|Home | About | Journals | Submit | Contact Us | Français|
Recent evidence demonstrates that HIV-1 infection leads to the attenuation of cellular immune responses, which has been correlated with the increased expression of programmed death 1 (PD-1) on virus-specific CD8+ T cells. PD-1 is induced upon T cell activation and its prolonged expression facilitates CD8+ T cell inhibitory signals when bound to its B7 family ligands, PD-L1/2, which are expressed on APCs. Importantly, early reports demonstrated that blockade of the PD-1/PD-L interaction by antibodies may help to counter the development of immune exhaustion driven by HIV viral persistence. To better understand the regulation of the PD-1 pathway during HIV infection, we examined the ability of the virus to induce PD-L expression on macrophages and dendritic cells. We found a direct relationship between the infection of APCs and the expression of PD-L1, in which virus-mediated upregulation induced a state of non-responsiveness in uninfected HIV-specific T cells. Furthermore, this exhaustion phenotype was revitalized by the blockade of PD-L1 after which T cells regained their capacity for proliferation and the secretion of proinflammatory cytokines IFN-γ, IL-2, and IL-12 upon restimulation. Additionally, we identify a critical role for the PI3K/Akt signaling pathway in PD-L1 upregulation of APC’s by HIV, as inhibition of these intracellular signal transducer enzymes significantly reduced PD-L1 induction by infection. These data identify a novel mechanism by which HIV exploits the immunosuppressive PD-1 pathway and suggest a new role for virus-infected cells in the local corruption of immune responses required for viral suppression.
The HIV-1 epidemic continues to be a major issue worldwide with around 41.3 million adults and 2.1 million children currently living with the virus and approximately 16,000 new infections per year. This retrovirus preferentially infects and kills CD4+ T cells and macrophages (1–4) resulting in the progressive dysfunction of the host immune system and increased susceptibility to various opportunistic infections and neoplasms. While slowing disease progression is typically achieved with Highly Active Antiretroviral Therapy (HAART), there is growing evidence that the immune system, on its own, can limit HIV replication in some cases. For instance, a subset of HIV-infected patients termed “elite controllers” have viral loads maintained below the detectable limit (< 50 copies HIV RNA/mL) without HAART (5–7). It has been suggested that this enhanced viral control directly correlated with CD8+ T cell activation and function as HIV specific CD8+ T cells from human controllers (long-term progressors) exhibit greater degree of activation and function (8) than those from non-controllers. Furthermore, in SIV infection models, CD8+ T cells are necessary for control of viremia, and vaccines that induce the most potent CD8+ T cell responses have proven to be the most effective in controlling disease progression (3, 9, 10).
Recent reports from several groups suggested that HIV persistence may be caused in certain instances by the inability of host HIV-specific CD8+ T cells to mount effective immune responses (11, 12). This deficit in HIV-specific CD8+ T cells correlated with increased expression of programmed death 1 (PD-1; also CD279), a receptor that inhibits T cell activation on HIV-specific CD8+ T cells (9, 13–19), and a decrease of CD4 T cell help (20). Moreover, PD-1 upregulation correlated with impaired immune function and disease progression (12, 15, 21). Induced expression of PD-1 on CD4+, CD8+ and natural killer T cells, engagement of its ligands, and subsequent signaling attenuates T-cell function via the inhibition of cellular kinases that signal through the T cell receptor and CD28 to promote cytokine production and cell proliferation (21). Through recruitment of phosphatases such as protein-tyrosine phosphatase SHP2, PD-1 decreases the phosphorylation and activation of kinases such as Spleen-tyrosine kinase (22), phosphatidylinositol-3-OH kinase (PI3K), and Serine-threonine kinase (Akt) (21). The essential role of PD-1 in suppressing T-cell activation and promoting immune homeostasis is underscored by the observation that pcdc −/− mice develop spontaneous, late-onset lupus-like disease and a dilated cardiomyopathy characterized by auto-Abs to troponin (23).
The effect of PD-1 in T-cell regulation and its role in the maintenance of a chronic viral infection was shown in the lymphocytic choriomeningitis virus (LCMV) murine infection model where PD-1 antibody blockade restored the function of Ag-specific T cells and led to clearance of the chronic infection (17–19, 24). A role for PD-1 in retroviral infection has also been suggested by several studies (13, 15, 17, 18, 25). During HIV infection, PD-1 expression on HIV-specific CD8+ T-cells correlates with disease progression as measured by viral load and CD4+ T cell counts (9, 17), and in chronically infected individuals PD-1 expression is high on HIV-specific CD8+ T-cells. Furthermore, administration of Abs which interfere with PD-1/PD-L1 (also CD274) binding leads to an increase in the activity of HIV-specific, PD-1high CD8+ T cells (18). Finally, in a SIV infection model, direct in vivo interruption of PD-1 signaling using a specific mAb increased the number of SIV-specific CD8+ T cells, decreased the viral load, and prolonged the survival of chronically infected animals (19). Together, these data suggest that interruption of PD-1/PD-L1 binding and presumably signaling can augment the host’s immune system to a chronic viral infection and possibly help mediate control.
While the expression of PD-1 on HIV-specific CD4+ and CD8+ T cells has been well documented, the natural source of PD-L1 is currently unknown. Furthermore, the signals regulating PD-L1 expression are also not well studied, and, most importantly, it is not known whether HIV infection can directly influence these. While few studies have directly investigated the role of PD-L1 in HIV-induced immune suppression, several have shown a clear correlation between HIV infection and PD-L1 upregulation (26–28). Baosso et al demonstrated that direct HIV-1 infection of monocytes in patients derived samples showed increased levels of PD-L1 expression (26). Trabattoni et al. reported that expression of PD-L1 in HIV-1 infected typical progressors and AIDS patients correlates with viral load and is inversely correlated with CD4+ T cell count (29).
We hypothesized that HIV infection directly disrupts the generation and/or activation of HIV-specific T cells by attenuating/disrupting Ag presentation via up- regulation of PD-L1/2 expression on APCs and hence signaling through PD-1 on T cells. In an effort to understand the role of PD-L1 in mediating HIV-induced immune system suppression, we explored the effect on HIV specific T-cells of disrupting the PD-1/PD-L1 interaction between HIV specific T-cells and APCs derived from the T-cell donor using PD-L1 siRNA to modulate PD-L1 expression. Additionally, we demonstrated that APCs including monocytes and dendritic cells (DCs) from HIV-positive humans have higher PD-L1 expression compared to normal donors. Infection of APCs was sufficient to directly induce PD-L1 expression, and interruption of PD-L1 in vitro by siRNA resulted in significant enhancement of proliferation and antiviral cytokine secretion by HIV-specific T cells. Understanding the mechanism of HIV-induced PD-L1 may provide new therapeutic targets for enhancing the host’s immune response and the results reported here have significant implications for the development of novel immunotherapeutic approaches (e.g. by blocking negative T-cell regulators) that will restore T-cell function in this widespread disease.
Leukopacks from individual healthy donors were obtained from the University of Pennsylvania School of Medicine Immunology Clinical Core (UPENN-ICC) and PBMCs were isolated by Ficoll-Hypaque (Pharmacia, Piscataway, NJ) density centrifugation. PBMCs, serum, and plasma from HIV-infected individuals were obtained from University of Pennsylvania-Center for AIDS Research (UPENN-CFAR) and UPENN-Immunology core. Samples were obtained in accordance with protocols approved by the Institutional Review Board of the Hospital of the University of Pennsylvania (UPENN-IRB). Some patients were receiving HAART containing at least one protease inhibitor (Table 1). Leukophoresis procedures were conducted in accordance with protocols approved by the UPENN-IRB. PBMCs were separated and cryopreserved in liquid N2. HIV-1 RNA level was determined from plasma using the Roche Amplicor 1.5 Kit (Roche Diagnostic Systems, NJ), as per manufacturer’s recommendations (Table 2). All primary cells were maintained in LPS-free RPMI-1640 medium supplemented with 10% FBS, 2mM L-glutamine, 20 mM HEPES, 100 U/ml penicillin, 100 µg/ml streptomycin, and in the presence of recombinant human IL-2 (50U/ml). CD3 MicroBeads were used for the positive selection of T cells from PBMCs using MACS® Technology isolation kit reagents and protocol (Miltenyi Biotec, CA). The resulting preparation was stained with anti-CD3-FITC (BD Biosciences) to confirm <95% purity of T cells. Human monocytic leukemia cell lines THP-1 and U937 were purchased from the American Type Culture Collection (ATCC; Manassas, VA) and maintained in their appropriate growth medium (30).
Monocytes were isolated from PBMCs using the human monocyte isolation kit (Automacs; Miltenyi Biotec, USA) according to the manufacturer's specifications. The isolated monocytes were washed four times with ice cold PBS containing 0.3% (w/v) BSA and 0.6% (w/v) Na3-citrate·3H2O, and the preparations were > 90% pure as determined by flow cytometry using anti-human CD14 Abs. Elutriated monocytes were cultured in RPMI 1640 supplemented with 10% FBS, 2 mM L-glutamine, 100 U/mL penicillin, 100 µg/mL streptomycin, 20 µg/mL gentamicin, (Gibco/Invitrogen), 50 ng/mL GM-CSF (R&D Systems, USA), and 25 ng/mL IL-4 (Peprotech, USA) at 37°C for 2 days; and then for an additional 5 days in medium lacking M-CSF before use in transfection, infection, ELISpot, and other assays (26, 30–33).
To assess modulation of PD-L1 transcription by viral infection, luciferase reporter plasmids expressing PD-L1 were assembled from synthetic oligonucleotides (GeneArt, Germany) and were cloned by inserting PD-L1 promoter sequences into specific restriction cloning sites (Xho 1 and BamH1) of the pMet-Luciferase reporter vector (Clontech, Mountain View, CA) (34). For the PD-L1 reporter assay, THP-1 cells were seeded onto a 6-well culture plate at a density of 500,000 cells/ml of medium and transiently co-transfected as previously described (34, 35) with a constant amount of the luciferase reporter PD-L1 promoter plasmid and various plasmids expressing viral gene-GFP fusion proteins. Furthermore, the total amount of plasmid DNA was kept constant by adding empty vectors. Cells were harvested 48 hr post-transfection, lysed, and then tested for luciferase activity with the luciferase assay kit (Promega, Madison WI) using LUMAT-LB9501 (Berthold, Bad Wildbad, Germany). Transfection efficiency was normalized by co-transfection with pLacZ and assayed for β-galactosidase expression (34).
Pools of three to five target-specific 19 – 25 nt siRNAs designed to knock down gene expression were used. Three different PD-L1 and PD-L2 Stealth™ siRNA duplexes and one recommended negative control siRNA duplexes were synthesized (Invitrogen, Carlsbad, CA) and displayed in Table 2. These target sequences were submitted to a BLAST search to ensure that only the PD-L1 or PD-L2 genes were targeted (36, 37). Similarly human Akt-1 siRNA and non-specific control pool (sc-29195) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). All siRNAs were dissolved or diluted to a concentration of 25 µM in DEPC-treated water and stored at −20°C. siRNAs were transfected into 5×106 experimental cells, using the Amaxa Nucleofector device (Amaxa, USA) with cell-specific Nucleofector kit V, according to the manufacturer's recommendations. After electroporation, cells were transferred to a complete RPMI-1640 medium for 48 hr before analysis for protein knockdown.
HIV-infected patient samples were added to Lysing Matrix D tubes (BIO 101, Carlsbad, CA) containing TRIzol Reagent and stored at −70°C. Total RNA was later extracted per the manufacturer’s protocol. To eliminate any residual genomic DNA, the RNA samples were incubated with DNase I for 15 min at room temperature. The DNase I was later inactivated by adding 1uL of 25mM EDTA to the reaction at 65°C for 10 min. The yield and purity of the total RNA was determined from the A260/280 absorbance value.
PCR primers for the genes of interest and β-actin (used as an internal control) were designed using Primer Express Software version 2.0, and a software program specially provided with the 7700 SDS (Applied Biosystems). The following gene-specific primers and Taqman probes were used: PD-L1 (XM-219775) are: 5’-TGTACCACGTCT CCCACATAACAG-3’ and reverse primer 5’-ACCCCACGATG AGGAACAAA-3’ and PD-L2 (XM-219777) are: 5’-TGACCCTCTGAGTT GGATGGA-3’ and reverse primer 5’-GCCGGGATGAAAGCATGA-3’ (38). For β-actin (NM-031144): 5’-TGCTGACAGGATGCAGAAGGA-3’ and reverse primer 5’-CGCTCAGGAGGAGCAATGAT-3’.
RT-PCR assays using SYBR Green were performed with the ABI PRISMER 7700 Sequence Detection System (Perkin/Elmer Applied Biosystems, Foster City, CA). For each primer pair, the dissociation curve analysis was conducted to ensure the specificity of the amplification at the end of PCR. Different concentrations of primers (50–100 nM) were tested to determine the optimal PCR conditions. Briefly, 90 µL Master Mix, containing 200 nM primers, 45 µL 2XqPCRTM Mastermix Plus for SybrTM Green I, and 2 µL cDNA were mixed before aliquoting in triplicate to a 96-well microtitre plate. The cDNA was amplified under the following universal conditions: one cycle at 50°C for 2 min and 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. To correct for variations in input RNA amounts and the efficiency of reverse transcription, an endogenous ‘housekeeping’ gene, β-actin, was also quantified and used to normalize the results (39–41).
The full-length proviral DNA constructs used for production of HIV-1 were as follows: HIV-1 NL4-3Wt, a chimeric VSV-G envelope (NL43ΔEnv/VSV-G Env), macrophage-tropic envelope AD8, 89.6 and Bal-mac (33, 34). These proviral vectors were obtained from the NIH AIDS Research and Reference Reagent Program, MD. HIV-1 env, vif, vpu, vpr and nef were individually cloned using PCR and characterized as previously described (34). For production of viruses, 293T cells were transfected in 10-cm-diameter dishes with 10ug each of proviral vectors as previously described. Briefly, we used 10µg of defective HIV-1 genomic vector, under control of the LTR promoter. HIV-1 stocks were produced in 293T cells (ATCC, Manassas, VA) were cultured at 37°C in 5% CO2 and Dulbecco’s modified Eagle’s medium supplemented with fetal bovine serum (10%), penicillin (50 IU/ml), and streptomycin (50 mg/ml) and pseudotyped using 5ug of VSV-G or AD8 to replace Env (29, 32) with FuGENE 6 transfection reagent (Roche Applied Science, Nutley, NJ) (30, 33). Twelve hours after transfection, the cells were washed with 5 ml of phosphate-buffered saline and 4 ml of complete medium was added. Control viruses were produced in the presence of equivalent concentrations of control vector. Three days post transfection, virus stocks were prepared by collecting culture supernatants and passing them through 0.45-mm-pore-size cellulose acetate syringe filters (Millex 0.45 µm filters; Millipore, USA) and after adding FBS (10%) stored at −80°C. Viral titers were determined by infection of the human Jurkat T or U937 cells and p24gag Ag was measured by capture ELISA at the Viral and Molecular Core Services, CFAR, University of Pennsylvania School of Medicine. For infection studies, human PBMCs were isolated from normal, HIV-1 negative donors as described above. PBMCs (2×105 cells/well) were mock infected (with media from the cell cultures used to grow cells) or infected with cell-free HIV-1 virus at a concentration of 100 TCID50/106 cells/ml. After 4–6 hr of incubation at 37°C, cells were gently washed, resuspended with complete medium, and maintained for the indicated time periods. Culture supernatants and cells were then harvested for p24 Gag by ELISA and by flow cytometry as described below.
ELISpot was performed as previously described (42, 43). Briefly, 96-well ELISpot plates (Millipore) were coated with anti-human IFN-γ capture Ab (R&D Systems) and incubated for 24 hr at 4°C. The following day, plates were washed with PBS and blocked for 2 hr with 1% BSA. Two hundred thousand PBMCs from HIV-1 positive patient were plated in triplicate and stimulated with overlapping HIV-1 peptide consisting of 15-mers overlapping by 9 amino acids and spanning the length of the appropriate protein (Gag or Env). Autologous monocytes with and without PD-L1 depletion using specific siRNA were then added to the corresponding wells and incubated overnight at 37°C in 5% CO2 in the presence of media alone (negative control), media with Con A (positive control), or media with peptide pools (Gag/Env) (10 µg/ml). As an additional control, anti–PD-L1 was added at 10 µg/mL. After 24 hr, the cells were washed and then incubated for an additional 24 hr at 4°C with biotinylated anti-human IFN-γ Ab (R&D Systems). Streptavidin–alkaline phosphatase (R&D Systems) was added to each well after washing and then the plate was incubated for 2 hr at room temperature. The plate was then washed, and 5-bromo-4-chloro-3′-indolylphosphate p-toluidine salt and nitro blue tetrazolium chloride (chromogen color reagent; R&D Systems) were added. Lastly, the plate was rinsed with distilled water, dried at room temperature, and spot forming units (SFU) were quantified by an automated ELISpot reader (CTL Limited, OH), and the raw values were normalized to SFU per million cells.
Anti–human Abs were used in cell surface staining and intracellular staining: CD3 (Clone UCHT1; BD Pharmingen), CD4 (clone RPA-T4; BD Pharmingen), CD8 (cloneRPA-T8; BD Pharmingen), CD14 (clone M5E2; BD Pharmingen), CD80 (L307.4; Pharmingen), CD86 (clone IT2.2; Pharmingen), PD-L1 (clone MIH1; eBioscience), and PD-L2 (clone MIH18; eBioscience). The PD-L1-blocking mAb (10F.9G2) was purchased from Biolegend. For staining, the experimental cells were incubated with combinations of fluorochrome-labeled Abs for 30 min on ice. Cell surface staining was performed at 4°C in FACS Buffer (1% FCS, PBS, 2mM EDTA; Invitrogen). For intracellular staining, cells were permeabilized using BD FixPerm (44) following staining. The percentages of cells expressing intracytoplasmic HIV-1 Gag-related products were evaluated using KC57-RD1/PE- or KC57/FITC-conjugated anti HIV-1 Gag mAb (Beckman Coulter, Miami, FL). Electronic compensation was conducted with Ab capture beads (BD Biosciences, San Jose, CA) stained separately with individual mAbs. Forward scatter area (FSC-A) versus forward scatter height (FSC-H) was used to gate out the cell aggregates. In addition, ViViD dye staining was used to exclude dead and dying cells (45). Cells were analyzed with a modified LSRII flow cytometry (BD Immunocytometry Systems, San Jose, CA) or Coulter EPICS® Flow Cytometer (Beckman Coulter, Miami, FL) using FlowJo software (TreeStar, Ashland, OR).
For CFSE labeling, T cells (0.5 – 2×106) were labeled with 0.2–1 µM CFSE (CFDA-SE; Invitrogen) just before stimulation (45). Cell division accompanied by CFSE dilution was analyzed by flow cytometry. Cell-free culture supernatants were stored at −80°C, thawed, and subsequently analyzed for the presence of IFN-γ, IL-2, IL-12 and IL-10 by ELISA (R&D Systems, MN), according to the manufacturer’s instructions. ELISA for phospho- and total Akt was performed using PathScan® kits (7160 and 7170) from Cell Signaling Technologies (Danvers, MA); the phospho-Akt specific ELISA detects Akt which is phosphorylated on serine 473. Although equal sample volumes were added to the wells, correction for differing protein concentrations was made before the calculation of relative Akt.
Infected and mock-infected cells were lysed in 60-mm dishes for 5 min at 4°C by using 250 µl of NP-40 lysis buffer (Invitrogen) supplemented with a phosphatase inhibitor cocktail (PhosSTOP) and a protease inhibitor cocktail (Complete) as directed by the manufacturer (Roche Applied Science). Lysates were collected and spun at 10,000 rpm for 5 min at 4°C, and then 100 µl of the supernatant was added to 20 µl of 6X sample buffer (Invitrogen) for 12% SDS-PAGE using equal volumes of lysate. Afterwards, proteins were transferred to nitrocellulose using iBlot® Dry Blotting System (Invitrogen) and immunodetection was performed with the SNAP i.d. Protein Detection System (Millipore, MA, USA) using specific mouse antiserum, and the expressed proteins were visualized with HRP-conjugated goat anti-mouse IgG using an ECL detection system (Amersham Pharmacia Biotech, Piscataway, NJ). Primary Abs were diluted in 5% (wt/vol) BSA (fraction V)-TBS-T as recommended by the Ab manufacturer. PI3K and Akt IV inhibitors were used and were purchased from Invitrogen and Calbiochem respectively. Detection Abs for Akt and phospho-AktSer473 were purchased from Cell Signaling Technologies (Danvers, MA), and the Ab against β-actin (1:5,000) was purchased from Santa Cruz, Inc.
Data was collected from cellular assays and presented as the mean +/− standard deviation which was calculated from triplicate wells of pooled samples from each experimental group. Prior to all statistical analysis, the normality of the data was confirmed with Levine's test. Analysis between groups were performed using independent samples t-test. Comparisons among three groups were performed with ANOVA with a post-hoc Fisher’s Least Significant Difference (LSD) test to correct for multiple comparisons between groups. In each case, p ≤ 0.05 was considered to be significant. All statistical analysis was carried out using the Statistical Package for the Social Sciences (SPSS).
It is has been reported that different viruses induce unique combinations of co-stimulatory ligands that determine the fate of antiviral activity (46). In this manner, viruses that specifically induce expression of PD-L1/2 (PD-Ls) on APCs might downregulate the adaptive immune response and favor viral persistence. Although the expression of PD-1 on HIV-specific CD4+ and CD8+ T cells has been extensively investigated, neither the expression nor the source of PD-Ls in activating PD-1 on HIV-specific T cells is known. For example, we still do not know if HIV infection can directly influence the expression of not only PD-1, but also the PD-Ls. If this is not the case, then we also do not know the source of the PD-Ls which activate PD-1. Although very few studies have directly investigated the importance of PD-L1 in HIV-induced immune suppression, some recent work suggests a clear association (26, 27). For example, Trabattoni et al. reported that the expression of the PD-Ls in HIV+ patients is directly correlated with viral load, and inversely correlated with CD4+ T cell counts (29), and Meier et al. (41) illustrated that uridine-rich HIV-derived siRNA is sufficient to stimulate PD-Ls expression through a MyD88-dependent mechanism. Unfortunately, the latter experiment failed to show if HIV-derived siRNA or MyD88 is required for the PD-Ls increase in an appropriate infection setting. To determine whether HIV-1 infection modulates the expression of PD-Ls, we compared the expression of PD-L1 and PD-L2 on infected (CD4+CD14+p24gag+) and uninfected (CD4+CD14+p24gag-) monocytes from HIV-1-infected patients (Fig. 1). Both PD-L1 and PD-L2 were expressed at much higher levels on infected cells (Fig. 1A) when compared to uninfected monocytes. This was further validated when we compared total monocytes from healthy donors to those from HIV-infected patients (Fig. 1B). In order to determine whether this increased expression of PD-L1 & L2 on infected monocytes is a post-translation effect, RT-PCR analysis of total RNA extracted from the HIV-1-infected purified CD14+ monocytes was performed. As shown in Figure 1C, HIV infection also increased the amount of PD-L1 & L2 mRNA transcript when compared to levels from uninfected control cells (Patient ID#9; u.i). Together, these data demonstrate that HIV infection significantly increased both the transcription and surface expression of PD-L1/2 on monocytes from HIV infected individuals.
It is known that numbers of circulating myeloid DCs (mDCs) and plasmacytoid DCs (pDCs), the two main subsets of blood DCs, are modified during different viral infections (47). The capacity of DCs to stimulate T cells depends on the net costimulatory signal delivered by DCs to T cells. Increased costimulatory marker expression (such as CD83, CD86, and CD40) can facilitate T cell activation (48). Thus, in addition to CD14+ cells, we also analyzed PD-L1 expression on DCs. Monocytic CD80 and CD86 expression in PBMCs was determined by dual-color staining analysis both in HIV-1 patients and uninfected controls. Most CD14+ monocytes expressed CD80 and CD86 (Fig. 1D) and there was no statistically significant difference in these levels (Fig. 1E) as determined by MFI values. We next examined whether this could also be the case during HIV infection by analyzing the frequency of mDC and pDC in seven HIV positive patients. PBMCs were stained, acquired, and gated during analysis to exclude debris and most polymorphonuclear cells and DC’s are identified as negative for the lineage markers CD3, CD14, CD16, CD19, CD20, and CD56 and positive for HLA-DR. Further, high lineage low DC were then phenotyped as myeloid (CD11chigh, CD123low) or plasmacytoid (CD11clow, CD123high) (Fig. 1F). As shown in the Figure 1G, MFI of PD-L1 on peripheral blood pDCs and mDCs from HIV-infected patients was significantly greater than that from HIV-uninfected patients. Together, these data show that HIV-1 infection increases the expression of PD-L1 on blood DCs from HIV-infected patients.
Although it appears that upregulation of PD-1 correlates with disease progression, we wondered if this effect was involved in the upregulation of PD-Ls in HIV-1 positive patients. As shown in Fig. 2A, there is a direct correlation between PD-1 on CD3+/CD8+ T cells and PD-L1 on CD14+ cells (p=0.0274; r=0.6895). Furthermore, a direct inverse correlation between CD4+ T cell counts and PD-L1 expression on CD14+ cells (p=0.0329; r=−0.6731) was observed as well as production of IFN-γ on CD8+ T cells and PD-L1 expression in APCs population (CD4+/CD14+) among HIV-1 positive patients (p=0.0016; r=−0.8563) (Fig. 2B & C). These data suggest that upregulation of PD-1/PD-L1 may be involved in immune exhaustion in vivo, which is likely to contribute to the CD4+ T cell depletion.
Since our data had shown that APCs from HIV-infected patients exhibit increased PD-Ls expression, we decided to investigate whether in vitro infection of PBMCs was sufficient to increase PD-L1/2 expression. As shown in Fig. 3, we analyzed PD-L1 &-L2 expression in infected (p24Gag+) and uninfected (p24Gag−)/CD14+ cells during in vitro HIV-infection of APCs using four different HIV virus strains (NL4-3 (VSV-G env), NL4-3 (AD8 env), Ba-L, and 89.6). We found that HIV-1 p24 positive cells exhibited a greater level of PD-L1/2 expression compared to that of negative cells (Fig. 3A and B). The consequence of this increase may be that these APCs may activate PD-1 signaling on HIV-specific T cells to repress their function. Therefore, understanding how HIV increases the expression of PD-L1/2 may be critical for understanding the mechanism of HIV-specific T cell repression. PD-L1 is also necessary for the effective generation and maintenance of inhibitory T regulatory cells (iT reg), which also limit HIV-specific T cell activation (31). Furthermore, this effect was not specific to any particular virus strain as we observed similar results with infection with the four different strains. Consistent with the data from HIV-infected patients, direct infection of PBMCs with HIV increased the expression of PD-L1 &-L2. Altogether, these results demonstrate that HIV infection of APCs in vitro is sufficient to induce PD-L1/2 expression.
Previous reports showed that Abs against PD-Ls that blocked signaling were capable of restoring T cell function/activation in vitro (16–18). However, the source of PD-Ls responsible for repressing T cell activation during HIV infection has not yet been described. Based on our observations that APCs from HIV-infected patients express more PD-Ls, we hypothesized that interruption of this signal could reverse exhaustion observed in PD-1-positive HIV-specific T-cells. To determine the role of HIV-induced PD-L upregulation on APCs, we performed functional assays following PD-Ls knock down by siRNA. To begin, we tested different siRNAs and determined their specificity in silencing PD-L1 expression on monocytes using cells electroporated with 0.25 nmol siRNA and subsequently cultured. PD-L1 siRNAs appreciably reduced PD-L1 mRNA levels 2 days after electroporation, while the negative control siRNAs had no effect (Fig. 4). PD-L1 siRNA clone 2 induced the most pronounced reduction of PD-L1 as determined by Western blot analysis (Fig. 4A) and flow cytometry (Fig. 4B); a decrease in the number of PD-L1 positive cells from 12.7% to 2.0% was observed after treatment with clone 2 siRNA. Furthermore, in a similar manner, we also knocked down PD-L2 in macrophages and DCs from HIV-infected patients with at least 2 different siRNA pools (data not shown). Together, these data demonstrate that the treatment of monocytes with siRNA specifically and significantly decreased the expression of PD-L1 protein.
To determine the contribution of PD-L1 expression on monocytes/DCs from HIV-1 positive patients on T-cell function, protein was specifically knocked-down using siRNA silencing. Monocytes (5×106) isolated from PBMCs were first treated for 48 h with the siRNA constructs specifically targeting PD-L1. Next, these cells were re-incubated overnight with the autologous purified T cells from which they were originally isolated along with HIV-specific peptides (Gag and Env) in an IFN-γ ELISpot assay. Co-culture with PD-L1 siRNA-silenced cells led to a significant increase of IFN-γ secreting cells among the PBMCs pool after re-stimulation with HIV-1 peptides in vitro (Fig. 4C and D). Likewise, HIV-1 infected PBMCs treated with a PD-L1 blocking Ab also showed a higher number of peptide specific CD8+T cells capable of secreting IFN-γ. This result was further confirmed by the detection of cytokines (IL-12, IFN-γ, IL-2 and IL-10) that were released into the supernatant during the above culture systems using the ELISA assay (Fig. 4E). Treatment with siRNAs that knocked-down PD-L1 resulted in a significant increase in the production of the pro-inflammatory cytokines IFN-γ, IL-2, and IL-12, while levels of the anti-inflammatory IL-10 were significantly decreased. In order to gain more insight into the levels of various cytokines produced by the different cell subsets upon stimulation with PD-L knockdown monocytes, co-culture cells were collected and analyzed by intracellular flow cytometry for delineation of cell-specific contributions of the observed cytokines. PD-L knockdown evidently elevated IL-2 and IFN-γ production by CD4+ as well as CD8+ T cells, whereas CD11c+ and CD14+ produced IL-12. Furthermore, production of IL-10 by CD14+ cells was decreased (data not shown). Altogether, these data confirm previous results showing that increased PD-1 signaling results in T cell suppression and demonstrate that siRNA-mediated silencing of PD-L1 on APCs was sufficient for significant improvement in T cell immune function.
Another hallmark of T cell function is the capacity to mount Ag-specific proliferative responses. Since pro-inflammatory cytokine responses were markedly improved in T cells from HIV-infected patients after PD-L1 silencing in peptide-presenting APCs, we next analyzed the proliferative ability of HIV-specific CD8+ T cells from HIV-infected PBMCs upon siRNA treatment (Fig. 5). Purified T cells from infected cells were labeled with CFSE and mixed with autologous monocytes (at a ratio of 1:10) that were untreated or electroporated with PD-L1 or control siRNA. Mixed cells were incubated in medium alone or stimulated with Gag or Env peptides and, 5 days later, cells were stained and analyzed by flow cytometry. While only modest peptide-specific proliferative CD4+ and CD8+ T cell responses were observed in the absence of siRNA treatment, knockdown of PD-L1 on APCs resulted in a marked increase in the number of responding T cells (Fig. 5A & B); CD4+ and CD8+ T cell responses were 3.0- and 4.0-fold higher for Gag stimulation and 4.4- and 2.7-fold for Env (Fig. 5C). These data show a clear increase in the proliferation of HIV-specific CD8+ T cells in the presence of PD-L1 knockdown on APCs as compared with expansion following stimulation with peptides in untreated cells. Altogether, our findings confirm previous reports that PD-1/PD-L1 signaling inhibits HIV-specific CD8+ T-cell proliferation and that the blockade of the PD-L signal pathway can rescue T cell function.
Our results thus far indicate that viral infection is capable of modulating PD-Ls regulatory components of the immune system. We next investigated the mechanism by which HIV induces PD-Ls upregulation in vivo. We transfected HIV accessory genes such as Vpr, Nef, Vif, Vpu and structural genes Env, and determined whether or not these proteins were fully capable of inducing PD-L1 expression. U937 cells were transiently transfected with the accessory genes fused with GFP expression plasmid and transfection efficiency was determined by flow cytometry. Thereafter, we measured PD-L1 expression with FACS analysis (Fig. 6A) and quantified the PD-L1 expression by MFI (Fig. 6B).
To further assess whether specific HIV accessory genes were sufficient in inducing PD-L1/2 expression on APCs, we constructed a promoter-based luciferase assay to delineate the specific enhancers that are required for HIV-induced transcription of PD-L1. As shown in Table 3, there are several transcription factor-binding sites including those for IRFs, NF-κB, SP1, etc. First, we transfected U937 cells with the promoter plasmid; at 48 hours post transfection; lysed and measured the luciferase activity (30) to validate the specific enhancers in driving HIV-induced PD-L1 transcription. Significantly, HIV-1 Nef increased the expression of PD-L1 in these monotropic cells as measured by increased luciferease production (Fig. 6C).
Further, we investigated if Nef is necessary to induce the expression of PD-L1. To address this question, we deleted Nef gene and developed a pseudoviral system as a tool to investigate the PD-L1 expression. We infected purified monocytes with this HIV-1 pseudovirus, and measured the expression of PD-L1/2 at 96 hours post infection. Specifically, cells were infected with VSV-G pseudotyped HIV-1WT or HIV-1ΔNef, with 50ng CAp24 equivalent of virus/105 cells. Loss of Nef, led to an attenuation of HIV-mediated PD-L1 upregulation upon infection (Fig. 6D). This effect did not appear to be a consequence of efficiency of infection as the mutant viruses all exhibited similar levels of p24Gag production (data not shown). Therefore, in the context of HIV infection of target cells, Nef activity may be a primary cause of the surface expression of PD-L1 on antigen presenting cells. Together, these experiments strongly suggest that HIV-1 Nef genes are necessary for HIV to induce PD-Ls upon infection.
Several viruses exploit the PI3K–Akt pathway to support their replication in host cells. However, involvement of this pathway in immune modulation of viral infections has not been conclusively demonstrated for HIV infection. We investigated the effect of a PI3K–specific signaling inhibitor on PD-Ls expression during infection by HIV-1. Proteins were harvested from NL4-3WT-inected cells at various times following infection, homogenized, and then analyzed for total Akt and pAkt (Ser473) by ELISA and Western blot. We observed that HIV-1 infection induced phosphorylation of Akt1 on serine 473 in U937 monocytic cells supporting a possible role in the upregulation of PD-L1 (Fig. 7A and B). To extend this observation we investigated whether a PI3K/Akt inhibitor could suppress the level of PD-L1 induction on macrophages during HIV infections. Accordingly, U937 macrophages were treated with a PI3K inhibitor and infected with the virus for 24 hr, and PD-L1 expression was analyzed by flow cytometry. Compared to untreated cells, the upregulation of PD-L1 in HIV-positive cells treated with PI3K inhibitor was greatly reduced (Fig. 7C). Furthermore, a similar level of inhibition of PD-L1 upregulation was observed when a specific Akt inhibitor was used. These data support that the PI3K–Akt pathway plays a role in PD-L1 upregulation in macrophage cells upon HIV infection.
Next, we studied the ability of PI3K/Akt inhibitors to inhibit PD-L1 expression on APCs by constructing a promoter-based luciferase assay to delineate the specific enhancers that are required for HIV-induced transcription of PD-L1. As shown in Fig. 6C, there are several transcription factor binding sites including those for IRFs, NF-κB, SP1, etc. For this study, we constructed a luciferase promoter reporter vector containing the PD-L1 promoter region to determine whether PD-L1 induction could be blocked by the PI3K/Akt inhibitors. U937 cells were transiently transfected with the PD-L1/Luc reporter construct and treated with various concentrations of LY294002 or Akt-IV inhibitor. Forty-eight hr post-transfection, cells were stimulated with 500 U/ml of IFN-γ for 12 hr then lysed for the measurement of luciferase activity to determine whether the specific enhancers driving HIV-induced PD-L1-transcription could be blocked by the PI3K/Akt inhibitors. We found that addition of IFN-γ increased the PD-L1 activation of these monotropic cells as measured by increased luciferase production (Fig. 8A). In contrast, a marked decrease of PD-L1 expression was detected when PI3K/Akt inhibitors were used in combination with IFN-γ stimulation. Together, these experiments demonstrated a role for PI3K/Akt signaling pathways in the regulation of the PD-L1 promoter and their necessity for HIV infection to induce PD-L1 in APCs.
Since phospho-Akt was involved in PD-L1 expression during HIV infection, the presence of this protein may contribute to the expression of PD-L1 in these cells. To further validate our results for the inhibitor experiments, we tested adapted Akt1 siRNA constructs targeting Akt into macrophages. U937 cells were transfected with Akt1 siRNA or control siRNA and 72 hours after transfection, whole cell lysates were prepared for Western blot analysis. This treatment resulted in a 50 – 60% reduction in Akt1 protein expression levels; however, Akt2 levels were not affected by Akt1 siRNA (Fig. 8B), further demonstrating the specificity of this effect. To better explore the role of Akt in regulating PD-L1 expression, we used U937 cells that are silenced for active Akt and have been infected with HIV virus. When Akt expression was suppressed by siRNA, PD-L1 levels were blocked significantly upon HIV infection (Fig. 8C). To our knowledge, these data are the first to demonstrate that downregulation of Akt leads to decreased expression of PD-L1. Collectively, the data summarized herein indicate that PI3K/Akt signaling regulates PD-L1 expression, and that Akt is a regulator of PD-L1 expression that is targeted by HIV during infection.
A hallmark of chronic viral infections, including HIV infection of humans, is the inability of the host to sustain effector immune responses sufficient to clear infection (12). Recent reports suggest that this observation may be explained in part by PD-1 upregulation and signals delivered to Ag-specific CD8+ T cells continually exposed to their antigen through PD-1 (9, 13, 18, 49). Indeed, steric inhibition of PD-1 ability to engage its ligands can increase the activity of Ag-specific CD8+ T cells (12, 18).
A poorly understood aspect of the relationship between PD-1 and HIV infection is the role of the natural ligands of PD-1. In particular, it is currently unknown whether HIV-induced immune suppression via PD-1 signaling is a direct or indirect effect mediated by APCs. Two naturally occurring ligands for PD-1 have been described thus far (23) and little is known about the tissue distribution, expression, and distribution of these in persons with chronic viral infections.
Several studies have demonstrated APC dysfunction during ineffective HIV-specific T cell immune responses and viral persistence (13, 18, 49). Our studies demonstrate a connection between HIV and PD-1 ligands on APC including monocytes and both major subsets of DC. Monocytes infected with HIV express high levels of the PD-1 ligands and DC isolated from infected host also have higher levels of PD-1 (Fig. 1). Further, the upregulation of PD-L1/2 was a direct effect of HIV infection in primary monocytes (Fig. 3). The expression of PD-1 ligands by these cells changes the outcome of the interaction between HIV specific T-cells and APC, rendering the T-cells non-functional.
We hypothesized that the upregulation of PD-L1 in pDC during HIV infection is due to the molecule’s pivotal role in Ag presentation and elicitation of T cell effectors. To examine this, we silenced PD-L’s expression on monocytes using siRNA duplexes which are highly specific for PD-L1 mRNA and stably limit up-regulation of PD-L1 allowing us to study the role PD-1/PD-L’s in the APC-T cell interaction. Prevention of this interaction by knock-down of PD-L1 and PD-L2 expression with siRNA recovered the functionality of HIV specific T-cells presented with HIV peptides (Fig. 4). We observed that PD-L1 silencing in APCs from patients with longstanding HIV infection resulted in the enhanced production of pro-inflammatory antiviral cytokines (IL-2, IL-12, and IFN-γ) by uninfected HIV-specific T cells (Fig. 5). Furthermore, this recovery was closely associated with a decreased secretion of the anti-inflammatory interleukin, IL-10. This cytokine profile has been shown to play a critical role in the regulation of virus-specific immune responses, and elevated pro-inflammatory titers measured here following siRNA treatment might help to bolster T-cell function in the midst of a chronic infection. The effect of interrupting the PD-1/PD-L’s pathway is another functional property of virus-specific T cells, and their proliferative capacity upon restimulation was also found to be greatly enhanced following siRNA treatment. Further, we tested whether Nef is necessary for the upregulation of PD-L1. Our study suggested that HIV’s accessory protein Nef is both necessary and sufficient to regulate PD-Ls upregulation in infected cells in vitro (Fig. 6). Together, these data showed that inhibiting PD-L’s expression on monocytic cells significantly influences HIV-specific CD8+ T cell responses and further supports the growing data demonstrating that PD-1/PD-L’s interactions play an important role in attenuating immune responses and that interruption of these interactions can reverse T cell dysfunction. What remains unclear and requires further investigation is whether or not PD-Ls expressed on APC in vivo are the major tolergenic mechanism for HIV specific T-cells in vivo.
Finally, we investigated the PI3K/Akt signaling pathway in HIV-induced PD-L expression, a pathway prominent for the activation of many cell types and one that has long been recognized for its significance during virus infection. Previously, Parsa et al. reported that the loss of PTEN, which negatively controls PI3K activation, increased the expression of PD-L1 (44) suggesting that PI3K activity correlates with PD-L1 expression. Based on our previous observations, we anticipated that the mechanism by which HIV induces PD-L’s expression utilized the PI3K/Akt pathway and that PD-Ls dysregulation can be blocked by inhibiting this signaling pathway. Indeed, we observed that inhibitors of PI3K and Akt caused similar changes in PD-L1 expression characterized by reduced expression (Fig. 7 and Fig. 8). Early activation of the PI3K/Akt pathway by HIV may also have implications in the pathogenesis of macrophages, which are among the primary targets of HIV and respond to infection by providing a number of important immune signals.
In conclusion, data here indicate that in patients with longstanding chronic HIV infection, blocking the PD-1 pathway may help to rescue the capacity of peripheral HIV-specific T cells and profoundly exhausted T-cells to mount effective antiviral immune responses (Fig. 9) without CD4+T-cell help. The possible benefit of understanding the mechanism of HIV-induced PD-L regulation would be the identification of new therapeutic targets for reducing CTL suppression during chronic viral infection. Results from this investigation may help lead to new and more specific immune therapeutic interventions for HIV infection and may also provide information that is relevant for other chronic infections such as HCV, an elusive goal that current drug therapy has not yet attained.
The National Institutes of Health AIDS Reagent Program provided reagents for this study.