Immune-checkpoint protein VISTA critically regulates the IL-23/IL-17 inflammatory axis

V-domain Immunoglobulin Suppressor of T cell Activation (VISTA) is an inhibitory immune-checkpoint molecule that suppresses CD4+ and CD8+ T cell activation when expressed on antigen-presenting cells. Vsir−/− mice developed loss of peripheral tolerance and multi-organ chronic inflammatory phenotypes. Vsir−/− CD4+ and CD8+ T cells were hyper-responsive towards self- and foreign antigens. Whether or not VISTA regulates innate immunity is unknown. Using a murine model of psoriasis induced by TLR7 agonist imiquimod (IMQ), we show that VISTA deficiency exacerbated psoriasiform inflammation. Enhanced TLR7 signaling in Vsir−/− dendritic cells (DCs) led to the hyper-activation of Erk1/2 and Jnk1/2, and augmented the production of IL-23. IL-23, in turn, promoted the expression of IL-17A in both TCRγδ+ T cells and CD4+ Th17 cells. Furthermore, VISTA regulates the peripheral homeostasis of CD27− γδ T cells and their activation upon TCR-mediated or cytokine-mediated stimulation. IL-17A-producing CD27− γδ T cells were expanded in the Vsir−/− mice and amplified the inflammatory cascade. In conclusion, this study has demonstrated that VISTA critically regulates the inflammatory responses mediated by DCs and IL-17-producing TCRγδ+ and CD4+ Th17 T cells following TLR7 stimulation. Our finding provides a rationale for therapeutically enhancing VISTA-mediated pathways to benefit the treatment of autoimmune and inflammatory disorders.

monoclonal antibody (mAb) enhanced T responses and anti-tumor immunity 1,8 . Our recent study has further demonstrated that VISTA and another B7 family immune-checkpoint PD-1 play non-redundant roles in controlling T cell responses 9 . Mice deficient for both genes developed the most severe inflammatory phenotypes, accompanied by spontaneous activation of CD4 + and CD8 + T cells. Combinational blockade of both VISTA and PD-1 proteins using blocking mAb led to synergized anti-tumor immune responses in murine models.
Irrespective of these findings, whether VISTA regulates innate immune responses is not known. To address this question, we have employed the imiquimod (IMQ)-induced murine model of psoriasis, where topical application of IMQ stimulates a network of innate immune cells, such as DCs and IL-17-producing γδ TCR + T cells, leading to psoriasiform skin inflammation 10 . This murine model bears strong relevance to human psoriasis, which is also mediated by the IL-23/IL-17 inflammatory axis. Multiple studies have reported the presence of both IL-17-producing TCRγδ + T cells (γδ T cells) and CD4 + Th17 cells in human psoriatic skin [11][12][13][14][15] . Treatment in human cancer patients with IMQ (Aldara ® ) has resulted in similar psoriasiform dermatitis, manifested as epidermal acanthosis and parakeratosis [16][17][18] .
In this study, using the IMQ-induced psoriasis model, we have demonstrated that VISTA plays a key role in suppressing the IL-23/IL-17-mediated inflammatory axis. VISTA inhibits the activation of DCs and the production of IL-23 following TLR7 stimulation. VISTA also regulates the activation of IL-17-producing γδ T cells and CD4 + Th17 T cells, as well as the peripheral homeostasis of CD27 − γδ T cell subsets that are pre-committed to produce IL-17A. Consequently, VISTA deficiency exacerbated psoriasiform inflammation. Taken together its role in suppressing CD4 + and CD8 + T cell activation, our study indicates that VISTA is a unique immune-checkpoint that regulates both innate and adaptive immune responses.

Results
Vsir −/− mice developed exacerbated psoriasiform inflammation. To address the role of VISTA in regulating innate immunity, we examined IMQ-induced psoriasiform dermatitis in wild type (WT) and Vsir −/− mice that were topically treated with 3.5% IMQ on both ears. Skin inflammatory response was quantified by measuring ear thickness. Our previous study reported chronic inflammatory phenotypes in aged (>10 month of age) Vsir −/− mice 7 . We first examined untreated naïve Vsir −/− mice (7-8 weeks of age) but did not observe any spontaneous skin inflammation (unpublished data). IMQ treatment in the Vsir −/− mice resulted in more severe ear swelling when compared to WT mice (Fig. 1a). Histological analyses confirmed the development of severe epidermal acanthosis in the Vsir −/− ear skin (Fig. 1b) and increased epidermal thickness (Fig. 1c). Furthermore, Vsir −/− ears showed ~20 fold increase in the area of neutrophilic abscesses (Munro's abscess), which is a histological hallmark in human psoriasis 18 (Fig. 1d).
In addition to Gr1 + neutrophils, further examinations show that multiple other cell types were present in the WT psoriatic skin lesions, including CD11c + DCs, CD4 + T cells, and γδ T cells (Fig. 1e). Similar lymphocyte populations were present in the ears of the Vsir −/− mice (data not shown). VISTA was highly expressed on these cells types (Fig. 1f), all of which are known to regulate the development of IMQ-induced psoriasiform inflammation [17][18][19][20] . Both γδ low and γδ high T cells were present in the ear skin, whereas only γδ low T cells were present in the ear-draining lymph nodes (LN). VISTA expression on γδ low T cells was higher from inflamed ear skin than cells from the draining LN (Fig. 1f). To further determine whether the exacerbated psoriasiform inflammation was due to the pre-existing inflammatory environment in the Vsir −/− mice, we treated mice with a VISTA-specific mAb 8 . Consistent with results from the Vsir −/− mice, VISTA-specific mAb treatment significantly enhanced IMQ-induced psoriasiform inflammation (Fig. 1g), although the magnitude of disease was not as severe as those seen in the Vsir −/− mice.
Exaggerated neutrophil infiltration may potentially result from an augmented production of inflammatory cytokines and chemokines 18 . To investigate the inflammatory milieu, mRNA from IMQ-treated ear skin was harvested and examined by quantitative RT-PCR (Q-PCR). A panel of psoriasis-associated genes was induced in both WT and Vsir −/− skin following IMQ treatment (Fig. 2a). Vsir −/− skin showed significantly higher expression of inflammatory cytokine genes Il23p19, Il1β, Il6, Il17a, Il22, Tnfα, and Ifnγ, chemokine gene Cxcl2, and neutrophil chemotactic gene S100a9 (Fig. 2a). Serum protein levels of IL-1β, IL-6, IL-17A, IFN-γ, and CXCL2 were higher in the Vsir −/− mice (Fig. 2b). Serum levels of S100A9, TNF-α, IL-22, and IL23 were very low and were not reliably detected before or after IMQ treatment (unpublished results). Serum levels of IL-22 and TNF-α have been previously reported in mice following treatment with 5% IMQ cream on the back skin 20 . To further confirm that VISTA deficiency resulted in enhanced protein production of IL-22 and TNF-α, WT and Vsir −/− mice were treated with 3.5% IMQ cream on the shaved back skin. Serum was harvested six hours after the treatment, and the concentration of IL-22 and TNF-α in the serum was examined by ELISA. Consistently, IMQ treatment of back skin led to accumulation of higher levels of IL-22 and TNF-α in the serum of the Vsir −/− mice ( Supplementary  Fig. 1). Together, these results indicate that VISTA inhibits the expression of inflammatory cytokines and chemokines in response to IMQ.
VISTA regulates the production of IL17 by both γδ T cells and CD4 + Th17 cells. The IL-23/IL-17 inflammatory axis has a well-established role in the development of inflammatory and autoimmune diseases 21,22 . In psoriatic lesions, high levels of IL-17A induce the release of neutrophil chemoattractants from keratinocytes, thus amplifying the inflammatory cascade. IL-22, another key Th17 cytokine, promotes sustained epidermal acanthosis and parakeratosis 18,23,24 .
Both γδ T cells and CD4 + Th17 cells are IL-17-producing effector cells that drive human psoriasis and the IMQ-induced model of psoriasiform inflammation 10,[12][13][14] . To investigate the contribution of γδ T cells and CD4 + Th17 cells during the development of psoriasiform inflammation in the Vsir −/− mice, we first examined their cell number in the ear and ear-draining LN. No significant difference was observed between WT and Vsir −/− mice (Fig. 3a). Since Il17a mRNA level was significantly augmented in the Vsir −/− skin (Fig. 2), we examined IL-17A protein expression in γδ T cells and CD4 + cells isolated from IMQ-treated WT and Vsir −/− mice. Vsir −/− γδ T cells in ear skin and draining LN produced significantly higher amount of IL-17A, but similar levels of IFN-γ and TNF-α when compared to WT cells ( Fig. 3b and Supplementary Fig. 2). Vsir −/− CD4 + T cells also expressed higher amount of IL-17A than WT cells in the ear and ear-draining LN, though the percentage of IL-17A positive cells was much lower than γδ T cells (Fig. 3c). Thus, both γδ T cells and Th17 cells contributed to the higher IL-17A production in IMQ-treated Vsir −/− mice. In addition to γδ T cells and CD4 + T cells, CD11b + myeloid cells expressed IFN-γ and TNF-α ( Supplementary Fig. 3a). Higher numbers of IFN-γ and TNF-α-expressing CD11b + cells were found to infiltrate IMQ-treated ears in the Vsir −/− mice, which may contribute to the higher levels of IFN-γ and TNF-α within the inflamed ear tissues ( Supplementary Fig. 3b). Similar numbers of IFN-γ and TNF-α-expressing Cd11b + cells were found in the ear-draining LN ( Supplementary Fig. 3c).
To further understand the mechanisms whereby VISTA regulates the homeostasis and activation of γδ T cells, we examined the subsets of γδ T cells in naïve WT and Vsir −/− mice 25,26 . The CD27 + and CD27 − γδ T cell subsets are developed in thymus before exiting to the periphery 25 . The CD27 − γδ T cells express higher levels of IL-1R and IL-23R, and are pre-committed to produce IL-17A upon TCR-or cytokine-mediated activation, whereas CD27 + γδ T cells predominantly produce IFN-γ 25,27 . To determine if loss of VISTA altered the development and peripheral homeostasis of γδ T cells, thymic and splenic γδ T cell subsets in WT and Vsir −/− mice were examined. Similar percentages of total γδ T cells and the Vγ4 + subset were observed in the spleen, indicating an overall normal development of γδ T cells in the absence of VISTA (unpublished data). On the other hand, higher percentage of the CD27 − subset was present within the splenic but not thymic Vsir −/− γδ T cells (Fig. 4a). This result indicates that VISTA regulates the peripheral homeostasis of CD27 − γδ T subsets.
It has been shown that IL-7 preferentially expands the IL-17A-producing CD27 − γδ T subset and promotes their peripheral homeostasis 28,29  with IL-7 for 4 days, and the number of CD27 − γδ T cells was examined. IL-7 treatment expanded both WT and Vsir −/− CD27 − γδ T subsets. However, Vsir −/− CD27 − γδ T cells were hyper-proliferative than WT cells, resulting in a ~10 fold increase in viable cell number and a ~5 fold increase in percentage within the total expanded γδ T cell population (Fig. 4b).
To determine how VISTA regulates the IL-7 receptor signaling, the phosphorylation status of STAT3 and STAT5 was examined, since both proteins are known to mediate IL-7 receptor signaling 30 . Our results show that higher level of phosphorylated STAT5 was induced in Vsir −/− γδ T cells than WT cells following IL-7 stimulation (Fig. 4c). On the other hand, similar level of phosphorylated STAT3 was observed (data not shown). These results indicate that VISTA restricts the activation of STAT5 but not STAT3 downstream of IL-7R signaling in γδ T cells. γδ T cells are activated by both TCR-specific stimuli and inflammatory cytokines such as IL-1β and IL-23 26 . Since VISTA is expressed on γδ T cells (Figs 1f and 4d), it is possible that VISTA suppresses the activation of γδ T cells in an autonomous manner. To test this, naïve splenic WT and Vsir −/− γδ T cells were isolated and stimulated with either TCR crosslinking, or cytokines IL-1β and IL-23. Vsir −/− γδ T cells produced more IL-17A and IL-22 than WT cells (Fig. 4e). Activated Vsir −/− γδ T cells consistently expressed higher level of RORγt, a ROR family transcription factor that binds to and activates the Il-17a promoter (Fig. 4f) 31,32 . Both CD27 + and CD27 − Vsir −/− γδ T subsets expressed more IL-17A than WT cells following IL-1β and IL-23 stimulation. These data indicate that in addition to regulating the peripheral homeostasis of CD27 − γδ T cells, VISTA directly controls the activation of γδ T cells (Fig. 4g).

VISTA expression on dendritic cells suppresses IMQ-induced TLR7 signaling and IL-23 production.
In both human psoriasis and murine models of psoriasiform inflammation, IL-23 is predominantly produced by myeloid DCs and promotes the expansion of pathogenic IL-17A-producing γδ T cells and CD4 + Th17 cells 17,20,24 . Since an elevated expression of IL23 gene was observed in IMQ-treated ear skin from Vsir −/− mice (Fig. 2), we hypothesize that VISTA expression on DCs suppresses IMQ/TLR7-induced IL-23 production. To test this hypothesis, WT and Vsir −/− mice were treated with 3.5% IMQ on the ears for 4 days. Ear tissues were harvested and the expression of IL-23p19 in ear CD11c + DCs were examined by flow cytometry. Significantly higher  percentages of Vsir −/− DCs expressed IL-23p19 protein than WT cells in ears and ear-draining LNs (Fig. 5a), whereas the number of DCs present in ears and the draining LNs was similar (Fig. 5b).
Although there was similar expression of DC activation markers such as CD80, CD86, and CD40 on naive WT and Vsir −/− splenic DCs ( Supplementary Fig. 4), it could not be formally excluded that an altered DC development in Vsir −/− mice may contribute to the hyper-response of DCs. To directly demonstrate the role of VISTA in suppressing DC cytokine production, we ectopically expressed either full-length VISTA, or a mutant VISTA lacking the cytoplasmic tail (deltaC), or GFP control protein in GM-CSF cultured Vsir −/− BM-derived DCs. Following stimulation with a TLR7/8 agonist R848, secreted IL-23 was examined by ELISA. Expression of both FL-VISTA and deltaC-VISTA significantly suppressed IL-23 production in Vsir −/− BMDCs (Fig. 5c). This result strongly supports the role of VISTA in inhibiting TLR7-mediated DCs activation and IL-23 expression. Furthermore, since the cytoplasmic tail is not required for the suppressive activity of VISTA, this result indicates that VISTA engages an unknown receptor, which in turn delivers an inhibitory signal.
The Il-23 promoter contains binding sites for AP-1 and NF-κB 33 . It has been shown that TLR4 stimulation in macrophages and DCs activates MAP kinases (Erk1/2, Jnk1/2, and p38), which are critical for the activation of transcription factor AP1 and the expression of Il-23p19 gene 33,34 . Furthermore, Erk1/2 inhibitor suppressed IL-23 production in DCs stimulated with TLR agonists 34 . To determine if VISTA regulates the activation of NF-κB and MAPK pathways, total cell lysates were prepared from WT and Vsir −/− splenic DCs that have been stimulated with R848 and examined by western blotting (Fig. 5d). R848 stimulation induced significantly higher levels of Erk1/2 phosphorylation and a moderately increased phosphorylation of Jnk1/2 in Vsir −/− DCs (Fig. 5d). On the contrary, similar levels of Iκ-B degradation and phosphorylation of NF-κB p65 were observed, indicating that the NF-κB pathway was not significantly altered in the absence of VISTA (Fig. 5e). Similar levels of phosphorylated p38 were present in lysates from WT and Vsir −/− BMDCs (unpublished data).

Discussion
The IL-23/IL-17-mediated inflammatory axis plays a critical role in many inflammatory disorders and autoimmune diseases such as psoriasis, rheumatoid arthritis, multiple sclerosis, and inflammatory bowel disease 22 .
In the current study we have demonstrated a novel role of VISTA in regulating this inflammatory axis. In the IMQ-induced psoriasis model, VISTA deficiency augmented the inflammatory responses of DCs, γδ T cells, and Th17 cells, resulting in exacerbated psoriasiform dermatitis.
In both human psoriasis and murine model of psoriasiform inflammation, one of the main initial responders are IL-23-producing DCs 17 . IL-23 promotes the expansion and activation of IL-17-producing CD4 + Th17 cells and γδ T cells. This inflammatory milieu recruits and activates additional effector cells such as inflammatory monocytes and neutrophils, which amplify inflammation and drive epidermal hyperplasia. Our results indicate that VISTA controls the production of IL-23 in DCs via inhibiting the activation of Erk1/2. We predict that strategies that enhance VISTA-regulated inhibitory signaling will dampen IL-23-mediated inflammatory axis and benefit the treatment of not only human psoriasis, but also other inflammatory diseases driven by IL-23.
In addition to regulating the activation of DCs, VISTA negatively regulates IL-7-mediated homeostasis of CD27 − γδ T cells, as well as the activation of γδ T cells in response to TCR-mediated or IL-23/IL-1β-mediated stimuli. These effects collectively contribute to the exaggerated psoriasiform inflammation in the Vsir −/− mice. It is noted that VISTA expression on γδ T cells was upregulated within the psoriatic skin when compared to the draining LN, indicating a potential feedback mechanism whereby inflammatory cytokines or other mediators may upregulate VISTA expression to dampen inflammation.
In addition to VISTA, other immune-checkpoint proteins including Programmed death-1 (PD-1) and B and T lymphocyte attenuator (BTLA) also regulate IL17 expression in γδ T cells 35,36 . Both receptors are expressed on γδ T cells and restrict their activation. PD-1 and BTLA knockout mice developed more severe psoriasiform dermatitis in the IMQ model 35,36 . These results warrant future efforts to determine whether these immune-checkpoint proteins act synergistically to regulate the function of γδ T cells.
In addition to psoriasis, the IL-23/IL-17 inflammatory axis regulates disease development in murine experimental autoimmune encephalomyelitis (EAE) and human autoimmune disease multiple sclerosis 26,37,38 . Previous studies have shown that VISTA genetic deletion or VISTA-blocking mAb treatment exacerbated disease in the EAE model 1,7,9 . Since both IL-17-producing γδ T cells and Th17 cells have been implicated as effector cells during EAE 39,40 , our current study provides additional mechanisms whereby VISTA regulates this disease.
In the context of cancer therapy, the IL-23/IL-17 inflammatory axis regulates the inflammatory tumor microenvironment (TME). Earlier studies have demonstrated the tumor-promoting role of IL-23 41,42 , whereas both tumor-promoting and tumor-inhibitory roles of IL-17 have been reported [43][44][45][46][47] . Results from this study indicate that blocking VISTA promotes the inflammatory responses mediated by IL-23/IL-17, particularly in the context of TLR stimulation. Our previous study has shown that VISTA-blocking mAb synergized with a tumor peptide vaccine and TLR agonists as adjuvants 8 . Future studies are warranted to determine whether the exacerbated IL23/IL17 inflammatory axis positively or negatively contributes to the anti-tumor immunity following VISTA blockade.
In conclusion, this study reveals a novel anti-inflammatory role of VISTA through regulating the IL-23/IL-17 inflammatory axis. Our findings distinguish VISTA from other immune-checkpoint proteins CTLA-4 and PD-1, and establish VISTA as a regulator of both innate and adaptive immunity 1,4,6-9 . Therapeutic agents have been developed to harness the immune-suppressive functions of immunecheckpoint proteins. For example, a fusion protein CTLA4-Ig (Abatacept) has been used in the clinic for treating autoimmune diseases such as rheumatoid arthritis 48 . Similarly, local overexpression of PD-L1-Ig or administration of purified PD-L1-Ig has been shown to promote allograft survival in murine models [49][50][51][52] . Our study indicates that enhancing the anti-inflammatory function of VISTA may benefit the treatment of a variety of inflammatory and autoimmune disorders.

Materials and Methods
Mice. C57BL/6 mice were purchased from Charles River Laboratories. Vsir −/− mice on a fully backcrossed C57BL/6 background were as described 7,9 . All animals were maintained in a pathogen-free facility at the Medical College of Wisconsin (Milwaukee, WI). All animal protocols were approved by the Institutional Animal Care and Use Committee of the Medical College of Wisconsin. All methods were performed in accordance with the relevant guidelines and regulations.
with R848 (5 μg/ml) for indicated amount of time. Total cell lysates were prepared and examined for the levels of phosphorylated Erk1/2 and Jnk1/2 by western blotting (d). The ratio of phosphorylated versus total Erk1/2 and Jnk1/2 was calculated based on the total protein level from the same lysate run on a parallel gel (d). The activation of NF-κB signaling was examined by western blotting the level of phosphorylated and total IκB, as well as phosphorylated and total NF-κB p65 subunit (e). To determine whether Erk and Jnk were required for the production of IL-23, Splenic DCs were isolated from naïve WT and Vsir −/− mice, and stimulated with R848 (5 μg/ml) in the presence of Erk1/2 inhibitor (SCH772984, 10 μM), or Jnk1/2 inhibitor (SP600125, 10 μM), or DMSO solvent control for overnight. Culture supernatant was collected and secreted IL-23p19/p40 was quantified by ELISA. Values from triplicated cultures are shown as mean ± SEM in (f). Representative results from two to three independent experiments were shown.