Disruption of the immune-checkpoint VISTA gene imparts a proinflammatory phenotype with predisposition to the development of autoimmunity

Significance The discovery of V domain-containing Ig suppressor of T-cell activation (VISTA) as a novel immune-checkpoint regulator comes at an exciting time, as the field of cancer immunotherapy has made significant progress owing to the clinical success of targeting immune-checkpoint proteins such as cytotoxic T lymphocyte-associated antigen 4 and programmed death 1 and ligand. Recent studies also show the promise of monoclonal antibody-mediated VISTA targeting for enhancing antitumor immunity in murine tumor models. The current study demonstrates the spectrum of immune alterations upon genetic disruption of VISTA in mice in the context of self-tolerance as well as immune response against neoantigen. These results enhance the understanding of the immune-regulatory role of VISTA and form the foundation for designing future clinical applications that target VISTA in treating human diseases. V domain-containing Ig suppressor of T-cell activation (VISTA) is a negative checkpoint regulator that suppresses T cell-mediated immune responses. Previous studies using a VISTA-neutralizing monoclonal antibody show that VISTA blockade enhances T-cell activation. The current study describes a comprehensive characterization of mice in which the gene for VISTA has been deleted. Despite the apparent normal hematopoietic development in young mice, VISTA genetic deficiency leads to a gradual accumulation of spontaneously activated T cells, accompanied by the production of a spectrum of inflammatory cytokines and chemokines. Enhanced T-cell responsiveness was also observed upon immunization with neoantigen. Despite the presence of multiorgan chronic inflammation, aged VISTA-deficient mice did not develop systemic or organ-specific autoimmune disease. Interbreeding of the VISTA-deficient mice with 2D2 T-cell receptor transgenic mice, which are predisposed to the development of experimental autoimmune encephalomyelitis, drastically enhanced disease incidence and intensity. Disease development is correlated with the increase in the activation of encephalitogenic T cells in the periphery and enhanced infiltration into the CNS. Taken together, our data suggest that VISTA is a negative checkpoint regulator whose loss of function lowers the threshold for T-cell activation, allowing for an enhanced proinflammatory phenotype and an increase in the frequency and intensity of autoimmunity under susceptible conditions.

V domain-containing Ig suppressor of T-cell activation (VISTA) is a negative checkpoint regulator that suppresses T cell-mediated immune responses. Previous studies using a VISTA-neutralizing monoclonal antibody show that VISTA blockade enhances T-cell activation. The current study describes a comprehensive characterization of mice in which the gene for VISTA has been deleted. Despite the apparent normal hematopoietic development in young mice, VISTA genetic deficiency leads to a gradual accumulation of spontaneously activated T cells, accompanied by the production of a spectrum of inflammatory cytokines and chemokines. Enhanced T-cell responsiveness was also observed upon immunization with neoantigen. Despite the presence of multiorgan chronic inflammation, aged VISTA-deficient mice did not develop systemic or organspecific autoimmune disease. Interbreeding of the VISTA-deficient mice with 2D2 T-cell receptor transgenic mice, which are predisposed to the development of experimental autoimmune encephalomyelitis, drastically enhanced disease incidence and intensity. Disease development is correlated with the increase in the activation of encephalitogenic T cells in the periphery and enhanced infiltration into the CNS. Taken together, our data suggest that VISTA is a negative checkpoint regulator whose loss of function lowers the threshold for T-cell activation, allowing for an enhanced proinflammatory phenotype and an increase in the frequency and intensity of autoimmunity under susceptible conditions. immune suppression | inflammation | autoimmunity | peripheral tolerance I mmune responses against foreign pathogens or self-antigens are regulated by multiple layers of positive and negative molecules and pathways, as exemplified by molecules of the B7 family. B7-1/2 and B7-H2 provide critical costimulatory signals for T-cell activation, whereas multiple "negative checkpoint" regulators, involving cytotoxic T lymphocyte-associated antigen 4 (CTLA-4), programmed death 1 (PD-1) and ligand (PD-L1), B7-H3, and B7-H4, down-regulate T-cell responses (1,2). Disruption of these pathways leads to loss of peripheral tolerance and development of autoimmunity (3). For example, CTLA-4 genetic deficiency leads to a fatal lymphoproliferative disorder (4,5), whereas PD-1-deficient mice develop autoimmune dilated cardiomyopathy or lupus-like autoimmune phenotypes depending upon the genetic background (6,7). In addition, PD-L1 or PD-1 blockade either by antibody or genetic deletion, on autoimmune-susceptible backgrounds, promotes autoimmune diabetes (8)(9)(10) and exacerbates autoimmune kidney disease (11), autoimmune hepatitis (12), and experimental autoimmune encephalomyelitis (EAE) (13,14).
V domain-containing Ig suppressor of T-cell activation (VISTA) is a member of the B7 family that bears homology to PD-L1 and is exclusively expressed within the hematopoietic compartment (15). VISTA is highly expressed on CD11b high myeloid cells, and is also expressed at lower densities on CD4 + and CD8 + T cells and Foxp3 + regulatory T cells. A soluble VISTA-Ig fusion protein or VISTA expressed on antigen-presenting cells (APCs) acts as a ligand that suppresses T-cell proliferation and cytokine production via an unidentified receptor. VISTAspecific monoclonal antibody reversed VISTA-mediated T-cell suppression in vitro and in vivo (15,16). The human homolog shares 90% homology with murine VISTA, and similar expression patterns and suppressive function were reported for human VISTA (17).
It is hypothesized that VISTA is an immune-checkpoint regulator that negatively regulates immune responses. To gain a comprehensive perspective on the immune-regulatory role of VISTA, we examined the impact of the genetic deletion of VISTA on the maintenance of self-tolerance as well as T-cell responses against neoantigens. The results show that VISTAdeficient mice demonstrate an age-related proinflammatory signature, spontaneous T-cell activation, as well as enhanced cell-mediated immune responses to neoantigen, and greatly promoted autoimmunity when interbred onto an autoimmunesusceptible background.

Significance
The discovery of V domain-containing Ig suppressor of T-cell activation (VISTA) as a novel immune-checkpoint regulator comes at an exciting time, as the field of cancer immunotherapy has made significant progress owing to the clinical success of targeting immune-checkpoint proteins such as cytotoxic T lymphocyte-associated antigen 4 and programmed death 1 and ligand. Recent studies also show the promise of monoclonal antibody-mediated VISTA targeting for enhancing antitumor immunity in murine tumor models. The current study demonstrates the spectrum of immune alterations upon genetic disruption of VISTA in mice in the context of self-tolerance as well as immune response against neoantigen. These results enhance the understanding of the immune-regulatory role of VISTA and form the foundation for designing future clinical applications that target VISTA in treating human diseases.

Results
Spontaneous T-Cell Activation and Chronic Multiorgan Inflammation in VISTA Knockout Mice. VISTA knockout (ko) mice were obtained from the Mutant Mouse Regional Resource Centers (www. mmrrc.org; stock no. 031656-UCD) (18). The original VISTA ko mice on a mixed genetic background were fully backcrossed onto the C57BL/6 background. VISTA ko mice were born at normal size, maturation, and fertility, with normal thymic development and with populations of lymphocytes [T, B, natural killer (NK), and NK T cells] in the bone marrow, spleen, and lymph nodes (LNs) indistinguishable in number and frequency from their WT counterparts.
Changes in a wide variety of immunological parameters were compared in VISTA ko and WT mice (7-10 mo of age). VISTA ko mice showed moderate increases in spleen size, indicating heightened homeostasis of certain hematopoietic cell populations (Fig. 1A). Surface phenotypes of splenic cells confirmed that there were increased percentages of CD11b + MHCII + myeloid populations as well as CD11b + CD11c + myeloid dendritic cells (DCs) and CD11b + Ly6G + neutrophils (Fig. 1B). Such an increase of myeloid APCs and inflammatory cells such as neutrophils might indicate certain levels of ongoing inflammation in the naïve VISTA ko mice.
In contrast to WT mice, VISTA ko mice showed significantly higher frequencies of activated peripheral T cells in both spleen and peripheral blood. This is manifested as increased populations of CD44 hi CD62L lo T cells ( Fig. 1D and Fig. S1A) as well as cytokine production, such as IFN-γ, TNF-α, and IL-17A from VISTA ko T cells upon polyclonal stimulation ( Fig. 1E and Fig. S1 B and C). The total T-cell percentage in the spleens of WT and VISTA ko mice is similar (Fig. 1C).
To further determine whether aged VISTA ko mice developed any systemic or organ-specific autoimmune disease, a comprehensive multiorgan pathological analysis was performed. H&E sections from heart, lung, liver, kidney, pancreas, salivary glands, small and large intestines, and brain were quantified for immune cell infiltration and potential structural damage due to immune cell-mediated destruction. Three organs (lung, liver, and pancreas) were found to contain significant numbers of immune cell infiltrates (Fig. 3 A and B). However, despite the presence of chronic inflammation in multiple tissues, aged VISTA ko mice (∼1-y-old) did not develop overt organ-specific autoimmune disease. There is no statistically significant increase of antinuclear autoantibodies in the serum, nor immunecomplex deposition in the glomeruli. Islets from inflamed VISTA ko pancreas were mostly intact, even when surrounded by infiltrating immune cells. Levels of IgM and IgA, but not IgG subtypes (i.e., IgG1, IgG2a, IgG2b, and IgG3), were moderately elevated in the serum of 1-y-old VISTA ko mice (Fig. 3C). Together, these data indicate that VISTA ko does not result in any overt autoimmunity in the absence of other predisposing factors.

VISTA Deficiency Enhances the Development of Th1/17-Mediated
Autoimmune Disease. The activated T-cell phenotypes observed in aging VISTA ko mice pointed to the fact that VISTA ko may lower the threshold of T-cell receptor (TCR) activation to selfantigens. As such, it was hypothesized that VISTA ko would greatly increase the predisposition to the development of autoimmunity when interbred onto a background where there was a high frequency of self-reactive T cells. VISTA ko mice were interbred with the 2D2 TCR transgenic strain that expresses a TCR recognizing the self-antigen myelin oligodendrocyte glycoprotein (MOG 35-55 ) (19). Previous studies reported that 4% of 2D2 TCR mice develop spontaneous EAE between 3 and 5 mo of age. A similar incidence and onset time of spontaneous EAE (∼3%; 1 out of 30 mice) was observed in our colony of WT 2D2 mice (Fig. 4A). In sharp contrast, VISTA deficiency accelerated disease onset, such that ∼61% (21 out of 34 mice) of 2D2 TCR x VISTA ko mice rapidly succumbed to complete hind limb paralysis within 2-3 mo of age (Fig. 4A). As summarized in Table  1, the age of disease onset in the knockout 2D2 mice was between 5 and 13 wk (SEM 0.48). Among diseased knockout 2D2 mice (a total of 21 mice), 19 mice reached the maximum disease score (3.5) approved by the institutional animal care and use committee (IACUC). Two diseased VISTA ko mice developed atypical EAE, manifested as unilateral limb paralysis, rather than typical bilateral ascending paralysis. Two diseased mice developed hind limb paresis around 8 wk of age, but failed to progress and partially recovered from the disease by 20 wk of age. Analysis of CNS and spleen further confirmed heightened frequencies of activated IFN-γ +and IL-17A + -producing CD4 + T cells in the CNS of VISTA ko 2D2 TCR mice compared with WT controls (Fig. 4 B and C). Histological analysis confirmed significant leukocyte infiltration into the CNS (Fig. 4D) and demyelination (Fig. S2) in the VISTA ko 2D2 mice at the peak of the disease. On the other hand, VISTA deficiency moderately enhanced the incidences of spontaneous optic neuritis in the 2D2 transgenic mice ( Table 1). The overall disease incidences were low, most likely due to the younger age group (within 20 wk of age) in our analysis than previously described (19).
VISTA is expressed on both T cells and myeloid cells. To gain a perspective on the relative contribution of VISTA in each of these lineages to controlling the development of autoimmunity, the functional role of VISTA in these lineages was evaluated in a passivetransfer model of EAE, where naïve MOG-specific CD4 + 2D2 T cells were adoptively transferred into the immune-deficient RAG ko hosts (Fig. 4E). This model offered the opportunity to restrict VISTA deficiency on either the transferred T cells or the host, neither, or both. Upon transfer of 2D2 T cells into a RAG ko host, T cells were activated and expanded, resulting in hind limb paralysis over a period of 1-2 mo. Disease progression was observed without the need for immunization with adjuvant or exogenous administration of peptide. Transfer of VISTA ko 2D2 cells into VISTA ko RAG ko hosts recapitulated the occurrence of aggressive disease observed in the intact 2D2 VISTA ko mice, so that the medium disease score was reached with much faster kinetics in the VISTA ko RAG ko hosts (23-32 d) compared with VISTA wt RAG ko hosts (38-40 d), regardless of whether WT or VISTA ko 2D2 cells were transferred (Fig. 4E). Such an acceleration of disease development established an important negative immune-regulatory role of VISTA expressed on myeloid cells. Loss of VISTA on T cells also impacted on disease progression, as transfer of VISTA ko 2D2 T cells into VISTA ko RAG ko hosts led to rapid disease progression, with a medium disease score reached around 23 d, versus 32 d when WT 2D2 T cells were transferred. Together, these data indicate that VISTA deficiency on both T and myeloid cells contributes to the control of autoimmunity.
Enhanced T-Cell Responses in VISTA-Deficient Mice in Response to Acute Immune Challenge. Spontaneous T-cell activation in aged VISTA ko mice and the enhanced autoimmunity observed in the absence of VISTA expression in autoimmune-prone mice indicate that VISTA contributes to the control of T-cell responsiveness toward either self-antigens or environmental antigens. As such, the question of whether VISTA deficiency could enhance T-cell activation in response to acute immune challenge with neoantigens was evaluated. The cell-mediated immune responses of WT and VISTA ko mice to a soluble model antigen peptide mixed with Toll-like receptor (TLR) agonist polyI:C as adjuvant were compared (Fig. 5 A and B). 2W1S is an MHC class II-restricted peptide, and SIN is an MHC class I-restricted peptide. On day +7 postimmunization, spleen cells were examined for the production of IFN-γ when restimulated with peptide. Significantly increased numbers of IFN-γ-producing cells were observed in cells derived from VISTA ko mice in response to 2W1S (Fig. 5A) and SIN peptide immunization (Fig. 5B).
In addition to soluble antigens, the immune responses against tumor-associated antigens were measured in WT and VISTA ko mice ( spleen were restimulated with tumor antigen and the number of IFN-γ-producing cells was quantified by the enzyme-linked immunospot (ELISPOT) assay. Significantly enhanced tumorspecific IFN-γ responses were observed in T cells from spleen and tumor-draining LNs, as well as within the tumor tissue (tumor-infiltrating T cells; TILs) of VISTA ko tumor-bearing mice compared with WT mice (Fig. 5C). However, tumor growth was not significantly reduced (Fig. 5D). Consistently, there was no significant increase of granzyme B and TNF-α production or of the proliferation (Ki67 expression) of tumor-infiltrating T cells in VISTA ko mice (Fig. 5E). In an attempt to further boost the magnitude of tumor-specific T-cell responses, B16 melanoma tumor-bearing mice were treated with a peptide vaccine using TLR agonists as adjuvant (16). Consistent with our hypothesis, tumor growth was significantly inhibited in the VISTA ko mice upon vaccine treatment compared with WT mice (Fig. 5F). Collectively, these data suggest that VISTA deficiency enhances T-cell responses against neoantigens in the absence and presence of induced inflammation such as via TLR agonists.

Discussion
The spontaneous activation of CD4 + and CD8 + T cells in aged VISTA ko mice, as well as the chronic inflammation in multiple tissues, indicates the loss of peripheral tolerance upon disruption of the VISTA pathway. In contrast to the lethal phenotype of CTLA-4 ko mice, VISTA deficiency did not lead to severe systemic autoimmune pathology, suggesting that other immuneregulatory pathways mitigated the development of inflammation and autoimmunity induced by VISTA deficiency. In that regard, multiple other immune checkpoints, namely CTLA-4, PD-1, B7-H3, and B7-H4, might compensate for the genetic loss of VISTA and might play a role in suppressing the ongoing inflammation in VISTA ko mice. VISTA deficiency overcame additional immune-regulatory mechanisms, and significantly enhanced the development and progression of EAE on a disease-prone transgenic background (Fig. 4). This result is similar to studies showing that PD-L1 deficiency exacerbates autoimmunity only under disease-prone conditions (8,11). Our study therefore affirms the notion that  VISTA is a nonredundant immune-checkpoint regulator that maintains peripheral tolerance and controls autoimmunity. Similar to PD-L1, VISTA is expressed on both APCs and T cells. Although our previous study demonstrated that VISTA expressed on APCs inhibited T-cell activation by engaging an unidentified receptor on T cells (15), the physiological role of VISTA expressed on T cells remains unclear. We established an adoptive-transfer model of EAE to address the relative contribution of VISTA expressed on T cells and non-T cells (Fig. 4E). Our data show a dominant role of VISTA expressed on non-T cells, because VISTA expression in the host effectively suppressed disease progression, regardless of whether WT or knockout 2D2 encephalitogenic T cells were transferred. On the other hand, VISTA expression on 2D2 T cells played a clear role in regulating disease progression, particularly in VISTA ko hosts. These data indicate that VISTA expression on both APCs and T cells contributes to its immune-regulatory role. In that regard, a recent study showed that VISTA might be an inhibitory receptor on T cells (20). Alternatively, VISTA on T cells might exert its immunosuppressive function via engaging a receptor expressed on APCs. Future studies will focus on distinguishing these potential mechanisms.
Enhanced T-cell response, manifested as IFN-γ production, was observed toward neoantigen immunization in the VISTA ko mice (Fig. 5 A and B). Although a similar increase of IFN-γ production was observed in the B16 melanoma-bearing VISTA ko mice, KO TILs failed to produce significantly higher amounts of granzyme B and TNF-α, and tumor growth was not significantly impaired either (Fig. 5 C-E). We speculate that the relatively weak production of TNF-α and granzyme B contributed to the lack of tumor control. In addition, it has been shown that IFN-γ might limit antitumor T-cell responses via up-regulating noncognate MHC class I on tumor cells (21). This result is, however, in contrast to our previous study demonstrating the efficacy of a VISTA-specific monoclonal antibody in the B16 melanoma model (16). We hypothesize that the engagement of Fc receptors by a VISTA-specific antibody might further enhance tumorspecific immune responses. Similar Fc receptor-dependent mechanisms have been shown to be critical for the therapeutic efficacies of antibodies targeting CTLA-4 and GITR (22,23). On the other hand, a recent study described the impaired growth of murine brain glioma cells in VISTA ko mice, as well as the role of VISTA-deficient CD4 + T cells in mediating tumor suppression (20). We speculate that the unique immunogenicity of the glioma tumor model might have contributed to the stronger tumorspecific T-cell responses than those seen in the B16 melanoma model. In conclusion, our study using VISTA genetic-deficient mice has confirmed the immune-suppressive function of VISTA and warrants further studies to tease apart the cellular and molecular mechanisms of VISTA-mediated immune regulation.

Materials and Methods
Mice. C57BL/6 mice were purchased from the National Cancer Institute. VISTA-deficient mice on a fully backcrossed C57BL/6 background were obtained from the Mutant Mouse Regional Resource Centers (18). 2D2 TCR transgenic mice were purchased from The Jackson Laboratory. All animals were maintained in a pathogen-free facility at The Geisel School of Medicine at Dartmouth. All animal protocols were IACUC-approved at Dartmouth College.
Mice Necropsy and Semiquantitative Pathological Analysis. Age-and sexmatched WT and VISTA ko mice were killed by CO 2 asphyxiation. Organs were harvested and fixed in 10% (vol/vol) buffered formalin. Hematoxylin and eosin staining was performed on tissue sections. Tissue inflammatory status was scored in a blind manner by a pathologist using the following semiquantitative scoring criteria: 0, normal; 1, mild/small foci of dense lymphocytic infiltrate; 2, moderate/multiple foci of dense/large activated lymphocytic infiltrate with/without a germinal center; 3, marked reactive/ activated or atypical lymphocytic infiltrate.
Monitoring the Spontaneous Disease Phenotypes in 2D2 Transgenic Mice. Spontaneous development of EAE and optic neuritis phenotypes was monitored as previously described (19). Clinical assessment of EAE was performed according to the following criteria: 0, no disease; 1, limp tails; 2, hind limb paresis or partial paralysis; 3, complete hind limb paralysis; 3.5, front limb paresis; 4, front and hind limb paralysis and moribund state. Disease onset was recorded when mice showed limp tails. The clinical assessment of optic neuritis is based on the following criteria: eyelid redness and swelling associated with tearing, and signs of atrophy of the eye.
Flow Cytometry and Analysis. Flow cytometry analysis was performed either on a FACScan using CellQuest software (BD Biosciences) or on a MACSQuant 7 color analyzer (Miltenyi Biotec). Data analysis was performed using FlowJo software (Tree Star).
Graphs and Statistical Analysis. All graphs and statistical analysis were generated using Prism 4 (GraphPad Software). A Student t test (two-tailed) was used for the data analyses. ***P < 0.005, **P < 0.025, *P < 0.05.