Decoupling the role of RORγt in the differentiation and effector function of T_H17 cells

Abstract

RORγt is known to instruct the differentiation of T helper 17 (T_H17) cells that mediate the pathogenesis of autoimmune diseases. However, it remains unknown whether RORγt plays a distinct role in the differentiation and effector function of T_H17 cells. Here, we show that mutation of RORγt lysine-256, a ubiquitination site, to arginine (K256R) separates the RORγt role in these two functions. Preventing ubiquitination at K256 via arginine substitution does not affect RORγt-dependent thymocyte development, and T_H17 differentiation in vitro and in vivo, however, greatly impaired the pathogenesis of T_H17 cell–mediated experimental autoimmune encephalomyelitis (EAE). Mechanistically, K256R mutation impairs RORγt to bind to and activate Runx1 expression critical for T_H17-mediated EAE. Thus, RORγt regulates the effector function of T_H17 cells in addition to T_H17 differentiation. This work informs the development of RORγt-based therapies that specifically target the effector function of T_H17 cells responsible for autoimmunity.

RORγt is required for both the effector function of T_H17 cells and T_H17 differentiation.

INTRODUCTION

Interleukin-17 (IL-17)–producing CD4⁺ T helper, or T helper 17, (T_H17) cells participate in immune responses against pathogens and the pathogenesis of diverse immunological diseases such as autoimmune diseases and even autism (1–6). The transcription factor retinoid-related orphan receptor γt (RORγt), encoded by the gene Rorc, instructs the differentiation of T_H17 cells (7–9). Mutations in Rorc affect IL-17 production and lead to severe immune deficiency in both mouse (7) and human (10). Thus, therapies that control the pathogenic T_H17 responses are needed clinically (11, 12). Numerous pharmacological RORγt inhibitors have been developed for clinical application (5, 11–13). Those inhibitors are believed to prevent T_H17-dependent autoimmunity by inhibiting the generation of T_H17 cells due to the essential function of RORγt in T_H17 differentiation.

The strength of the T_H17 immune responses is determined by the overall number of T_H17 cells and their effector function. T_H17 cells are derived from naïve CD4⁺ T cells upon activation in the presence of an appropriate cytokine milieu including IL-6, transforming growth factor–β (TGF-β), and/or IL-23 (7, 14). Although the function of RORγt in T_H17 differentiation has long been demonstrated, it remains unknown whether RORγt plays a role in the effector function of T_H17 cells and whether RORγt has a distinct role in the differentiation versus effector function of T_H17 cells. In addition to regulating T_H17 cells, RORγt enhances thymocyte survival (15–18) and is required for lymph node development (8, 18–20). Previously, we have generated a mutation in RORγt that disrupts T_H17 differentiation but not thymocyte development (8), indicating that RORγt uses different mechanisms to regulate the function of T_H17 cells and thymocytes.

Ubiquitination is a posttranslational modification that regulates many aspects of cellular function (21). Ubiquitin is conjugated to the lysine residues of the proteins to modulate protein function by regulating protein stability and/or protein-protein interactions. Thus, cellular context-dependent ubiquitination of RORγt may be a mechanism to modulate the diverse RORγt functions. In vitro studies indicate the regulation of RORγt stability by ubiquitination (22–25). However, in vivo function of RORγt ubiquitination is difficult to prove, as it requires the generation of the mice expressing mutant RORγt incapable of being ubiquitinated. Previous studies used mice deficient in ubiquitin ligases or deubiquitinases to understand the role of ubiquitination in immune responses (26). Such an approach prevents the ubiquitination of all the substrates; thus, it is impossible to determine the function of a specific substrate and even less about the function of a specific ubiquitination site on the substrate in immunological function.

In this study, three in vitro assays were developed to dissect RORγt function in thymocyte development, T_H17 differentiation, and effector function in experimental autoimmune encephalomyelitis (EAE) induction. A RORγt mutation at a ubiquitination site, lysine (K)–256 to arginine (RORγt^K256R), was found to specifically impair the effector function of T_H17 cells in inducing EAE without interfering with RORγt function in T_H17 differentiation and thymocyte development. A strain of mice was established to express RORγt^K256R (RORγt^K256R/K256R), which cannot be ubiquitinated at this site. This strain of mice allows us to determine the in vivo function of ubiquitination of RORγt at a specific site. RORγt^K256R/K256R mice have normal thymocyte development, lymph node development, and T_H17 differentiation. However, RORγt^K256R/K256R mice display greatly impaired T_H17 immune responses leading to EAE. Further, RORγt^K256R/K256R T_H17 cells showed decreased RORγt^K256R binding to the promoter region of the Runx1 gene and reduced expression of Runx1. Forced expression of Runx1 in RORγt^K256R/K256R T_H17 cells restored the ability to induce EAE. Therefore, RORγt regulates the effector function of T_H17 cells in addition to T_H17 differentiation.

RESULTS

Three assays are developed to dissect RORγt function

RORγt^−/− mice display defects in thymic T cell development, T_H17 differentiation, and development of T_H17-dependent EAE (7, 18). To dissect the function of RORγt, we developed three assays to separate these three functions. To determine the function of RORγt in thymocytes, we used an in vitro thymocyte differentiation system (27). CD4⁻CD8⁻ thymocytes from wild-type (WT) but not RORγt^−/− mice could differentiate into CD4⁺CD8⁺ and CD4⁺ cells (Fig. 1A). Consistently, RORγt^−/− CD4⁻CD8⁻ thymocytes transduced with a retrovirus expressing a RORγt (RORγt) but not green fluorescent protein (GFP) alone [empty virus (EV)] rescued the development of CD4⁺CD8⁺ and CD4⁺ from RORγt^−/− thymocyte (Fig. 1B and fig. S1A for gating strategy). We next used an in vitro assay to determine the function of RORγt in T_H17 differentiation. Under T_H17 polarization conditions, RORγt^−/− naïve CD4⁺ T cells could not differentiate into T_H17 cells unless exogenous RORγt was provided via retrovirus transduction (Fig. 1C and fig. S1B for gating strategy). Last, adoptive transfer EAE model enabled the testing of RORγt T_H17 effector function by using 2D2 T cell receptor (TCR) transgenic mice (Tg^Tcr2D2) that recognize myelin oligodendrocyte glycoprotein (MOG_35–55) (28, 29). RORγt^−/−/Tg^Tcr2D2 CD4⁺ T cells were transduced with virus expressing RORγt, polarized under T_H17 conditions and adoptively transferred into Rag1^−/− recipients for inducing EAE (30). RORγt^−/−/Tg^Tcr2D2 CD4⁺ T cells expressing RORγt induced very severe EAE (RORγt^−/−/Tg^Tcr2D2 + RORγt); these mice had the highest disease score of 5 (the Institutional Animal Care and Use Committee protocol does not permit disease development beyond this point) equivalent to Rag1^−/− recipients with WT Tg^Tcr2D2 CD4⁺ T cells transduced with GFP alone (WT Tg^Tcr2D2 + EV) (Fig. 1D). In contrast, RORγt^−/−/Tg^Tcr2D2 CD4⁺ T cells expressing GFP alone (RORγt^−/−/Tg^Tcr2D2 + EV) resulted in greatly delayed and impaired EAE with the highest disease score of 3. RORγt deficiency did not affect the proliferation and survival of Tg^Tcr2D2 CD4⁺ T cells in vitro (fig. S1, C and D) and in vivo (fig. S1E), which is also indicated by an equivalent number CD4⁺ T cells recovered from the spleen after adoptive transfer (fig. S1F). The successful establishment of above assays allowed us to dissect three RORγt-regulated functions.

Fig. 1. Three assays are developed to dissect RORγt function.

(A) Representative flow cytometric analysis (left) of CD4⁺ and CD8⁺ thymocytes ex vivo developed from indicated genotypes of CD4⁻CD8⁻ thymocytes placed on the OP9-DL4 stroma cells for 3 days (n = 3 per genotype). The number indicates the percentage of cells in the gated area throughout. Right: Percentage of CD4⁺CD8⁺ and CD4⁺ cells. (B) Representative flow cytometric analysis (left) of CD4⁺ and CD8⁺ thymocytes ex vivo developed from sorted RORγt^−/−CD4⁻CD8⁻ thymocytes transduced with retrovirus expressing GFP alone (EV) or with RORγt, and cultured on OP9-DL4 stroma cells for 3 days (n = 3 per genotype). Right: Percentage of CD4⁺CD8⁺ + CD4⁺ cells. (C) Representative flow cytometric analysis of IL-17A (left) and percentage of IL-17A⁺ cells (right) among indicated genotypes of CD4⁺ T cells transduced with retrovirus expressing GFP alone (EV) or with RORγt and polarized under T_H17 conditions for 3 days (n = 3 per genotype). (D) Mean clinical EAE score of Rag1^−/− mice adoptively transferred with same number of Tg^Tcr2D2 or RORγt^−/−/Tg^Tcr2D2 CD4⁺ T cells transduced with retrovirus expressing GFP alone (EV) or with RORγt and polarized under T_H17 conditions for 3 days. Bars are means ± SE. ***P < 0.001 (two-tailed Student’s t test).

RORγt-K256, a ubiquitination site, is critical for the effector function but not the differentiation of T_H17 cells

RORγt has been shown to regulate T_H17 differentiation; it, however, remains unknown whether and how RORγt regulates the effector function of T_H17 cells. To address this question, we aimed to identify RORγt mutations that specifically disrupt RORγt function in effector T_H17 cell–mediated EAE but not in T_H17 differentiation and thymocyte development. Previously, we mutated lysine residues (K) on RORγt to arginine (R) to study the function of posttranslational modification of RORγt (8). Thus, we first compared the ability of WT and RORγt mutants to rescue RORγt^−/− thymocyte development. Only RORγt-K31R mutation moderately affected thymocyte development (Fig. 2A). In terms of T_H17 differentiation, K31R, K69R, and K313R impaired T_H17 differentiation (Fig. 2, B to D), consistent with our published results (8, 9, 31). Further, most RORγt mutations either impaired or potentiated T_H17 differentiation (Fig. 2B), suggesting that RORγt uses very different mechanisms to regulate the function of thymocytes and T_H17 cells.

Fig. 2. RORγt-K256, a ubiquitination site, is critical for the effector function but not the differentiation of T_H17 cells.

(A) The ratio of rescued thymocyte development of RORγt^−/−CD4⁻CD8⁻ by indicated RORγt K-R mutants relative to that by WT RORγt (1 = 100%). (B) Ratio of rescued T_H17 differentiation of RORγt^−/−CD4⁺ T cells by indicated RORγt K-R mutants relative to WT RORγt (1 = 100%). (C) Flow cytometric analysis of IL-17A⁺ cells among RORγt^−/− CD4⁺ T cells transduced with retroviruses expressing GFP alone (EV) or indicated RORγt mutants and subsequently polarized for 3 days under T_H17 conditions. (D) Percentages of IL-17A⁺ cells rescued by indicated RORγt mutants shown in (C) (n = 5 to 12). (E) Immunoblot (IB) of ubiquitinated (Ub) RORγt^K256(K1) immunoprecipitated (IP) from human embryonic kidney (HEK) 293T cells expressing RORγt^K256(K1) alone (none) or together with hemagglutinin (HA)–tagged WT or mutated ubiquitin (K0, all K mutated to R). WCL, whole-cell lysates. (F) Immunoblot of ubiquitinated WT RORγt or RORγt^K256R (only K256 mutated to R) enriched by Tandem Ubiquitin Binding Entities 2 (TUBE2) from RORγt^−/− T_H17 cells expressing indicated RORγt type. (G) Mean clinical score of Rag1^−/− mice adoptively transferred with sorted RORγt^−/−/Tg^Tcr2D2 T_H17 cells expressing GFP alone (EV) or with indicated RORγt (n = 5 to 7 mice per group). (H) Total number and percentage of indicated immune cells recovered from the CNS of EAE-induced mice from (G). (I) Flow cytometric analysis of IL-17A⁺ cells among CD4⁺ T cells recovered from the CNS of EAE-induced mice shown in (G). (H and I) Box plots show median (central line), maximum, minimum (box ends), and outliers (extended lines). *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant (P > 0.05); two-tailed Student’s t test.

To separate RORγt function in T_H17 differentiation versus effector T_H17 cells, we focused on mutants that did not perturb T_H17 differentiation and thymocyte development, such as K99R, K256R, and K288R (Fig. 2, A to D). K256 was identified as a prominent ubiquitination site by mass spectrometry analysis of immunoprecipitated RORγt (fig. S2A). To validate the K256 ubiquitination site, we generated a RORγt mutant with all K mutated to R except K256 (RORγt^K256(K1)) so that only K256 can be ubiquitinated. In the presence of WT ubiquitin, RORγt^K256(K1) was ubiquitinated, whereas the ubiquitination signals were absent in the presence of a mutant ubiquitin that had all K mutated to R (K0) so that it cannot be attached to the substrates (Fig. 2E). We next used a RORγt mutant carrying a single K256 to R mutation (RORγt^K256R) that cannot be ubiquitinated only at the K256 site. WT RORγt or RORγt^K256R was then expressed in RORγt^−/− T cells under T_H17-polarizing conditions. Ubiquitinated RORγt was readily detected in T_H17 cells (Fig. 2F). Compared to the WT RORγt, RORγt^K256R had obviously reduced ubiquitination signals. RORγt-K256 is thus ubiquitinated in T_H17 cells.

Next, we tested the effector function of T_H17 cells in the induction of EAE. As described above (Fig. 1D), in vitro–differentiated RORγt^−/−/Tg^Tcr2D2 T_H17 cells expressing exogenous RORγt, RORγt^K99R, and RORγt^K288R induced severe EAE after adoptive transfer into Rag1^−/− recipients (Fig. 2G), suggesting that K99R and K288R do not affect the effector function of T_H17 cells in the induction of EAE, whereas RORγt^K256R-expressing RORγt^−/−/Tg^Tcr2D2 T_H17 cells induced greatly impaired EAE. RORγt^K69R also induced impaired EAE, as we previously observed because of reduced T_H17 differentiation (8). There was no difference in the proliferation of the RORγt^−/−/Tg^Tcr2D2 cells expressing various RORγt (fig. S2B). The central nervous system (CNS) immune cell infiltrate was also analyzed (Fig. 2H and fig. S2C for gating strategy). Consistent with the impaired EAE, the total number of CD4⁺ T cells and monocytes/macrophages in the CNS was significantly reduced in recipients with RORγt^K256R or RORγt^K69R mutants, indicating reduced inflammation. In addition, recipients with RORγt^K256R and RORγt^K69R cells also showed reduced infiltrate IL-17A⁺ cells in the CNS (Fig. 2I and fig. S2D for gating strategy), whereas not obvious changes in interferon-γ–positive (IFN-γ⁺) T_H1 cells from recipients with RORγt^K256R compared to that with WT RORγt were observed (fig. S2, E and F). Therefore, the RORγt-K256 ubiquitination site, although dispensable for T_H17 differentiation, is required for the effector function of T_H17 cells in the pathogenesis of EAE.

RORγt-K256R mutation does not affect RORγt-dependent development of thymocytes and lymph nodes

To investigate the function of RORγt-K256 ubiquitination in vivo, we generated homozygous mice for RORγt^K256R (RORγt^K256R/K256R) under the control of endogenous RORγt locus (fig. S3A for genetic engineering strategy and fig. S3B for confirming the K256R mutation by sequencing). We first examined RORγt-dependent thymocyte development (18–20, 32). WT RORγt and RORγt^K256R had the same expression patterns in thymocytes (Fig. 3A); our gene-targeting strategy thus did not affect RORγt^K256R expression. Furthermore, RORγt^K256R was as stable as WT RORγt (fig. S3, C to D), suggesting that K256 ubiquitination site does not affect RORγt stability. Thymic cellularity (Fig. 3B) and distribution of different developmental stages of thymocytes, CD4⁻CD8⁻ double-negative (DN), CD4⁺CD8⁺ double-positive (DP), and CD4⁺ or CD8⁺ single-positive thymocytes (Fig. 3C) were comparable between RORγt^K256R/K256R and WT mice, which were different from RORγt^−/− mice, suggesting normal thymocyte development in RORγt^K256R/K256R mice. Furthermore, the percentage of natural killer T (NKT) and γδ T cells in the thymus (Fig. 3, D and E) and spleens (fig. S3, E and F) was equivalent between WT and RORγt^K256R/K256R mice. The accelerated CD4⁺CD8⁺ thymocyte apoptosis (Fig. 3F and fig. S3G for analysis of apoptotic cells) accounted for the reduced percentage of CD4⁺CD8⁺ thymocytes (Fig. 3C) and decreased thymic cellularity (Fig. 3B) in RORγt^−/− mice (18), whereas the apoptosis of CD4⁺CD8⁺ thymocytes from RORγt^K256R/K256R mice was the same as that of the WT mice (Fig. 3F). In addition, thymocytes from RORγt^−/− mice had a higher percentage of cells with >2N of DNA (Fig. 3G), indicating more cells in the DNA synthesis phase of the cell cycle (18), while thymocytes from RORγt^K256R/K256R mice did not show an increased percentage of the cells in the DNA synthesis phase compared to the WT mice. Therefore, RORγt-K256R mutation does not affect RORγt function in thymocyte development. Furthermore, unlike RORγt ^−/− mice that lack all the peripheral lymph nodes (18), RORγt^K256R/K256R mice developed lymph nodes including inguinal lymph nodes (Fig. 3H) and Peyer’s patches (Fig. 3I), same as the WT mice. Together, prevention of RORγt-K256 ubiquitination does not affect the development of thymocytes and lymph nodes.

Fig. 3. RORγt-K256R mutation does not affect RORγt-dependent development of thymocytes and lymph nodes.

(A) Flow cytometric analysis of RORγt in thymocytes obtained from indicated mice (n = 5 to 8). MFI, mean fluorescence intensity. (B) Total thymocyte numbers of the indicated mice (n = 5 to 8). The graph shows mean ± SD. (C) Flow cytometric analysis of CD4 and CD8 on thymocytes from indicated mice (n = 4 per genotype). (D) Representative flow cytometric analysis of NKT cells (n = 4) from the thymus of the indicated mice. (E) Representative flow cytometric analysis of γδ T cells (n = 4) from the thymus of the indicated mice. (F) Flow cytometric analysis of the survival of DP thymocytes cultured in vitro for the indicated time (n = 4 per genotype). (G) Representative flow cytometric analysis of DNA content of indicated thymocytes stained by propidium iodide (PI) (n = 4 to 6 per genotype). (H and I) Number of inguinal lymph nodes (LN) (H) and Peyer’s patches (I) in the indicated mice (n = 9 to 12 per genotype). **P < 0.01; ***P < 0.001 (two-tailed t test). (A, C, E, F, and G) Box plots or scatter plots show median (central line), maximum, minimum (box ends), and outliers (extended lines).

RORγt^K256R/K256R mice have normal T_H17 differentiation but develop impaired T_H17-dependent EAE

Upon maturation in the thymus, T cells migrate to the periphery to mediate immune responses. RORγt^K256R/K256R and WT mice had comparable splenocytes (fig. S4A), CD4⁺, CD8⁺ (fig. S4B), CD62L^hiCD44^lo naïve, and CD62L^loCD44^hi memory-like T cell counts (fig. S4, C to D). We next examined T_H17 differentiation using a GFP reporter mouse line of IL-17 (Il17a^GFP). We confirmed that in vitro–differentiated RORγt^K256R/K256R T_H17 cells showed comparable RORγt (Fig. 4A) and IL-17 (Fig. 4B) expression compared to WT cells, consistent with the notion that RORγt-K256R mutation does affect T_H17 differentiation (Fig. 2, B and D). In addition, RORγt^K256R/K256R mice showed a normal percentage of splenic regulatory T cells (T_regs) (fig. S4E) and normal differentiation of T_regs from naïve CD4⁺ T cells (fig. S4F). To test whether RORγt-K256R affects the effector function of T_H17 cells responsible for the induction of EAE, we induced EAE by immunization with MOG_35–55 peptide. Before immunization, there were almost no T_H17 cells detected in the spleens from WT and RORγt^K256R/K256R mice (Fig. 4C, top). Six days after immunization, equivalent T_H17 (fig. S4, G and H) as well as T_H1 (fig. S4, I to J) cells were induced in spleens (fig. S4, G and I) and lymph nodes (fig. S4, H and J) of WT and RORγt^K256R/K256R mice. Again, on day 12 after immunization, when EAE symptoms started to develop, the percentage of T_H17 cells still showed no obvious difference in the spleens of WT and RORγt^K256R/K256R mice (Fig. 4C, bottom), confirming that RORγt-K256R does not affect the generation of T_H17 cells in vivo. However, compared to WT mice, RORγt^K256R/K256R mice developed greatly impaired EAE (Fig. 4D), supporting impaired RORγt^K256R/K256R T_H17 effector function. Consistently, adoptive transfer of in vitro–differentiated RORγt^K256R/K256R T_H17 cells also induced less severe EAE in Rag1^−/− recipients compared to that induced by WT T_H17 cells (fig. S4K). In addition, impaired EAE induction was associated with reduced immune cell infiltrate including CD4⁺, CD8⁺, B cells, and monocytes in the CNS of RORγt^K256R/K256R mice, although no obvious changes were observed in neutrophil numbers (Fig. 4E and fig. S4L for gating strategy). Further analysis of CNS lymphocyte infiltrate in the RORγt^K256R/K256R mice showed a decreased percentage of IL-17A⁺CD4⁺ cells (Fig. 4F), particularly pathogenic IL-17A⁺ granulocyte-macrophage colony-stimulating factor–positive (GM-CSF⁺) cells that play an important role in EAE development (Fig. 4, G and H, and fig. S4M for gating strategy) (2, 3). These results suggest that RORγt-K256R mutation, which prevents the ubiquitination, impairs the effector function but not the differentiation of T_H17 cells responsible for the pathogenesis of EAE.

Fig. 4. RORγt^K256R/K256R mice have normal T_H17 differentiation but develop impaired T_H17-dependent EAE.

(A) Flow cytometric analysis of RORγt in WT and RORγt^K256R/K256R CD4⁺ T cells polarized for 3 days under T_H17 conditions (n = 4 to 6 per genotype). Bars are means ± SEM. (B) Flow cytometric analysis of GFP⁺ (IL-17A⁺) cells among CD4⁺ T cells from indicated genotypes of cells differentiated under T_H17 conditions as described in (A) (n = 4 per genotype). (C) Representative flow cytometric analysis of IL-17A⁺ cells among CD4⁺ T cells in the spleen of indicated genotype of mice either before (naïve) or 12 days after MOG_35–55 immunization (n = 4). (D) Mean clinical score of indicated mice different days after MOG_35–55 immunization. Bars are means ± SEM. (E) Number of indicated immune cells recovered from the CNS of indicated EAE-induced mice from (D). (F) Representative flow cytometric analysis of GFP⁺ (IL-17A⁺) among CD4⁺ T cell infiltrate to the CNS of indicated mice shown in (D) (n = 10 per genotype). (G) Representative flow cytometric analysis of GM-CSF or IFN-γ expression in the lymphocyte infiltrated to the CNS of indicated mice shown in (D). (H) Quantification of GM-CSF⁺/CD4⁺ % and IFN-γ⁺/CD4⁺ % cells in the CNS (n = 5 per genotype). **P < 0.01; ***P < 0.001 (two-tailed t test). (E, F, and H) Box plots: Median (central line), maximum, minimum (box ends), and outliers (extended lines).

RORγt-K256R impairs the pathways critical for the pathogenesis of EAE but not for the T_H17 differentiation

To understand the mechanisms responsible for the RORγt-K256R mutation-disrupted effector function of T_H17 cells responsible for EAE, GFP⁺ T_H17 cells derived from WT and RORγt^K256R/K256R/IL-17^GFP CD4⁺ T cells were sorted to high purity (>98%) (fig. S5A for gating strategy) and subjected to RNA sequencing (RNA-seq) analysis. On the basis of the computational principal components analysis, the six RNA-seq samples were divided into two groups: WT and RORγt^K256R/K256R T_H17 cells (fig. S5B), indicating reproducible gene expression patterns within each group and thus the high quality of RNA-seq results. We identified 1375 differentially expressed genes (DEGs) [P < 0.05 and fold change (FC) > 1.5], with 943 up-regulated and 432 down-regulated genes, between WT and RORγt^K256R/K256R T_H17 cells (Fig. 5A, fig. S5C, and tables S2 and S3). Subjection of DEGs to pathway analysis did not find significant changes in the T_H17 differentiation pathway (Fig. 5, B and C) between WT and RORγt^K256R/K256R cells, confirming that RORγt-K256R mutation does not affect T_H17 differentiation. Down-regulated pathways in RORγt^K256R/K256R cells include IL-23 signaling (Fig. 5, B and D) and glycolysis pathways (Fig. 5, B and E); both play important roles in the pathogenesis of EAE (33, 34). In addition, inflammatory cytokines such as IL-9, IL-17, and IL-22 signaling pathways were also down-regulated, likely reflecting a reduced ability to induce inflammation responsible for the tissue damages. We next subjected DEGs to gene set enrichment analysis (GSEA) using the gene set specifically expressed in T_H17 cells responsible for the development of pathogenic EAE (fig. S5D) (35) and found that RORγt^K256R/K256R T_H17 cells had significantly reduced enrichment of the genes important for the pathogenesis of EAE when compared to the WT T_H17 cells (Fig. 5F). This result is consistent with the impaired EAE observed in RORγt^K256R/K256R mice. To determine the DEGs that are directly regulated by RORγt, we performed chromatin immunoprecipitation sequencing (ChIP-seq) analysis to detect genome-wide RORγt occupancy. ChIP-seq analysis in T_H17 cells revealed obvious RORγt-binding peaks at Il17a and Il17f loci (Fig. 5G), consistent with published results (8, 36). RORγt^K256R-binding peaks at the Il17a and Il17f loci were comparable to that of WT RORγt, supporting that K256R did not affect RORγt binding to Il17 gene and thus its expression. Using our RNA-seq and ChIP-seq data, we cross-examined genes that were down-regulated in RORγt^K256R/K256R T_H17 cells with the genes that had reduced RORγt^K256R-binding signals (see table S4 for the full list), identifying 31 genes (Fig. 5H). These 31 genes are considered directly regulated by RORγt, and their reduced expression in RORγt^K256R/K256R T_H17 cells is likely due to reduced RORγt^K256R binding and activating their expression. The 31 genes were then cross-examined with the gene set specifically expressed in T_H17 cells and responsible for the pathogenesis of EAE (fig. S5D) (35) and identified Runx1 (Fig. 5H). Furthermore, Runx1 was found to be a core regulator for the IL-23 signaling pathway that was down-regulated in RORγt^K256R/K256R T_H17 cells by a protein-protein interaction network analysis (Fig. 5I), indicating that down-regulated IL-23 signaling pathway is likely due to down-regulated Runx1 expression. Therefore, computational analysis of the transcriptome is consistent with the phenotypes observed in RORγt^K256R/K256R mice that RORγt-K256R mutation does not affect T_H17 differentiation but impairs the effector function of T_H17 cells responsible for the development of EAE. Runx1 is thus a possible RORγt-regulated gene that is down-regulated in RORγt^K256R/K256R T_H17 cells and responsible for the observed defective effector function of RORγt^K256R/K256R T_H17 cells in the induction of EAE.

Fig. 5. RORγt-K256R impairs the pathways critical for the pathogenesis of EAE but not for the T_H17 differentiation.

(A to F) Computation analysis of RNA-seq data obtained from GFP⁺/(IL-17A⁺) WT and RORγt^K256R/K256R CD4⁺ T cells polarized under T_H17 conditions. (A) The number of DEGs (black) including up-regulated (red) and down-regulated (blue) genes with a cutoff at P value of <0.05 and FC > 1.5. (B) Ingenuity pathway analysis (IPA) canonical signaling pathway analysis of down-regulated pathways (z-score < 0) in RORγt^K256R/K256R versus WT T_H17 cells. The dots denote P value, the vertical dotted line marks P value of 0.05 (log₁₀P = 1.3). JAK, Janus kinase; STAT, signal transducer and activator of transcription. (C to F) GSEA plots showing the enrichment of genes critical for T_H17 differentiation (C), IL-23 signaling (D), glycolysis (E), and pathogenesis of T_H17-mediated EAE (F) between RORγt^K256R/K256R versus WT T_H17 cells. NES, normalized enrichment score. Gene sets were derived from the molecular signaling database (MSigDB) for (C) to (E) and the previous report for (F) (35). (G) ChIP-seq analysis of RORγt DNA-binding peaks at the Il17a and Il17f loci in RORγt^−/− T_H17 cells expressing indicated RORγt. (H) Venn diagram of gene overlapping among 432 down-regulated genes in RORγt^K256R/K256R T_H17 cells, 1277 genes with decreased RORγt^K256R-binding activity, and 224 T_H17 genes critical for the pathogenesis of EAE. (I) Protein-protein network analysis of differentially expressed regulators critical for the IL-23 signaling pathway between RORγt^K256R/K256R versus WT T_H17 cells. Genes critical for IL-23 signaling pathway were customized on the basis of IPA-curated pathways.

RORγt-K256R mutation impairs RORγt to bind and activate Runx1 gene critical for EAE development

Our computational analysis identified Runx1 as a potential gene responsible for impaired effector function of RORγt^K256R/K256R T_H17 cells, because (i) Runx1 was down-regulated in RORγt^K256R/K256R/Tg^Tcr2D2 T_H17 cells, which was also confirmed by individual quantitative polymerase chain reaction (qPCR) analysis (Fig. 6A); (ii) three prominent RORγt-binding peaks close to the transcription start site of Runx1 gene were identified by ChIP-seq analysis (Fig. 6B), and RORγt^K256R binding signals at those peaks were substantially decreased (Fig. 6, B and C). RORγt-binding sites were identified within the peak region by sequencing analysis (fig. S6A). Furthermore, individual ChIP assays also confirmed RORγt binding to the peak region, whereas RORγt^K256R-binding signals to this region were greatly decreased (Fig. 6D), which correlates to the decreased levels of Runx1 mRNA in RORγt^K256R/K256R T_H17 cells (Fig. 6A); (iii) Runx1 was a gene expressed in T_H17 cells and responsible for the pathogenesis of EAE, as deletion of Runx1 impairs the EAE development (35, 37). To determine the effects of decreased RORγt^K256R-binding peak signals on Runx1 gene transcription, we cloned the DNA fragment covering the region with two potential RORγt-binding sites (fig. S6A) to a luciferase reporter gene (pGL3) driven by a basic thymidine kinase (TK) promoter. The reporter activity was greatly stimulated by WT RORγt but not as much by RORγt^K256R (Fig. 6E). Therefore, reduced RORγt^K256R binding to the Runx1 gene correlates well with the reduced ability of RORγt^K256R to stimulate Runx1 gene expression. In contrast, WT RORγt and RORγt^K256R equivalently stimulated IL-17 promoter-luciferase report activity, which correlates with the equivalent binding of RORγt and RORγt^K256R to and activation of the Il17 gene (Figs. 4B and 5G). To further determine whether identified RORγt-binding peaks are important for the expression of endogenous Runx1 gene, the region containing the RORγt-binding peaks was deleted using CRISPR-Cas9 with two guiding RNAs (ΔRgn1) in CD4⁺ T from mice expressing Cas9 (fig. S6A). At the same time, we used a nontargeting control (NTC) and a deleted adjacent region (ΔRgn2) as negative controls, whereas deleted Runx1 gene itself as a positive control (ΔRunx1) (fig. S6B for gating strategy for detecting Runx1 expression after deletion). Deletion of the Runx1 gene prevented Runx1 expression [Fig. 6F and fig. S6C for mean fluorescence intensity (MFI) of Runx1], demonstrating a successful deletion strategy with CRISPR-Cas9. Furthermore, deletion of the RORγt-binding region, but not the adjacent region or NTC, greatly reduced expression of Runx1, strongly indicating the critical function of the RORγt-binding region in the stimulation of Runx1 expression. Together with the results that RORγt^K256R had decreased binding signals to Runx1 (Fig. 6, B to D), these results suggest that reduced Runx1 expression in RORγt^K256R/K256R T_H17 cells is due to impaired RORγt^K256R-binding and activating Runx1 gene expression. To determine whether the reduced level of Runx1 in RORγt^K256R/K256R cells is responsible for the impaired development of EAE, we force-expressed Runx1 in RORγt^K256R/K256R/Tg^Tcr2D2 CD4⁺ T cells (Fig. 6G) and adoptively transferred them to Rag1^−/− mice to induce EAE (Fig. 6H). Forced expression of Runx1 significantly enhanced RORγt^K256R/K256R/Tg^Tcr2D2 CD4⁺ T cell function in the induction of EAE comparable to that of WT Tg^Tcr2D2 T cells, which was also confirmed by increased infiltration of CD4⁺ T cells to the CNS (Fig. 6I). Furthermore, the percentage of pathogenic IL-17⁺GM-CSF⁺ and IL-17⁺IFN-γ⁺ cells in CNS was reduced in Rag1^−/− mice transferred with RORγt^K256R/K256R/Tg^Tcr2D2 CD4⁺ T cells compared to that transferred with WT Tg^Tcr2D2 T cells but restored to WT levels in RORγt^K256R/K256R/Tg^Tcr2D2 CD4⁺ T cells expressing Runx1 (Fig. 6J), consistent with rescuing the effector function of RORγt^K256R/K256R T cells by expressing Runx1. Collectively, our results demonstrated that the ubiquitination site of RORγt-K256, which is not essential for T_H17 differentiation, regulates the effector function of T_H17 cells required for inducing EAE via up-regulating Runx1. By decoupling the function of RORγt in the differentiation and effector function of T_H17 cells, we demonstrated that RORγt also regulates the effector function of T_H17 cells.

Fig. 6. RORγt-K256R mutation impairs RORγt to bind and activate Runx1 gene critical for EAE development.

(A) qPCR of Runx1 mRNA in indicated genotypes of T_H17 cells (n = 3). Bars are means ± SEM. (B) ChIP-seq analysis of RORγt DNA-binding peaks at the Runx1 locus in RORγt^−/− T_H17 cells transduced with indicated vectors. (C) FC of RORγt-binding peaks indicated in (B). FDR, false discovery rate. (D) ChIP analysis of RORγt-binding signals at the Runx1 gene locus shown in (B) (n = 3). Bars are means ± SEM. (E) Luciferase (luc) activity from HEK 293T cells transfected with a pGL3 basic luciferase vector, a luciferase reporter with RORγt-binding sites (RORBS)–containing DNA fragment cloned from Runx1 locus and a basic TK promoter or IL-17 promoter (IL-17P) luciferase reporter under indicated conditions. (F) Flow cytometric analysis of Runx1 among GFP⁺(Cas9⁺) T_H17 cells retrovirally transduced with indicated single-guide RNAs (n = 4). (G) Flow cytometric analysis of Runx1 among indicated genotypes of T_H17 cells retrovirally transduced with GFP alone (EV) or together with Runx1 (n = 3). (H) Mean clinical score of Rag1^−/− mice adoptively transferred with indicated genotypes of T_H17 cells expressing GFP alone or with Runx1 shown in (G). (I) Number of infiltrating CD4⁺ T cells recovered from the CNS of the EAE-induced mice shown in (H). (J) Flow cytometric analysis of GM-CSF, IFN-γ, and IL-17A in CD4⁺ T cells recovered from the CNS of EAE-induced mice shown in (H). *P < 0.05; **P < 0.01; ***P < 0.001 (two-tailed t test). (I and J) Box plot: Median (central line), maximum, minimum (box ends), and outliers (extended lines).

DISCUSSION

The transcription factor RORγt, which is encoded by gene Rorc, is well known for instructing the differentiation of T_H17 cells. Activation of naïve T cells in the presence of TGF-β, IL-6, and/or IL-23 is sufficient to up-regulate RORγt, which instructs the differentiation into T_H17 cells (14, 38–41). The hallmark of T_H17 differentiation is the activation and expression of IL-17. RORγt directly binds to Il17a and Il17f gene loci to stimulate their expression (8, 36), which explains the essential function of RORγt in the differentiation of IL-17–producing T_H17 cells. Differentiated T_H17 cells are able to induce tissue inflammation involved in the pathogenesis of autoimmune diseases, including psoriasis, inflammatory bowel disease, and multiple sclerosis (2, 3, 14). Previous studies using RORγt^−/− mice demonstrated the essential function of RORγt in T_H17 cell–mediated autoimmunity such as EAE (7, 9). RORγt^−/− mice are resistant to EAE and other T_H17-mediated autoimmunity, which is due to the lack of RORγt-dependent generation of T_H17 cells. The question remains whether RORγt plays a role in the effector function of T_H17 cells involved in autoimmunity. The mutation RORγt^K256R we identified at the ubiquitination site does not affect the generation of IL-17–producing T_H17 cells but impairs the effector function of T_H17 cells in the induction of EAE. Further, RORγt^K256R/K256R mice have normal T_H17 differentiation both in vitro and in vivo but have greatly impaired T_H17 immune responses that led to EAE. In addition, RORγt^K256R/K256R mice have normal RORγt-dependent thymocyte development and lymph node genesis including Peyer’s patches. This study thus reveals a previously unidentified and essential RORγt function in effector T_H17 cells in addition to T_H17 differentiation. This study informs the development of RORγt-based therapies that specifically target the effector function of T_H17 cells responsible for the pathogenesis of autoimmunity. This will have a significant impact on the clinical treatment of T_H17-mediated autoimmunity, as usually medical treatment is performed after a diagnosis of autoimmune diseases resulting from the effector function of already developed T_H17 cells.

Runx1 is a transcription factor known to regulate hematopoiesis (42–44) and oncogenesis (45). Runx1 was reported to play a role in T_H17 cells. Particularly relevant to T_H17 cells involved in the pathogenesis of EAE, Runx1 has been shown to be up-regulated in T_H17 cells and work together with T-bet to stimulate IFN-γ expression that is believed to be responsible for the induction of EAE (37). However, it is not clear how Runx1 is up-regulated in T_H17 cells. Our results demonstrate that Runx1 expression is stimulated by RORγt, as RORγt-binding peaks were detected by ChIP-seq on the promoter region of the Runx1 gene. The conserved RORγt-binding sequence was identified within the detected peaks, and the deletion of the DNA fragment containing the RORγt-binding site greatly reduced Runx1 expression. Moreover, RORγt^K256R has reduced binding signals at the RORγt-binding region, which correlates with the impaired ability of RORγt^K256R in stimulating a luciferase reporter gene driven by the RORγt-binding region identified in the Runx1 locus. These results demonstrate the mechanisms for how RORγt regulates the effector function of T_H17 cells via up-regulating Runx1 expression and why RORγt-K256R mutation impairs the effector function of T_H17 cells. An in vitro study showed that Runx1 can stimulate RORγt expression, which is, however, inhibited by T-bet (46). Because T-bet is required for IFN-γ expression, this seems to suggest that T-bet and RORγt inhibit each other. In addition, Runx1 has been shown to be required for forkhead box P3 (Foxp3) expression (47), and Runx1 is able to stimulate itself expression via an autoregulation mechanism (48). The function of Runx1 is thus complicated and dependent on the microenvironment. In our study, both in vitro and in vivo, we did not find obvious changes in the levels of RORγt expression in RORγt^K256R/K256R T_H17 cells that have lower levels of Runx1, thus not supporting the role of K256 ubiquitination in the regulation of RORγt and T_H17 differentiation. Furthermore, even forced expression of Runx1 in RORγt^K256R/K256R CD4⁺ T cells does not affect T_H17 differentiation. Therefore, RORγt-regulated Runx1 expression does not affect T_H17 differentiation but is required for the effector function of T_H17 cells that mediate pathogenic EAE.

Dysregulated T_H17 cells are often associated with autoimmune diseases such as EAE and psoriasis resulting from a reaction to self-antigens (49). In addition to IL-17, IL-23 also plays an important role in T_H17 cell–dependent autoimmune diseases (1, 50) such as EAE and psoriasis (51, 52). Neutralizing antibodies for IL-23 and IL-17 or their receptors are used for the treatment of these autoimmune conditions (53–55). Therefore, inhibiting the T_H17 pathway is effective for treating autoimmune conditions (52, 56). Our results show that the IL-23 signaling pathway is down-regulated in RORγt^K256R/K256R T_H17 cells. Further, network analysis supports that Runx1 is a core regulator for the IL-23 pathway and down-regulated IL-23 pathway thus likely resulting from down-regulated Runx1. Therefore, RORγt-regulated Runx1 seems to control the effector function of T_H17 cells at least partially through regulating the IL-23 pathway known to be critical for T_H17-mediated autoimmunity.

Considering the essential function of RORγt in T_H17 cells, RORγt inhibitors are being developed to treat T_H17-dependent autoimmunity (5, 11–13, 57). However, these RORγt inhibitors mostly target T_H17 differentiation and RORγt-dependent thymocyte development. Inhibition of RORγt-dependent thymocyte development leads to a high frequency of thymic lymphoma (8, 58, 59). Our results demonstrate that a posttranslational ubiquitination event can dictate RORγt function in T_H17-dependent responses involved in autoimmunity. However, this ubiquitination event is dispensable for thymocyte development and T_H17 differentiation. Therefore, targeting this RORγt ubiquitination event is a potential treatment for T_H17-dependent autoimmune disease without induction of thymocyte lymphoma. Currently, it remains unknown about the RORγt ubiquitination pathway including the enzymes involved in the ubiquitination of RORγt at K256. Illustrating the detailed mechanisms responsible for ubiquitination of RORγt will facilitate the development of novel treatments that target the RORγt-dependent effector function of T_H17 cells responsible for autoimmunity while minimizing the other toxic effects such as lymphoma.

MATERIALS AND METHODS

Experimental design

The objective of this study was to determine whether RORγt plays a role in effector function of T_H17 cells in addition to its known function in T_H17 differentiation. To achieve this goal, we dissected the function of RORγt with K-R mutations in thymus development, T_H17 differentiation, and induction of EAE. RORγt-K256R mutation did not affect T_H17 differentiation but impaired the effector function of T_H17 cells responsible for inducing EAE, which was also confirmed by the in vivo studies using RORγt^K256R/K256R mice. RNA-seq and ChIP-seq assays identified Runx1 as a direct target of RORγt in the regulation of effector function of T_H17 cells.

Mice

All male and female mice used for experiments were between 6 and 12 weeks old; age-matched littermates were used. The RORγt^−/− (Rorc^tm1Litt, stock no. 007571) mouse strain was described previously (18). The RORγt^K256R/K256R point-mutated mice were designed and generated by Biocytogen LLC. Rag1^−/− (Rag1^tm1Mom, stock no. 002216), Tg^Tcr2D2 (Tcra2D2 and Tcrb2D2, stock no. 006912), IL-17A–GFP (Il17a^tm1Bcgen, stock no. 018472), CRISPR-Cas9– enhanced GFP (EGFP) [Gt(ROSA)26Sor^{em1.1(CAG-cas9}*^,-EGFP)Rsky, stock no. 028555], and C57BL/6J (stock no. 000664) mice were purchased from the Jackson Laboratory. For some assays, the mice were crossed to generate RORγt^−/−/Tg^Tcr2D2, RORγt^K256R/K256R/Tg^Tcr2D2, and RORγt^K256R/K256R/IL-17A-IRES-GFP-KI mice. All mice were bred at the C57BL/6J background and maintained in a pathogen-free animal facility at City of Hope. All animal experiments were conducted per the protocols approved by the Institutional Animal Care and Use Committee at City of Hope. Statistical tests were not used to predetermine sample sizes. The sample sizes were chosen on the basis of previous studies of our own and by others in the field (8). The sample sizes are indicated in the figure legends or figures. Allocation of mice to experimental groups was random.

Induction and assessment of EAE

Active EAE was induced and assessed as previously described (8). Briefly, mice were immunized with 200 mg of MOG_35–55 (Hooke Laboratories) in complete Freund’s adjuvant by subcutaneous injection at two dorsal sites at day 0, followed by two intraperitoneal injections of 80 ng of pertussis toxin at days 0 and 1. For passive EAE, Rag1^−/− mice were adoptively transferred with 1 × 10⁵ TCR^MOG-expressing (Tg^Tcr2D2) T_H17 cells that were differentiated under T_H17 polarization condition, followed by an immunization with MOG at 7 days after injection. In certain experiments, WT RORγt cells were virally transduced with retrovirus expressing GFP alone, while RORγt^K256R/K256R cells were transduced with retrovirus expressing Runx1 and GFP before in vitro T_H17 differentiation. In other experiments, RORγt^−/− cells were transduced with an empty vector or vectors encoding WT RORγt, RORγt^K87R, RORγt^K99R, RORγt^K256R, or RORγt^K288R. All transduced cells were sorted for CD4 and GFP expression before adoptive transfer into mice. Severity of EAE was monitored, and a clinical score of 0 to 5 was assigned (30): 0 = no disease, 0.5 = partially limp tail, 1 = paralyzed tail, 2 = hindlimb weakness, 3 = hindlimb paralysis, 4 = hindlimb and forelimb paralysis, and 5 = moribundity and death.

Isolation of naïve CD4⁺ T cell isolation and in vitro T_H17 differentiation

Murine CD4⁺ T cells were isolated from spleens by negative selection using the Naive CD4⁺ T Cell Isolation Kit (Miltenyi Biotec). Suspensions of 4 × 10⁵ cells/ml in RPMI 1640 medium (Corning Inc.) containing 2 mM l-glutamine, 50 μM β-mercaptoethanol, penicillin (100 U/ml), streptomycin (100 mg/ml), and 10% fetal bovine serum (FBS) were activated with hamster anti-CD3 (0.25 μg/ml; 145-2C11, BioLegend) and hamster anti-CD28 (1 μg/ml; 37.51, BioLegend) antibodies overnight in 24-well plates precoated with rabbit anti-hamster immunoglobulin G fraction (0.1 mg/ml; catalog no. 55398, MP Biomedicals). The following T_H17 differentiation was carried out by supplementing to the culture medium mentioned above with TGF-β (2 ng/ml; Miltenyi Biotec), IL-6 (20 ng/ml; Miltenyi Biotec), anti–IL-4 (2 μg/ml; 11B11, BioLegend), and anti–IFN-γ (2 μg/ml; XMG 1.2, BioLegend), and additional IL-23 (20 ng/ml; Miltenyi Biotec) was also added for the induction of pathogenic T_H17 cells.

Retroviral transduction

The retroviral vector murine stem cell virus (MSCV)–internal ribosomal entry site (IRES)–GFP (MIGR1, a gift from W.S. Pear, University of Pennsylvania) was used to clone WT or mutated RORγt. MSCV vector for expressing Runx1 was a gift from I. Taniuchi (RIKEN Center for Integrative Medical Sciences, Japan). Vectors were first transfected to Platinum-E (Cell Biolabs) retroviral packaging cells using BioT transfection reagent (Bioland Scientific), followed by a change of fresh medium at 24 hours. The virus-containing supernatant was collected at 48 and 72 hours, filtered with a 0.45-μm polyvinylidene difluoride (PVDF) syringe filter (Millipore), and used to transduce T cells or stored for future use at −80°C. Transduction of activated CD4⁺ T was performed by spin infection with viral supernatants (two, 500g, 30°C for 2 hours) in the presence of polybrene (8 μg/ml; Sigma-Aldrich). After spinning, the plates were incubated at 37°C for 3 hours. The viral supernatant was replaced with fresh culture medium with polarizing cytokines for in vitro differentiation.

CRISPR-Cas9–mediated genomic DNA deletion

Single-guide RNA (sgRNA) of Runx1, targeting the exon region (Addgene, library 67988), was cloned to pMSCV-U6sgRNA(Bbs I)-PGKpuro2ABFP (Addgene, 102796) with modification of Bbs I sites (table S1) for using universal primer design through this study. To generate plasmids for the deletion of large fragments of genomic DNA, PCR products of two U6 promoter–sgRNA cassettes and a phosphoglycerate kinase (PGK) promoter–TagBFP cassette were assembled using the Golden Gate assembly method and inserted into the MIGR1 vector with disrupted Bsp MI site. Bbs I sites and newly introduced Bsp MI sites were used for the insertion of gRNAs into each cassette. sgRNAs were delivered to the cells by retroviral transduction. The U6 promoter–driven transcription of sgRNAs in each cassette was confirmed by examining sgRORγt expression in T_H17 cells together with a simultaneously expressed NTC sgRNA in another cassette. Three sgRNAs (sgRNA1, sgRNA2, and sgRNA3; fig. S6A) targeting the sequence of the Runx1 gene were designed by using an online tool (CRISPOR, http://crispor.tefor.net/). A simultaneous expression of sgRNA1 with sgRNA2 or sgRNA2 with sgRNA3 in Cas9-expressing cells was designed for deletion of the RORγt binding region (Rgn1) and the adjacent region (Rgn2) without RORγt biding site. sgRNA primers are listed in table S1.

In vitro T cell–development assay

Murine thymocytes from RORγt^−/− mice were subjected to fluorescence-activated cell sorting for isolating DN (Thy1.2⁺CD4⁻CD8⁻) cells. Sorted cells at 5 × 10⁵ cells/ml were cultured overnight on an 80% confluent bone marrow-derived stromal cell line (OP9) expressing the delta like canonical Notch ligand 4 (Dll4/DL4) (OP9-DL4) monolayer (a gift from E.V. Rothenberg) in 24-well culture plates with α-modified minimum essential medium (Invitrogen Life Technologies) supplemented with 20% FBS, penicillin-streptomycin (100 U/ml), 2 mM l-glutamine (Invitrogen Life Technologies), and recombinant mouse IL-7 (5 ng/ml; PeproTech). The cells were then transduced with RORγt carrying K/R mutations as described above. Cocultures were maintained for an additional 3 days in the fresh medium containing murine IL-7 (5 ng/ml). Cells were harvested for flow cytometry analysis.

Apoptosis assays

Murine thymocytes were collected by smashing the thymus in a 40-μm cell strainer. Cells were suspended in RPMI 1640 medium supplemented with 10% FBS, 1% penicillin-streptomycin, and 2 mM l-glutamine at 1 × 10⁶ cells/ml and cultured for 0, 3, 6, 12, 18, or 24 hours. Thymocytes were then incubated with anti-Thy1.2 antibody and a fixable LIVE/DEAD near-infrared dye (Thermo Fisher Scientific). After two washes, the cells were stained with 5 μl of phycoerythrin–annexin V in 100 μl of binding buffer containing 0.01 M Hepes (pH 7.4), 0.14 M NaCl, and 2.5 mM CaCl₂ for 15 min. An additional 400 μl of binding buffer was added to the suspension before analysis.

RNA-seq and analysis

CD4⁺ T cells isolated from RORγt^WT/IL-17A-GFP^+/− and RORγt^K256R/K256R/IL-17A-GFP^+/− mice were differentiated into pathogenic T_H17 cells as described above. RNA was extracted from sorted ~1 × 10⁶ GFP-expressing T_H17 cells (CD4⁺GFP⁺) using an RNAeasy mini kit (QIAGEN). Each group contained three replicates from different mice. Quality control, library preparation, and sequencing were performed at Novogene. The analysis was performed through Partek Flow. Briefly, the sequence reads were aligned to the mouse whole genome (GRCm38) with validation of quality through prealignment and postalignment quality assurance/quality control (QA/QC). Aligned reads were further subjected to quantification using the Partek expectation/maximization (E/M) algorithm and normalization to counts per million with 0.001 added to each. The identification of differentially expressed features was performed through the Partek gene specific analysis (GSA) algorithm that applies multiple statistical models to each gene. Genes with total counts over 30 were considered to be statistically expressed in the cells. The expression values of pathogenic genes were extracted and subjected to ingenuity pathway analysis (IPA), gene set enrichment assay (GSEA), and network analysis.

Chromatin immunoprecipitation sequencing

In vitro–activated (see above) RORγt^−/− CD4⁺ T cells that were transduced with retroviruses carrying GFP, RORγt-3xFlag/GFP, or RORγt^K256R-3xFlag/GFP were used. After T_H17 polarization, 2 × 10⁷ cells were fixed in 1% formaldehyde at room temperature for 10 min to cross-link proteins with chromatin. The reaction was stopped with incubation in glycine for 5 min. Genomic DNA was fragmented with enzyme cocktail (ChIP-IT Express Enzymatic kit, Active Motif) for 10 min as directed. Cell lysates were centrifuged at 15,000 rpm for 10 min to remove debris, and the supernatant was used for immunoprecipitation. An equal amount of DNA was incubated with anti-FLAG (M2, Sigma-Aldrich) overnight, followed by precipitation with protein G agarose beads. Beads complexed with DNA fragments were extensively washed five times, and DNA was eluted, followed by reverse cross-linking. Recovered DNA was subjected to NovaSeq with 51–base pair (bp) paired-end sequencing length. Primers used in reverse-transcription quantitative PCR (RT-qPCR) are listed in table S1. Reads were analyzed using Partek Flow through alignment to the mm10 mouse genome using the Burrow-Wheeler aligner (BWA). Peaks were identified with the model-based analysis of ChIP-seq 2 (MACS2) tool (version 2.1.1) and quantified with a minimum region size of 50 bp.

Flow cytometry

For surface staining, cells isolated from mice or in vitro culture were directly stained with antibodies in phosphate-buffered saline (PBS) with 2% FBS and 1 mM EDTA at 4°C for 15 min. A blocking of Fc receptors with anti-CD16/32 antibody was carried out in case monocytes/macrophages were present. For staining transcription factors, cells were prestained for surface markers, fixed, and permeabilized in TF Fix/Perm buffer (BD Biosciences) at 4°C for 20 min and washed once with TF Perm/Wash buffer. Cells were stained for target proteins (see antibody list below) in the TF Perm/Wash buffer at 4°C for 15 min. For cytokine staining, cells were prestimulated with phorbol 12-myristate 13-acetate (50 ng/ml; Sigma-Aldrich) and ionomycin (750 ng/ml; Sigma-Aldrich) for 3 hours ahead of staining. Meanwhile, GolgiStop (BD Biosciences) was cotreated to block protein transport. In certain experiments, cells were stained with surface markers and/or fixable live/dead dye (Thermo Fisher Scientific). Cells were fixed and permeabilized with CytoFix/CytoPerm buffer (BD Biosciences), followed by staining for cytokines in the Perm/Wash buffer (BD Biosciences) after washing. To measure cell proliferation, either naïve CD4⁺ T cells or in vitro–differentiated T_H17 cells for adoptive transfer were stained with CellTrace Violet dye (Thermo Fisher Scientific) in PBS (1:5000) at room temperature for 20 min. After washing, naïve CD4⁺ T cells were subjected to anti-CD3/anti-CD28 stimulation and T_H17 differentiation for measuring in vitro proliferation, and T_H17 cells were sorted out and injected to Rag1^−/− mice for measuring hemostatic proliferation at day 3. Subsequent analysis was performed in the BD LSRFortessa flow cytometer.

The following antibodies were used for flow cytometric assay: anti-CD45 (BioLegend, clone 30-F11), anti-CD3 (BioLegend, 145-2C11), CD4 (BioLegend, RM4-5), anti-CD8 (BioLegend, 53-6.7), anti-CD19 (BioLegend, 1D3), anti-lymphocyte antigen 6 complex locus G6D (Ly6G) (BioLegend, 1A8), anti-CD62L (BioLegend, MEL-14), anti-CD44 (BioLegend, IM-7), anti–IFN-γ (BioLegend, XMG-1.2), anti–GM-CSF (BioLegend, MP1-22F9), killer cell lectin-like receptor subfamily B member 1C (Klrb1c/NK1.1) (BioLegend, PK136), anti-CD11b (eBioscience, M1/70), anti-Ly6C (eBioscience, HK1.4), anti-Thy1.2 (eBioscience, 53-1.2), anti–IL-17A (eBioscience, eBio17B7), anti-Runx1 (eBioscience, RXDMC), anti–IL-22 (eBioscience, 1H8PWSR), anti-Foxp3 (eBioscience, FJK-16 s), anti-RORγt (BD Biosicences, Q31-378), and CD1d-tetramer [National Institutes of Health (NIH), PBS-57].

Western blotting

A total 1.5 × 10⁷ T_H17 cells were lysed in radioimmunoprecipitation assay buffer containing 20 mM tris-HCl (pH 7.4), 150 mM NaCl, 1 mM Na₂EDTA, 1 mM EGTA, 1% NP-40, 1% sodium deoxycholate, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na₃VO₄, and leupeptin (1 μg/ml). Ubiquitinated proteins were precipitated and enriched with 20 μl of equilibrated agarose-Tandem Ubiquitin Binding Entities 2 (TUBE2) (Lifesensors) at 4°C for 4 hours. Agarose-TUBE2-protein complex were washed with Tris buffered saline containing 0.1% Tween-20 (TBST) and subjected to heating in 2× Laemmli sample buffer (Bio-Rad) with β-mercaptoethanol at 90°C for 5 min. The supernatant containing precipitated proteins was subjected to SDS–polyacrylamide gel electrophoresis, and the protein was transferred to the PVDF membrane. Target proteins were sequentially immunoblotted with relevant primary antibodies and fluorescent secondary antibodies (LI-COR Biosciences), followed by measuring fluorescent intensity with a LI-COR Odyssey blot imager (LI-COR Biosciences). The eventual samples for Western blotting were pooled from three different experiments. Quantification of ubiquitination signals of blots showing in Fig. 2 (E and F) was performed to the area above 50 kDa.

Reverse-transcription quantitative PCR

Total RNA was extracted using the RNeasy mini kit (QIAGEN) as directed. A Tetro complementary DNA synthesis kit (Bioline) was used for reverse transcription. Subsequent qPCR was performed using PowerUp SYBR Green Master Mix (Applied Biosystems) in the QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific). The primers used for qPCR are listed in table S1. The amplification efficiency of all primers was tested and optimized. Gene expression was calculated with the delta-delta Ct (∆∆Ct) method normalized to the control gene encoding β-actin and glyceraldehyde-3-phosphate dehydrogenase, and all measurements were performed in triplicate.

Luciferase assay

Human embryonic kidney 293T cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% FBS, 2 mM glutamine, penicillin (100 U/ml), and streptomycin (100 mg/ml). A total of 8 × 10⁵ cells were seeded to each well of a six-well plate and transfected with the reporter vectors (400 ng), pSV40-Renilla luciferase vector (200 ng), and expression vectors (2 μg) using BioT transfection reagent (Bioland Scientific, Paramount, CA). The same amount of plasmid DNA was used by adjusting with an empty vector. Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI) per the manufacturer’s instruction and normalized against Renilla luciferase activities. “Relative luciferase activities” were plotted with further normalization of luciferase activities of each group to the pGL3-basic reporter vector plus the empty vector group. The generation of reporter plasmids was done by PCR amplification of Runx1 genomic DNA containing the RORγt-binding region and a subsequent insertion upstream of a mini TK promoter that was cloned to pGL3-basic vector (Promega) for a minimal expression of luciferase. Cloning primers and mini TK promoter sequence are listed in table S1.

Statistical analysis

The ratio of rescue for thymocyte development and T_H17 differentiation in Fig. 2 (A and B) was calculated as relative ratio of the extent of RORγt^−/− cells transduced with RORγt mutants to that of cells transduced with WT RORγt. The statistical parameters are indicated in the figure legends. The results were analyzed for statistical significance with unpaired Student’s t test. Bodyweights are presented as means ± SD, and other data are shown as means ± SEM. P values are calculated using GraphPad Prism and presented where the statistical significance (P < 0.05) was found.

Supplementary Materials

This PDF file includes:

Figs. S1 to S7

Tables S1 to S4

View/request a protocol for this paper from Bio-protocol.