Env-induced syncytia at the premitochondrial level, upstream of the upregulation of Bax. To address the involvement of p53 in syncytial apoptosis in a completely different experimental setting, MEFs with different p53 genotypes were subjected to short-term exposure to PEG, resulting in their fusion. In wild-type MEFs (p53^+/+), syncytium formation led to phosphorylation of p53S15, positive staining for Bax (Fig. 8 A), and apoptotic cell death (Fig. 8 B), whereas syncytia formed from p53^−/− MEFs (which obviously do not stain for p53S15P) failed to activate Bax (Fig. 8 A) and to undergo apoptosis (Fig. 8 B). In conclusion, p53 is rate-limiting for apoptosis triggered by syncytium formation.

HeLa CD4 cells cocultured with H9 lymphoid cells chronically infected with the HIV-1 strain IIIb can form syncytia, which undergo apoptosis. Syncytial apoptosis was inhibited by rapamycin, both at the mitochondrial and the nuclear levels (Fig. 9 A and B). To further substantiate the rapamycin-mediated inhibition of syncytial cell death, CEM T lymphoma cells engineered to contain a GFP gene under the control of the HIV-1 LTR promoter (normally inactive) were cocultured with HIV-1 IIIb-infected H9 cells, in the presence of absence of rapamycin. Upon formation of syncytia, HIV-1–encoded Tat transactivates LTR and induces GFP expression 33. In the presence of rapamycin (or Z-VAD.fmk, as a positive control) the frequency of viable syncytia was increased, leading to an enhanced GFP-dependent fluorescence (Fig. 9 C). When individual nuclei from such syncytia were subjected to cytofluorometric DNA content analysis, the frequency of hypoploid (apoptotic) nuclei was found to be strongly reduced by rapamycin (Fig. 9 D). In a further series of experiments, primary CD4⁺ lymphoblasts from healthy donors were infected by HIV-1 IIIb in vitro, a manipulation that induced syncytium formation and subsequent apoptosis. Addition of rapamycin 48 h after infection did not affect viral replication, as determined by measuring the production of p24 (Fig. 9 E). However, rapamycin did inhibit chromatin condensation and led to an increase in the number of nuclei per syncytium, similar to Z-VAD.fmk (Fig. 9 E). Thus, the cytoprotective effects of rapamycin on lymphoblasts infected with HIV-1 cannot be attributed to an inhibition of viral replication or syncytium formation. Rapamycin inhibited the phosphorylation of p53S15 in HIV-infected cells (Fig. 9 B). Taken together, these data confirm that the rapamycin-inhibited mTOR/FRAP→p53S15P pathway selectively controls apoptosis of syncytia in human primary CD4⁺ lymphocytes infected by HIV-1.

As shown here, HIV-1-Env–induced syncytium formation leads to apoptosis via a pathway that involves p53S15 phosphorylation by mTOR/FRAP. This corroborates the role of p53 as a major point of integration of proapoptotic signaling and damage sensing 21. The data presented here unravel that p53 is rate limiting for the apoptotic demise of syncytia. It is tempting to speculate that p53 could be generally required for the clearance of nonphysiological polyploid cells (syncytia or plasmodia). Indeed, HeLa cells transfected with antisense p53 spontaneously form giant multinucleate cells 41; antisense oligonucleotides suppressing p53 expression induce polyploidy in regenerating rat liver 55; and p53^−/− mice accumulate multinucleate cells in the testis 56 as well as in the prostate epithelium 57. However, further studies are required to distinguish the relative impact of p53 on the prevention of nuclear endoreplication and the illicit survival of polyploid cells in these model systems.

The role of mTOR/FRAP in apoptosis control is far less established than p53. The mTOR/FRAP analogues of S. cerevisae and Drosophila are known to play a major role in the control of cell size 24,25. External constraints on cell size and shape have a profound effect on HIV-1-Env–induced syncytial apoptosis 58. It thus may be possible that the abnormal size or the distorted volume/surface ratio of syncytia, sensed via yet unknown mechanisms, activate the mTOR/FRAP pathway and consequent cellular demise. Alternatively or in addition, signals triggered via the Env-CD4/CXCR4 interaction might stimulate the mTOR/FRAP pathway. In some cell types, the mTOR/FRAP antagonist rapamycin induces apoptosis 59 via a p53-independent pathway 60, indicating that mTOR/FRAP can also act as an endogenous cell death antagonist. Indeed, mTOR/FRAP mediates survival signals received through the receptors for IL-3, insulin, and insulin-like growth factor 61,62 and may participate in oncogenic transformation 26. mTOR/FRAP is constitutively active in mammalian cells, at least in the presence of serum and nutrients 24,52, suggesting that the cellular context rather than the expression level of mTOR/FRAP dictates whether its proapoptotic or antiapoptotic function prevails.

In the context of HIV-1-Env–induced syncytia, mTOR/FRAP-mediated p53S15 phosphorylation appears to contribute to apoptosis by activating Bax (and perhaps other proapoptotic proteins) and consecutive MMP. Intriguingly, a fraction of lymph node cells from HIV-1⁺ donors are p53S15P⁺ (this paper), express (p53-inducible, reference 63) tTgase 15, and exhibit a ΔΨ_m loss 64,65, suggesting that they die via a pathway similar to that delineated for Env-induced syncytial apoptosis. Pharmacological targeting of the mitochondrial or premitochondrial events involved in HIV-1–associated apoptosis may thus offer a novel opportunity to intervene on AIDS pathogenesis.

We thank Drs. Jie Chen (Illinois University), Thierry Soussi (Institut Curie), Bert Vogelstein (Howard Hughes Medical Institute), Klas Wiman (Karolinska Cancer Center) for plasmids, John C. Reed (The Burnham Institute, La Jolla, CA) for recombinant Bcl-2, Jacques Corbeil (University of California at San Diego), Tyler Jacks (MIT), and the National Institutes of Health AIDS reagents program (Bethesda, MD) for cell lines.

This work has been supported by a special grant from the Ligue Nationale contre le Cancer, Comité Val de Marne de la Ligue contre le Cancer, as well as grants from Agence Nationale de Recherches sur le Sida, Fondation pour la Recherche Médicale (to G. Kroemer), the European Commission (QLG1-CT-1999-00739) (to G. Kroemer and M. Piacentini), Ricerca Corrente e Finalizzata from the Italian Ministry of Health (to M. Piacentini), Fundació irsiCaixa, MCYT BFM2000-1382 (to J.A. Este), FIS 00/0893 (to J. Barretina), and the Picasso Program (to J.A. Este and G. Kroemer). K.F. Ferri receives a fellowship from the French Ministry of Science. J. Blanco is a researcher at the Fundació per a la Recerca Biomèdica Germans Trias i Pujol, FIS 98/3047.