TIGIT Marks Exhausted T Cells, Correlates with Disease Progression, and Serves as a Target for Immune Restoration in HIV and SIV Infection

HIV infection induces phenotypic and functional changes to CD8+ T cells defined by the coordinated upregulation of a series of negative checkpoint receptors that eventually result in T cell exhaustion and failure to control viral replication. We report that effector CD8+ T cells during HIV infection in blood and SIV infection in lymphoid tissue exhibit higher levels of the negative checkpoint receptor TIGIT. Increased frequencies of TIGIT+ and TIGIT+ PD-1+ CD8+ T cells correlated with parameters of HIV and SIV disease progression. TIGIT remained elevated despite viral suppression in those with either pharmacological antiretroviral control or immunologically in elite controllers. HIV and SIV-specific CD8+ T cells were dysfunctional and expressed high levels of TIGIT and PD-1. Ex-vivo single or combinational antibody blockade of TIGIT and/or PD-L1 restored viral-specific CD8+ T cell effector responses. The frequency of TIGIT+ CD4+ T cells correlated with the CD4+ T cell total HIV DNA. These findings identify TIGIT as a novel marker of dysfunctional HIV-specific T cells and suggest TIGIT along with other checkpoint receptors may be novel curative HIV targets to reverse T cell exhaustion.


Introduction
During chronic viral infections, high antigenic loads continually stimulate T cells leading to progressive loss of function termed "T cell exhaustion" [1]. Throughout this period, T cells increase expression of several inhibitory immune receptors that raise the threshold for activation, resulting in suppressed immune responses. While Programmed Death Receptor-1 (PD-1) was one of the earliest surface markers of immune exhaustion identified [2][3][4][5][6][7], we have shown that the surface glycoprotein, T cell immunoglobulin-and mucin domain-containing molecule (Tim)-3, defines a state of T cell exhaustion with diminished proliferative and cytokine capacities in chronic viral infection [8,9]. Thus, the upregulation of these and other negative checkpoints receptors may serve as potential targets for the reversal of T cell exhaustion.
Indeed, blocking the interaction of T cell negative checkpoint receptor pathways using targeted reagents against PD-1/Programmed Death-Ligand 1 (PD-L1), Tim-3, Lymphocyte-activation gene 3 (Lag-3) and CD160 has shown promise in reversing CD8 + T cell exhaustion [7,8,[10][11][12]. Reagents targeting many of these receptors are rapidly advancing in the clinic and are showing efficacy in the control of viral infectious disease [13] as well as anti-tumor immunity [14][15][16][17][18][19]. A single dose of an antibody against PD-1 led to sustained clearance of hepatitis C virus infection in a small subset of individuals [13]. Blockade of the PD-1/PD-L1 axis in vivo demonstrated efficacy in restoring simian immunodeficiency virus (SIV)-specific T cell and humoral immunity, and led to a reduction of SIV viremia and in immune activation. However, this did not completely control virus, suggesting that additional therapies are needed. Importantly, not all features of the exhausted T-cells are restored by interfering with single pathways [2][3][4]8,20]. Synergistic simultaneous dual blockade has yielded more promising responses suggesting these co-inhibitory molecules are non-redundant [10,19,21,22].
T cell immunoreceptor with immunoglobulin and ITIM domains (TIGIT) is a recently described immune checkpoint receptor that belongs to the CD28 family and contains an extracellular IgV domain, a transmembrane domain, and a cytoplasmic tail containing two-immunoreceptor tyrosine-based inhibitory motif (ITIM) [23]. It has been reported to be expressed on natural killer (NK) cells, CD8 + T cells and CD4 + T cell subsets [23] and is induced upon activation [23][24][25][26][27]. TIGIT competes with DNAM-1, a co-stimulatory molecule, and TACTILE, a co-inhibitory molecule, for the poliovirus receptor (PVR) a member of the nectin family of adhesion molecules that is expressed on dendritic cells (DCs) [23,24,28]. Several murine and Resources (NCCR) grant P51 OD011092 and the NIAID grant P30 AI027763. This Funding was also made possible by collaborative efforts of the Office of AIDS Research, the NIAID, and the International AIDS Society (LCN). Additional support was provided in part with federal funds from the NIAID and NIH contract Number NOI-AI-85341. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the NIH. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: I hereby declare that MM (Scientist, Immuno-Oncology) and AJK (Vice president, Immuno-Oncology) are employed by Bristol-Myers Squibb, and have no significant competing financial, professional or personal interests that might have influenced the performance or presentation of the work described in this manuscript. This does not alter our adherence to all PLOS Pathogens policies on sharing data and materials.
human studies strongly suggest that TIGIT is a negative modulator of T cell and NK cell function [25,[29][30][31]. A number of plausible mechanisms exist by which TIGIT can mediate inhibition of T and NK cell activation. Signaling through the TIGIT/PVR pathway with the standard recruitment of phosphatases via the intracellular ITIM domain of TIGIT can curtail T cell and NK cell responses [26]. This interaction has been shown to induce tolerogenic DCs to release the immunosuppressive cytokine IL-10 [25,31]. Furthermore, disruption of DNAM-1 homodimerization by TIGIT can abrogate the positive co-stimulatory signals required for activation [18]. Recently, potent anti-viral and anti-tumor responses related to enhanced CD8 + T cell effector activity were generated following synergistic dual blockade of PD-L1 and TIGIT in the mouse model of chronic lymphocytic choriomeningitis virus (LCMV) infection [18] and exvivo in patients with advanced melanoma [19].
To date, these results have not been replicated in any human viral disease, but over-expression of both TIGIT and PD-1 on virally exhausted T cells suggests that this is a promising avenue of exploration as a viable strategy to increase control or eliminate viral infections through T cell modulation. Given the potential to restore anti-HIV-specific CD8 + T cell responses by synergistic modulation of negative checkpoint receptors, we investigated the expression and function of TIGIT in HIV disease pathogenesis, and in the SIV non-human primate model of HIV/AIDS.

Expansion of TIGIT + T cells during HIV infection and correlations with clinical parameters of HIV disease progression, T cell activation, and the CD4 + cell-associated HIV DNA
We assessed the surface expression of TIGIT on T cells from peripheral blood mononuclear cells (PBMCs) from HIV-infected individuals that were either acutely infected (AI), non-controllers (NC), cART suppressed (AS), or elite controllers (EC), and compared these results to age-matched HIV-uninfected healthy donors (HD) ( Table 1; Figs 1A-1D, S1A and S1B). We observed a significant expansion in the frequency of TIGIT + CD8 + T cells in HIV-infected participants (AS; 44.95%; EC 56.7%; NC, 64.5%), even among those with viral suppression, relative to HD (median: 28.05%; Fig 1C). We observed a non-significant trend in the expansion of TIGIT + CD8 + T cells in AI (40.4%) relative to HD (Fig 1C). TIGIT + CD4 + T cells were significantly elevated among NC (24.5%) compared to HD (16.05%) (Fig 1D).
Among the HIV-infected NC, TIGIT + CD8 + T cells inversely correlated with CD4 cell counts, but not with CD8 + T cell activation or plasma viral load (Figs 1E and S1E). TIGIT + CD4 + T cells did not correlate with CD4 cell counts in NC (Fig 1F). Among EC, TIGIT + CD8 + T cells trended to correlate with CD8 + T cell activation, while frequencies of TIGIT + CD4 + T cells correlated with CD4 + T cell activation (Fig 1G and 1H). We did not observe any other significant correlations with TIGIT + T cells (S1C-S1F,S1G and S1I Fig). Given the high levels of TIGIT in the midst of viral suppression, we assessed the relationship between TIGIT and the cellular HIV content in highly purified CD4 + T cells among HIV-infected "cART initiators" who met strict selection criteria of well documented and long-term persistent viral suppression (L-AS; Table 1). We did not observe a correlation with the frequency of CD8 + T cell or CD4 + T cell TIGIT expression and HIV RNA from purified CD4 + T cells (S1H and S1J Fig). However, the frequency of TIGIT + CD4 + T cells positively correlated with purified CD4 + T cell HIV DNA content, but not with frequency of TIGIT + CD8 + T cells (Fig 1I and 1J). These data indicate that TIGIT expression on CD4 + T cells may be linked to chronic HIV disease pathogenesis, residual immune activation, and the cellular HIV DNA content among those with viral suppression.

Phenotype and function of TIGIT on CD8 + T cells during HIV infection
HIV infection leads to an expansion of intermediately differentiated memory CD8 + T cells that are not fully mature effectors [32][33][34]. We profiled the pattern of TIGIT expression in the heterogeneous CD8 + T cell subpopulations and found TIGIT was significantly expanded on the CD8 + T cell intermediate/transitional and effector subsets with the highest expression of TIGIT on the effector CD8 + T cell subset (Figs 2A and S2A-S2E) compared to AS. In the naïve population TIGIT expression was relatively stable with only a significant difference seen between HD and the non-controllers (Fig 2A). We did observe a statistically significant difference in TIGIT expression between the HD and AI group in the memory CD8 + T cell population (Fig 2A). Thus, TIGIT is expanded on the intermediate/transitional and effector CD8 + T cell subsets during chronic HIV infection, consistent with a role for TIGIT as potential regulator of intermediate/transitional and effector T cell responses.
We next profiled the expression of TIGIT on viral specific CD8 + T cells from chronically HIV-infected participants using matched HLA-I restricted pentamers for various HIV and CMV peptide epitopes. TIGIT was expressed on over half of all CD8 + T cells for specific for HIV-1 Gag (55.3%), Polymerase (54.7%), Envelope (54.3%), Nef (52%), and also for CMV pp65 (57.8%) ( Fig  2B and 2C). Comparable levels of TIGIT on HIV and CMV specific CD8 + T cells were observed on a per cell basis as measured by Geometric Mean Fluorescence Intensity (GMFI) (Fig 2C).  Targeting TIGIT Enhances HIV/SIV T Cell Immunity We next assessed the effector phenotype and functional properties of TIGIT expressing CD8 + T cells to determine whether they retain features of immune exhaustion. We found that most of the TIGIT expressing CD8 + T cells co-expressed PD-1 with the frequency of TIGIT + PD-1 + CD8 + T cells significantly expanded in chronic HIV infection (AS, 18.65%; EC, 20.85%; NC, 38.15%) compared to HD (13.65%) (Fig 3A and 3B). The frequency of TIGIT + PD-1 + CD8 + T cells inversely correlated with CD4 counts ( Fig 3C) and positively correlated with plasma viral load levels ( Fig 3D) among all chronically HIV-infected individuals. We observed significantly higher frequencies of TIGIT + PD-1 + co-expression on HIV-Gag-specific CD8 + T cells compared to non-HIV-Gag-specific CD8 + T cells derived from PBMCs (Fig 3E-3G). Furthermore, the majority of the TIGIT + PD-1 + HIV-Gag-specific CD8 + T cells retained a transitional/intermediate memory (CD45RA -CCR7 -CD27 + ) phenotype (Fig 3H-3J). These results suggest TIGIT may render a large fraction of viral specific CD8 + T cells vulnerable to negative regulation.
Given the high expression of PD-1 among TIGIT + CD8 + T cells, we evaluated the functional status of the TIGIT expressing cells. We stained T cells with the nuclear antigen Ki-67, which is associated with cellular proliferation, and observed that TIGIT + cells expressed significantly more Ki-67 than TIGIT -CD8 + T cells (Fig 4A and 4B). However, in contrast, Ki-67 expression was equivalently distributed between PD-1 + and PD-1 -CD8 + T cells (Fig 4C and 4D). Using intracellular cytokine staining, in response to stimulation with an overlapping 15mer HIV-1 Gag peptide pool, we observed that TIGIT + CD8 + T cells produced significantly less IFN-γ, TNF-α and IL-2 compared to TIGIT -CD8 + T cells (Fig 4E and 4F). We observed phenotypically the majority of the HIV specific cytokine responsive CD8 + T cells lacked TIGIT and PD-1 dual expression and were minimally represented in the TIGIT + PD-1 + subset. However, single expressing cells retain some functional responses ( Fig 4G).
To directly assess the functionality of the TIGIT + PD-1 + subset in HIV infected individuals, CD8 + T cells expressing TIGIT and/or PD-1 on their surface were sorted to high purities (S3A and S3B Fig TIGIT + PD-1cells produced less cytokines than TIGIT -PD-1 + cells. These data are partially in alignment with results observed in Fig 4G. However, it was notable that IL-10 production was almost exclusively produced by the TIGIT -PD-1 + cell subset. These data suggest that TIGIT + CD8 + T cells, particularly TIGIT + PD-1 + co-expressing CD8 + T cells exhibit distinguishing features of exhausted T cells. . P values were calculated using one-way ANOVA, followed by Tukey's multiple comparisons test. Graphs show correlation of total chronic infected (+: AS, EC, and NC; left panel, n = 60) and non-controllers (right panel, n = 20) frequency (%) of (E) TIGIT + CD8 + and (F) TIGIT + CD4 + T cells against clinical CD4 Count (cells/mm 3 ). Graphs show correlation of total chronic infected (+: AS, EC, and NC; left panel, n = 60) and elite controllers (right panel, n = 20) frequency (%) of (G) TIGIT + CD8 + and (H) TIGIT + CD4 + T cells against frequency (%) of T cell activation (CD38 + HLA-DR + ). Graphs show correlation of frequency (%) of (I) TIGIT + CD8 + and (J) TIGIT + CD4 + T cell among aviremic HIV infected "ART initiators" with known duration of long-term viral suppression from the SCOPE cohort (L-AS, n = 19, open inverted triangles) versus copies of CD4 + T cell associated HIV DNA per million CD4 + T cells (log 10 ). Spearman's rho tests were performed for correlations.  Next, we evaluated the intracellular granular content of TIGIT expressing cells. We observed that TIGIT expressing CD8 + T cells contained significantly more perforin and granzyme B compared to non-TIGIT expressing CD8 + T cells (Fig 4H and 4I). We observed no difference in the ability of TIGIT + CD8 + T cells to degranulate compared to TIGIT -CD8 + T cells when stimulated with HIV-1 Gag peptide pool (Fig 4J and 4K). However, upon stimulation with anti-CD3 + anti-CD28 Dynabeads, TIGIT + cells degranulated less than TIGITcells (Fig 4J-4L).

Induction of TIGIT on CD8 + T cells during HIV infection
To explore the regulation of TIGIT expression we stimulated HIV-specific CD8 + T cells from chronically HIV-infected individuals with HIV-1 Gag peptides. HIV-1 Gag peptide stimulation did not significantly increase the expression of TIGIT on HIV-specific CD8 + T cells, although we did observe an upregulation of TIGIT in a subset of individuals (Fig 5A-5C). Several common gamma-chain (γ-chain) cytokines have been shown to directly upregulate negative checkpoint receptors on CD8 + T cells during retroviral infections [35]. To further understand the mechanism driving TIGIT upregulation, we explored the capacity of γ-chain and non-γ-chain cytokines to regulate TIGIT expression (Fig 5D-5F). We found that IL-2 and IL-15 prominently led to a significant increase in TIGIT expression on CD8 + T cells from HIV-infected individuals unlike non-γ-chain cytokines IL-12 and IL-18 ( Fig 5E). This effect was not evident on CD8 + T cells derived from HIV-uninfected participants ( Fig 5F).

Effects of anti-TIGIT and anti-PD-L1 mAb blockade on HIV-specific CD8 + T cell cytokine and proliferative responses
Since TIGIT and PD-1 are co-expressed, and dual blockade in the mouse model limits in vivo LCMV replication [18] and elicits anti-tumor CD8 + T cell responses [19], we evaluated the effects of TIGIT and PD-L1 blockade on HIV-Gag-specific CD8 + T cells using cells from chronically HIV-infected individuals at various stages of infection (Table 2). To evaluate the ex vivo HIV-specific T cell cytokine restoration, we used a modified version of our previously published in vitro short-term primary recall blockade assay [8]. Incubation with either anti-TIGIT mAb alone or anti-PD-L1 mAb alone significantly increased IFN-γ production, however dual blockade of both TIGIT and PD-L1 did not enhance IFN-γ responses over anti-TIGIT or anti-PD-L1 alone (Fig 6A and 6B). We also observed that only dual blockade of TIGIT and PD-L1 significantly increased IL-2 production by CD8 + T cells (S5A and S5B Fig).

Targeting TIGIT Enhances HIV/SIV T Cell Immunity
Single blockade of PD-L1 significantly enhanced HIV-specific CD8 + T cell proliferation while single blockade of TIGIT did not improve CD8 + T cell proliferation (Fig 6C and 6D). When both anti-TIGIT and anti-PD-L1 were combined there was significant increased CD8 + T cell proliferation compared to PD-L1 blockade alone (Fig 6C and 6D). Though donor OM115 had the highest baseline levels of TIGIT + CD8 + T cells among the group, no significant association was seen between the magnitude of IFN-γ production and proliferation by TIGIT blockade and baseline TIGIT + CD8 + T cell expression (r = 0.24, p = 0.257). These data suggest that HIV-specific CD8 + T cell proliferation can be markedly improved with simultaneous combination blockade of TIGIT and PD-L1.

rhTIGIT is elevated on dysfunctional effector CD8 + T cells in the SIV rhesus macaque model of HIV/AIDS and is associated with SIV disease progression
To explore the role of TIGIT in the rhesus macaque model of HIV/AIDS we cloned rhesus TIGIT (rhTIGIT) (GenBank: KR534505) and observed that it shares 88.11% sequence homology with human TIGIT (S6A Fig). We reasoned that rhTIGIT expression and function would mimic our human HIV studies and be replicable in the SIV-infected rhesus macaque model of HIV/AIDS. RhTIGIT expression was significantly increased on CD8 + T cells derived from the lymph node (LN) (38.6%) and splenic (60.9%) compartments when compared to SIV-uninfected macaques (LN 10.82% and spleen 25.55%), but not in PBMCs (Figs 7A and S6B). Similar to what we observed in HIV-infected participants, the frequency of rhTIGIT + CD8 + T cells from PBMC did not correlate with plasma SIV viral load. However, we did observe a significant correlation with the frequencies of rhTIGIT + CD8 + T cells in LN and viral load ( Fig 7B).
As observed in human HIV infection, rhTIGIT expression was more prominently expressed in SIV infection on effector memory (EM, CD28 -CD95 + ), and central memory (CM, CD28 + CD95 + ) CD8 + T cells when compared to naïve (N, CD28 + CD95 -) CD8 + T cells from PBMCs, LN and from the spleen (S6C Fig). In the tissues, it was notable that TIGIT expression was highest on the central memory CD8 + T cells when compared to PBMCs (S6C Fig). As in HIV infection, stimulation with γ-chain cytokines such as IL-2 and IL-15 upregulated rhTIGIT on CD8 + T cells from SIV-infected animals (S6D Fig). rhTIGIT was also expressed on~40% of SIV Gag or Tat tetramer specific CD8 + T cells derived from PBMCs or secondary lymphoid tissues, even in animals with full cART suppression of peripheral SIV viremia (Fig 7C). This was more prominently found in the tissues of SIV-infected animals where higher frequency of SIV-specific CD8 + T cells co-expressed both rhTIGIT and rhesus macaque PD-1 (rhPD-1) (S6E- S6H Fig).
Given the similarities of rhTIGIT and human TIGIT, we evaluated TIGIT and PD-L1 blockade on SIV peptide stimulated CD8 + T cell responses. We found that dual blockade of rhPD-L1 and rhTIGIT enhanced SIV-specific CD8 + T cell proliferation in PBMCs and spleen while single blockade of rhPD-L1 enhanced SIV-specific proliferation in the spleen (Fig 7F and  7G). Taken together, rhTIGIT pathway is active in the rhesus macaque model of HIV/AIDS and partially mimics human TIGIT expression and function during HIV infection.

Discussion
In this report we profiled TIGIT expression on T cells in HIV-infected participants with various degrees of viral control and in SIV-infected rhesus macaques. We (1) unveil a role for TIGIT + CD8 + T cells in HIV disease progression and demonstrate its relation to T cell exhaustion, (2) observe that TIGIT appears to associate with the cellular viral reservoir in CD4 + T cells, (3) we found that co-blockade of TIGIT and PD-L1 lead to a greater restoration of T cell function compared with a single blockade, and (4) by successfully cloning rhTIGIT (GenBank: KR534505) we reveal the similarities in expression and function of rhTIGIT on T cells in the non-human primate model of HIV/AIDS. Our findings reveal a novel inhibitory pathway involved in the suppression of T cell responses during chronic viral infection, the blockade of which may contribute to the reversal of T cell dysfunction in the control or elimination of HIV infection.
While TIGIT levels on CD8 + T cells tracked disease progression (depletion of CD4 T cells and T cell activation), this was not evident across the various HIV-infected groups. There was a significant difference in TIGIT + memory CD8 + T cells in acute infection compared to the uninfected group, but not in the global CD8 + T cell population. This suggests there may be a gradient, with an increase in global TIGIT + CD8 + T cells in acute infection that becomes greater over time, which is distributed among the differentiated CD8 + T cell populations and may differ when compared to other negative checkpoint receptors which are found elevated during the early stages of HIV infection [4,38,39].
TIGIT induction appears to be driven by polyclonal TCR stimulation and this is a common feature among immune checkpoint receptors [38,40]. We observed HIV-1-Gag-SL9-specific CD8 + T cells did not increase TIGIT expression after HIV-1 Gag peptide stimulation as a group, however a subset of individuals with moderate levels of TIGIT increased expression after stimulation, and individuals that expressed high levels of TIGIT retained expression after Targeting TIGIT Enhances HIV/SIV T Cell Immunity Targeting TIGIT Enhances HIV/SIV T Cell Immunity stimulation. TIGIT remained elevated despite antigen in cART-suppressed individuals; previous studies have also shown that common γ-chain cytokines maintain the ability to regulate immune checkpoint receptor expression in the absence of antigenic stimulation [35,41]. Our studies align with these previous observations and suggest that a cytokine milieu conducive for the maintenance of an exhausted T cell profile persists in HIV and SIV infection even during viral suppression.
TIGIT expression was found to be associated with T cell activation principally among EC who represent a small population of HIV-infected individuals able to spontaneously suppress their viral load (<50 copies/ml) in the absence of cART for prolonged periods of time [42]. However, over time a subset of EC will lose virologic control and develop viremia and CD4 + T cell loss [43,44]. In addition, EC maintain elevated levels of T cell activation despite viral control [45,46]. High TIGIT expression may reflect ongoing immune activation in the EC population. The institution of cART in those EC has led to a reduction in immune activation [47,48]. Given our finding, in addition to cART, some EC may benefit from TIGIT blockade to alleviate the persistent T cell immune activation thereby reducing the risk of adverse non-AIDS events that have been documented to occur in this population, however such strategies need to be considered carefully given the risk of autoimmunity as described in anti-PD-1 clinical trials in the oncology field [49,50]. Viral clearance of the chronic strain of LCMV (clone 13) in mice by combined blockade of TIGIT and PD-L1 provided the first evidence of the advantages of targeting these two pathways [18]. In addition, targeting TIGIT and PD-L1 on CD8 + tumor infiltrating lymphocytes in patients with advanced melanoma synergistically improves potent anti-tumor responses [19].
Here we extend these finding to human and simian retroviral infections. This was significant given the expansion of dual expressing TIGIT and PD-1 CD8 + T cells in HIV infection despite pharmacological or immunological viral suppression. Our data shows the presence of TIGIT and PD-1 dual expressing HIV and SIV-specific CD8 + T cells and co-blockade of TIGIT and PD-L1 better enhanced proliferation of HIV and SIV-specific CD8 + T cell responses compared to single blockade. Although we see a significant increase among all HIV-infected individuals, it was evident that subsets of weak-responders exist and appear heterogeneous irrespective of stage of infection, viral load levels or viral suppression. Indeed, combinational blockade of CTLA-4 and PD-1 revealed a subset of weak-responders to anti-tumor activity. Different or expanded combinations of immune checkpoint blockers with anti-TIGIT may need to be considered in the arsenal to improve anti-viral T cell immunity in all individuals.
Persistence of the cellular latent HIV reservoir has been a major barrier to the eradication of HIV [51]. One proposed strategy is to 'Shock' the latently infected cells to flush out virus with latency reversal agents (LRAs) [52][53][54]. The development of the 'Shock' strategies is advancing at a rapid pace with in vivo studies yielding activity in reactivating of latent virus. However, the 'Kill' component is less well developed. Shan and colleagues demonstrate that after reactivation of latent virus from CD4 + T cells, CD8 + T cells' activity had the capacity to kill latently infected CD4 + T cells if appropriate pre-stimulation of HIV peptide and IL-2 was provided an in vitro latency assay using Bcl-2 as a survival signal to prolong the longevity of the latent CD4 + T cells [55]. Furthermore, recent studies show that HIV-infected individuals on cART retain broad HIV-specific cytotoxic T-lymphocyte responses that are able to target the mutated latent virus [56]. Blocking immune checkpoint pathways such as TIGIT and PD-1/PD-L1 can be harnessed to boost HIV-specific CD8 + T cells responses given that these pathways persist in the setting of viral suppression. Furthermore, given our findings showing the relationship with TIGIT expression on CD4 + T cells and the total cell associated HIV DNA, it remains unclear what role TIGIT may play in the establishment of the reservoir in CD4 + T cells, however it is likely related to the capacity of TIGIT's ability control T cell activation or proliferation.  Our study provides a novel role for TIGIT during HIV disease pathogenesis and our demonstration of a role of rhTIGIT in the non-human primate model of HIV/AIDS provides a platform to investigate our understanding of the complex networks of co-inhibition that can be tailored to each individual or viral infection. Improving CD8 + T cell functions may further aid in the 'Shock and Kill' approaches being considered to eliminate latent virus and improve T cell mediated vaccine responses to prevent or limit infection [57].

Study participants
We recruited participants from the following cohorts: University of California, San Francisco (UCSF) SCOPE and OPTIONS cohorts [8] the Hawaii HIV-1 (HHC) cohort, the Torontobased cohort CIRC (Maple Leaf Clinic and St. Michael's Hospital, Toronto, Canada) [58] and the Duke Human Vaccine Institute (DHVI) tissue repository. SCOPE specimens (n = 80) were selected from the following groups: untreated non-controllers (n = 20) (participants who had never been treated with antiretroviral agents or who had been off therapy for at least six months), treated virologic controllers (participants who had an undetectable plasma HIV-1 RNA level for the previous six months while on cART) (n = 20), spontaneous "elite" virologic controllers (participants who are untreated and who have at least three documented plasma HIV-1 RNA levels <2,000 copies/ml over at least a 12-month period) (n = 20), Some of these participants had persistent plasma HIV-1 RNA levels <75 copies/ml) and HIV-infected "cART initiators" (n = 20) who meet strict selection criteria and well documented persistent viral suppression for over 1.5 years. Participants with acute HIV infection (n = 24) were obtained for the OPTIONS cohort of primary HIV infection [8] and age-matched HIV-uninfected (n = 20) (Table 1) and chronically infected virally suppressed leukapheresesed individuals were obtained from the HAHC cohort [59]. Additional participants with chronic infection at various stages of infection were obtained from participants with various levels of viral control from the Toronto-based cohort CIRC cohort, HHC, and DHVI.

Ethics statement
All persons gave written informed consent to participate in the study and approval for the study was obtained from the University of Hawaii Committee of Human Subjects. Samples were obtained from Indian rhesus macaques (Macaca mulatta) housed at the Oregon National Primate Research Center (ONPRC), which were SIV infected for other ongoing, unrelated studies. Oregon Health and Science University (OHSU) Institutional Animal Care and Use Committee (IACUC) Protocol #: 0989. The OHSU Institutional Animal Care and Use Committee reviewed and approved all study protocols. All macaques in this study were managed according to the ONPRC animal husbandry program, which aims at providing consistent and excellent care to nonhuman primates. This program is based on the laws, regulations, and representative Mamu-A*01 animal with full cART suppression. Compiled data of rhTIGIT expression frequency (%) on tetramer specific CD8 + T cells (n = 4) from PBMCs (circle), LNs (square), and spleen (triangle) from Mamu-A*01 + macaques with full cART suppression. (D) Representative flow cytometry plots of PBMCs (n = 12) or lymph nodes (n = 18) stimulated without or with PMA + Ionomycin. Graphs show frequency (%) of IFN-γ from CD8 + T cells expressing TIGIT or not expressing TIGIT. P values were calculated by Wilcoxon matched-pairs signed ranked test. (E) Representative flow cytometry plots of PBMCs from a Mamu-A*01+ macaque stimulated without or with SIV-Gag 181-189 CM9 peptide. Graph shows frequency (%) of IFN-γ from SIV-Gag 181-189 CM9-specific CD8 + T cells expressing TIGIT or not expressing TIGIT from Mamu-A*01 + macaques (n = 6). P values were calculated by Wilcoxon matchedpairs signed ranked test. (F) Representative flow cytometry plots of CD8 + T cells from two separate SIV-infected macaques showing CD8 + T cell CFSE dilution in response to AT2-inactivated SIV with either no antibody, anti-TIGIT, anti-PD-L1 or dual blockade (anti-TIGIT + anti-PD-L1). (G) Graphs show compiled data of CD8 + T cell CFSE dilution from PBMC (left panel), Lymph node (middle panel), or Spleen (right panel) as percent of max (n = 4). P values were calculated by Wilcoxon matched-pairs signed ranked test. doi:10.1371/journal.ppat.1005349.g007 Targeting TIGIT Enhances HIV/SIV T Cell Immunity guidelines set forth by the United States Department of Agriculture (e.g., the Animal Welfare Act and its regulations, and the Animal Care Policy Manual), Institute for Laboratory Animal Research (e.g., Guide for the Care and Use of Laboratory Animals, 8 th edition), Public Health Service, National Research Council, Centers for Disease Control, and the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) International. The nutritional plan utilized by the ONPRC is based on National Research Council recommendations and supplemented with a variety of fruits, vegetables, and other edible objects as part of the environmental enrichment program established by the Behavioral Management Unit. Paired/grouped animals exhibiting incompatible behaviors were reported to the Behavioral Management staff and managed accordingly. All efforts were made to minimize suffering through the use of minimally invasive procedures, anesthetics, and analgesics when appropriate. Animals were painlessly euthanized with sodium pentobarbital and euthanasia was assured by exsanguination and bilateral pneumothorax, consistent with the recommendations of the American Veterinary Medical Guidelines on Euthanasia (2013) Quantifying cellular HIV-1 DNA and RNA Cryopreserved PBMCs were rapidly thawed and enriched for CD4 + T cells to high purities with an EasySep Human CD4 + T cell enrichment kit (Stemcell Technologies, Vancouver, British Columbia, Canada). Cellular RNA and DNA from PBMC T-cell subsets cells were purified using the AllPrep DNA/RNA kit (Qiagen, Ventura CA) as specified by the manufacturer, quantified using a Nanodrop (ND-1000) spectrophotometer and normalized to cell equivalents by qPCR using human genomic TERT for DNA and GAPDH or RPLP0 expression for RNA (Life Technologies, Grand Island NY). Total cellular HIV-1 DNA (integrated and unintegrated) and RNA (unspliced and multiply spliced) was quantified with a qPCR TaqMan assay using LTR-specific primers F522-43 (5' GCC TCA ATA AAG CTT GCC TTG A 3'; HXB2 522-543) and R626-43 (5' GGG CGC CAC TGC TAG AGA 3'; 626-643) coupled with a FAM-BQ probe (5' CCA GAG TCA CAC AAC AGA CGG GCA CA 3) [60] on a StepOne Plus Real-time PCR System (Applied Biosystems Inc, Foster City CA). Cell associated HIV-1 DNA copy number was determined using a reaction volume of 20 μl with 10 μl of 2x TaqMan Universal Master Mix II including UNG (Life technologies), 4 pmol of each primer, 4 pmol of probe, and 5 μl of DNA. Cycling conditions were 50°C for 2 min, 95°C for 10 min, then 60 cycles of 95°C for 15s and 59°C for 1 min. Cell associated HIV-1 RNA copy number was determined in a reaction volume of 20 μl with 10 μl of 2x TaqMan RNA to Ct 1 Step kit (Life Technologies), 4 pmol of each primer, 4pmol of probe, 0.5 μl reverse transcriptase, and 5μl of RNA under identical cycling conditions. For HIV-1 DNA measurements, external quantitation standards were prepared from pNL4-3 in a background of HIV-1 negative human cellular DNA, calibrated to the Virology Quality Assurance (VQA, NIH Division of AIDS) cellular DNA quantitation standards. For HIV RNA measurements, external quantitation standards were prepared from full length NL4-3 virion RNA followed by copy number determination using the Abbott RealTime assay (Abbott Diagnostics, Des Plains Ill) and calibrated to VQA HIV-1 RNA standards. Patient specimens were assayed with up to 800 ng total cellular RNA or DNA in replicate reaction wells and copy number determined by extrapolation against a 7-point standard curve (1-10,000 cps) performed in triplicate.

Pentamer analysis
We used the following Biotin labeled pentamers: A Ã 02:01 SLYNTVATL HIV-1-Gag, A Ã 02:01 ILKEPVHGV HIV-1-Pol, B Ã 07:02 IPRRIRQGL HIV-1-Env, B Ã 07:02 TPGPGVRYPL HIV-1-Nef, and A Ã 02:01 NLVPMVATV CMV-pp65-NV9. All pentamers were obtained from Proimmune Ltd, Oxford, UK. Using protocol outlined previously [8] and stained with antibodies against CD3, CD8, TIGIT, PD-1, TIGIT isotype control or PD-1 isotype control and acquired on the flow cytometer as above. In some experiments PBMCs were stimulated with HIV-1 Gag Peptide pool and evaluated for pentamer phenotype anti-TIGIT and anti-PD-L1 monoclonal antibodies The TIGIT antibody clones 11G11 and 23G8 were generated in HuMab mice [61,62] immunized with a TIGIT-Fc fusion protein and selected based on their high affinity for TIGIT and ability to block TIGIT/PVR interaction. Clone 11G11 is a fully human IgG1 antibody that was engineered to contain a well-characterized set of mutation in the Fc that eliminate FcR interaction [63]. Clone 23G8 is a fully human IgG2 antibody that cross-reacts with macaque TIGIT. The anti-human PD-L1 antibody, clone 12A4, is a fully human IgG4 (S228P) that was generated in HuMab mice immunized with PD-L1-Fc. This antibody was selected based on its ability to block the binding of PD-L1 to both PD-1 and CD80. 12A4 cross-reacts with macaque PD-L1.

Peptides and polyclonal stimulation
123 Overlapping~15mer HIV-1 clade B gag peptides obtained from the National Institute of Health AIDS Reagent Program. Stimulations were performed with a final concentration of 10 μg/ml peptide. T cell activator (anti-CD3 + anti-CD28 mAb Dynabeads) (Life Technologies) followed recommended manufacture procedure.

Functional assays
In the intracellular cytokine stimulation assay studies, thawed cryopreserved PBMCs were stimulated for 12 hours in an incubator at 37°C with 5% CO 2 with 5 μg/ml brefeldin A and 5 μg/ml monensin (Sigma-Aldrich) culture media, DMSO alone, pooled HIV-1 Gag peptides, or anti-CD3/CD28 dynabeads (Life Technologies) in the presence or absence of purified isotype IgG control, anti-TIGIT and/or anti-PD-L1 mAbs. After stimulation, the cells were washed and stained for viability with AARD and cultured with surface phenotype panel against CD8, TIGIT or an isotype control antibody, followed by intracellular staining of CD3 and IFNγ and acquisition on the flow cytometer as above. For the proliferation assay, thawed PMBCs were washed two times and resuspended in PBS supplemented with 0.01% BSA at a concentration of one million cells per milliliter. Cells were labeled with 1mM Carboxyfluorescein succinimidyl ester (CFSE) violet trace (Invitrogen) using protocols previously described [8]. Cells were stimulated for seven days with DMSO alone, pooled HIV-1 gag peptides, or anti-CD3/ CD28 Dynabeads in the presence or absence of purified isotype IgG control, anti-TIGIT or anti-PD-L1 at 37°C with 5% CO 2 . At the end of the stimulation, cells were washed and stained for viability with AARD and cultured with surface phenotype panel against CD3, CD8, TIGIT or an isotype control antibody and acquired on the flow cytometer as above.

In vitro cytokine stimulation
PBMCs were thawed and one million cells were plated per stimulation condition. Stimulation conditions included media alone, 25 ng/ml IL-2 (Roche), 50 IU/ml of IL-12 (MBL international, Woburn, Massachusetts), 50 IU/ml of IL-18 (MBL international) and 25 ng/ml IL-15 (R&D Systems, Minneapolis, Minnesota). Cells were stimulated for six days in an incubator at 37°C with 5% CO 2 . After stimulation, the cells were washed and stained for viability with AARD cultured with surface phenotype panel against CD3, CD4, CD8, TIGIT or isotype control antibody and acquired on the flow cytometer as above.

T cell function in vitro antibody blockade studies
HIV-infected cryopreserved PBMC from individuals identified in Table 2 were stimulated for 12 hours in an incubator at 37°C with 5% CO 2 . Stimulation conditions contained culture media, DMSO alone or pooled HIV-1 gag peptides, in the presence of brefeldin A (Sigma-Aldrich), monensin (Sigma-Aldrich), anti-TIGIT mAb anti-PD-L1 mAb or mouse IgG1 isotype control. After stimulation, cells were washed and stained for cellular viability with AARD and conjugated antibodies against CD8 and CD4, followed by intracellular staining of CD3 and IFN-γ and acquired on a flow cytometer as described above.

Statistical analysis
The repeated-measures, one-way ANOVA followed by Tukey's multiple comparison, Wilcoxon matched-pairs signed ranked test was used for paired tests and the Spearman's rho test was used for correlation analyses. Measures of central tendency are expressed as medians and interquartile ranges (IQRs; given in the form 25th percentile, 75th percentile). Statistical analyses were conducted using Prism Graphpad release 5.0d (Graphpad Software, San Diego, California) or SPSS 22.0 (IBM, Armonk, New York) and the statistical significance of the findings was set at a p-value of less than 0.05.

Accession numbers
Indian rhesus macaque (Macaca mulatta) TIGIT (rhTIGIT): GenBank KR534505. Mary-Margaret Byron MS (UH) for their assistance in sample preparation. We thank Ms. Lindsay Kohorn MPH (Hawaii Center for AIDS, UH) for her consultation in statistical analysis and Dr. Alexandra Gurary PhD (UH) for assistance in FACS sorting. We thank Dr. Louis Picker PhD (VGTI, OHSU) for providing PBMC samples and helpful discussions. We would like to further extend a profound appreciation to the participants of the SCOPE, OPTIONS, Toronto, Hawaii, and EQAPOL cohorts for their participation in this study.