Dispensable roles of Gsdmd and Ripk3 in sustaining IL-1β production and chronic inflammation in Th17-mediated autoimmune arthritis

Programmed necrosis, such as necroptosis and pyroptosis, is a highly pro-inflammatory cellular event that is associated with chronic inflammation. Although there are various triggers of pyroptosis and necroptosis in autoimmune tissue inflammation and subsequent lytic forms of cell death release abundant inflammatory mediators, including damage-associated molecular patterns and IL-1β, capable of amplifying autoimmune Th17 effector functions, it remains largely unclear whether the programs play a crucial role in the pathogenesis of autoimmune arthritis. We herein report that Gasdermin D (Gsdmd) and receptor interacting serine/threonine kinase 3 (Ripk3)—key molecules of pyroptosis and necroptosis, respectively—are upregulated in inflamed synovial tissues, but dispensable for IL-1β production and the development of IL-17-producing T helper (Th17) cell-mediated autoimmune arthritis in SKG mice. Gsdmd−/−, Ripk3−/−, or Gsdmd−/− Ripk3−/− SKG mice showed severe arthritis with expansion of arthritogenic Th17 cells in the draining LNs and inflamed joints, which was comparable to that in wild-type SKG mice. Despite the marked reduction of IL-1β secretion from Gsdmd−/− or Ripk3−/− bone marrow-derived DCs by canonical stimuli, IL-1β levels in the inflamed synovium were not affected in the absence of Gsdmd or Ripk3. Our results revealed that T cell-mediated autoimmune arthritis proceeds independently of the pyroptosis and necroptosis pathways.

activation. However, it remains elusive whether programed necrosis plays a crucial role in the pathogenesis of T cell-mediated autoimmune diseases such as rheumatoid arthritis (RA).
RA, one of the most common human autoimmune diseases, affects approximately 1% of the world population, and is characterized by bone-destructive chronic polyarthritis 15 . Although the precise pathophysiology of RA remains to be determined, it is widely appreciated that CD4 + T helper (Th) cells play a pivotal role in the pathogenesis of RA; this is supported by accumulating evidence, for example, MHC class II haplotype human leukocyte antigen (HLA)-DRB1 was identified as the strongest disease susceptibility gene and a high efficacy of cytotoxic T lymphocyte antigen 4 (CTLA4)-Ig blocking T cell co-stimulatory molecules for the treatment of RA [15][16][17] . Chronic synovial inflammation can accompany the cell death of proliferated fibroblast-like synoviocytes (FLSs) and recruited immune cells as well as pro-inflammatory cytokine production and the release of DAMPs, presumably by pyroptosis and necroptosis due to various triggers that are present in inflamed joints, including TNF and endogenous inflammasome ligands. Although it is considered that these events can amplify chronic inflammation and joint pathology 18 , it is unclear how a type of programed necrosis contributes to autoimmune arthritis.
SKG mice, a murine model of RA, spontaneously develop IL-17-producing T helper (Th17) cell-dependent autoimmune arthritis, resembling the immunopathology of chronic synovial inflammation in RA patients 19,20 . We have previously shown that DAMPs, including the alarmin IL-33, augment chronic inflammation in the inflamed joints of SKG mice, at least in part by stimulating tissue-resident synovial innate lymphoid cells (ILCs) 21 . Effector Th17 cells together with activated FLSs and ILCs orchestrate chronic arthritis by secreting abundant proinflammatory cytokines, including IL-6, IL-17, GM-CSF, and TNF, and these inflammatory cellular responses and their signaling pathways potentially trigger necrosis. Because lytic forms of programed necrosis can release abundant inflammatory mediators, including DAMPs and IL-1β, which is capable of amplifying the autoimmune Th17 effector functions and joint inflammation, in this study, SKG mice were used as a suitable model to investigate whether programed necrosis contributed to the differentiation and expansion of arthritogenic Th17 cells and the development of autoimmune arthritis.

Results
The generation of Gsdmd −/− SKG mice and CD4 + T cell profiles in Peyer's patches under a homeostatic state. To investigate the role of Gsdmd in the pathogenesis of autoimmune arthritis, we first evaluated the Gsdmd expression in whole synovial tissues from healthy and arthritic SKG mice. When arthritis was induced in SKG mice, the expression of Gsdmd tended to be upregulated in the inflamed joints (Fig. 1A). We next assessed the Gsdmd expression in CD4 + T cells, CD11b + Ly6G + neutrophils, CD11b + Ly6G -Ly6C high inflammatory monocytes and CD45 -CD31 -Podoplanin + FLSs sorted from the inflamed joints, each of which are the key cellular components that can augment chronic joint inflammation. The expression of Gsdmd in neutrophils and monocytes was higher in comparison to CD4 + T cells and FLSs, indicating the possible relevance of Gsdmd in these myeloid cells to the pathophysiology of SKG arthritis (Fig. 1B). We then generated Gsdmd −/− SKG mice on a BALB/c background by using CRISPR/Cas9 system to examine the effects of Gsdmd on the differentiation and maintenance of CD4 + T cell subsets in homeostatic and inflammatory conditions ( Supplementary Fig. 1A). As reported previously in a C57BL/6 background 5,10,11 , Gsdmd −/− SKG mice grew up normally like wild-type (WT) SKG mice and the deletion of Gsdmd had no effect on their development, lifespan, or fertility (data not shown). Loss of the Gsdmd function in pyroptosis was confirmed by in vitro assay of LPS transfection into bone marrow-derived dendritic cells (BMDCs) from Gsdmd −/− Rag2 −/− mice, which showed a remarkable reduction of the capacity for IL-1β release in comparison to BMDCs from WT Rag2 −/− mice (Fig. 1C), as previously reported using other Gsdmd −/− strains 5 . To explore whether the physiological Gsdmd expression levels have any impact on CD4 + T cell profiles under a homeostatic condition, we assessed the Gsdmd expression in normal tissues from BALB/c mice maintained under specific-pathogen-free (SPF) conditions and found that the small intestine is an organ that constitutively expresses the high levels of Gsdmd, and the expression was found to be independent of the gut microbiota and bacterial components based on the comparable expression of Gsdmd in germ-free (GF) mice (Fig. 1D). Because various CD4 + T cell subsets reside in Peyer's patches of the small intestine, we examined the proportions of physiological Th1, Th17 and regulatory T (Treg) cells in Peyer's patches of WT and Gsdmd −/− SKG mice under SPF conditions. Gsdmd −/− SKG mice showed almost the same CD4 + T cell profiles in terms of the expression of IL-17, GM-CSF, IFN-γ and Foxp3 in comparison to WT SKG mice (Fig. 1E, F). Taken together, these results indicated that Gsdmd had little impact on the differentiation and maintenance of physiological CD4 + T cell subsets under a homeostatic state.
Gsdmd is dispensable for the induction and development of SKG arthritis. We induced arthritis in WT and Gsdmd −/− SKG mice and assessed the clinical phenotypes and the development of arthritogenic Th17 cells. The severity of arthritis did not differ between the two groups of mice ( Fig. 2A). Notably, the levels of IL-1β in the synovium from the inflamed joints of the two groups were comparable, suggesting that IL-1β could be produced in a Gsdmd-independent mechanism (Fig. 2B). We then assessed the effects of Gsdmd on CD4 + T cell profiles after the induction of arthritis. The frequencies of IFN-γ-, GM-CSF-, IL-17-, and Foxp3-expressing subsets in CD4 + T cells from inflamed joints or draining lymph nodes (LNs) were almost unchanged between the groups (Fig. 2C, D). We also evaluated the onset and severity of arthritis induced by the adoptive transfer of CD4 + T cells from WT or Gsdmd −/− SKG mice into Rag2 −/− or Gsdmd −/− Rag2 −/− mice, which enables us to dissect the specific role of Gsdmd in CD4 + T cells and non-CD4 + T cells in recipient mice, separately (Fig. 2E). Like the arthritis phenotype in Gsdmd −/− SKG mice, the disease progression curves of arthritis were comparable in the four groups. Furthermore, the differentiation and expansion of arthritogenic Th17 cells in inflamed joints and draining LNs were not affected in Gsdmd-deficient conditions (Fig. 2F-H). These data demonstrated that The generation of Ripk3 −/− SKG mice and CD4 + T cell profiles in Peyer's patches under a homeostatic state. Next, we similarly investigated a role of Ripk3 in CD4 + T cell homeostasis and the SKG arthritis model. Consistent with the previous report 22 , the expression of Ripk3 was significantly increased in inflamed joints (Fig. 3A). We further assessed the Ripk3 expression in CD4 + T cells, CD11b + Ly6G + neutrophils, CD11b + Ly6G -Ly6C high inflammatory monocytes and CD45 -CD31 -Podoplanin + FLSs sorted from the inflamed joints. Ripk3 was widely expressed by those key inflammatory cells in SKG arthritic joints, and the expression pattern was different from Gsdmd (Fig. 3B). We then generated Ripk3 −/− SKG mice on a BALB/c background using a CRISPR/Cas9 system to examine the effects of Ripk3 on the differentiation and maintenance of CD4 + T cell subsets in homeostatic and inflammatory conditions ( Supplementary Fig. 1B). Like Gsdmd -/-SKG mice, Ripk3 −/− SKG mice showed no overt phenotypic changes in development or longevity. We confirmed the loss   (Fig. 3C), as previously reported 23 . When the physiological Ripk3 expression levels in normal tissues from BALB/c mice maintained under SPF or GF conditions were assessed, the Ripk3 expression was relatively high in the gut, LNs, and spleen (Fig. 3D). Among them, the expression of Ripk3 in the small intestine was likely controlled by the gut microbiota because of the specific reduction of its expression in GF mice (Fig. 3D). However, we found no difference in the frequency of the IL-17A, GM-CSF, IFN-γ and Foxp3 expression in the CD4 + T cells of Peyer's patches between WT and Ripk3 −/− SKG mice (Fig. 3E, F). Taken together, these data indicated that Ripk3 did not affect the differentiation or maintenance of physiological CD4 + T cell subsets under a homeostatic state.
Ripk3 is dispensable for the induction and development of SKG arthritis. We next investigated the role of Ripk3 in the induction and development of SKG arthritis. We induced arthritis in WT and Ripk3 −/− SKG mice. There was no significant difference in the clinical arthritis scores of the two groups of mice (Fig. 4A). Unlike the in vitro assay, IL-1β could be secreted in the inflamed synovium of Ripk3 −/− SKG mice (Fig. 4B). The frequencies of IFN-γ-, GM-CSF-, IL-17-, and Foxp3-expressing subsets in CD4 + T cells in inflamed joints and draining LNs were almost comparable between the groups (Fig. 4C, D). In addition, adoptive transfer experi-  www.nature.com/scientificreports/ ments using CD4 + T cells from WT or Ripk3 −/− SKG mice also revealed no significant differences in the clinical arthritis scores or the proportions of arthritogenic Th17 cells in draining LNs and inflamed joints from the recipient Rag2 −/− or Ripk3 −/− Rag2 −/− mice, albeit a marginal decrease in the proportions of IFN-γ-and GM-CSF-producing Th cells in the inflamed joints ( Fig. 4E-H). Taken together, these data indicated that Ripk3 is dispensable in the induction of pathogenic Th17 cells and the development of chronic arthritis in SKG mice.

Crosstalk or the compensatory role between Gsdmd and Ripk3 is not required for the induction and development of SKG arthritis. Because a compensatory role and context-dependent interac-
tion between Gsdmd and Ripk3 have been described 24,25 , we generated Gsdmd −/− Ripk3 −/− SKG mice by crossing Gsdmd −/− SKG and Ripk3 −/− SKG mice to assess whether Ripk3 or Gsdmd could compensate for a Gsdmd-or Ripk3-deficient condition, respectively, in the development of autoimmune arthritis, and evaluated their arthritis phenotypes and the development of Th17 cells in Gsdmd and Ripk3 double-deficient conditions. The arthritis scores of Gsdmd −/− Ripk3 −/− SKG mice after the induction of arthritis were unchanged from those in WT SKG mice (Fig. 5A). The differentiation and expansion of IFN-γ-, GM-CSF-, and IL-17-producing Th subsets from the inflamed joints and draining LNs of Gsdmd −/− Ripk3 −/− SKG mice were similar to those of WT SKG mice, albeit with a marginal decrease in the proportions of Foxp3 + Treg cells (Fig. 5B, C). Collectively, these results further strengthened the concept that both Gsdmd and Ripk3 and their interactions are dispensable for the induction and development of chronic autoimmune arthritis in SKG mice.

Discussion
Although accumulated evidence strongly supports that pyroptosis and necroptosis have evolved to combat against infectious pathogens that escape detection by infection-induced apoptosis, particularly intracellular bacteria such as Shigella flexneri, the subsequent acute tissue inflammation and pro-inflammatory cytokine production triggered by the programmed cell death could mediate the immunopathology through the aberrant activation of various inflammatory cells [1][2][3][4][9][10][11][12][13][14] . It is noteworthy that the Ripk3 and MLKL-molecules that are specifically involved in necroptosis-were reported to be upregulated in the inflamed synovium of collageninduced arthritis, indicating a possible contribution of necroptosis to the formation of severe synovitis 22 . How programmed necrosis contributes to T cell-mediated autoimmune arthritis remains poorly understood. Thus, one of the key questions we addressed in the present study was whether the pyroptosis and necroptosis pathways would exacerbate chronic inflammation, in particular sterile synovial inflammation. Using Gsdmd −/− , Ripk3 −/− , or Gsdmd −/− Ripk3 −/− mice, we demonstrated that despite abundant cell death in inflamed joints, pyroptosis and necroptosis are not required for the expansion of arthritogenic Th17 cells or the progression of Th17-mediated autoimmune arthritis. Inflammatory mediators, including DAMPs and pro-inflammatory cytokines released by pyroptosis and necroptosis could amplify Th17 functions; however, their modulation of pathogenic Th17 cells seems to be dependent on the type of inflammation and the target organ. Lytic forms of cell death seemed to augment systemic auto-inflammation in an FMF model, and acute tissue injuries in ischemia and transplantation models through the excess production of IL-1β and DAMPs, which could mediate pathology by the activation of innate immune cells 10,[12][13][14] . Given that the pathogenesis of these models is less dependent on adaptive T cell functions, pyroptosis and necroptosis are probably involved in the exacerbation of acute inflammation and tissue injuries are often autonomously amplified by the activation of innate immune and mesenchymal cells. On the other hand, the role of pyroptosis and necroptosis in T cell-mediated autoimmune pathology may be dispensable or limited in a specific inflammatory context. EAE is an exceptional model which relies on Gsdmd for the development of Th17-mediated autoimmune neuroinflammation. Mechanistically, Gsdmd-dependent pyroptosis of the peripheral myeloid cells is necessary for the activation and differentiation of pathogenic Th cells in peripheral lymphoid tissues, which in turn migrate into the target spinal cord to cause chronic tissue inflammation 11 . Although pathogenic Th17 cells predominantly drive SKG arthritis and EAE, the differential requirement of Gsdmd in the models may be attributed how these diseases are induced. For EAE, immunization with complete Freund's adjuvant and pertussis toxin, potent bacterial compounds, may be a strong inducer of pyroptosis in myeloid cells, following the activation of inflammasomes. In contrast, SKG mice are a spontaneous arthritis model and can be induced by the injection of mannan, which is a fungal compound distinct from that used in EAE and may not be involved in the activation of the pyroptosis pathway. Therefore, T cell priming in peripheral LNs between the models is probably different and Gsdmd was not required for the induction of arthritogenic Th17 cells in the LNs. However, it is tempting to speculate that persistent IL-1β production and Th17-mediated chronic inflammation in the target tissues of EAE or SKG mice may be similarly independent of pyroptosis and necroptosis, despite abundant cell death occurring in the inflamed tissue.
Crosstalk among different forms of programmed cell death, including apoptosis and necrosis, has recently been highlighted as a new aspect of interactions and compensatory roles between the signaling pathways of the distinct programs [24][25][26] . Although there are various triggers (TNF, endogenous inflammasome ligands, etc.) of pyroptosis and necroptosis in inflamed joints, our data using Gsdmd −/− Ripk3 −/− SKG mice clearly excluded the compensatory role between Gsdmd and Ripk3 for the development of autoimmune arthritis and further strengthened our claim that they are not involved in Th17 cell-mediated chronic inflammation. However, considering the IL-1β production capability in Gsdmd −/− and Ripk3 −/− mice and its crucial role in SKG arthritis 27 , unappreciated cellular mechanisms that can also process IL-1β maturation and release may be present in sterile tissue inflammation (e.g., the synovium and central nervous system) 28,29 .
In contrast, given that Gsdmd −/− and Ripk3 −/− mice do not show an overt phenotype in development or physiology, inhibitors of Gsdmd or Ripk3 may be a potential therapeutic target for the treatment of innate immune cell-mediated tissue injuries and systemic auto-inflammation because the inhibitors could have limited adverse

Materials and methods
Mice. SKG and Rag2 −/− mice were described previously 21 . Gsdmd −/− and Ripk3 −/− SKG mice were generated using a CRISPR/Cas9 system. Gsdmd −/− Rag2 −/− , Ripk3 −/− Rag2 −/− , and Gsdmd −/− Ripk3 −/− SKG were generated by crossing the above described strains. BALB/c mice housed under SPF and GF conditions were purchased from CLEA Japan. All animal experiments were approved by the Ethical Committee of Institute for Frontier Life and Medical Sciences and Graduate School of Medicine, Kyoto University, and were performed in compliance with the institutional and ARRIVE guidelines.

Induction of arthritis.
Arthritis was induced by a single intraperitoneal injection of mannan (20 mg; Sigma-Aldrich) in SKG mice, or by adoptive transfer of SKG CD4 + T cells (intravenously) into recipient mice, and was scored as described previously 19,20,30 . For the adoptive transfer of CD4 + T cells, CD4 + T cells were positively selected from the spleen and peripheral LNs using MACS CD4 microbeads and an LS column (Miltenyi Biotec) according to the manufacturer's instruction, and 2 × 10 6 CD4 + T cells were transferred into each recipient mouse intravenously via the tail vein.

Generation of bone marrow-derived dendritic cells (BMDCs). Bone marrow cells were collected
from the femur and tibia and red blood cells were lysed with RBC lysis buffer (Sigma-Aldrich). Resultant cells were cultured at 5 × 10 5 /ml in RPMI containing 10% FBS, 2-mercaptoethanol and penicillin-streptomycin with 20 ng/ml GM-CSF for 6 days to differentiate into BMDCs for in vitro assays.
Measurement of IL-1β from BMDCs and joint tissues. BMDCs were cultured in 96-well plates at 1 × 10 5 cells per well and either treated with 1 μg/ml LPS (InvivoGen), or transfected with 1 μg/ml LPS using 0.3% v/v FuGENE HD (Promega) after priming with 1 μg/ml Pam3CSK4 (InvivoGen) for 6 h. At 16 h after LPS stimulation, the IL-1β concentrations in the culture supernatants were measured using a Cytometric Bead Array (BD Biosciences) in accordance with the manufacturer's instructions. As for the measurement of IL-1β in the synovium, synovial tissues were resected from the inflamed joints, cut into small pieces, and the tissue weight was measured. The resultant synovial tissues were placed into 500 μl of D-PBS (−) (Nacalai Tesque), then vortexed and centrifuged. The collected supernatants were then used for the measurement of the IL-1β concentration, which was performed using a Cytometric Bead Array, and the results were normalized with tissue weights of individual samples.
Statistical analysis. Statistical analyses were performed using the GraphPad PRISM7 software program (GraphPad Software). A two-tailed t-test was used for the statistical analysis. A two-way ANOVA followed by Tukey's multiple comparisons test was used for the analysis of grouped data. Data are shown as the mean ± standard error of mean (SEM) or mean ± standard deviation (SD). P values of < 0.05 were considered to indicate statistical significance.

Data availability
The datasets analyzed during the current study are available from the corresponding author on reasonable request.