|Home | About | Journals | Submit | Contact Us | Français|
The human immunodeficiency virus type 1 (HIV-1) viral protein R (Vpr) causes cell cycle arrest in G2. Vpr-expressing cells display the hallmarks of certain forms of DNA damage, specifically activation of the ataxia telangiectasia mutated and Rad3-related kinase, ATR. However, evidence that Vpr function is relevant in vivo or in the context of viral infection is still lacking. In the present study, we demonstrate that HIV-1 infection of primary, human CD4+ lymphocytes causes G2 arrest in a Vpr-dependent manner and that this response requires ATR, as shown by RNA interference. The event leading to ATR activation in CD4+ lymphocytes is the accumulation of replication protein A in nuclear foci, an indication that Vpr likely induces stalling of replication forks. Primary macrophages are refractory to ATR activation by Vpr, a finding that is consistent with the lack of detectable ATR, Rad17, and Chk1 protein expression in these nondividing cells. These observations begin to explain the remarkable resilience of macrophages to HIV-1-induced cytopathicity. To study the in vivo consequences of Vpr function, we isolated CD4+ lymphocytes from HIV-1-infected individuals and interrogated the cell cycle status of anti-p24Gag-immunoreactive cells. We report that infected cells in vivo display an aberrant cell cycle profile whereby a majority of cells have a 4N DNA content, consistent with the onset of G2 arrest.
Human immunodeficiency virus type 1 (HIV-1) infection results in the killing of CD4+ T lymphocytes, ultimately causing AIDS. Work from the past decade has demonstrated that T-cell killing is mediated, in part, by direct viral infection (25, 41, 57). Several mechanisms have been proposed to explain HIV-1-mediated cytopathicity, including the presence of unintegrated proviral DNA (49), plasma membrane disruptions due to viral egress (17), activation-induced cell death (58), and the proapoptotic function of HIV-1 gene products (22, 47). The HIV-1 accessory proteins, Nef, Vif, Vpr, and Vpu, which are dispensable for viral replication in vitro, are essential for viral replication and pathogenesis in vivo (reviewed in reference 16). It is thought that accessory proteins, collectively, manipulate host cell biology in order to promote viral replication, persistence, and immune escape.
The host cell proteins and pathways with which HIV-1 Vpr interacts have been explored extensively in vitro (reviewed in references 3, 34, and 51). However, how Vpr constitutes a determinant of viral pathogenesis in vivo is less well understood. Evidence supporting a cytopathic function for Vpr stems from variations in proviral sequence in HIV-1-infected individuals that display attenuated disease progression (long-term nonprogressors) (37, 44, 52, 56, 59). These studies, however, do not provide any specific details about the activity of Vpr in vivo.
In cell culture, Vpr mediates nuclear import of the preintegration complex, cell cycle arrest in G2, apoptosis, and transactivation of the viral promoter (reviewed in references 2, 3, and 34). Our previous studies have focused on elucidating the mechanism by which Vpr induces G2 arrest. We and others have reported that Vpr induces G2 arrest by activating the ATR kinase (33, 46, 60). ATR is a DNA damage sensor kinase that initiates the G2 checkpoint under conditions of genotoxic stress, preventing entry into mitosis (1). Genotoxic stresses that can activate ATR include stalled DNA replication forks and DNA double-strand breaks (DSBs), both of which result in abnormally long and persistent stretches of single-stranded DNA (1). It is thought that ATR is recruited to these sites of stress via protein-protein interactions with the single-stranded DNA (ssDNA) binding protein complex, replication protein A (RPA), and the ATR-interacting protein (ATRIP) (11, 62). We and others have shown that the primary DSB sensor kinase, the ataxia telangiectasia mutated (ATM) protein, is dispensable for Vpr-induced G2 arrest (6, 60). Thus, we hypothesized that Vpr activates ATR in a DSB-independent manner via replication stress.
In addition to T cells, HIV-1 infects macrophages. Tissue macrophages infected with HIV-1 are resistant to the viral cytopathic effects and are thought to persist throughout the course of infection (19, 21, 31). The macrophage compartment is viewed as one of the reservoirs for HIV-1 infection in the presence of antiretroviral treatment, and in this context, the half-life of infected macrophages was calculated to be about 2 weeks (whereas that of CD4+ T lymphocytes is in the range of 1 to 2 days) (reviewed in reference 42). In vitro infected macrophages can harbor and produce virus for more than a month. Thus, macrophages are an important reservoir for HIV-1, capable of disseminating the virus to various tissues, including the brain (20, 40). Based on the above findings and also on the knowledge that differentiated macrophages are largely nondividing, we hypothesized that Vpr is unable to activate ATR in these cells.
The present study addresses the following questions: (i) does Vpr induce G2 arrest in the context of HIV-1 infection?; (ii) are primary cells sensitive to the cytostatic effect of Vpr, including peripheral blood T cells and monocyte-derived macrophages (MDM)?; (iii) is ATR the mediator of HIV-1-induced G2 arrest in primary cells?; (iv) can we detect the formation of RPA foci, the hallmark of ATR activation, in primary cells?; and (v) are infected CD4+ T lymphocytes arrested in G2 in vivo?
Lentiviral vectors, pHR-Vpr-IRES-GFP (herein referred to as pHR-Vpr) and pHR-GFP, were produced and titered as previously described (46, 60). Cells were infected by spin infection as follows. Cells (106) were diluted in viral stocks with 10 μg/ml polybrene and centrifuged at 1,700 × g for 2 h at 25°C, and cells were then washed and resuspended in normal growth medium.
HIV-1 molecular clones, HIV-1NL4-3, HIV-1NL4-3VprX, HIV-1AD8, and HIV-1AD8VprX, were transfected into 2 × 107 HEK293FT cells by calcium phosphate transfection. At 48 h after transfection, transfected HEK293FT cells were cocultured with 107 MT-2 cells (HIV-1NL4-3 and HIV-1NL4-3VprX) or 107 CEM-CCR5 cells (CEM.NKR-CCR5, catalog no. 4376; AIDS Research and Reference Reagent Program, NIAID, NIH) for 5 h. MT-2 and CEM-CCR5 cells were then cultured alone until approximately 75% of cell clumps showed syncytia. Virus-containing supernatants were then cleared of cells and debris by centrifugation at 2,000 rpm for 10 min. Viral stocks were then frozen at −80°C. Spin infections were performed as described above.
Cells were lysed in lysis buffer (0.1% sodium dodecyl sulfate, 0.1% NP-40, and 0.5% sodium deoxycholate in phosphate-buffered saline [PBS]). Protein concentration was determined by a bicinchoninic acid assay (Pierce, Rockford, IL). Equal amounts of protein were boiled and loaded onto Criterion XT precast polyacrylamide gels (Bio-Rad, Hercules, CA) and separated by electrophoresis. Proteins were transferred to HyBond polyvinylidene difluoride membranes (Amersham, Piscataway, NJ) by semidry transfer (Bio-Rad) and blocked for 1 h at 25°C with 5% nonfat milk in 0.1% Tween 20 in PBS. Membranes were incubated with rabbit anti-ATR (Genetex, San Antonio, TX), rabbit anti-ATM (Bethyl, Montgomery, TX), rabbit anti-phospho-Chk1 (Ser345; Cell Signaling, Eugene, OR), rabbit anti-phospho-RPA32 (Ser4/Ser8; Bethyl), rabbit anti-actin (Santa Cruz, Santa Cruz, CA), rabbit anti-Rad17 (Santa Cruz), or rabbit anti-ATR (gift from Paul Nghiem, University of Washington) primary antibody at a dilution of 1:1,000 in 0.1% Tween 20 in PBS overnight at 4°C. Secondary antibody goat anti-rabbit immunoglobulin G (IgG)-horseradish peroxidase conjugate (Santa Cruz) was used at a 1:7,000 dilution for 1 h at 25°C. Total Chk1 protein was visualized using a monoclonal antibody (1:250 dilution; Santa Cruz) followed by a horseradish peroxidase-linked anti-mouse secondary antibody (1:1,000 dilution; Amersham). Blots were visualized using ECL Plus (Amersham).
Cells were washed with fluorescence-activated cell sorter (FACS) buffer (2% fetal bovine serum and 0.02% sodium azide in PBS) and fixed with 0.1% paraformaldehyde in PBS buffer for 15 min on ice. Cells were then permeabilized with 0.1% Tween 20 in PBS. Primary human anti-p24Gag monoclonal antibody (monoclonal antibody to HIV-1 p24, clone 71-31; obtained from Susan Zolla-Pazner, AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH) was used at 1:100 in FACS buffer for 30 min on ice. Cells were then washed with FACS buffer and incubated with secondary goat anti-human IgG antibody conjugated to fluorescein isothiocyanate (FITC) at a 1:100 dilution for 30 min on ice. Cells were then washed with FACS buffer, resuspended in DNA stain buffer (10 μg of propidium iodide/ml and 11.25 kU of RNase A/ml in FACS buffer), and analyzed for FITC and propidium iodide content by flow cytometry (FACScan; Becton Dickinson, Franklin Lakes, NJ). Cell cycle distribution was modeled using ModFit software (Verity Software House, Topsham, ME).
Cells were stained and imaged as described in reference 60. Primary antibodies against HIV-1 Gag p24 (NIH) and RPA32 (Abcam, Cambridge, MA) were used at 1:100 dilutions. Goat anti-human IgG Alexa 488 (Invitrogen) and goat anti-mouse IgG Alexa 568 (Invitrogen) secondary antibodies were used at 1:1,000 dilutions. Cells were adhered to glass coverslips precoated with Cel-Tak cell adhesive (Becton Dickinson).
Peripheral blood mononuclear cells (PBMC) were obtained from Leukopaks from unidentified healthy donors (American Red Cross), and CD4+ lymphocytes were purified using anti-CD4 magnetic beads (Dynal/Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. CD4+ T cells were activated by culture in RPMI-10% fetal bovine serum-1 mM l-glutamine with three anti-CD3/anti-CD28 beads per cell (Invitrogen) for 2 days, with the medium changed daily. After 2 days, recombinant interleukin-2 was added to the culture medium at a concentration of 100 units/ml. For patient analysis, PBMC were processed within 2 h of venipuncture and isolated as described above. PBMC were then fixed, permeabilized, and immunostained directly without activation or culturing.
Monocyte-derived macrophages were prepared as described by Elstad et al. (15). Briefly, mononuclear cells were isolated from the blood of healthy donors by dextran sedimentation, hypotonic lysis of red blood cells, and density gradient centrifugation. The monocyte layer was extracted from the tube and centrifuged at 300 × g for 5 min at room temperature. The supernatant was discarded, and the cell pellet was washed four times with medium 199 without serum (Cambrex, East Rutherford, NJ). Washed monocytes were plated in 1 ml of medium (medium 199) without serum in six-well plates (Becton Dickinson) at 8 × 106 cells/ml. Monocytes were incubated at 37°C and allowed to adhere to the plastic for 1 to 2 h. The medium was aspirated and replaced with 1.5 ml medium 199 supplemented with 10% human serum, 1% penicillin-streptomycin-l-glutamine (Invitrogen), and 2.5 μg/ml amphotericin B (Invitrogen). One to 2 days after plating, the monocytes were washed four times with medium 199 without serum, with the culture medium replaced after the fourth wash. Cell medium was replaced every 3 to 4 days while the monocytes differentiated to macrophages (7 to 10 days).
Cells were incubated with 10 mM hydroxyurea (HU) or 1 μM mitomycin C (MMC) for 2 h or 1 μM aphidicolin for 12 h prior to immunostaining or lysis.
Activated primary CD4+ T cells (3 × 106) were suspended in CD4+ T-cell nucleofection buffer (Amaxa, Gaithersburg, MD) with 1.5 μg SmartPool small interfering RNA (siRNA) (Dharmacon, Lafayette, CO) and added to Amaxa nucleofection cuvettes. Cells were electroporated using program T-23 and resuspended in CD4+ culture medium. The transfection procedure was repeated on the same cells 24 h later.
Subjects recently infected with HIV were recruited through the UCSF Options Project. These subjects had been infected for less than 12 months. Recent infection status was determined based on having a negative or indeterminate HIV-1 antibody test with an HIV-1 RNA level of >2,000 copies/ml at enrollment or, if HIV-1 antibody positive at presentation, having a documented negative HIV-1 antibody test within 12 months, or having a history compatible with recent HIV-1 infection, confirmed with an enzyme immunoassay antibody test using a cutoff of 0.75 with an Organon Technica Vironostica-based assay (27). All participants were at least 18 years of age (or at least 16 years of age for emancipated minors). Subjects who had received antiretroviral therapy for HIV prior to this study were excluded. Patient samples were analyzed in a blinded fashion. The study protocol was approved by the Committee on Human Research at the University of California, San Francisco.
A patient under antiretroviral treatment at the University of Utah (patient SL1) was also included in the study. This sample was provided to us without identifiers. PBMC from this patient were cultured in vitro under standard conditions until virus replication was detectable, at which point cells were fixed and stained for cell cycle profiles and RPA nuclear foci. The study protocol was approved by the University of Utah Internal Review Board.
PBMC were fixed and permeabilized in 1% paraformaldehyde, 1 mg/ml human immunoglobulin G (Gemini Bio-Products, Woodland, CA), and 0.1% Tween 20 in FACS buffer for at least 1 h at 4°C, washed, and then incubated with unconjugated mouse immunoglobulin G (Coulter) at a 1:10 dilution in phosphate-buffered saline solution containing 10% fetal calf serum for at least 1 h at room temperature. Cells (106 per 50 μl of FACS buffer) were immunostained with KC57 (Lot 13, 1:50; Coulter), an FITC-labeled anti-p24 monoclonal antibody, and a phycoerythrin-labeled anti-CD4 antibody (1:50; Coulter, Fullerton, CA). The cellular DNA content was assessed by staining with 0.01 mM To-Pro-3 iodide (Molecular Probes, Eugene, OR) in the presence of 1 mg/ml RNase A, followed by analysis with a FACScan flow cytometer set to acquire linear fluorescence in the FL4 channel. DNA profiles were analyzed with FlowJo software (Tree Star, San Carlos, CA). An FITC-conjugated immunoglobulin G control of matched isotype (1:50; Caltag, Burlingame, CA) was used to determine the degree of nonspecific antibody binding in each sample. When these staining parameters were applied to PBMC freshly isolated from 10 HIV-seronegative donors, the level of anti-p24Gag lymphocytes was no more than 3% higher than that in isotype controls. Cells were sorted with a FACS Vantage SE.
Sorted lymphocytes were collected in DNase- and RNase-free siliconized Eppendorf tubes (Ambion, Austin, TX) and incubated with proteinase K at 60°C for 30 min and then at 70°C for 10 min. For volumes greater than 500 μl, cells were initially filtered onto a culture plate insert (Millicell-CM; Millipore, Billerica, MA) and lysed as described above. Total DNA was extracted with DNAzol BD reagent (Invitrogen, Carlsbad, CA) as recommended by the manufacturer, with 5 μg of linear polyacrylamide (Ambion, Austin, TX) as the carrier, and precipitated with ethanol.
Cell numbers for each sample were quantified by using primers and probe for the 18S rRNA gene (Applied Biosystems, Foster City, CA) with the ABI Prism 7700 sequence detection system, using the manufacturer's conditions except that the primer and probe concentrations were 100 nM each in the presence of template nucleic acid (DNA and linear polyacrylamide). A DNA standard curve was prepared from human DNA of known concentration (Clontech, Palo Alto, CA). HIV-1 viral DNA was quantified by using primers and probe complementary to the HIV-1 NL4-3 gag gene as follows: forward, 5′ CCA TCA ATG AGG AAG CTG CAGA 3′; reverse, 5′ TGC TAT GTC ACT TCC CCT TGG 3′; probe, 5′ VIC CAG GGC CTA TTG CAC MGB 3′ (MGB, minor groove binder; ABI). A standard curve of the measured amplicon was generated with an HIV-1 control reagent (GeneAmplimer; ABI) diluted in uninfected human placental DNA (10 ng/μl). Cycling conditions were as specified by the manufacturer except that annealing/extension was performed for 30 s at 62°C.
To begin to examine the cytostatic function of HIV-1 Vpr in vivo, we first examined the effects of HIV-1 infection on the cell cycle properties of primary CD4+ T cells obtained from peripheral blood, and the potential role of ATR was tested by RNA interference. Given the availability of methods to detect intracellular p24 core antigen by flow cytometry (39), it was possible to perform dual staining that allowed for simultaneous measurement of DNA content and p24 antigen. This dual staining allowed us to specifically study the DNA content of infected cells, irrespective of their frequency. We then used RNA interference to reduce the levels of ATR and exposed cells to HIV-1 and evaluated the effect of ATR knockdown on cell cycle profiles.
Transfections of siRNAs were performed using an Amaxa nucleofector, and we utilized ATR-specific siRNA, ATM-specific siRNA, or nonspecific siRNA. Following two serial transfections, 24 h apart, cells were infected with either an X4 (HIV-1NL4-3 (Fig. (Fig.1A)1A) or an R5 isolate (HIV-1AD8) (data not shown). At 48 h after infection, cells were stained for intracellular p24Gag and DNA content. As shown in Fig. Fig.1A,1A, HIV-1NL4-3-infected cells displayed characteristic levels of G2 arrest, whether they had been mock transfected or transfected with nonspecific or ATM-specific siRNA [(G2 + M)/G1 ratio of 3.6, 3.8, or 4.1, respectively]. In contrast, cells pretreated with ATR-specific siRNA displayed a significant relief of HIV-1NL4-3-induced G2 arrest [(G2 + M)/G1 = 1.7]. Densitometry scanning of Western blots showed that ATR and ATM were knocked down by 89% and 92%, respectively, compared to bands corresponding to mock-transfected cells (Fig. (Fig.1B).1B). These data support a role for ATR in HIV-1 Vpr-induced G2 arrest in primary CD4+ T lymphocytes.
To confirm the role of Vpr as the main viral determinant of cell cycle arrest, we also infected cells with a Vpr-deleted HIV-1 clone (HIV-1NL4-3VprX) (Fig. (Fig.1C).1C). Cells infected with the Vpr-deficient virus, HIV-1NL4-3VprX, displayed a modest but reproducible increase in cells with G2 DNA content [(G2 + M)/G1 = 0.83] (Fig. (Fig.1).1). In previous studies, we have consistently observed a small but significant increase in the G2 population when infecting cells with Vpr-deficient viruses (5). When performing experiments with lentiviral vectors, however, we do not observe any detectable increase in the G2 population after infection with green fluorescent protein (GFP)-expressing lentivirus vectors or vectors encoding a truncated Vpr reading frame (5). Therefore, it is likely that expression of other HIV-1 genes may result in moderate activation of the G2 checkpoint or a slight delay in the transition from G2 to mitosis. In support of the previous notion, two reports have suggested that Env and Vif contribute to the onset of G2 arrest by HIV-1 (32, 48).
Cells infected with HIV-1AD8 exhibited large amounts of cells in G2 [(G2 + M)/G1 = 2.9], whereas cells infected with HIV-1AD8VprX exhibited a small degree of G2 arrest [(G2 + M)/G1 = 1.5] (data not shown). Therefore, the cell cycle effects of HIV-1 on primary CD4+ T cells were essentially identical when CXCR4 or CCR5-tropic viruses were used and were largely, but not entirely, dependent on the presence of Vpr.
Macrophages also constitute an important target for HIV-1 in vivo (19, 21, 31). Since macrophages are fully differentiated and postmitotic, we expected that HIV-1 infection would not result in any change in their DNA content. To investigate this prediction, we obtained monocytes from healthy donors and cultured them under conditions that led to differentiation into macrophages in vitro (MDM). MDM were then infected with HIV-1AD8 or HIV-1AD8VprX, and dual staining for DNA content/p24 was performed as described above. As expected, HIV-1AD8- and HIV-1AD8VprX-infected macrophages had cell cycle profiles indistinguishable from those of uninfected cells (data not shown). The nonproliferating status of MDM was verified by 5-bromo-2′-deoxyuridine labeling (data not shown).
The failure of macrophages to undergo G2 arrest was expected because of their postmitotic status. We previously demonstrated that in tumor cell lines, the signaling molecule that is the cellular trigger for Vpr-induced G2 arrest is the ATR protein. ATR is a serine-threonine protein kinase that has approximately 15 known targets in the human proteome (1, 50). The ATR target that activates the G2 checkpoint in the presence of replication stress or Vpr expression is Chk1 (36, 46). The lack of activation of the G2 checkpoint in macrophages does not exclude the possibility that Vpr may still activate ATR in these cells, leading to phosphorylation of various ATR targets. To explore this possibility, we infected macrophages with a lentivirus vector that expresses Vpr and GFP (pHR-Vpr) or a control vector that expresses only GFP (pHR-GFP). Unlike replication-competent HIV-1, lentivirus vectors can be grown to high titers, allowing for efficient infections of most cell types, including MDM. We then asked whether H2AX, a known target of ATR and ATM, became phosphorylated at Ser139. As a positive control for H2AX phosphorylation at Ser139 (γ-H2AX), we exposed macrophages to MMC or ionizing radiation (IR), both inducers of double-strand breaks, which potently activate ATM and should result in ATM-dependent H2AX phosphorylation and focus formation. While MMC and IR induced nuclear focal γ-H2AX staining patterns, transduction with pHR-Vpr or pHR-GFP or mock transduction failed to cause γ-H2AX foci (Fig. (Fig.2A2A).
The apparent inability of Vpr to activate ATR in macrophages is highly unusual, as ATR is a widely expressed protein. Therefore, we wished to examine another ATR target, Chk1, for phosphorylation status. Unlike H2AX, Chk1 is only a target for ATR and not for ATM. We exposed MDM to IR (a known activator of ATM and ATR) or N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) (a known activator of ATR) and tested for phosphorylation of Chk1 at Ser345, an ATR-specific target residue. Although treatment with 10 nM MNNG induced Chk1 phosphorylation in control dividing HeLa cells (Fig. (Fig.2B,2B, lane 4), it failed to induce detectable Chk1 phosphorylation in MDM (Fig. (Fig.2B,2B, lane 3). Treatment with IR induced Chk1 phosphorylation in HeLa cells (Fig. (Fig.2B,2B, lane 2) because IR typically activates both ATM and ATR. In contrast, IR failed to induce phosphorylation of Chk1 in MDM (Fig. (Fig.2B,2B, lane 1). Since Chk1 is a target for ATR but not for ATM, these results further support that ATR cannot be activated in MDM.
Two possible events could account for the above results. First, perhaps ATR is refractory to activation in MDM. Second, it is also possible that ATR is not expressed in MDM. Thus, we tested for the presence of ATR protein by Western blot analysis (Fig. (Fig.2C).2C). We compared ATR protein levels in lysates from freshly isolated PBMC, activated/proliferating CD4+ T cells, and MDM. We found that fresh (unactivated) PBMC and macrophages did not express detectable levels of ATR protein (Fig. (Fig.2D).2D). In contrast, activated CD4+ T cells (Fig. (Fig.2C)2C) expressed ATR. We detected a similar expression profile for Rad17 (Fig. (Fig.2C),2C), which is required for ATR-dependent activation of the G2 checkpoint (61).
Using a phosphorylation-independent antibody to Chk1, we also tested whether this protein can be detected in MDM (Fig. (Fig.2D).2D). Chk1 is not present in MDM at detectable levels. Taken together, these data suggest that quiescent, terminally differentiated, or otherwise nondividing cells do not express ATR, Chk1, or Rad17, proteins essential to the cellular response to DNA replication stress and certain forms of DNA damage.
In previous work, we have established that three (Chk1, H2AX, and BRCA1) of the approximately 15 known targets for the ATR protein kinase are phosphorylated in the presence of Vpr expression (4, 46, 60). Little is known, however, regarding the steps leading to the initial ATR activation by Vpr. A well-characterized stimulus leading to ATR activation is the presence of long stretches of RPA-coated ssDNA (62). RPA-coated ssDNA is typically present at stalled replication forks and processed DSBs. When cells undergo DNA replication stress, RPA forms intense nuclear foci, corresponding to stalled replication forks. These RPA-rich regions recruit ATR via direct interaction of RPA with the ATR-interacting protein, ATRIP, which is strongly bound to ATR (62).
To directly examine whether HIV-1 infection leads to RPA focus formation, normal donor-derived CD4+ T lymphocytes were infected in vitro with HIV-1NL4-3 or HIV-1NL4-3VprX. Cells were infected, mock infected, treated with 10 mM HU for 1 h to induce replication stress, or treated for 12 h with 2 μg/ml aphidicolin, an inhibitor of DNA replication polymerases α and δ. Infection levels were initially low (3 to 7%) but were allowed to spread until approximately 25% of the cells stained positive with intracellular anti-p24Gag antibody. When 25% of the cells were infected, samples were fixed and stained for confocal microscopy, using RPA32- and HIV-1 p24Gag-specific antibodies followed by staining with Alexa 568 and Alexa 488 fluorochrome-conjugated secondary antibodies, respectively (Fig. (Fig.3A).3A). Cells containing RPA foci and anti-p24Gag-immunoreactive cells were visually counted (Fig. (Fig.3B).3B). We found that HIV-1NL4-3-infected p24Gag+ cells exhibited RPA foci (64%) reminiscent of those observed following HU (47%) or aphidicolin (52%) treatment, while HIV-1NL4-3VprX-infected p24Gag+ cells displayed punctate RPA staining patterns less frequently (20%), as did uninfected cells (7.6%). Similar results were observed for cells infected with HIV-1AD8 or its Vpr-deficient counterpart, HIV-1AD8VprX (53% or 11% infected cells with RPA-rich foci, respectively, versus 7.6% in mock-infected cells) (Fig. (Fig.3B3B).
We next tested whether Vpr, expressed by a lentiviral vector (pHR-VPR), in the absence of any other HIV-1 genes was sufficient to induce RPA focus formation (Fig. (Fig.3C).3C). Our observations demonstrate that Vpr is not only necessary but also sufficient for inducing RPA focus formation, ATR activation, and ATR target phosphorylation.
Accumulation of RPA at stalled replication forks leads to subsequent recruitment and activation of ATR. RPA is a heterotrimer consisting of subunits of 70, 32, and 14 kDa (RPA70, RPA32, and RPA14, respectively) (7). Activation of ATR at local sites of DNA damage, in turn, promotes phosphorylation of RPA32 by ATR. This phosphorylation serves to amplify the DNA damage signal, enhancing further recruitment and activation of ATR (1, 54). Once localized to these sites of stalled replication, the ATR kinase activity is triggered, leading to downstream activation of DNA repair, cell cycle arrest, and/or proapoptotic signaling.
To examine the phosphorylation state of RPA32 in the context of HIV-1 infection, we performed Western blot analysis on protein extracts from the above infections. Infections with HIV-1NL4-3 or HIV-1NL4-3VprX were passaged until 25% of the cells were positive for intracellular staining with anti-p24Gag antibody by flow cytometry. Compared to mock-infected cells (Fig. (Fig.3D,3D, lane 1), HU- and aphidicolin-treated cells (Fig. (Fig.3D,3D, lanes 2 and 3, respectively) as well as HIV-1-infected cells (Fig. (Fig.3D,3D, lanes 4 and 5) exhibited elevated levels of RPA32 phosphorylation on Ser4/Ser8. Additionally, infection with HIV-1NL4-3VprX induced less RPA32 phosphorylation than did HIV-1NL4-3 (Fig. (Fig.3D,3D, compare lanes 4 and 5). Since HIV-1NL4-3VprX did induce a minor but reproducible degree of G2 arrest, the moderate level of RPA32 phosphorylation observed in HIV-1NL4-3VprX-infected cells was not unexpected. Additionally, since only 25% of the population was infected with either virus prior to lysis, the lower level of RPA phosphorylation in the infected cells relative to HU or aphidicolin treatment is not surprising, since these drugs presumably affect most or all cycling cells in the culture.
Since the above results were observed using HIV-1 molecular clones, which do not reflect the viral diversity observed during an in vivo infection, we wished to confirm these findings in a sample from an HIV-1-infected individual. We purified CD4+ lymphocytes from peripheral blood of an HIV-1-infected donor (patient SL1) undergoing antiretroviral therapy. Since the plasma viral load was low (240 copies/ml) at the time of PBMC extraction, we stimulated the isolated CD4+ lymphocytes with anti-CD3/anti-CD28 beads and interleukin-2 to allow for T-cell activation and viral replication. Intracellular p24Gag levels of the culture were monitored daily following stimulation. When the level of infection reached approximately 10%, cells were immunostained with anti-p24Gag and anti-RPA32 antibodies. The p24Gag+ population exhibited a significant increase in G2-arrested cells (Fig. (Fig.4A),4A), and RPA foci were readily detectable in the nuclei of these infected cells (Fig. (Fig.4B).4B). Indeed, a significant percentage (60%) of infected cells exhibited RPA-rich nuclear foci (Fig. (Fig.4B),4B), while only 13% of the p24Gag− cells were positive for RPA foci.
Taken together, the above results indicate that Vpr causes accumulation of RPA-rich nuclear foci in the context of HIV-1 infection.
In view of the above results, we hypothesized that cycling HIV-1-infected CD4+ T cells, isolated from patients and without subsequent in vitro culture, would also be found in G2. In order to directly test this hypothesis, we applied the anti-p24Gag staining to PBMC isolated from patients identified as newly infected within a 12-month time period. These patients would be predicted to have a relatively high frequency of infected circulating target cells, concomitant with a high viral load prior to the initiation of antiretroviral therapy.
We first asked whether anti-p24Gag-reactive cells could be detected in the blood of recent seroconverters. We isolated blood by venipuncture from three patients (Options Project samples 712, 671, and 719). PBMC from these samples were stained with anti-CD4 antibodies and gated based on forward scatter, side scatter, and CD4 staining. (Fig. (Fig.55 and Table Table1).1). We found a significant number of p24Gag+ cells from the peripheral blood of these patients, ranging from 0.53% to 8.7% (Fig. (Fig.55 and Table Table1),1), both in the CD4-positive and in the CD4-negative compartments. To verify that p24Gag+ cells were truly harboring HIV-1 sequences, CD4-positive and -negative cells were then sorted into p24Gag+ and p24Gag− gates and then analyzed for viral DNA content by using quantitative PCR (Table (Table1).1). These data show a remarkable consistency from patient to patient, where two to seven copies/cell of HIV-1 Gag DNA could be found in the CD4− or CD4+ cell populations that were positive for anti-p24Gag staining. This represents an enrichment of HIV-1 DNA in the range of 10-fold to nearly 1,000-fold based on p24Gag protein staining. Thus, these cells appeared to be productively infected by two different criteria.
To ascertain the cell cycle profiles of p24Gag+ cells in vivo, we then analyzed CD4+ lymphocytes from seven additional HIV-1-infected patients (Table (Table2).2). Cells were initially separated based on forward and side scatter properties and CD4 expression, as shown in Fig. Fig.5.5. Electronic gates were then established that would separate CD4+ lymphocytes into cycling and noncycling populations, and the cycling population was then interrogated for p24Gag expression (Fig. (Fig.6,6, top two panels). These cells were sorted and then stained for DNA content (Fig. (Fig.6).6). In each subject, the p24Gag+ lymphocytes had a higher (G2 + M)/G1 ratio than the p24Gag− lymphocytes (results for this experiment are summarized in Table Table2).2). Gag+ cells were also detected in the noncycling (resting) population (data not shown). This eliminated the possibility that cells in the G2/M phase were more easily immunostained for p24Gag than those in G0/G1, thereby falsely increasing the appearance of G2 arrest. Together, these findings demonstrate that HIV-1 induces G2 cell cycle arrest in productively infected lymphocytes in vivo.
Cell killing via direct viral infection is one mechanism underlying the CD4+ T-cell depletion that occurs during HIV-1 infection. The mechanisms by which the virus-encoded proteins can cause cytopathicity have been described in great detail using in vitro models. However, few studies confirm these virus-dependent phenotypes in vivo or even in primary CD4+ T cells infected with full-length HIV-1 in vitro. It has previously been shown that Vpr expression is necessary and sufficient to activate the G2 checkpoint when Vpr is expressed in transformed cell lines and activated CD4+ T lymphocytes (23, 29, 43, 45). Such arrested cells ultimately die as a result of apoptosis (4, 53). In the present study, we demonstrate that anti-p24Gag-immunoreactive, activated CD4+ T cells from HIV-1-infected individuals are arrested in G2. We also report Vpr-dependent G2 arrest in primary CD4+ T cells infected in vitro with HIV-1 molecular clones. Furthermore, in these primary CD4+ T cells, G2 arrest was relieved by RNA interference-mediated knockdown of ATR.
We also extended our studies to another physiologically relevant target cell for HIV-1 infection, the macrophage. As expected, Vpr did not alter the DNA content of primary monocyte-derived macrophages, due to their nondividing status. To our surprise, however, HIV-1 Vpr failed to induce phosphorylation of the known ATR targets, H2AX and Chk1. For H2AX, we demonstrated that the defect in MDM is at the level of ATR signaling and not H2AX itself, since an ATM-activating stimulus results in H2AX phosphorylation and focus formation. We found that the failure to activate ATR in MDM was due to the lack of protein expression of at least three essential proteins in the signaling axis: ATR, Chk1, and Rad17. Based on these findings and on our previous report that Vpr-induced apoptosis is ATR dependent (4), we hypothesize that macrophages are refractory to Vpr-induced apoptosis. If confirmed, the predicted resistance to Vpr-induced apoptosis could be one contributing factor to the ability of macrophages and, perhaps, other postmitotic cells to serve as viral reservoirs. Sensitizing postmitotic cells to the cytopathic properties of Vpr by reactivating the ATR pathway could provide a means to eliminate these reservoirs.
The apparent absence of ATR in quiescent lymphocytes was previously reported by Jones et al. (28), and we herein report, for the first time, that ATR is also absent in MDM. The absence of ATR in the previous two cell types is surprising in light of previous experiments that demonstrate an absolute requirement for ATR during mouse development (8, 12). In addition, experiments with a conditionally deleted form of the ATR gene indicated that it is essential for the viability of somatic cells (11). Therefore, the regulation of ATR gene expression and its relationship with the dividing status of cells deserve further investigation.
Virion-bound Vpr is required for efficient infection of nondividing cells, including monocyte-derived macrophages (9, 14, 24, 26). Since induction of G2 arrest and infection of nondividing cells are independent separable functions of Vpr (13, 18, 55), our results do not negate an immediate-early role of Vpr in facilitating infection of nondividing cells.
ATR is thought to be a sensor of DNA replication stress through binding to RPA-rich ssDNA regions in the genome (10). To test whether Vpr activates ATR in a canonical manner, we assayed for the presence of RPA-rich foci and RPA phosphorylation in the context of HIV-1 infection. Our results conclusively demonstrate that HIV-1 infection of primary lymphocytes induces DNA replication stress and that removal of vpr from the virus genome relieves this effect.
To ascertain whether the ability of Vpr to induce G2 arrest can be observed in vivo, we performed experiments with human samples. We first observed that in vitro culture of lymphocytes from an infected patient would result in virus amplification and then induction of G2 arrest and RPA foci. This experiment indicates that a primary virus, which is not molecularly or biologically cloned and is, therefore, presumably a quasispecies, maintains the cell cycle-perturbing properties previously ascribed to Vpr in overexpression systems.
In a second experiment with in vivo samples, freshly isolated CD4+ lymphocytes from recent seroconverters were interrogated for DNA content in the absence of any in vitro treatment, other than cell sorting. Our results show, for the first time, that p24Gag+ lymphocytes, but not p24Gag− ones from the same patient, can be found arrested in G2. We observed this for all patients from whom sufficient p24Gag+ cells could be recovered.
It is somewhat surprising that the CD4-negative compartments from the in vivo samples showed levels of infection similar to those of the CD4-positive compartment (Table (Table1).1). This infected CD4− population likely represents HIV-1-infected cells that have downregulated CD4 surface expression due to expression of Nef and other viral gene products (reviewed in reference 35). There is strong precedent for the existence of significant levels of infected cells in the CD4-negative compartment. In a study by Marodon et al. (38), HIV-1 DNA copies were analyzed by quantitative PCR in sorted CD4+ CD8−, CD4− CD8+, and CD4− CD8− lymphocytes from HIV-infected individuals. In this study, high levels of viral DNA were found in 16 out of 19 patients analyzed. When the surface phenotypes of these cells were analyzed, they were shown to contain typical markers of normal lymphocytes, such as rearranged T-cell receptors, and were always negative for CD8. These results suggests that infected cells were originally CD4+ CD8− and had efficiently downregulated CD4 following infection.
The levels of detected viral DNA copies (two to seven HIV-1 Gag DNA copies per p24Gag+ cell) we report (Table (Table1)1) are higher than we originally anticipated. Since the frequency of HIV-1-infected cells in peripheral blood is typically low, we expected viral DNA copy numbers in the vicinity of 1 per cell. A previous study, however, found that HIV-1-infected human splenocytes harbored integrated viral DNA copies that ranged from one to eight copies per infected splenocyte, with a mean around three, as evidenced by fluorescence in situ hybridization (30). Jung et al. have in fact argued that copy numbers higher than 1 would help explain the high frequencies of in vivo recombination (30).
In separate studies, we have shown that the ATR signaling cascade is required for Vpr-induced apoptosis (4). Thus, we hypothesize that the attenuated cytopathicity of HIV-1 in macrophages may be explained, in part, by the failure of Vpr to activate ATR in macrophages. Several other questions, which compel further investigation, are raised by our findings. What is the precise molecular mechanism by which Vpr causes DNA replication stress? How are the expressions of ATR, Rad17, and Chk1 extinguished in postmitotic noncycling cells and stimulated in cycling cells? Will reconstitution of the ATR pathway in macrophages or other nondividing cells sensitize them to HIV-1-induced cytopathicity?
We thank Susan Zolla-Pazner for the monoclonal antibody to HIV-1 p24, clone 71-31, and Alexandra Trkola for the CEM.NKR-CCR5 cells, which were obtained through the AIDS Research and Reference Reagent Program, NIAID, NIH. We thank Eric Daar (Cedars-Sinai Medical Center, Los Angeles, CA) for useful discussions.
This work was supported by NIH R01 AI45234, NIH R01 AI49057, and UCSF-GIVI Center for AIDS Research grant NIH P30 MH59037. M.P.S. was supported by NIH/NIAID grant KO8 AI01866. E.S.Z. was supported by NIH Genetics Training Grant T32 GM07464.
Published ahead of print on 6 September 2006.