ABSTRACT
Along with other immune checkpoints, T cell immunoglobulin and mucin domain-containing protein 3 (Tim-3) is expressed on exhausted CD4+ and CD8+ T cells and is upregulated on the surface of these cells upon infection by human immunodeficiency virus type 1 (HIV-1). Recent reports have suggested an antiviral role for Tim-3. However, the molecular determinants of HIV-1 which modulate cell surface Tim-3 levels have yet to be determined. Here, we demonstrate that HIV-1 Vpu downregulates Tim-3 from the surface of infected primary CD4+ T cells, thus attenuating HIV-1-induced upregulation of Tim-3. We also provide evidence that the transmembrane domain of Vpu is required for Tim-3 downregulation. Using immunofluorescence microscopy, we determined that Vpu is in close proximity to Tim-3 and alters its subcellular localization by directing it to Rab 5-positive (Rab 5+) vesicles and targeting it for sequestration within the trans- Golgi network (TGN). Intriguingly, Tim-3 knockdown and Tim-3 blockade increased HIV-1 replication in primary CD4+ T cells, thereby suggesting that Tim-3 expression might represent a natural immune mechanism limiting viral spread.
IMPORTANCE HIV infection modulates the surface expression of Tim-3, but the molecular determinants remain poorly understood. Here, we show that HIV-1 Vpu downregulates Tim-3 from the surface of infected primary CD4+ T cells through its transmembrane domain and alters its subcellular localization. Tim-3 blockade increases HIV-1 replication, suggesting a potential negative role of this protein in viral spread that is counteracted by Vpu.
INTRODUCTION
Upon viral infection of the host, robust immune responses are mounted to eliminate invading pathogens. These events include the proliferation and activation of various immune cells (reviewed in reference 1). Although these immune responses are usually effective in clearing the viral antigen, they may also induce damage to surrounding cells (1). Avoidance of these deleterious effects necessitates a tightly regulated immune response. A prominent example of negative immune regulation is observed in acute infections, where many antigen-specific immune cells undergo apoptosis during an attrition phase following the clearance of viral antigens, while a small minority differentiate into long-lived memory cells (2). Conversely, during chronic viral infections and certain cancers, antigens are not readily cleared, thereby preventing the ability of immune cells to undergo apoptotic regulation (2, 3). Specifically, continuous signaling through the T cell receptor (TCR) and chronic inflammation induce an exhausted phenotype in both CD4+ and CD8+ T cells (4, 5).
T cell exhaustion is a stepwise process leading to T cell dysfunction that corresponds to a distinct epigenetic program (6, 7). T cell exhaustion can be characterized as a decreased ability to produce proinflammatory cytokines such as interferon gamma (IFN-γ), tumor necrosis factor alpha (TNF-α), and interleukin 2 (IL-2) (6, 8). Furthermore, exhausted T cells lose their ability to proliferate and differentiate into memory subsets (9, 10). In severe states of exhaustion, T cells can undergo apoptosis (11). Notably, T cell exhaustion is marked by an upregulation of multiple receptors on the cell surface, such as programmed cell death protein 1 (PD-1), lymphocyte activation gene 3 (Lag-3), T cell immunoreceptor with Ig and ITIM domains (Tigit), and T cell immunoglobulin and mucin domain-containing protein 3 (Tim-3) (12–14). Blockade of these immune checkpoints can partially reverse T cell exhaustion, as shown by the remarkable recent successes of cancer immunotherapy (15, 16).
Specifically, Tim-3 is a transmembrane protein expressed on the surface of several immune cell types, notably exhausted CD4+ and CD8+ T cells (14, 17). Although the precise mechanisms are unknown, Tim-3 inhibits or stimulates signaling through the TCR in a ligand-dependent manner (18, 19). In addition to regulating T cell activation, Tim-3 exhibits antiviral activity along with other members of the Tim protein family (Tim-1 and Tim-4) (20). To achieve this, the conserved N-terminal variable immunoglobulin-like (IgV) domain of these proteins binds phosphatidylserine (PS), a component of enveloped viral particles, thereby inhibiting virion release (20). Interestingly, Tim-3 is upregulated on the surface of human immunodeficiency virus type 1 (HIV-1) productively infected cells (21). However, the HIV-1 molecular determinants modulating cell surface Tim-3 expression on infected cells remain largely unknown.
HIV-1 encodes four accessory proteins, Vpu, Nef, Vif, and Vpr, which often hijack conserved host cellular trafficking pathways to sequester or degrade host cellular proteins (22, 23). Among these accessory proteins, Vpu enhances viral release and replication by inducing the degradation of several host proteins. For example, Vpu mediates the downregulation of bone marrow stromal antigen 2 (BST-2) (tetherin) from the cell surface, thereby allowing newly synthesized virions to egress from the cell surface (24, 25). Vpu also prevents superinfection by directing the degradation of CD4 via the endoplasmic reticulum-associated degradation (ERAD) pathway (26). Furthermore, Vpu impairs T cell activation by interfering with the activation of nuclear factor kappa light chain enhancer of activated B cells (NF-κB) (27, 28). Vpu has also been implicated in the downregulation of the costimulatory molecule CD28 from the surface of CD4+ T cells, suggesting that Vpu also regulates cell surface levels of receptors that affect T cell activation (29).
Since Vpu modulates cell surface levels of multiple receptors critical for viral release and T cell activation, we hypothesized that Vpu could regulate cell surface levels of Tim-3. Here, we show that HIV-1 Vpu from the transmitted/founder (TF) viruses CH58 and CH77 downregulates cell surface levels of Tim-3 in primary CD4+ cells, attenuating the overall increasing effect of HIV-1 on Tim-3 cell surface expression. This Tim-3 antagonism is dependent on the Vpu transmembrane domain (TMD). Using immunofluorescence microscopy, we provide evidence that Vpu redirects Tim-3 to the trans- Golgi network (TGN) and to Rab 5-positive (Rab 5+) vesicles. Furthermore, Tim-3 depletion or blockade enhanced viral replication in primary CD4+ T cells.
RESULTS
HIV-1 accessory protein Vpu downregulates Tim-3 from the surface of HIV-1-infected cells.We first evaluated the levels of expression of different Tim family proteins on the surface of activated primary CD4+ T cells. Accordingly, primary CD4+ T cells from HIV-negative donors were activated with phytohemagglutinin-L (PHA)/IL-2 and mock infected or infected with the TF HIV-1 strain CH58. Cell surface levels of Tim-1, Tim-3, and Tim-4 were then evaluated by flow cytometry. Of the three Tim family members tested, only Tim-3 was significantly expressed on the surface of primary CD4+ T cells (Fig. 1A and B). Since HIV-1 infection resulted in a significant increase of Tim-3 at the cell surface (Fig. 1A and B), we then sought to investigate the effects of the HIV-1 accessory protein Vpu on the cell surface expression of Tim-3. Primary CD4+ T cells were infected with primary CH58 and CH77 TF viruses expressing the wild-type (WT) Vpu protein or not (Vpu−), and Tim-3 cell surface levels were measured by flow cytometry. Cell surface levels of Tim-3 were significantly increased upon vpu deletion in both the CH58 and CH77 strains, suggesting that Vpu expression diminishes the increase in cell surface Tim-3 (Fig. 1C and D).
HIV-1 Vpu from transmitted/founder viruses downregulates the checkpoint molecule Tim-3 from the cell surface. (A and B) Primary CD4+ T cells were mock infected or infected with the transmitted/founder virus CH58 WT. At 48 h postinfection, cells were stained with antibodies recognizing Tim-1, Tim-3, Tim-4, or their isotype controls. (A) Histograms depicting representative staining of the Tim proteins on cells mock infected or infected with CH58 WT. (B) Graphs representing the mean fluorescence intensities (MFIs) of cell surface Tim family proteins obtained on mock-infected cells or on infected (p24+) cell populations. These data were obtained in 3 independent experiments using cells from 3 different HIV-negative donors. (C to G) Primary CD4+ T cells were mock infected or infected with the transmitted/founder viruses CH58 and CH77, either WT or defective for Vpu expression. At 48 h postinfection, cells were surface stained with antibodies recognizing the checkpoint molecule Tim-3, PD-1, Tigit, or Lag-3. (C) Histograms depicting representative staining of the inhibitory receptors on the surface of cells infected with the different CH58 virus variants. (D to G) Graphs representing the MFIs obtained on mock cells or on the infected (p24+) population for cell surface checkpoint molecules Tim-3 (D), PD-1 (E), Tigit (F), and Lag-3 (G). These data were obtained in 5 independent experiments using cells from 5 different HIV-negative donors. Error bars indicate means ± standard errors of the means (SEM). Statistical significance was tested using a paired t test or a Wilcoxon signed-rank test based on statistical normality (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, nonsignificant).
Knowing that Vpu regulates cell surface levels of Tim-3, we next tested if Vpu also modulated other well-established inhibitory immune checkpoints (PD-1, Lag-3, and Tigit). In agreement with previous reports (21), PD-1 and Lag-3 cell surface levels were increased upon HIV-1 infection, but no differential effects were observed upon vpu deletion (Fig. 1C, E, and G). Therefore, these results suggest that PD-1 and Lag-3 upregulation upon HIV-1 infection is independent of Vpu expression. Furthermore, we observed that the cell surface levels of Tigit were unchanged after HIV infection (Fig. 1C and F). Therefore, the HIV-1 accessory protein Vpu downregulates cell surface levels of Tim-3, but not PD-1, Lag-3, or Tigit, in primary CD4+ T cells.
The transmembrane domain of Vpu is required for Tim-3 downregulation.To characterize the relationship between Vpu and Tim-3, we next sought to determine which residues of Vpu are responsible for Tim-3 antagonism. Since the TMD of Vpu is required for the downregulation of several transmembrane proteins, including BST-2; NK, T cell, B cell antigen (NTB-A); poliovirus receptor (PVR); CCR7; HLA-C; and CD62L (25, 30–35), we hypothesized that Vpu’s TMD is also required for Vpu-mediated Tim-3 downregulation. To test this, site-directed mutagenesis was used to introduce point mutations in two highly conserved residues of Vpu’s TMD critical for the downregulation of multiple Vpu substrates (A14L/A18L in CH58 and A15L/A19L in CH77 [36]). Primary CD4+ T cells from HIV-negative donors were infected with WT CH58 or CH77 TF viruses, isogenic viruses encoding the mutant Vpu proteins (CH58 Vpu A14L/A18L or CH77 Vpu A15L/A19L), or vpu-deleted (Vpu−) viruses. Subsequently, cell surface levels of Tim-3 were measured by flow cytometry. Consistent with our previous results, infection with Vpu− viruses resulted in increased cell surface levels of Tim-3 (Fig. 2A and B, CH58 Vpu− and CH77 Vpu−). Mutation of the TMD of Vpu abrogated Vpu’s capacity to downregulate Tim-3, indicating that Vpu’s TMD is required for Tim-3 downregulation (Fig. 2A and B, CH58 Vpu A14L/A18L and CH77 Vpu A15L/A19L). As expected, these mutations had no effect on the capacity of Vpu to downmodulate cell surface CD4 (Fig. 2C and D). In contrast, mutations introduced into the phosphoserine motif of Vpu (S52A/S56A) abrogated Vpu-dependent CD4 downregulation but had no significant effect on Tim-3 cell surface levels compared to WT Vpu, suggesting that this domain is not required for Vpu-mediated Tim-3 downregulation (Fig. 2A to D, CH58 Vpu S52A/S56A). Albeit significant, Tim-3 downregulation by Vpu was found to be less pronounced than Vpu-mediated downregulation of CD4 and BST-2 (Fig. 2E).
The HIV-1 Vpu transmembrane domain is required for efficient downregulation of Tim-3. Primary CD4+ T cells were mock infected or infected with the transmitted/founder viruses CH58 and CH77 expressing wild-type Vpu (WT), Vpu A14L/A18L (or A15L/A19L), or Vpu S52A/S56A or defective for Vpu expression (Vpu−). At 48 h postinfection, cells were stained with antibodies recognizing the cell surface receptor Tim-3 or CD4. (A and C) Histograms depicting representative staining of Tim-3 (A) and CD4 (C) on the surface of cells infected with the different CH58 variants. (B and D) Graphs representing the MFIs obtained on mock cells or on the infected (p24+) population for Tim-3 (B) or CD4 (D). These data were obtained in 5 independent experiments using cells from 5 different HIV-negative donors. (E) Graph representing the MFIs obtained with the different CH58 Vpu variants on the infected (p24+) population relative to cells infected with CH58 Vpu− for different Vpu substrates (CD4, BST-2, and Tim-3). Error bars indicate means ± SEM. Statistical significance was tested using a paired t test (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, nonsignificant).
Vpu is in close proximity to Tim-3 within cells.As protein-protein interactions are required for the regulation of other cell surface proteins by Vpu (30), we next evaluated whether Vpu and Tim-3 can associate within cells. To this end, we used a bimolecular fluorescence complementation (BiFC) assay that has been previously described (37–39). Briefly, in our assay, BiFC requires the fusion of two proteins of interest to either half of the Venus fluorophore (VN or VC) (Fig. 3A and B). If these proteins come within close proximity, the fluorophore reconstitutes and emits a green fluorescence signal (38). However, in the absence of an interaction, green fluorescence is not observed (38). In our BiFC assay, FLAG-tagged Tim-3 and HIV-1 CH58 Vpu were expressed from plasmids containing the respective split Venus fluorophore moieties (VN and VC, respectively) (Fig. 3A and B). Accordingly, CD4+ HeLa cells were cotransfected with Tim-3-FLAG-VN and Vpu-VC, immunostained for FLAG (magenta) and Vpu (red) to determine transfected cells, and then imaged using fluorescence microscopy (Fig. 3C). We observed a positive BiFC signal, suggesting that Tim-3 and Vpu are in close proximity in cells (Fig. 3C, Tim-3 VN + Vpu VC, green). Importantly, fluorescence was not observed in nontransfected cells (Fig. 3C, NT) or when Vpu-VC (Fig. 3C, Vpu Vc) or Tim-3-FLAG-VN (Fig. 3C, Tim-3) was singly transfected and was decreased in the presence of a cytoplasmic tail-deleted Tim-3 protein (not shown), ensuring the specificity of the BiFC signal and the absence of background fluorescence (Fig. 3C). Altogether, these results demonstrate that Vpu is in close proximity to Tim-3 in cells.
HIV-1 Vpu and Tim-3 are in close proximity within cells. (A) Schematic of the bimolecular fluorescence complementation (BiFC) assay. Each half of the Venus fluorophore is fused to two proteins of interest (Tim-3 and Vpu). If the two proteins come within close proximity, the fluorophore reconstitutes, and the interaction is visualized using a fluorescence microscope. (B) DYKDDDDK-tagged (FLAG-tagged) mouse Tim-3 (Tim-3-FLAG) and HIV-1 CH58 Vpu were cloned into plasmids containing the VN and VC halves of the Venus fluorophore, respectively. (C) Tim-3-FLAG-VN and Vpu-VC were cotransfected into CD4+ HeLa cells. Cells were fixed, permeabilized, stained with anti-FLAG antibody and Vpu antiserum, and imaged at a ×63 magnification using a confocal microscope with settings optimized for each fluorescent signal. Cells were mounted on ProLong Diamond antifade mountant with DAPI for nuclear staining (blue). Cells that were not transfected (NT) and singly transfected were included as negative controls. Representative images from 5 independent experiments are shown. Bar, 10 μm.
Vpu relocalizes Tim-3 to the trans-Golgi network and Rab 5+ vesicles.As Vpu alters the intracellular trafficking of multiple transmembrane proteins, including BST-2, NTB-A, and PVR, by inducing their sequestration in intracellular compartments, including the TGN, we next investigated whether Vpu affects the trafficking of Tim-3 in a similar manner (32, 40–42). To test this, CD4+ HeLa cells were transfected with Tim-3-FLAG-VN or cotransfected with Tim-3-FLAG-VN and Vpu-VC, fixed, and immunostained for TGN46 (a marker of the TGN) or Rab 5. We observed a striking increase in the colocalization of Tim-3 with TGN46 in the presence of Vpu (Fig. 4A and B) (Pearson’s coefficient = 0.627) relative to Tim-3 in the absence of Vpu (Fig. 4A and B) (Pearson’s coefficient = 0.322). In addition, we also observed an increased overlap of the Tim-3 and Rab 5 fluorescence signals in the presence of Vpu (Fig. 4C and D) (Pearson’s coefficient without Vpu = 0.469; Pearson’s coefficient with Vpu = 0.663). In contrast, the presence of Vpu did not affect the overlap between Tim-3 and CD63 (a marker of multivesicular bodies [MVBs]) or lysosome-associated membrane protein 1 (LAMP-1) (a marker of lysosomes) (Fig. 5). Furthermore, to ensure that the BiFC signal observed in lysosomes was not attenuated due to potential protein degradation within lysosomes, we treated cells with ammonium chloride to inhibit lysosomal acidification. Similar to our observations with untreated cells, we did not observe any difference in the colocalization of Tim-3 with LAMP-1 in the presence of Vpu relative to Tim-3 in the absence of Vpu (Fig. 5A to D). Together, these observations suggest that Vpu directs Tim-3 toward the TGN and Rab 5-positive vesicles.
Vpu localizes Tim-3 to the trans-Golgi network and Rab 5-positive vesicles. (A) Tim-3-FLAG-VN was transfected into CD4+ HeLa cells alone or was cotransfected with Vpu-VC. Cells were fixed, permeabilized, and stained with anti-FLAG and anti-TGN46 antibodies. Cells were subsequently imaged using a wide-field microscope. BiFC (green) was imaged under the FITC channel, FLAG staining was imaged under the far-red channel (modified postimaging to appear green), and staining of the various compartments was imaged under the Cy3 channel (red). Cells were mounted on DAPI-Fluoromount G for nuclear staining (blue). Representative images from 3 independent experiments are shown. Bar, 10 μm. (B) Colocalization of Tim-3 or BiFC fluorescence signals (green) with TGN46 fluorescence (red) was quantified using the JACoP plug-in in ImageJ. Results are presented as mean Pearson’s coefficients ± SEM. A two-tailed paired t test was used to evaluate differences in colocalization between Tim-3 alone and Tim-3 and Vpu with TGN46. Over 30 cells were quantified over 3 independent experiments. (C) Cells were transfected, fixed, and permeabilized as described above for panel A. Cells were immunostained with FLAG and Rab 5 antibodies and imaged as described above for panel A. (D) Colocalization of Tim-3 or BiFC fluorescence signals (green) with Rab 5 fluorescence (red) was quantified and analyzed as described above for panel B (*, P < 0.05; **, P < 0.01).
Vpu does not relocalize Tim-3 to lysosomes or multivesicular bodies. (A) Tim-3-FLAG-VN was transfected into CD4+ HeLa cells alone or was cotransfected with Vpu-VC. Cells were fixed, permeabilized, stained with FLAG and lysosome-associated membrane protein 1 (LAMP-1) antibodies, and subsequently imaged using a wide-field microscope. BiFC (green) was imaged under the FITC channel, FLAG staining was imaged under the far-red channel (modified postimaging to appear green), and LAMP-1 staining was imaged under the Cy3 channel (red). Cells were mounted on DAPI-Fluoromount G for nuclear staining (blue). Representative images from 3 independent experiments are shown. Bar, 10 μm. (B) Colocalization of Tim-3 or BiFC fluorescence signals (green) with LAMP-1 fluorescence (red) was quantified using the JACoP plug-in in ImageJ. Results are presented as mean Pearson’s coefficients ± SEM. A two-tailed paired t test was used to evaluate differences in colocalization between Tim-3 alone and Tim-3:Vpu with LAMP-1. Over 30 cells were quantified over 3 independent experiments. (C) Cells were transfected as described above for panel A. Three hours before fixing, 25 mM ammonium chloride (NH4Cl) in Dulbecco’s modified Eagle medium (DMEM) was added to inhibit lysosomal acidification. Cells were then imaged as described above for panel A. (D) Colocalization of Tim-3 or BiFC fluorescence signals (green) with LAMP-1 fluorescence (red) was quantified and analyzed as described above for panel B. (E) Cells were transfected, fixed, and permeabilized as described above for panel A. Cells were immunostained with FLAG and CD63 antibodies and imaged as described above for panel A. (F) Colocalization of Tim-3 or BiFC fluorescence signals (green) with CD63 fluorescence (red) was quantified and analyzed as described above for panel B (ns, nonsignificant).
Tim-3 expression modulates HIV-1 replication in a Vpu-dependent manner.We next sought to determine the consequence of Tim-3 expression on HIV-1 replication. To assess this, primary CD4+ T cells mock infected or infected with CH58 TF virus were electroporated with Tim-3-targeting or nontargeting (NT) small interfering RNA (siRNA). Reduced Tim-3 expression (Fig. 6A and B) resulted in a modest but significant enhancement of viral replication (Fig. 6D). Furthermore, since Tim family proteins were shown to restrict HIV-1 release through their interaction with PS (20) present on the outer leaflet of virion lipid membranes (43), we evaluated the effect of blocking the Tim-3:PS interaction using an antibody (Ab) known to specifically block PS binding by Tim-3 (anti-Tim-3, clone F38-2E2) (44). Briefly, primary CD4+ T cells were infected with CH58 TF viruses or CH58 TF viruses not expressing Vpu (CH58 TF Vpu−), and HIV-1 replication was measured by assessing the percentage of p24+ cells every 24 h over a 4-day period postinfection. We observed that Tim-3 blockade moderately increased the viral replication of WT virus after 4 days of ongoing replication (Fig. 7A). In the absence of Vpu, Tim-3 blockade increased viral replication after only 3 days of ongoing replication (Fig. 7B). The presence of blocking antibodies on cells treated with anti-Tim-3 was confirmed by staining with an anti-mouse IgG secondary antibody (Fig. 7C). Furthermore, treatment of WT-infected cells with Tim-3-blocking antibodies decreased the overall level of cell surface Tim-3 compared to that in cells treated with the isotype control (Fig. 7D and E). Inversely, cells infected with a vpu-defective virus displayed an increased amount of cell surface Tim-3 upon Tim-3 blockade (Fig. 7D and E). Overall, these results indicate that Tim-3 expression modestly restricts HIV-1 replication in primary CD4+ T cells, but whether an interaction with PS is required for this phenotype remains to be established.
Reduction of Tim-3 expression enhances HIV-1 replication in primary CD4+ T cells. Primary CD4+ T cells were mock infected or infected with CH58 WT. At 16 h postinfection, infected cells were electroporated with siRNAs targeting Tim-3 mRNA or nontargeting (NT) siRNAs. (A) At 72 h postinfection, cells were stained with anti-Tim-3 or its matched IgG isotype control. The histogram depicts representative Tim-3 staining upon siRNA electroporation on mock-infected cells or infected (p24+) cells. (B and C) Graphs representing the MFIs obtained on mock-infected cells or on the infected (p24+) population using anti-Tim-3 (B) or its matched IgG isotype control (C). (D) The percentage of infected (p24+) cells was evaluated at 0 h, 24 h, 48 h, and 72 h postinfection. These data were obtained in 5 independent experiments using cells from 5 different HIV-negative donors. Error bars indicate means ± SEM. Statistical significance was tested using a paired t test (*, P < 0.05; **, P < 0.01; ns, nonsignificant).
Tim-3 blockade increases HIV-1 replication in primary CD4+ T cells. Primary CD4+ T cells were infected with CH58 viruses, either WT or defective for Vpu expression. Following spin infection, cells were treated with a Tim-3-blocking antibody or its matched IgG isotype control (10 μg/ml). (A and B) The percentage of infected (p24+) cells was evaluated at 0 h, 24 h, 48 h, 72 h, and 96 h postinfection. (C to E) At 96 h postinfection, cells were stained with anti-mouse secondary antibodies that were used to measure levels of anti-Tim-3 antibody still associated at the cell surface (C) or with anti-Tim-3 followed by anti-mouse secondary antibodies (D and E). (D) Histogram depicting representative Tim-3 staining upon blocking antibody treatment on mock-infected cells or infected (p24+) cells. (C and E) Graphs representing the MFIs obtained on mock-infected cells or on the infected (p24+) population using the indicated antibodies. These data were obtained in 5 independent experiments using cells from 5 different HIV-negative donors. Error bars indicate means ± SEM. Statistical significance was tested using a paired t test (*, P < 0.05; **, P < 0.01; ns, nonsignificant).
IFN-β impairs the ability of Vpu to downregulate cell surface Tim-3.We recently reported that type I IFNs decrease Vpu’s polyfunctionality by upregulating BST-2 (35). We demonstrated that the specific occupation of Vpu’s TMD by BST-2 affects its capacity to target other transmembrane proteins, including NTB-A, PVR, and CD62L (35). Since we demonstrated that the TMD of Vpu is required for Vpu-mediated Tim-3 downmodulation, we next investigated the impact of type I IFN treatment on the capacity of Vpu to downmodulate Tim-3. Primary CD4+ T cells were infected with CH58 or CH77 TF viruses, either WT or deficient in Vpu expression. At 24 h postinfection, cells were treated or not with IFN-β, and cell surface levels of Tim-3 were monitored 24 h after IFN-β treatment by flow cytometry. As demonstrated in Fig. 8A to C, treatment with type I IFNs significantly enhanced cell surface levels of Tim-3 in cells infected (p24+) with WT viruses. In that context, Tim-3 cell surface levels were restored to the levels detected in the context of infections with isogenic Vpu-defective viruses (Fig. 8A and C). This enhancement was specific for infected (p24+) cells, indicating that type I IFNs prevent HIV-1-mediated Tim-3 downmodulation (Fig. 8B and D). In contrast, no significant increase was observed when cells were infected with a Vpu-defective virus, suggesting that type I IFNs’ enhancement of cell surface Tim-3 is dependent on Vpu expression (Fig. 8C and E). Taken together, these data suggest that type I IFNs impair the ability of Vpu to downmodulate Tim-3.
IFN-β impairs the ability of Vpu to downregulate cell surface Tim-3. Primary CD4+ T cells were mock infected or infected with the transmitted/founder virus CH58 or CH77, WT or defective for Vpu expression, and either mock treated or treated for 24 h with IFN-β. At 48 h postinfection, cells were stained with anti-Tim-3 antibody. (A) Histograms depicting representative staining with or without IFN-β treatment. (B to E) Graphs representing the MFIs obtained on mock cells, uninfected p24− cells, or infected p24+ cells (B and C) or fold changes in MFI upon IFN-β treatment (D and E) for 5 independent experiments using cells from 5 different HIV-negative donors. Error bars indicate means ± SEM. Statistical significance was tested using a paired t test (**, P < 0.01; ***, P < 0.001; ns, nonsignificant).
We next sought to extend our assessment of IFN-β-dependent Tim-3 upregulation by testing this on ex vivo-expanded infected cells obtained from five HIV-1-infected individuals. Accordingly, CD4+ T cells were isolated from HIV-1-positive donors and treated with PHA/IL-2 to expand endogenous virus. Subsequently, cells were treated with IFN-β for 24 h, and cell surface levels of Tim-3 were quantified by flow cytometry. Consistent with our results upon exogenous infection of CD4+ T cells (Fig. 8), we observed a significant increase of Tim-3 on the surface of endogenously infected CD4+ T cells upon IFN-β treatment (Fig. 9A to C). Conversely, this IFN-β-dependent increase was not observed in uninfected (p24−) cells, consistent with our previous results upon exogenous infection (Fig. 8 and Fig. 9A and B). Therefore, treatment with IFN-β impairs Vpu’s ability to downregulate cell surface levels of Tim-3 on both endogenously and exogenously infected CD4+ T cells.
IFN-β upregulates Tim-3 expression on ex vivo-expanded endogenously infected primary CD4+ T cells. Primary CD4+ T cells from five different HIV-1-infected individuals were isolated and reactivated with PHA for 36 h, followed by incubation with IL-2 to expand the endogenous virus. Upon reactivation, cells were either mock treated or treated for 24 h with IFN-β, followed by staining with anti-Tim-3 antibody or its matched isotype control. (A) Histograms depicting representative staining with or without IFN-β treatment. (B and C) Graphs representing the MFIs obtained on uninfected p24− cells or infected p24+ cells (B) or fold changes in MFI upon IFN-β treatment (C) using cells from 5 different HIV-positive donors. Error bars indicate means ± SEM. Statistical significance was tested using repeated measures (RM) one-way analysis of variance (ANOVA) (B) or a paired t test (C) (*, P < 0.05; ns, nonsignificant).
The effect of IFN-β on Tim-3 cell surface levels is conserved in SIVcpz but not SIVmac.To determine if IFN-β inhibition of Tim-3 cell surface downregulation is a conserved function among lentiviruses, we infected cells with CH58TF WT viruses or simian immunodeficiency viruses (SIVs) naturally expressing Vpu (SIVcpzPtt EK505) or not (SIVmac239) and measured cell surface Tim-3 levels using flow cytometry. We observed that IFN-β treatment enhanced cell surface levels of Tim-3 in SIVcpzPtt EK505-infected cells but not in cells infected with SIVmac239, which lacks Vpu (Fig. 10A and B). We noticed that cells infected with SIVmac239 displayed more cell surface Tim-3 initially than cells infected with CH58 TF or SIVcpzPtt EK505; however, the addition of IFN-β abrogated this difference (Fig. 10B). Overall, this suggests that Vpu’s ability to downregulate Tim-3 is a conserved function in the HIV-1/SIVcpz lineage.
IFN-β upregulates Tim-3 expression on cells infected with SIVcpzPtt but not those infected with SIVmac. Primary CD4+ T cells were either mock infected or infected with HIV-1 CH58 TF, SIVcpzPtt EK505, or SIVmac239. At 48 h postinfection, cells were either mock treated or treated for 24 h with IFN-β, followed by staining with anti-Tim-3 antibody or its matched isotype control. (A) Histograms depicting representative staining with or without IFN-β treatment. (B) Graphs representing the MFIs obtained on uninfected p24− cells (mock), infected p24+ cells (CH58 TF and SIVcpzPtt EK505), or infected p27+ cells (SIVmac239) using cells from 4 different HIV-negative donors. Error bars indicate means ± SEM. Statistical significance was tested using a paired t test (*, P < 0.05; ns, nonsignificant).
DISCUSSION
Here, we identify Tim-3 as a new substrate downregulated by the HIV-1 accessory protein Vpu. We demonstrate that the TMD of Vpu is critical for Tim-3 downregulation and that Vpu comes in close proximity to Tim-3 within the cell and alters its subcellular localization. Furthermore, our data suggest that Vpu-mediated Tim-3 downregulation affects viral replication and that Vpu-mediated Tim-3 antagonism is a conserved function among lentiviruses.
A hallmark of viral accessory protein-host cell receptor interactions is the ability of the viral protein to reroute these receptors within the cell. We found that Vpu directs Tim-3 to both the TGN and Rab 5+ vesicles. This increased localization of Tim-3 within the TGN and Rab 5+ vesicles in the presence of Vpu could be due to a decreased rate of anterograde transport of Tim-3 from the TGN to the cell surface, which would explain its increased colocalization with the TGN. Interestingly, a similar mechanism was observed in the Vpu-mediated antagonism of NTB-A, a coactivating NK cell receptor (42). Bolduan et al. demonstrated that Vpu decreases, but does not completely inhibit, the transport of newly synthesized NTB-A from the TGN to the cell surface (42). However, it is also possible that Vpu is affecting the retrograde transport of Tim-3. Therefore, additional studies will be required to confirm the complete trafficking itinerary undertaken by Tim-3 in the presence of Vpu.
Vpu has been shown to downregulate host cellular proteins using multiple distinct mechanisms. For example, Vpu has been shown to sequester BST-2 within the TGN, thereby preventing its localization to sites of viral egress (41, 45). In addition, Vpu uses its phosphoserine motif (S52/56) to induce the proteasomal degradation of BST-2 (46). However, as cell surface levels of Tim-3 were not restored following mutation of the phosphoserine motif of Vpu (S52/56A), it is not likely that the recruitment of the SCFβTrCP E3 ubiquitin ligase is required for Tim-3 downregulation. Furthermore, as Tim-3 does not localize to lysosomes in the presence of Vpu, it is unlikely that Vpu is inducing its lysosomal degradation. Taken together, these results suggest that Vpu does not induce the degradation of Tim-3 and downregulates it primarily by sequestration.
Recently, it has been suggested that Tim-3 on the surface of HIV-1-infected monocyte-derived macrophages impairs the egress of HIV-1 virions (20). Furthermore, Tim-1, a transmembrane protein closely related to Tim-3, has also been demonstrated to impair HIV-1 release in transfected HEK 293T cells by tethering virions to the cell surface, similar to the effect of BST-2 (20). It is believed that Tim-1 and Tim-3 prevent HIV-1 release by binding PS, a component of the HIV-1 envelope, via the conserved IgV domain (20). A recent report by Li et al. demonstrated that the NL4-3 Nef protein induces the internalization of Tim-1, demonstrating that HIV-1 accessory proteins act on different Tim family receptors (47). Furthermore, recent evidence suggests that Nef counters Tim-3-mediated inhibition of viral release from the surface of macrophages; however, whether Nef modulates Tim-3 expression levels in CD4+ T cells remains to be determined (47). It is unsurprising that HIV-1 encodes two accessory proteins, Nef and Vpu, that both antagonize the Tim-3 restriction factor; indeed, Nef and Vpu have been previously demonstrated to antagonize the same protein, such as CD4 and CD28 (29, 48).
This work exemplifies the complex interactions between HIV-1, T cell immune checkpoints, and innate immunity and raises questions that will need to be addressed in further studies. Like other inhibitory receptors, Tim-3 thus has multifaceted effects: its expression correlates with disease progression (49), and it can inhibit HIV-specific T cell responses (14), dampen viral replication, and potentially facilitate the persistence of reservoirs with a proclivity for viral transcription (50). Determination of the net impact of any given coreceptor pathway therefore likely requires in vivo interventions, such as in the nonhuman primate (NHP) model of simian-human immunodeficiency virus (SHIV) infection or the humanized mouse model of HIV infection (51–54).
Why HIV-1 stimulates Tim-3 expression and then uses its Vpu accessory protein to moderately downregulate it from the cell surface is unclear. It is possible that the upregulation of Tim-3 is a natural T cell response to viral infections leading to apoptosis (55), and therefore, it must be downregulated to allow completion of the replication cycle. Alternatively, Vpu-mediated Tim-3 downregulation might be required to enhance viral production since its expression has been shown to block viral release (20). Distinctive characteristics that we identified for Tim-3, compared to the other inhibitory receptors examined, include its specific downregulation by Vpu and the selective counterregulation of this effect by type I IFN. It is notable that type I IFNs specifically modulate Tim-3 on infected CD4+ T cells, whereas common γ-chain cytokines broadly upregulate PD-1 on T cell subsets (56). The impact of these mechanisms on viral replication will likely depend on the tissue microenvironment. Indeed, as pleiotropic innate antiviral cytokines, type I IFNs are key for the control of acute infection, but prolonged signaling can lead to waning effects and/or immune dysfunction in chronic infections, including SIV (57, 58; reviewed in reference 59). Tissue studies will help clarify these interactions in anatomic compartments.
Overall, our results show that Vpu downregulates Tim-3 from the surface of primary CD4+ T cells and that the TMD of Vpu is involved in this conserved phenomenon by directing Tim-3 to the TGN. Furthermore, we provide evidence that cell surface Tim-3 impairs the HIV-1 replication capacity in primary CD4+ T cells. Additional studies are required to understand the role of Tim-3 expression in modulating HIV-1 replication and reservoir dynamics in vivo.
MATERIALS AND METHODS
Ethics statement.Informed consent was obtained from all subjects according to ethical guidelines of CRCHUM in accordance with Institutional Review Board approval.
Cell culture and isolation of primary cells.HEK 293T cells (ATCC, Manassas, VA) were grown as previously described (60). Primary human CD4+ T cells were isolated, activated, and cultured as previously described (60, 61). Briefly, peripheral blood mononuclear cells (PBMCs) were obtained by Ficoll density gradient centrifugation from whole-blood samples obtained from healthy donors. CD4+ T lymphocytes were purified from resting PBMCs by negative selection using immunomagnetic beads according to the manufacturer’s instructions (StemCell Technologies, Vancouver, BC, Canada). CD4+ T cells were activated with phytohemagglutinin-L (PHA) (10 μg/ml) for 48 h and maintained in RPMI 1640 complete medium supplemented with recombinant IL-2 (rIL-2) (100 U/ml).
CD4+ HeLa cells (ATCC) (62) were grown in Dulbecco’s modified Eagle medium (DMEM) containing 4 mM l-glutamine, 4,500 mg/liter glucose, and sodium pyruvate (HyClone, Logan, UT) and supplemented with 10% fetal bovine serum (Wisent, St. Bruno, QC, Canada) and 1% penicillin and streptomycin (HyClone). Cells were grown and subcultured according to the supplier’s suggestions.
DNA constructs.Transmitted/founder infectious molecular clones (IMCs) of patients CH58 and CH77 were inferred, constructed, and biologically characterized as previously described (63–67). Mutations were introduced using a QuikChange II XL site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA) as previously described (68, 69). Briefly, the vpu-defective CH58 IMC construct was generated by introducing a premature stop codon at position 2 of its open reading frame. All mutations were confirmed by sequencing. SIV IMC constructs (SIVmac239 and SIVcpzPtt EK505) were previously described (70, 71).
For microscopy, N-terminal DYKDDDK-tagged mouse Tim-3 (Tim-3-FLAG) was provided by Lawrence Kane (University of Pittsburgh, Pittsburgh, PA) and PCR amplified using primers with ApaI and BamHI cut sites. Tim-3-FLAG was cloned into a pN1 backbone (Clontech, Mountain View, CA) containing the VN-173 portion of the Venus fluorophore at the C terminus using ApaI and BamHI enzymes (72). HIV-1 CH58 Vpu was obtained by GeneArt gene synthesis (Invitrogen, Rockford, IL) and cloned into a pcDNA3.1(−) backbone (Life Technologies, Carlsbad, CA) containing the VC-155 portion of the Venus fluorophore at the C terminus using EcoRI and BamHI enzymes (Vpu-VC) (72). All cloning and mutations were confirmed by sequencing (London Regional Genomics Centre, London, ON, Canada).
Viral production, infections, and ex vivo amplification.To achieve similar levels of infection among all viruses, vesicular stomatitis virus G (VSVG)-pseudotyped HIV-1 viruses were produced in HEK 293T cells and titrated as previously described (73). Viruses were then used to achieve a level of infection of ∼10% of the total primary CD4+ T cells at 48 h postinfection. PHA/IL-2-activated primary CD4+ T cells from healthy HIV-1-negative donors were spinoculated at 800 × g for 1 h in 96-well plates at 25°C. In order to expand endogenously infected CD4+ T cells, primary CD4+ T cells were isolated from PBMCs from HIV-1-infected individuals. Purified CD4+ T cells were activated with PHA at 10 μg/ml for 36 h and then cultured for 6 to 8 days in RPMI 1640 complete medium supplemented with rIL-2 (100 U/ml) (61).
Antibodies.The following Abs were used as primary Abs for cell surface staining of primary CD4+ T cells: allophycocyanin (APC)-anti-human CD366 (Tim-3) (clone F38-2E2; BioLegend, San Diego, CA), APC-anti-human CD279 (PD-1) (clone EH12.2H7; BioLegend), APC-anti-human Tigit (clone A15153G; BioLegend), Alexa Fluor 647-anti-human CD223 (Lag-3) (clone 11C3C65; BioLegend), phycoerythrin (PE)-cyanine 7 (Cy7)-anti-human CD317 (BST-2) (clone RS38E; BioLegend), mouse anti-human CD365 (Tim-1) (clone 1D12; BioLegend), mouse anti-human Tim-4 (clone 9F4; BioLegend), and mouse anti-CD4 (clone OKT4; eBioscience, San Diego, CA). APC-mouse IgG1 (clone MOPC-21; BioLegend), APC-mouse IgG2 (clone MOPC-173; BioLegend), and Alexa Fluor 647-mouse IgG1 (clone MOPC-21; BioLegend) were used as matched IgG isotype controls. The following Ab was used for blockade experiments: mouse anti-human CD366 (Tim-3) (clone F38-2E2; BioLegend) or its matched IgG1 isotype control (clone MOPC-21; BioLegend). Goat anti-mouse antibodies precoupled to Alexa Fluor 647 (Invitrogen) were used as secondary antibodies for uncoupled mouse antibodies during flow cytometry.
The following Abs were used as primary Abs for microscopy: rat anti-FLAG (1:400) (clone L5; BioLegend), rabbit anti-lysosome-associated membrane protein 1 (LAMP-1) (1:200) (Invitrogen), rabbit anti-TGN integral membrane protein 2 (TGN46) (1:100) (Sigma-Aldrich, St. Louis, MO), rabbit anti-Rab 5 (1:200) (clone C8B1; Cell Signaling Tech, Danvers, MA), and mouse anti-CD63 (1:200) (clone H5C6; Developmental Studies Hybridoma Bank, University of Iowa).
The following Abs were used as secondary Abs for microscopy (Jackson ImmunoResearch, West Grove, PA): donkey anti-rat Alexa Fluor 647 (1:500), donkey anti-rabbit Cy3-conjugated Ab (1:400 [LAMP-1, TGN46, and Vpu] and 1:1,000 [Rab 5]), and donkey anti-mouse Cy3-conjugated Ab (1:400).
Vpu antiserum production.The immunogen (EMGHHAPWDVDDL) was designed by MediMabs (Montreal, QC, Canada) to maximize immunogenicity and structural availability and minimize nonspecific signals. It was synthesized and coupled with keyhole limpet hemocyanin (KLH) for immunization through the addition of an N-terminal cysteine. Two New Zealand White rabbits were immunized using MediMabs’ 77-day Canadian Council on Animal Care (CCAC)-accredited protocol. The first immunization was done using complete Freund’s adjuvant followed by 4 immunizations with incomplete Freund’s adjuvant. Rabbits were used solely for this project and were sacrificed by total exsanguination. Blood was processed, and serum was used directly without purification at a dilution of 1:100 for microscopy.
Flow cytometry analysis of cell surface staining.Cell surface staining was performed as previously described (61, 73). Binding of Abs to cell surface Tim-1 (5 μg/ml), Tim-3 (7.5 μg/ml), Tim-4 (5 μg/ml), PD-1 (5 μg/ml), Tigit (1.5 μg/ml), Lag-3 (2.5 μg/ml), and CD4 (0.5 μg/ml) was performed at 48 h postinfection. Infected cells were stained intracellularly for HIV-1 p24 (or p27 for SIVmac239) using the Cytofix/Cytoperm fixation/permeabilization kit (BD Biosciences, Mississauga, ON, Canada) and a fluorescent anti-p24 monoclonal Ab (mAb) (PE-conjugated anti-p24, clone KC57; Beckman Coulter/Immunotech, Brea, CA) or a fluorescent anti-p27 mAb (Alexa Fluor 488-conjugated anti-p27, clone 2F12). The percentage of infected cells (p24+) was determined by gating the living cell population using Aqua Vivid viability dye staining (Thermo Fisher Scientific, Waltham, MA). Samples were acquired on an LSRII cytometer (BD Biosciences), and data analysis was performed using FlowJo vX.0.7 (TreeStar, Ashland, OR).
Type I IFN treatments.IFN-β (Rebif; EMD Serono Inc., Mississauga, ON, Canada) (74) was added to the cells at 1 ng/ml, at 24 h postinfection, 24 h before cell surface staining, as described previously (35).
siRNA electroporation.Primary CD4+ T cells, mock infected or infected with CH58, were electroporated with pools of 4 siRNAs to silence Tim-3 expression (ON-TARGETplus human HAVCR2 siRNA-SMARTpool; Dharmacon, Lafayette, CO) (Table 1) or with pools of nontargeting (NT) siRNA (ON-TARGETplus nontargeting siRNA pool; Dharmacon). Infected or mock-infected primary CD4+ T cells were resuspended at a concentration of 5 × 107 cells/ml in Opti-MEM medium (Invitrogen) and transferred to an electroporation cuvette (Harvard Apparatus, Holliston, MA). Pools of NT siRNA or siRNA targeting Tim-3 sequences were added to the cells (150 pmol/3 × 106 cells). Cells were electroporated at 250 V for 2 ms using a BTX Gemini X2 electroporation system (Harvard Apparatus) and resuspended in RPMI 1640 complete medium supplemented with rIL-2 (100 U/ml). Cells were further cultured until day 3 postinfection, where Tim-3 knockdown was found to be optimal.
ON-TARGETplus siRNA SMARTpool sequences used in electroporation assays
Tim-3 blockade replication assay.Primary CD4+ T cells were mock infected or infected with CH58 WT or Vpu− viruses. Following spinoculation, cells were treated with Tim-3-blocking Ab or its matched IgG isotype control (10 μg/ml), and the percentage of infection was evaluated at 24 h, 48 h, 72 h, and 96 h postinfection using p24 staining. At 96 h postinfection, cells were stained with anti-mouse secondary Abs to evaluate the remaining level of blocking Ab bound at the surface of infected or mock-infected cells. Cells were also stained with anti-Tim-3 Ab followed by anti-mouse secondary Abs to evaluate the overall amount of cell surface Tim-3 proteins.
Microscopy.HeLa cells were seeded onto coverslips at 5 × 105 cells/coverslip. Twenty-four hours later, cells were transfected with 0.5 μg of each plasmid using the PolyJet transfection reagent (FroggaBio, Toronto, ON, Canada), according to the supplier’s protocol. For experiments requiring the inhibition of lysosomal acidification, 25 mM ammonium chloride in DMEM was added for 3 h prior to fixing. At 24 h posttransfection, the BiFC fluorophore matured at 22°C for 30 min, and cells were fixed for 20 min in 4% paraformaldehyde. Immunostaining was performed as described previously (75). Briefly, cells were blocked in 5% bovine serum albumin (BSA; Tocris Bioscience, Bristol, UK) in phosphate-buffered saline (PBS) containing 0.01% Triton X-100 (blocking buffer) for 2 h and incubated with primary Abs and secondary Abs for 2 h each. Abs were diluted in blocking buffer. Coverslips were mounted onto slides with Fluoromount-G containing 4′,6-diamidino-2-phenylindole (DAPI; SouthernBiotech, Birmingham, AL) and imaged on a Leica DMI6000 B wide-field microscope (Leica Microsystems, Wetzlar, Germany) at a ×63 or ×100 magnification (numerical aperture [NA], 1.4) using the fluorescein isothiocyanate (FITC), Cy3, Cy5, and DAPI filter settings and a Photometrics Evolve 512 Delta electron-multiplying charge-coupled device (EM-CCD) camera (Photometrics, Tucson, AZ). For confocal microscopy, coverslips were prepared as described above, mounted onto slides with ProLong Diamond antifade mountant with DAPI (Thermo Fisher Scientific), and imaged on a Leica TCS SP8 confocal laser scanning microscope at a ×63 magnification (NA, 1.4) using settings for the following fluorophores: DAPI (405 nm), enhanced green fluorescent protein (eGFP) (488 nm), Cy3 (552 nm), and Alexa Fluor 647 (638 nm).
Microscopy analysis.For colocalization analysis, images were deconvolved using the Advanced Fluorescence Deconvolution application on the Leica Application Suite software (Leica). Colocalization analysis was conducted on a minimum of 30 cells over 3 independent experiments using Pearson’s coefficient from the ImageJ plug-in JACoP, as described previously (76).
Statistical analysis.Statistics were analyzed using GraphPad Prism version 8.0.2 (GraphPad, San Diego, CA, USA). Data sets were tested for statistical normality, and this information was used to apply the appropriate statistical test. For microscopy analysis, unpaired, two-tailed t tests were used. Significance values are indicated in the figures (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001).
ACKNOWLEDGMENTS
We thank the BSL3 facility from CRCHUM, Dominique Gauchat from the CRCHUM Flow Cytometry Platform for technical assistance, Mario Legault from the FRQS HIV network for cohort coordination, and Alexa Galbraith for technical assistance. We thank MediMabs for their scientific and technical support during the development of the Vpu antiserum.
This work was supported by an operating grant from the Canadian Institutes of Health Research (CIHR) to J.D.D. (CIHR project grant 389413) and by infrastructure grants from the Canadian Foundation for Innovation and The University of Western Ontario. This work was also supported by CIHR Foundation grant 352417 and NIH grant R01 AI148379 to A.F., by an American Foundation for AIDS Research (amfAR) Mathilde Krim Fellowship in Basic Biomedical Research to J.R., and by CIHR project grant 377124 to D.E.K. J.P. is the recipient of a CIHR doctoral fellowship. S.P. and S.J.D.N. are supported by Wellcome Trust senior research fellowship WT098049AIA. D.E.K. is supported by a merit award of the Quebec Health Research Fund (FRQS). F.K. is funded by the DFG. A.F. is the recipient of a Canada Research Chair on Retroviral Entry (number RCHS0235-950-232424). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
FOOTNOTES
- Received 29 November 2019.
- Accepted 12 January 2020.
- Accepted manuscript posted online 15 January 2020.
- Copyright © 2020 American Society for Microbiology.
