Conceived and designed the experiments: J. Ward, J. DeHart, E. Zimmerman, D. Mavilio, V. Planelles, E. Barker. Performed the experiments: J. Ward, Z. Davis, A. Bosque, E. Barker. Analyzed the data: J. Ward, Z. Davis, V. Planelles, E. Barker. Contributed reagents/materials/analysis tools: J. DeHart, E. Brunetta, D. Mavilio. Wrote the paper: J. Ward, Z. Davis, V. Planelles, E. Barker.
The authors have declared that no competing interests exist.
Natural killer (NK) cells are stimulated by ligands on virus-infected cells. We have recently demonstrated that NK cells respond to human immunodeficiency virus type-1 (HIV-1)-infected autologous T-cells, in part, through the recognition of ligands for the NK cell activating receptor NKG2D on the surface of the infected cells. Uninfected primary CD4pos T-cell blasts express little, if any, NKG2D ligands. In the present study we determined the mechanism through which ligands for NKG2D are induced on HIV-1-infected cells. Our studies reveal that expression of
Natural killer (NK) cells are part of the innate immune response against virus infection and cancer. Recently we demonstrated that ligands for the NK cell activation receptor, NKG2D, trigger NK cell-mediated response to infected cells. These ligands are expressed on HIV-1-infected cells and not on uninfected cells. Despite the observation that NKG2D ligands are expressed on infected cells, it is unclear how HIV-1 induces their expression. In the present study, we demonstrate that HIV induces the ligands of the NKG2D receptor through the viral gene product Vpr. Vpr triggers a DNA damage response in infected cells, which in turn, increases virus production. We also demonstrate that by blocking the activity of ATR, a major component in the DNA damage response, we were able to prevent NKG2D ligand expression. When Vpr was removed from the virus genome, NK cells lost their ability to lyse the HIV-infected cells. Thus, HIV-1 actively triggers NK cells through the activity of its viral gene product, Vpr.
NK cells are involved in the immune response against tumor cells and virus-infected cells without the requirement for previous exposure to their targets or their products. The importance of NK cells in restraining viral infection was first shown in studies using murine models in which NK cells were depleted
One of the major roles of NK cells in controlling viruses is the destruction of the infected cells. Direct target cell killing by NK cells is mediated by the regulated release of granules containing perforin and granzymes
NK cell release of cytotoxic contents is regulated by a large array of signals provided by a variety of membrane-bound, activating receptors expressed on NK cells that interact with their corresponding ligands on target cells (see
Our prior studies
A recent study by Gasser
Initially, we determined the role of Vpr in the up-regulation of NKG2D ligands on primary CD4pos T-cells. To eliminate possible differences in replication kinetics due to the presence or absence of Vpr
We tested the binding of soluble NKG2D as a measure of ligand expression in PBMC from five individuals, in the presence or absence of
DHIV (A) and HIV-1NL4/3 (B) -infected primary T-cell blasts and uninfected CD4pos T-cells were surface stained with fluorochrome-conjugated anti-CD4 Ab and a fusion protein of human NKG2D and the Fc portion of human IgG1 along fluorochrome-conjugated goat anti-human IgG1. All cells were stained intracellularly for HIV-1 p24 antigen (Ag). Histograms were derived following acquisition on a flow cytometer of either 104 viable cells [(for uninfected cells) blue line] or 104 viable CD4neg cells and HIV-1 p24 Agpos infected cells (red line). As controls (green line), cells were stained with secondary Ab alone. This figure is representative of five separate experiments.
We then sought to determine whether Vpr was responsible for the expression of NKG2D ligands. We generated DHIV containing a truncation in Vpr. As controls, mutants of DHIV unable to express Vif, Vpu or Nef were also generated.
Infected primary T-cell blasts [(A) Wild-type, (B) ΔVpr, (C) ΔVif, (D) ΔVpu and (E) ΔNef] and uninfected CD4pos T-cells were surface stained with fluorochrome-conjugated anti-CD4 Ab and a fusion protein of human NKG2D and the Fc portion of human IgG1 along with fluorochrome-conjugated goat anti-human IgG1. All cells were stained intracellularly for HIV-1 p24 Ag. Histograms were derived following acquisition on a flow cytometer of either 104 viable CD4pos cells [(for uninfected cells) blue line] or 104 viable CD4neg and HIV-1 p24 Agpos infected cells (red line). The figure is representative data from three separate experiments.
To determine whether Vpr expression is sufficient to induce NKG2D ligands, we resorted to a lentiviral vector that encodes HIV-1 Vpr and GFP but no other viral gene (pPR-VIP). As a control, we used a similar lentiviral vector encoding only GFP. We observed that Vpr alone, but not the control lentiviral vector, was able to induce NKG2D ligands on CD4pos T-cells (
Primary CD4pos T-cell blasts were transduced with a lentivirus vector (pPR-Vip) coexpressing Vpr and GFP (red line). As a control the same cells were infected with a lentivirus vector expressing GFP only (green line). Following transduction cells were stained with a fusion protein of human NKG2D and the Fc portion of human IgG1 along with fluorochrome-conjugated goat anti-human IgG1. Histograms were derived following acquisition on a flow cytometer of either 104 viable cells [(for untransduced cells) blue line] or 104 viable GFPpos cells (green and red line). The figure is representative data from two separate experiments.
The studies illustrated in
Primary CD4pos T-cell blasts were infected with wild-type (WT) HIV-1. As a control, the same cells were infected with HIV-1 that was deficient in expression of Vpr (ΔVpr). As an additional control we evaluated uninfected CD4pos T-cells for expression of individual NKG2D ligands (blue lines). Following infection, cells were surface stained using fluorochrome-conjugated mAb specific for: CD4 (A-J), ULBP-1 (A–B), ULBP-2 (C–D), ULBP-3 (E–F), MIC-A (G–H) or MIC-B (I–J). The cells were then intracellularly stained for HIV-1 p24 antigen. Histograms were derived following acquisition of 104 viable uninfected CD4pos cells (blue line) or 104 viable CD4neg and HIV-1 p24 Agpos infected cells (red line). Green line represents the histogram of staining controls (isotype controls). The figure is representative data from two separate experiments.
DHIV-ΔVpr induced considerably lower levels of ULBP-1 (
Next we asked whether induction of NKG2D ligands on infected cells was accomplished through increased steady-state levels of ULBP-1 and ULBP-2 mRNA. We infected CD4pos T-cells with DHIV and select mutants, and then measured the mRNA levels of ULBP-1, ULBP-2, ULBP-3, MIC-A and MIC-B relative to the level of GADPH mRNA (
RNA was extracted from HIV-infected CD4pos T-cell blasts and uninfected cells and reverse transcribed. Relative copy numbers of NKG2D ligands were determined by real-time PCR using NKG2D-ligand specific primer pairs and normalized relative to GAPDH expression. The relative increase in copy numbers was calculated as described in
Vpr induces G2 arrest of infected cells through interaction with a Cul4a-based ubiquitin ligase that also contains the adaptor, DDB1, and the substrate receptor, DCAF1 (reviewed in
We first tested whether domains in Vpr that are involved in recruiting or activating the Cullin-4a-based E3 ubiquitin ligase were required for induction of ULBP-1 and ULBP-2. The Vpr Q65R and Vpr R80A mutations have been previously shown to abate the ability of Vpr to induce cell cycle arrest
Primary T-cell blasts infected with wild type HIV-1 and HIV-1 with various mutations in Vpr were surface stained with fluorochrome-conjugated anti-CD4 Ab and a fusion protein of human NKG2D and the Fc portion of human IgG1 along with fluorochrome-conjugated goat anti-human IgG1. Uninfected CD4pos T-cells were surface stained in a similar fashion. All cells were stained intracellularly for HIV-1 p24 Ag. The mean fluorescent intensity of NKG2D ligand staining was obtained from collection of 104 CD4neg p24pos cells for all infected cells, and of the 104 CD4pos p24neg cells for the uninfected control. This figure is a representative of three separate experiments.
To more directly assess the requirement of the Cullin-4a-based E3 ubiquitin ligase in inducing NKG2D ligand expression, we resorted to RNA interference-mediated depletion of DCAF1 (
Primary CD4pos T-cell blasts were transduced with shRNA of a DCAF1-specific sequence (C) or an shRNA with a scrambled sequence (B). Untransduced cells were used as a negative control (A). The cells were then infected and stained with anti-CD4 Ab and a fusion protein of human NKG2D and the Fc portion of human IgG1 along with fluorochrome-conjugated goat anti-human IgG1. Transduced cells were detected by GFP expression. The figure is representative data from two separate experiments.
Manipulation of the Cul4aDCAF-1 complex by Vpr results in ATR activation and G2 arrest
Primary CD4pos T-cells were infected in the presence (C) or absence (B) of the ATR inhibitor, caffeine. As controls CD4pos T-cells were not infected (A) but treated with caffeine or vehicle. Following exposure to 4 mM caffeine and HIV-1 infection, NKG2D ligand expression on the surface of infected cells was measured by staining with a fusion protein of human NKG2D and the Fc portion of human IgG1 along with fluorochrome-conjugated goat anti-human IgG1. The histograms are gated for either 104 viable CD4pos/HIV-1 p24neg cells (blue) or on 104 viable CD4neg HIV-1 p24pos cells (red). This figure is representative of three separate experiments.
Since caffeine inhibits both ATR and ATM
Primary CD4pos T-cells were infected in the presence (B) of the ATM-specific inhibitor KU55933 (10 µM). As a positive control infected cells were treated with 4 mM caffeine (C). Negative controls included vehicle-treated infected cells (A) or CD4pos T-cells infected with HIV-1 lacking Vpr [(ΔVpr) D]. Following exposure to inhibitors and HIV-1 infection, NKG2D ligand expression on the surface of infected cells was measured by staining with a fusion protein of human NKG2D and the Fc portion of human IgG1 along with fluorochrome-conjugated goat anti-human IgG1. The histograms are gated for either 104 viable CD4pos/HIV-1 p24neg cells (blue) or on 104 viable CD4neg HIV-1 p24pos cells (red). MFI = mean fluorescent intensity. This figure is representative of three separate experiments.
Our studies, thus far, demonstrate that HIV-1 induces the expression of ULBP-1 and ULBP-2 through Vpr. However, whether Vpr mediated-induction of ULBP-1 and ULBP-2 on infected cells constitutes a signal that will trigger NK cell lysis remains to be determined. To address this, we compared the ability of primary NK cells to lyse autologous T-cell blasts when infected with either WT or ΔVpr viruses. As a control, we blocked NKG2D on NK cells prior to exposure to target cells. If NKG2D interaction with its ligands on infected cells had any role in triggering NK cell lysis, then blocking NKG2D would abate the effect. As shown in
Primary CD4pos T-cell blasts were infected with HIV-1 that was deficient in expression of Vpr (ΔVpr) (A&B). As a control, the same cells were infected with wild-type (WT) HIV-1 (A&B). Following infection the infected cells were isolated, labeled with 51Cr and mixed with autologous NK cells at 2.5∶1 (open bars) and 5∶1 (closed) effector cell to target cell ratios. Each group was done in triplicate. Prior to the lytic assay some of the NK cells were exposed to blocking antibodies to NKG2D (A). At the end of the incubation period culture fluids were harvested and analyzed for the presence of 51Cr. Percent specific lysis was determined as described in the
Our previous studies demonstrated that primary T-cell blasts infected both
DNA damage leads to the expression of NKG2D ligands
The available literature on whether retroviruses activate DNA damage responses upon integration remains controversial. Several publications propose that HIV (and other retroviruses) do not cause DNA damage signaling when they integrate
Recent studies from other investigators indicate that soluble Vpr has an impact on NK cell activity
HIV-1 accessory gene products other than Vpr appear to have no effect on the expression of NKG2D ligands. Our observations here are in contrast to those of Cerboni
The Cul4aDCAF-1 ubiquitin ligase contributes to the ability of Vpr to arrest HIV-1-infected cells in G2 phase of the cell cycle
Our expectation was that induction of cell-surface expression of ULBP-1 and ULBP-2, would trigger NK cells to kill infected cells. This indeed is what we found, since NK cells have a decreased ability to lyse virus-infected cells lacking Vpr. Moreover, the reduction in killing activity of the ΔVpr virus was strikingly similar to that obtained against WT virus-infected cells in which we masked the NKG2D receptor. We have reported, in previous published studies, that variations exist between HIV-infected donors in the capacity of their NK cells to kill autologous HIV-infected cells
NK cell function in HIV-infected individuals is similar to that of uninfected individuals unless the patient is highly viremic
Based on our studies, the expected consequence of Vpr's induction of ULBP-1 and -2 would be that HIV-1-infected cells would become sensitive to killing by NK cells. This scenario would clearly be detrimental to virus replication. A logical explanation for this paradox is that up-regulation of ULBPs by Vpr is a downstream consequence of its biological activity (e.g., Vpr induces LTR transactivation in the G2 phase of the cell cycle
It is difficult to reconcile how HIV-1 could persist if it increases the likelihood of being destroyed by NK cells through induction of ULBP-1 and ULBP-2 (
The presence of ULBP-1, ULBP-2 on the infected cell surface would allow the subpopulation of NK cells not regulated by HLA-C and HLA–E to kill the HIV-1-infected cells. However, NK cells are also regulated at the level of coactivating receptors
Our observation that HIV-1 Vpr induces NK cell ligands opens a new set of questions that will need to be addressed in the near future. Specifically, it will be compelling to ascertain whether up-regulation of ULBP-1 and 2 by Vpr can be exploited therapeutically, such that HIV-1 infected cells can be manipulated to become more susceptible to NK lysis; (ii) whether other accessory or, in general, viral proteins can manipulate other aspects of the NK response; and (iii) the signaling steps linking activation of cell cycle checkpoint proteins and the increase in steady-state mRNA levels for NKG2D ligands.
The mouse anti-human CD4, CD16, CD56, CD112 and MIC-A/MIC-B monoclonal antibodies (mAbs) were obtained from BD Biosciences (
All primary cells used in this study were isolated from peripheral blood drawn from all healthy donors after informed written consent was obtained in accordance with the Declaration of Helsinki and the policies of the Institutional Review Board at Rush University Medical Center, Chicago, IL. Peripheral blood mononuclear cells (PBMC) were isolated by centrifugation (1000×
The envelope-defective DHIV vector is isogenic to the HIV molecular clone HIV-1NL4/3. To construct DHIV vectors with premature stop-codons in the
The FG12 vector system was graciously provided by Dr. Dong Sung An (University of California, Los Angles, CA). The DCAF1_3590 target sequences have been reported previously
Vesicular stomatitis virus (VSV)-G protein pseudotyped FG12, pPR-VIP and DHIV vectors were produced as previously described
Primary T-cells were activated using anti-CD3/anti-CD28 mAb coupled to magnetic beads for 48 hr for DHIV and pPR-VIP or 24 hr for the FG12 vector. Infection/transduction was done by spin-inoculated as described
Simultaneous detection of surface antigens and intracellular HIV-1 p24 antigen (Ag) was done as previously described
Forty-eight hrs after infection culture, CD4neg p24pos cells were purified by removing CD4pos p24neg cells with anti-CD4 mAb coated magnetic beads (Dynal). To eliminate dead cells the Dead Cell removal kit from Miltenyi was used (
Primer pairs for detection of ULBP-1, ULBP-2, ULBP-3, MIC-A and MIC-B were obtained from previous studies described in
The cytotoxicity of HIV infected cells by autologous NK was measured using the 51Cr release assay. The cytotoxicity assay was done according to the methods previously described in
Gating strategy used for detection of NKG2D ligands on infected cells. Infected primary T-cell blasts and uninfected CD4pos T-cells were surface stained with anti-CD4 Ab. All cells were stained intracellularly for HIV-1 p24 antigen (Ag). Cells were then incubated in the presence of Aquadead stain kit (Invitrogen) to distinguish viable and non-viable cells. Throughout the study NKG2D ligands were evaluated on either 104 viable uninfected (CD4pos HIV-1 p24 Agneg cells) or 104 viable infected cells (CD4neg HIV-1 p24 Agpos). FSC = forward scatter, SSC = side scatter. Gates in red indicate selection process for infected and uninfected cells.
(2.48 MB TIF)
NKG2D ligands are not expressed on CD4pos T-cells infected with ΔVpr HIV-1. Infected primary T-cell blasts and uninfected CD4pos T-cells were surface stained with a fusion protein of human NKG2D and the Fc portion of human IgG1 along with fluorochrome-conjugated goat anti-human IgG1. All cells were stained intracellularly for HIV-1 p24 antigen (Ag). Two-dimensional plots were derived following acquisition on a flow cytometer of 104 viable cells. Markers in dot plots were positioned based on the staining controls. The figure is representative data from three separate experiments.
(0.98 MB TIF)
Effect of point mutations in postions 65 and 80 of Vpr on the cell cylcle of HIV-infected cells. Infected primary T-cell blasts and uninfected CD4pos T-cells were stained with TO-PRO-3 in order to obtain the (G2+M)/G1 ratio.
(0.67 MB TIF)
Expression of DCAF1 is not affected by HIV-1 Vpr. The HeLa cell line was either treated with 10 µM aphidocolin (Aph), or infected with VSV-G pseudotyped HIV-1 with wild-type Vpr (Vpr) or HIV with Q65R and R80A mutations in Vpr (Vpr QR). Following treatment/infection cells lysates were made and western blotted. Western blots were probed with DCAF1 specific antibody.
(0.56 MB TIF)
Inhibition of ATR activity relieves Vpr-induced G2 arrest. Primary CD4pos T-cell blasts were exposed to 4 mM of the ATR inhibitor, caffeine (B and D) or vehicle (A or C) and either infected with HIV-1 (C and D) or left uninfected (A and B). Forty-eight hrs. following exposure to caffeine and HIV-1 infection the cell cycle profile of the uninfected and infected cells were detected by TO-PRO-3 staining.
(1.50 MB TIF)
Inhibition of ATM activity reduces NKG2D ligand expression on primary CD4pos T-cells treated with aphidocolin. Primary CD4pos T-cells were treated with 10 µM aphidicolin in the presence of 10 µM KU55933 (ATM-specific inhibitor) or 4 mM caffeine. As a negative control aphidicolin-treated cells were exposed to vehicles used to dissolve the inhibitor in solution. Following 48 h exposure to KU55933, caffeine or vehicle the cells were stained with a fusion protein of human NKG2D and the Fc portion of human IgG1 and fluorochrome conjugated-goat anti-human IgG Fc specific antibody (blue line) or secondary antibody alone [staining control (red line)]. The histograms are gated for 104 viable CD4pos cells. This is a representative of two experiments.
(0.71 MB TIF)
Ability of NK cells to lyse autologous T-cells infected with HIV-1 lacking Vpr. Primary CD4pos T-cell blasts were infected with HIV-1 that were deficient in expression of Vpr (ΔVpr). As a control, the same cells were infected with wild-type (WT) HIV-1. Following infection the infected cells were isolated, labeled with 51Cr and mixed with autologous NK cells at 2.5∶1 (A and C) and 5∶1 (B and D) effector cell to target cell ratios. Prior to the lytic assay some of the NK cells were exposed to blocking antibodies to NKG2D (C and D). At the end of the incubation period culture fluids were harvested and analyzed for the presence of 51Cr. Percent specific lysis was determined as described in the
(1.02 MB TIF)
Expression of viral proteins by various HIV-1 mutants. HIV-1-infected CD4pos cells were lysed in cell lysis buffer (1% NP-40, 150 mM NaCl, 50 mM Tris-HCl, pH 7.4, 0.25% Na-deoxycholate, 1 mM EDTA. 1 mM PMSF, 1 mM Na3VO4, 0.1% SDS, and protease inhibitors), run on 15% SDS-PAGE gels, transferred to PVDF, and probed for the indicated proteins with specific antibodies. (A) Lysates from CD4pos cells infected with DHIVΔVpr, DHIVΔVif, DHIVΔVpr,ΔVif or DHIV containing Vpr with point mutations in specific residues. (B) Lyates from CD4pos T-cells infected with DHIVΔNef.
(3.19 MB TIF)
Ability of shRNA expressing either scrambled or DCAF1 specific sequence to down modulate DCAF1 in transduced cells. Primary T-cell blasts were transduced with the shRNA with specific sequences, sorted by FACS to greater than 95% purity based on GFP expression, lysed in cell lysis buffer (1% NP-40, 150 mM NaCl, 50 mM Tris-HCl, pH 7.4, 0.25% Na-deoxycholate, 1 mM EDTA. 1 mM PMSF, 1 mM Na3VO4, 0.1% SDS, and protease inhibitors), run on 15% SDS-PAGE gels, transferred to PVDF, and probed for the either DCAF1 or β-actin proteins with specific antibodies.
(1.57 MB TIF)
The authors wish to thank Professor Alessandro Moretta (University of Genova, Genova, IT) for providing monoclonal antibodies used in the cytotoxicity assay and Dr. Dong Sung An (University of California, Los Angles, CA, USA) for providing the FG12 vector system. We would like to thank Dr. Linda Baum (Rush University Medical Center, Chicago, IL, USA) for her review and editing of the manuscript. We also thank Kelly Hudspeth and Bharatwaj Sowrirajan (Rush University Medical Center, Chicago, IL, USA) for their excellent technical assistance, and Dr. Alexander Steinle (University of Tübingen, DE) for helpful advice with the real-time RT-PCR.