SKI Expression Suppresses Pathogenic Th17 Cell Response and Mitigates Experimental Autoimmune Encephalomyelitis

Pathogenic Th17 cells are critically involved in many autoimmune diseases, while non-pathogenic Th17 cells are more immune regulatory. Understanding the mechanisms of the induction and maintenance of pathogenic Th17 cells will benefit the development of therapeutic treatments of related diseases. We have shown that the transforming growth factor-β (TGFβ) induced SKI degradation and dissociation from Smad4 complex is a prerequisite for TGFβ-induced Th17 cell differentiation. However, it is unclear whether and how SKI regulates pathogenic Th17 differentiation, which does not require TGFβ cytokine. Here we showed that SKI expression was downregulated during pathogenic Th17 cell differentiation and the ectopic expression of SKI abrogated the differentiation of pathogenic Th17 cells. Functionally, using a knock-in mouse model, we found ectopic SKI expression specifically in T cells prevented myelin oligodendrocyte glycoprotein peptide (MOG33–55) induced experimental autoimmune encephalomyelitis (EAE), an animal model of human multiple sclerosis. We further revealed that induced SKI expression in already differentiated pathogenic Th17 cells reduced the maintenance of Th17 program and ameliorated EAE in an adoptive T cell transfer model. Therefore, our study provides valuable insights of targeting SKI to modulate pathogenic Th17 cell function and treat Th17-related diseases.


INTRODUCTION
IL-17-producing CD4+ T helper 17 (Th17) cells are a heterogeneous population and can be classified into pathogenic and non-pathogenic Th17 cells based on their function (1,2). While the non-pathogenic Th17 cells are involved with tissue homeostasis (3)(4)(5), the pathogenic Th17 cells contribute to many human autoimmune diseases (6), including multiple sclerosis, inflammatory bowel disease, rheumatoid arthritis, psoriasis and type 1 diabetes (7)(8)(9)(10). The pathogenicity of Th17 cells is influenced by both genetic and environmental factors including cytokines. While the combination of IL-6, IL-1b and IL-23 instructs the differentiation program of pathogenic Th17 cells, TGFb1 plus IL-6 can induce the differentiation of nonpathogenic Th17 cells. However, the detailed mechanism that dictates pathogenic verses non-pathogenic Th17 program is not fully understood. The understanding of the genetic and molecular mechanisms of pathogenic Th17 cell differentiation and maintenance is important for the development of therapeutic targets to treat Th17-related diseases.
TGFb signaling plays important roles in the development, differentiation and function of T cells (11,12). TGFb is required to promote non-pathogenic Th17 cell differentiation (13,14). However, the understanding of how TGFb signaling is regulated in pathogenic Th17 cell differentiation is evolving. It was first reported that the generation of pathogenic Th17 was independent of TGFb1 (9). A later study showed that TGFb3, instead of TGFb1, was induced and important for the differentiation of pathogenic Th17 cells (10), indicating TGFb signaling may differentially regulated in pathogenic and nonpathogenic Th17 cells. Additionally, the TGFb superfamily member Activin has also been reported to be able to induce Th17 cells together with IL-6 (15,16) and Activin-A-induced Th17 cells are more pathogenic than those induced by TGFb1 (17), further diversifying the involvement of TGFb superfamily signaling in Th17 cells. Nevertheless, underlying mechanisms of how signaling from TGFb and its superfamily members dictate pathogenic Th17 cells and non-pathogenic Th17 cells is not fully understood. More importantly, it is unclear whether there is a central regulator that can sense and integrate TGFb superfamily signaling to differentially modulate pathogenic and nonpathogenic Th17 cells.
We have shown that TGFb signaling induced dissociation of SKI-Smad4 complex is a prerequisite for Th17 cell differentiation (16). As a suppressor of TGFb signaling, SKI interacts with Smad4 to restrain the expression of RORgt, the central regulator of Th17 program (18,19). During Th17 cell differentiation, TGFb induces the degradation of SKI and release the inhibition on RORgt expression, which induces the Th17 program. Interestingly, our previous results show that SKI can also be regulated and degraded by Activin-A, indicating SKI could be a central regulator that can sense and integrate TGFb superfamily signaling (16). However, whether and how SKI regulates pathogenic Th17 differentiation, which does not require TGFb cytokine, remains unknown. Importantly, the functional importance of SKI in the pathogenic Th17 cells related autoimmune diseases remains elusive.
In this study, we investigated the role of SKI in regulating the differentiation and maintenance of pathogenic Th17 cells. We found that SKI expression was downregulated during pathogenic Th17 differentiation, which can be reversed by the TGFb receptor inhibitor SB431542. By using a knock-in mouse model (SKI-KI) that express SKI specifically in T cells, we demonstrated that SKI expression blocked pathogenic Th17 cell differentiation but had less effect on TGFb-induced nonpathogenic Th17 cell differentiation. Functionally, SKI-KI mice were refractory to experimental autoimmune encephalomyelitis (EAE), an animal model of human multiple sclerosis (20,21). We further showed that ectopic expression of SKI in already differentiated pathogenic Th17 cells reduced the maintenance of Th17 program and ameliorated EAE symptoms in an adoptive T cell transfer model.

Animals
Rosa26 iSki , Rag1 −/− , MOG 35-55 -specific 2D2 TCR-transgene, Cd4Cre, ERCre and CD45.1 congenic wild-type mice were on C57BL/6 background. Rosa26 iSki mice were bred with Cd4Cre and ERCre mice to generate Rosa26 iSki Cd4Cre (SKI-KI) and Rosa26 iSki ERCre (SKI-ER) mice, respectively. All mice were housed and bred under specific pathogen-free conditions in the animal facility at the University of North Carolina at Chapel Hill. All mouse experiments were approved by the Institution Animal Care and Use Committee of the University of North Carolina at Chapel Hill.
For the in vitro maintenance of pathogenic Th17 cells, cells from 4 days of pathogenic Th17 cell differentiation were harvested and transferred into 24-well plated coated with 2 µg/ ml anti-CD3 and 2 µg/ml anti-CD28 antibodies and cultured in serum-free X-VIVO 20 medium for another 4 days. Samples for FACS, immunoblotting and realtime PCR were harvested at indicated time points.

Experimental Autoimmune Encephalomyelitis
For MOG  induced EAE, age (8-12 weeks old) and gender matched mice were subcutaneously immunized (s.c.) with 50 mg of MOG  peptide (MEVGWYRSPFSR VVHLYRNGK, AnaSpec) and 500 mg of M. tuberculosis (Difco) emulsified in IFA (Difco). In addition, the mice were administrated 200 ng of Pertussis Toxin (List Biological Laboratories) intraperitoneally (i.p.) on days 0 and 2, respectively. For adoptive transferred EAE, 2D2 CD4 + T cells were differentiated under pathogenic Th17 differentiation conditions for 4 days and re-stimulated in fresh medium in the presence of platebound anti-CD3 (2 µg/ml) and anti-CD28 (2 µg/ml) antibodies for 2 days. Live cells were purified as described. Approximately 1 × 10 6 purified cells were i.v. injected into Rag1 −/− mice. Approximately 200 ng of Pertussis toxin (List Biological) was injected intraperitoneal (i.p.) on days 0 and 2 after T cell transfer. For induced SKI expression in adoptive T cell transfer induced EAE, tamoxifen (TMA group) or corn oil (Control group) were injected intraperitoneally at days 1 and 3 after EAE elicitation as illustrated.
The severity of EAE was monitored and graded on a clinical score of 0 to 5 by the following standard: 0 = no clinical signs, 1 = Limp tail, 2 = Para-paresis (weakness, incomplete paralysis of one or two hind limbs), 3 = Paraplegia (complete paralysis of two hind limbs), 4 = Paraplegia with forelimb weakness or paralysis, 5 = Moribund or death. The statistical significance was analyzed by two-way multiple-range ANOVA test.
After EAE elicitation, mice were sacrificed and perfused with ice-cold phosphate buffered saline containing 20 U/ml of heparin. Spinal cords were separated from spine columns after removal of all tissues. The isolated spinal cords were minced and digested with 1 mg/ml collagenase D (Sigma) for 45 min at 37°C. The digested tissues were filtered with 40 mm strainer and centrifuged, thereafter resuspended into 38% percoll (Sigma) and centrifuged at 2,000 rpm for 20 min to separate lymphocytes. Lymphocytes were then isolated and subjected to FACS analysis.

RNA Preparation and Realtime PCR
Total RNA was prepared from T cells using TRIzol reagent (Invitrogen) per manufacturer's instructions and was reversetranscribed into cDNA with iScript ™ cDNA Synthesis Kit (BioRad, #1708891). Quantitative PCR was performed on ABI9700 real-time PCR system with Taqman-probe sets purchased from Applied Biosystems and Integrated DNA Technologies (IDT). The relative mRNA expression of indicated genes was calculated based on the expression of b-actin.

Statistical Analysis
Data analysis was processed and presented by Prism (GraphPad, San Diego). Two-tailed Student's t-test was used to compare two groups of samples. One-way ANOVA post Holm-Sidak's multiple comparisons test was used to compare data with multiple groups. Two-way multiple-range ANOVA test was used to analyze clinical score in EAE experiments. A p value of less than 0.05 was considered significant.

The TGFb Signaling Was Required for the Induction of Pathogenic Th17 Cells
To understand whether TGFb signaling is required for the differentiation of pathogenic Th17 cells, we blocked the TGFb signaling with a pharmacological inhibitor SB431542. We found that SB431542 nearly completely suppressed the differentiation of pathogenic Th17 cells induced by cytokines IL-6, IL-1b and IL-23 ( Figure 1A). Our result does not contradict with the previous reports that TGFb signaling is dispensable for the differentiation of pathogenic Th17 cells using anti-TGFb blocking antibodies (9), as SB431542 not only targets ALK5 (TGFb type I receptor), but also inhibits ALK4 (Activin type I receptor) and ALK7 (Nodal type I receptor, another TGFb superfamily member) (22). Additionally, we found SB431542 treatment could drastically increase SKI protein expression ( Figure 1B), indicating the presence of TGFb signaling and regulation of SKI protein level during pathogenic Th17 cell differentiation. Our result revealed that, although no TGFb cytokine was added during pathogenic Th17 cell differentiation, the endogenous TGFb signaling was activated from either TGFb cytokines or its superfamily members, which could be blocked by the inhibitor SB431542.

Ectopic Expression of SKI in SKI-KI Mice Blocked the Differentiation of Pathogenic Th17 Cells
Next, we asked whether ectopic expression of SKI in T cells is sufficient to block the generation of pathogenic Th17 cells by using Rosa26 iSki Cd4Cre (SKI-KI) mice (23) that ectopically express WT SKI in T cells (Supplementary Figure 1A). We first examined the protein expression of SKI during pathogenic Th17 and TGFb induced Th17 differentiation. We found while there was an increase of SKI expression in SKI-KI T cells under Th0 (IL-2 only) condition, the SKI expression was drastically downregulated in TGFb induced non-pathogenic Th17 cells but modestly downregulated in pathogenic Th17 cells ( Figure 1C). This result indicated that SKI functions differently in pathogenic and non-pathogenic T cells.
Indeed, the differentiation of pathogenic Th17 cells was severely blocked in SKI-KI group, and only less than 1% of IL-17A + cells were generated from SKI-KI CD4 + T cells compared to around 15% in WT controls ( Figure 1D). However, the blockade of Th17 cell differentiation in TGFb-induced nonpathogenic Th17 conditions was much less efficient, with 25% of IL-17A + cells in SKI-KI compared to 40% in WT controls ( Figure 1E). The defect of Th17 differentiation in SKI-KI T cells was less likely due to cell survival, as no significant increase of apoptotic cells was observed in either pathogenic Th17 cells or non-pathogenic Th17 cells in SKI-KI group ( Supplementary  Figures 1B, C). To examine whether the severe defect of pTh17 differentiation in SKI-KI T cells is cell intrinsic or caused by inflammatory cytokines secreted by effector or memory cells, naïve CD4 + cells from WT (CD45.2 -) and SKI-KI (CD45.2 + ) mice were purified, mixed together and differentiated under pathogenic Th17 conditions. Indeed, the defect of pathogenic Th17 differentiation persisted (Supplementary Figure 1D), indicating that SKI expression blocks pathogenic Th17 differentiation in a cell intrinsic manner. The preferential perturbation of pathogenic Th17 differentiation was further confirmed by the mRNA expression of Th17 related genes,  ) and (E) respectively. Relative expression levels were measured by quantitative real-time RT-PCR and were normalized to actin expression level using the standard curve method. The results were from three independent experiments; n = 3-6 mice of three independent experiments; means ± s.d., ns, not significant, **p < 0.01, ****p < 0.0001, by two-sided Student's t-test.
such as Il17a, Il17f, Rorc and Il23r with real-time RT-PCR ( Figures 1F, G). Thus, our results demonstrated that SKI expression could efficiently block the generation of pathogenic TH17 cells with less effect on TGFb-induced non-pathogenic Th17 cells.

SKI-KI Mice Were Refractory to EAE
The preferential blockade of pathogenic Th17 cells in SKI-KI T cells prompted us to investigate whether ectopic expression of SKI in T cells could inhibit EAE, where pathogenic Th17 cells are critically involved (9,(24)(25)(26). To elicit EAE, WT and SKI-KI mice were immunized with MOG 33-55 peptide on day 0 and treated with pertussis toxin (PTX) on days 0 and 2. We found all the WT mice rapidly developed very severe EAE symptoms from 10 days after immunization, whereas SKI-KI mice showed much less incidence of EAE (Figure 2A). Importantly, the average clinical score and regression clinical score were drastically lower in SKI-KI mice compared to WT controls ( Figures 2B, C). Accordingly, much fewer infiltrating cells were recovered from spinal cords of SKI-KI with EAE compared to WT controls, although there was an increase of lymphocytes in spleens and draining lymph nodes of SKI-KI mice at day 25 of EAE ( Figure 2D). Further analysis of T cells in draining lymph nodes, spleens, and spinal cords of SKI-KI mice revealed that IL-17A producing T cells, but not IFNg producing T cells, were greatly reduced in SKI-KI mice ( Figure 2E). Although we observed an increase of IFNg producing T cells in the spleens and lymph nodes from SKI-KI mice, it seemed that it was not sufficient to promote disease development without Th17 cells. As a result, the number of IL-17A producing T cells, as well as IFNg producing T cells was greatly reduced in the spinal cords of SKI-KI mice ( Figure 2F). Therefore, these results demonstrated that ectopic overexpression of SKI in T cells was sufficient to block pathogenic Th17-mediated autoimmunity in EAE model.

Induced SKI Expression Was Effective to Halt Pathogenic Th17 Cell Differentiation
As Cd4Cre mediated SKI activation occurred in early T cell development in Rosa26 iSki Cd4Cre (SKI-KI) mice, all naïve SKI-KI T cells already expressed high level of SKI before T cell activation and differentiation (Supplementary Figure 1A), making it difficult to study the inhibition role of SKI during pathogenic Th17 cell differentiation. To understand whether SKI could suppress pathogenic Th17 differentiation after T cell activation, we generated the Rosa26 iSki ERCre (SKI-ER) mice by breeding Rosa26 iSki mice with ERCre mice (27), so that the expression of SKI can be induced at the time of our choice with tamoxifen in vitro and in vivo (27). While comparable pathogenic Th17 cells were generated in SKI-ER group in the absence of 4-hydroxytamoxifen (4-OHT), 4-OHT treatment at day 0 effectively abrogated pathogenic Th17 cell differentiation in SKI-ER but not in WT controls ( Figure 3A), consistent with the increased protein expression of SKI in SKI-ER T cells as early as one day after 4-OHT treatment ( Figure 3B).
Next, we induced SKI expression at different time points during ongoing pathogenic Th17 differentiation by adding 4-OHT at the indicated time points ( Figure 3C). We found that SKI induction within 48 hours of T cell activation was sufficient to halt the ongoing differentiation of pathogenic Th17 cells (Figures 3D, E). Yet, no obvious suppression was observed when SKI was induced at 72 hours of T cell activation ( Figures 3D, E), which could be due to insufficient time to allow effective SKI expression in SKI-ER T cells.

SKI Expression Impaired the Maintenance of Pathogenic Th17 Cells In Vitro
To further investigate whether SKI controls the maintenance of already differentiated pathogenic Th17 cells, we first generated Relative expression levels were measured by quantitative real-time RT-PCR and were normalized to actin expression level using the standard curve method. The data were pooled from three to six mice per group of three independent experiments, n = 3-6 mice; means ± s.d., ns, not significant p > 0.05, ***p < 0.001, ****p < 0.0001, by one-way ANOVA post Holm-Sidak's multiple comparisons test.
pathogenic Th17 cells from SKI-ER T cells without 4-OHT treatment. Then we induced SKI expression in these cells by adding 4-OHT. After additional four days of culture, we found there was a decrease of the percentage of IL-17 + T cells in 4-OHT treated group compared to the control group ( Figure 4A). Moreover, we observed a downregulation of Il17a, Il17f, Rorc and Il23r in pathogenic Th17 cells two days after 4-OHT treatment by real-time RT-PCR ( Figure 4B), coinciding with induced SKI expression ( Figure 4C). These findings suggest that SKI suppressed not only the induction but also the maintenance of pathogenic Th17 cells.

Induced SKI Expression Impaired the Maintenance of Pathogenic Th17 Cells and Mitigated Pathogenic Th17 Cell Elicited EAE
The in vitro results encouraged us to examine whether SKI also controls the maintenance of pathogenic Th17 in vivo and if so whether induced SKI expression can ameliorate Th17-related diseases. We generated Rosa26 iSki ERCre 2D2 (SKI-ER-2D2) mice, where T cells harbor MOG 33-55 specific TCR transgene (28). SKI-ER-2D2 pathogenic Th17 cells were differentiated and then transferred to Rag1 −/− mice to elicit EAE ( Figure 5A) (29). Tamoxifen was then administrated into recipient mice via intraperitoneal injection (i.p.) to induce SKI expression in transferred SKI-ER-2D2 pathogenic Th17 cells. We found that IL-17 + T cells, but not IFNg+ T cells, were reduced in the blood from tamoxifen treated mice compared to the controls 8 days after T cell transfer ( Figure 5B), indicating SKI expression suppressed the maintenance of already differentiated pathogenic Th17 cells in vivo. Importantly, tamoxifen treatment effectively reduced the clinical scores of EAE induced by the transferred SKI-ER-2D2 pathogenic Th17 cells ( Figure 5C) and suppressed the immune cell infiltration in spinal cords ( Figure 5D). Further analysis revealed that tamoxifen treatment suppressed the population of IL-17 + IFNgand IL-17 + IFNg + T cells in both spleens and spinal cords. In contrast, IL-17 -IFNg + population was slightly increased in spinal cords ( Figure 5E), indicating SKI expression specifically affected Th17 cell population during EAE. In summary, our results showed that SKI induction in T cells can disrupt pathogenic Th17 cell response and to mitigate Th17 cell promoted disease.

DISCUSSION
Pathogenic Th17 cells drive the pathogenesis of many autoimmune diseases, including multiple sclerosis, inflammatory bowel disease, rheumatoid arthritis, psoriasis, and type 1 diabetes (7)(8)(9)(10). The study of the regulation of pathogenic Th17 cell differentiation and maintenance will provide novel insights into the development of therapeutic targets to treat these diseases. Unlike the wellestablished requirement of TGFb in non-pathogenic Th17 cells (3,10), the role of TGFb signaling in regulating pathogenic has been evolving. Evidence suggests that other TGFb superfamily members such as TGFb3 and Activin, instead of TGFb1; may be critical for the differentiation of pathogenic Th17 cells (10,17). It is important to address the question of how different signaling from TGFb family members are relayed by downstream pathways to control the functionally distinct pathogenic and non-pathogenic Th17 cells.
In this study, we showed that ectopic expression of SKI, a suppressor of TGFb signaling, in T cells could efficiently block pathogenic Th17 cell differentiation and the development of EAE, but only partially affected TGFb-induced non-pathogenic Th17 cells. We further showed that induced expression of SKI in already differentiated pathogenic Th17 cells impaired their maintenance and function in adoptive transferred EAE. Therefore, our study provided novel insights of how the TGFb signaling could be differentially regulated by SKI to modulate the differentiation and maintenance of pathogenic Th17 cells. Many approaches targeting TGFb signaling have been developed, including blocking antibodies, dominant negative mutations, and small molecule inhibitors (30)(31)(32). However, its application to Th17 cells is very limited. One reason could be that the TGFb signaling plays very important roles in many aspects of T cell biology including development, differentiation and homeostasis (11,12,33). Since there is a lack of clear molecular target that is specifically required for pathogenic Th17 cells, blocking TGFb signaling also affect the other immune cells, especially Treg cells, resulting in systemic lethal autoinflammation and autoimmunity. By using the Rosa26 iSki Cd4Cre (SKI-KI) mouse strain, we were able to efficiently block the differentiation of pathogenic Th17 cells, but only partially disrupted the TGFb-induced non-pathogenic Th17 cells. In addition, Treg cell populations are unperturbed in SKI-KI mice which are phenotypically normal without developing systemic autoimmune syndrome (23). These results revealed that SKI could be a new target to fine tune TGFb superfamily signaling to specifically modulate pathogenic TH17 cells. Thus, our study provides important clues to manipulate TGFb signaling for therapeutic purposes.
Our study reveals a critical role of SKI in regulating pathogenic Th17 cells and related autoimmune disease. The further study of the detailed molecular regulation of SKI protein will provide more clues of how SKI can be targeted by therapeutic approaches, in additional to ectopic expression.
While the expression of SKI protein is negatively regulated by TGFb signaling, the regulation is predominantly via posttranslational modification and ubiquitination mediated degradation (16,33,34). Our study has shown that both TGFb and Activin could induce the degradation of SKI protein in T cells (16). However, the detailed pathway that relays the signaling from TGFb receptor to polyubiquitination and proteasomemediated protein degradation remains elusive. While the E3 ubiquitin ligases Arkadia is reported to target SKI in nonlymphoid cells (34,35), whether it also regulates SKI stability in T cells remains to be examined. Additionally, whether there are other E3 ubiquitin ligases and proteasomal-independent mechanisms exist to modulate SKI expression need to be further investigated.
In summary, we demonstrated that expression of SKI could preferentially block the induction and maintenance of pathogenic Th17 cells, with less effect on the TGFb-induced non-pathogenic Th17 cells. We showed that the mouse models that express WT SKI specifically in T cells were refractory to pathogenic Th17 cell induced EAE. This study suggests targeting SKI could be a viable approach to control pathogenic Th17 cell response and to treat Th17 cell related diseases. The results were representative of two independent experiments; n = 4 mice, mean ± s.d., ns, not significant p > 0.05, * p < 0.05, ***p < 0.001 by two-sided Student's t-test.