Helicase A determines the transcription program of TH17 lineage differentiation and autoimmunity Authors: Yanhong Su, Anjun Jiao, Xiaoran Zhang, Yanan Xu, Huanxin Ping, Kun Zhu, Qiaohao Wang, Ning Yuan, Yang Gao, Bomiao Ju, Renyi Ding, Ni Wang, Lan He, Lei Lei, Yong Zhao, Mingzhen Zhang, Baojun Zhang Published: 2026 ABSTRACT T helper 17 (TH17) lineage is governed by canonical transcriptional factors and plays a pivotal role in the pathogenesis of autoimmune diseases in both mice and humans. However, the precise orchestration of this transcriptional program remains poorly understood. Here, we identified a positive correlation between the expression of Dhx9, a nuclear helicase, and TH17 lineage during the progression of autoimmune diseases. Conditional deletion of Dhx9 in T cells significantly reduced TH17 differentiation and ameliorated the symptoms in mouse models of experimental autoimmune encephalomyelitis (EAE) and rheumatoid arthritis (RA). Mechanistically, Dhx9 was essential for increasing chromatin accessibility at the Rorc and Il17 loci, thereby facilitating the binding of canonical factors SMAD2, SMAD3, STAT3 (signal transducer and activator of transcription 3), and IRF4 (interferon regulatory factor 4) to drive the transcription of these two core TH17-specific genes. We additionally identified Nono as an uncharacterized regulator of TH17 differentiation that acts dependent on its interaction with Dhx9. Meanwhile, IL-6 (interleukin-6)–STAT3 signaling promotes TH17 cell differentiation by up-regulating Dhx9 expression. We identified a potential small-molecule inhibitor of Dhx9, punicalagin, which effectively suppressed TH17 differentiation and EAE progression. These findings uncover a mechanism orchestrating the transcriptional program of TH17 lineage commitment and highlight its therapeutic potential for autoimmune disease intervention. INTRODUCTION T helper 17 (TH17) cells are a predominant subset of CD4+ T cells implicated in multiple autoimmune diseases, such as multiple sclerosis (MS), rheumatoid arthritis (RA), and type 1 diabetes (T1D) (1–4), although they also play a protective role in maintaining mucosal integrity (5–7). TH17 cells are characterized by the production of a distinct set of cytokines, such as interleukin-17A (IL-17A; also known as IL-17), IL-17F, IL-21, and IL-22 (1, 3). Aberrant differentiation and accumulation of TH17 cells contribute to excessive inflammation, which drive the onset and progression of autoimmune disorders (4, 8, 9). The IL-17A cytokine promotes inflammation by recruiting neutrophils, lymphocytes, and macrophages to inflamed sites and stimulating the production of pro-inflammatory cytokines, including IL-1β, IL-6, and tumor necrosis factor–α (TNF-α), primarily by neutrophils and macrophages (10–13). Genetic disruption of TH17 cell differentiation has been shown to prevent the development and progression of autoimmune diseases in animal models (14–17). Although neutralization of IL-17A or inflammatory cytokines such as TNF-α and IL-6 can temporarily alleviate disease symptoms, frequent relapse and nonresponse are commonly observed in both mouse models and clinical settings (18, 19). Given that TH17 cells secrete not only IL-17A but also a cascade of additional pro-inflammatory cytokines, targeting a single effector cytokine may be ineffective (18). Thus, elucidating the molecular mechanisms governing TH17 cell differentiation remains essential for identifying effective long-term therapeutic strategies. TH17 cells are induced from T cell receptor (TCR)–stimulated CD4+ T cells in the presence of IL-6 and transforming growth factor–β1 (TGF-β1) (20, 21). The retinoic acid receptor–related orphan receptor γt (RORγt) serves as the master transcription factor (TF) governing TH17 cell lineage specification (22). IL-6 signals through the activation of signal transducer and activator of transcription 3 (STAT3), whereas TGF-β1 functions via SMAD2/3. STAT3 and SMAD2/3 bind to the Rorc promoter and distinct cis-regulatory elements, conserved noncoding sequence 6 (CNS6) and CNS9, to regulate its transcription (23). In addition, STAT3 and SMAD2/3 directly bind to and transactivate the Il17a promoter (9). Interferon regulatory factor 4 (IRF4) and basic leucine zipper transcription factor ATF-like (BATF) also contribute to TH17 cell differentiation through both RORγt-dependent and RORγt-independent mechanisms (24, 25). BATF stabilizes TH17 cell differentiation by antagonizing STAT5-mediated recruitment of Ets1-Runx1 complexes to TH1- and regulatory T cell (Treg cell)–specific gene loci (26). In addition to these canonical regulators, other signaling axes, including EGR2-CCR6, USP1-TAZ, STAT3-SET9/SETD1, and RORα-RORγt, have also been implicated in TH17 cell differentiation (14, 15, 27–29). However, the precise coordination of these transcriptional programs in orchestrating TH17 lineage commitment remains poorly understood. Dhx9 is a nucleic acid sensor known to induce innate immunity and functions as an adenosine triphosphate (ATP)–dependent RNA helicase A (RHA) that regulates gene transcription, DNA replication, translation, RNA processing and transport, and genomic stability (30). For example, Dhx9 recognizes pathogenic RNA and activates nuclear factor κB (NF-κB) and IRF3 in an IPS-1–dependent manner, thereby promoting the production of type I interferons (IFNs) and pro-inflammatory cytokines in myeloid dendritic cells (DCs) (31). Dhx9 also assists NLRP9b in sensing short double-stranded RNA (dsRNA), facilitating inflammasome assembly and triggering gasdermin D–mediated pyroptosis to combat rotavirus infection in intestinal epithelial cells (32). Beyond its role as a cytoplasmic dsRNA sensor, Dhx9 regulates gene expression or protein activation by acting as a transcriptional factor or coactivator through interactions with other proteins. For instance, Dhx9 directly binds to the Id2 promoter to regulate its transcription, thereby promoting effector CD8+ T cell differentiation (33). It also interacts with STAT1 to facilitate the transcription of interferon-stimulated genes (ISGs), enhancing resistance to RNA virus infections (34). Moreover, Dhx9 interacts with NF-κB and facilitate the recruitment of RNA polymerase II (RNAPII) to NF-κB–binding sites, enhancing the production of inflammatory cytokines in response to DNA virus infections (35, 36). Dhx9 is essential for proper TCR signal transduction and protects CD8+ T cells from apoptosis by facilitating LCK-mediated phosphorylation of ZAP70 (33). Despite these diverse functions, whether and how Dhx9 regulates CD4+ T cell lineage differentiation through its helicase activity remain unknown. In this study, we observed sustained up-regulation of Dhx9 expression in CD4+ T cells during the progression of autoimmune diseases in both patients and mice. T cell–specific Dhx9-deficient mice exhibited significantly ameliorated symptoms in experimental autoimmune encephalomyelitis (EAE) and RA, accompanied by reduced frequencies of IFN-γ+ and IL-17A+ CD4+ T cells. The in vitro experiments confirmed that Dhx9 deficiency markedly attenuates TH17 cell differentiation. Mechanistically, Dhx9 modulates chromatin accessibility through its helicase activity, facilitating the binding of canonical factors SMAD2, SMAD3, STAT3, and/or IRF4 to Il17a and Rorc loci to promote their transcription. Furthermore, we found that Dhx9 acts as a binding partner and coactivator of Nono, which also contributes to TH17 cell differentiation. IL-6–STAT3 signaling promotes TH17 cell differentiation by up-regulating Dhx9 expression. Notably, we identified punicalagin, a potential small-molecule inhibitor of Dhx9, which effectively inhibits TH17 cell differentiation in both mouse and human cells while also ameliorates mouse EAE progression. Together, our findings reveal a previously unknown mechanism orchestrating the transcriptional program of TH17 lineage commitment and provide a potential molecular target for the effective treatment of autoimmune diseases. RESULTS Dhx9 expression is required for T cell–dependent autoimmunity To identify key regulators involved in T cell–mediated autoimmune diseases, we first analyzed gene transcription changes in spinal cord CD4+ T cells from different disease stages of EAE mice using single-cell RNA sequencing (scRNA-seq) data (GSE182803). Among the nine transcriptional patterns identified, the 34 genes in the C3 cluster showed a continuous increase from disease onset to peak (Fig. 1A), indicating their potential involvement in EAE development and progression. Then, we examined the expression of these genes in human CD4+ T cells from patients with Crohn’s disease (CD) (GSE134809), MS (GSE133028), primary Sjögren’s syndrome (PSS) (GSE157278), and RA (GSE159117). Thirteen genes were found to be consistently up-regulated across all four autoimmune diseases compared to healthy donors (Fig. 1, B and C, and fig. S1A). Among these genes, Dhx9 drew our immediate attention due to its fundamental role as a helicase, potentially orchestrating downstream gene transcription. Elevated DHX9 expression was validated in CD4+ T cells from patients with RA and SLE compared to heathy donors (fig. S1B), indicating a potential role of Dhx9 in the pathogenesis of autoimmune diseases. Dhx9 expression is required for T cell–dependent autoimmunity. (A) Transcriptomic profiling of CD4+ T cells in spinal cords from the mouse model of EAE at various stages of disease progression (GSE182803). (B) ScRNA-seq analysis of selected genes (from the C3 cluster) in CD4+ T cells from PBMCs of patients with CD (GSE134809), MS (GSE133028), PSS (GSE157278), and RA (GSE159117) compared to healthy controls (HC). (C) Venn diagram illustrating the shared up-regulated genes in CD4+ T cells from PBMCs in four types of autoimmune diseases (CD versus HC, MS versus HC, PSS versus HC, and RA versus HC). (D) WT (Dhx9f/f CD4Cre−) and KO (Dhx9f/f CD4Cre+) mice were immunized with MOG35–55 emulsified in FCA to induce EAE. Clinical scores were recorded daily from day 7 postimmunization (n = 10). (E) H&E staining of the spinal cord from WT and KO mice on day 19 postimmunization. (F) Cytokine levels in the serum from WT and KO mice with EAE were measured by Luminex assay on day 19 postimmunization (WT, n = 3; KO, n = 4). (G to J) WT and KO mice were immunized with type II collagen emulsified in FCA on day 0 and day 21 to induce CIA. (G) Representative images of joints from WT and KO mice. The severity scores (H) and incidence rates (I) of arthritis (WT, n = 9; KO, n = 10). (J) H&E staining of the ankle joints from WT and KO mice at the peak of disease progress. Data are shown as the means ± SEM. Statistical analysis: two-tailed Student’s t test. *P ≤ 0.05 and **P ≤ 0.01. To investigate the role of Dhx9 in CD4+ T cells in the context of autoimmune diseases, we generated T cell–specific Dhx9 knockout (KO) mice by crossing Dhx9-floxed mice with CD4Cre mice, as previously described (33). Then, wild-type (WT; Dhx9f/fCD4Cre−) and Dhx9 KO (Dhx9f/fCD4Cre+) mice were immunized with MOG35–55 emulsified in Freund’s complete adjuvant (FCA), followed by injections of pertussis toxin on days 0 and 2 to induce EAE. Whereas WT mice displayed severe clinical symptoms of EAE, Dhx9-deficient mice were almost free of disease (Fig. 1D). Hematoxylin and eosin (H&E) staining revealed notably greater infiltration of inflammatory cells in the spinal cord of WT mice than Dhx9-deficient counterparts (Fig. 1E). Consistently, Luminex assay showed a reduction of IL-17A and TNF-α levels in the serum of Dhx9-deficient mice (Fig. 1F). To further confirm the role of Dhx9 in autoimmunity, we induced collagen-induced arthritis (CIA) by immunizing WT and Dhx9-deficient mice with type II collagen emulsified in FCA. WT mice developed pronounced joint swelling and redness, whereas these symptoms were markedly ameliorated in Dhx9-deficient mice (Fig. 1G). Both the clinical score and incidence of CIA were significantly lower in Dhx9-deficient mice than that in WT controls (Fig. 1, H and I). H&E staining also revealed less inflammatory infiltration in the joints of Dhx9-deficient mice (Fig. 1J). Collectively, these findings demonstrate that Dhx9 in T cells contributes to the pathogenesis of autoimmune diseases. Dhx9 is essential for TH17 lineage differentiation both in vitro and in vivo To elucidate how Dhx9 deficiency alleviates the progression of autoimmune diseases, we carried out single-cell transcriptomic analysis of CD45+ immune cells isolated from the spinal cords of WT and Dhx9-deficient EAE mice. T cells were identified and clustered into nine distinct populations, including CD8+ T, Treg, CD4 Tem, CD4 Tcm like, TH17, inflammated T, natural killer T (NKT), gdTH17, and activated T cells. Notably, the proportions of Treg, CD4+ Tem, and TH17 cells were markedly decreased in Dhx9-deficient mice compared to WT controls (Fig. 2, A and B). Flow cytometry analysis confirmed reductions in both frequencies and numbers of CD4+ and CD8+ T cells (fig. S2, A and B), along with decreased percentages and numbers of IFN-γ+, IL-17A+, IFN-γ+IL-17A+, and TNF-α+ CD4+ T cells in the spinal cords of Dhx9-deficient mice (Fig. 2, C to F). In addition, the proportion and absolute number of Treg cells were reduced in the Dhx9-deficient mice compared to that in WT mice (fig. S2, C and D), indicating that Treg cells are not responsible for the alleviated disease phenotype. Therefore, in our subsequent analyses, we focused primarily on the effect of Dhx9 efficiency on effector CD4+ T cell population. Dhx9 is essential for TH17 lineage differentiation both in vitro and in vivo. (A) UMAPs of T cell clusters in the spinal cord of WT and KO mice, with clusters color-coded. (B) Percentages of T cell clusters in the spinal cord of WT and KO mice. (C to F) Representative FACS plots (C and D), frequencies (E), and numbers (F) of IFN-γ+, IL-17A+, IFN-γ+IL-17A+, and TNF-α+ cells in CD4+ T cells from the spinal cord of WT and KO mice (n = 4). (G) qPCR analysis of Dhx9 mRNA expression in WT naïve CD4+ T cells cultured under the polarizing conditions of TH0, TH1, TH2, TH17, and Treg cells for 24 hours (n = 3). (H) Frequencies of IFN-γ+, IL-4+, IL-17A+, and Foxp3+ cells in WT and KO CD4+ T cells cultured under the polarizing conditions of TH1, TH2, TH17, or Treg cells for 72 hours (n = 3). (I) qPCR analysis of Il17a and Rorc mRNA expression in WT and KO TH17 cells (n = 3). (J) Representative histogram plots and the mean fluorescence intensity (MFI) of RORγt in WT and KO TH17 cells (n = 3). (K and L) WT and KO naïve CD4+ T cells were activated for 16 hours and then infected with either Mock- or Rorc-overexpressing retrovirus, followed by culture under TH17-polarizing conditions for 48 hours. Representative FACS plots (K) and percentages (L) of IL-17A+ CD4+ T cells (n = 3). (M) Scatterplot depicting TH17 lineage-associated genes in WT and KO CD4+ T cells from TH17-polarizing conditions for 24 hours (n = 3). Data are shown as the means ± SEM and are representative of three independent experiments. Statistical analysis: two-tailed Student’s t test [(E), (F), and (H) to (J)] or one-way ANOVA (G) or two-way ANOVA (L). *P ≤ 0.05, **P ≤ 0.01, and ****P ≤ 0.0001; n.s., not significant. To determine whether Dhx9 affects TH17 cell differentiation, we first assessed the mRNA levels of Dhx9 in different TH subsets derived from WT mice. Quantitative polymerase chain reaction (qPCR) analysis revealed that Dhx9 expression was uniquely higher in TH17 cells compared to other TH subsets (Fig. 2G). Then, we sorted naive CD4+ T cells from WT and Dhx9-deficient mice and cultured them under TH1-, TH2-, TH17-, or Treg cell–polarizing conditions for 72 hours. Dhx9 deficiency significantly impaired TH17 lineage differentiation, without affecting the differentiation of other TH subsets (Fig. 2H and fig. S3). RORγt is the master TF governing TH17 cell differentiation. We measured the mRNA expression of Il17a and Rorc in WT and Dhx9-deficient CD4+ T cells cultured with IL-6 and TGF-β1 for 24 or 48 hours by qPCR and found that both Il17a and Rorc expression were significantly reduced in Dhx9-deficient TH17 cells (Fig. 2I). Consistently, the protein level of RORγt was also markedly decreased in Dhx9-deficient TH17 cells (Fig. 2J). Overexpression of RORγt partially rescued the impaired TH17 differentiation in Dhx9-deficient cells (Fig. 2, K and L). However, when we performed RNA-seq analysis of WT versus KO CD4+ T cells under TH17-polarizing condition for 24 hours, we found that Dhx9 deletion did not affect the transcription of other canonical regulators (Smad2, Smad3, Stat3, Batf, Irf4, etc.), expect for Rorc (Fig. 2M). Together, these data demonstrated that Dhx9 controls TH17 lineage differentiation primarily through regulating Rorc and Il17a transcription, suggesting that Dhx9 may function as an organizer of canonical regulators in this process. To further determine whether the reduced TH17 differentiation was due to impaired cell survival, as previously reported in activated CD8+ T cells (33), we cultured WT and Dhx9-deficient CD4+ T cells under TH17-polarizing conditions for 24, 48, and 72 hours. Although there was a slightly higher percentage of apoptotic cells in Dhx9-deficient cells under TH17 culture conditions compared to WT controls, treatment with pan-caspase inhibitor Z-VAD failed to restore IL-17A+ TH17 cell frequencies in the Dhx9-deficient group (fig. S4, A to E). These results approved that Dhx9 controls TH17 cell differentiation independent of apoptosis. Although IFN-γ+CD4+ T cells were reduced in the spinal cords of Dhx9-deficient EAE mice, no difference in TH1 differentiation was observed between WT and Dhx9-deficient CD4+ T cells in vitro (Fig. 2H and fig. S3A). We hypothesized that the decreased number of IFN-γ+CD4+ T cells in Dhx9-deficient mice was due to increased cell death, as previously reported in activated CD8+ T cells and TH17 cells (33). As expected, a significantly higher percentage of apoptotic cells (annexin V+ 7-AAD+) was observed in Dhx9-deficient TH1 cells compared to WT controls (fig. S5, A and B). Correspondingly, Dhx9-deficient TH1 cells exhibited notably increased caspase-3 expression whereas decreased Bcl-2 expression (fig. S5C). Treatment with the pan-caspase inhibitor (Z-VAD) was able to rescue the survival of Dhx9-deficient TH1 cells (fig. S5, A and B). Furthermore, we detected a reduction of ERK activation and ZAP70 phosphorylation in Dhx9-deficient CD4+ T cells upon stimulation with anti-CD3 and anti-CD28 antibodies for 5 to 30 min (fig. S5D). Together, these data demonstrate that Dhx9 supports the survival of activated CD4+ T cells by modulating TCR downstream signaling through a mechanism similar to that previously described in CD8+ T cells (33). Dhx9 is required for both IL-6 and TGF-β1 signals inducing TH17 differentiation Because IL-6 and TGF-β signals synergistically drive TH17 cell differentiation, we investigated whether their downstream signaling transduction depends on Dhx9 in this context. Naive WT and Dhx9-deficient CD4+ T cells were stimulated with plate-coated anti-CD3 and soluble anti-CD28 antibodies under varying concentration of IL-6 or TGF-β1. Under constant TGF-β1 (2 ng/ml), WT cells showed a dose-dependent increase in TH17 cell percentages with increasing IL-6, whereas KO cells exhibited consistently lower TH17 cell frequencies, with a modest increase in IL-6 (from 1 to 5 ng/ml) but no further increase up to 100 ng/ml (fig. S6, A and B). Likewise, under constant IL-6 (100 ng/ml), WT cells exhibited increased TH17 differentiation as the TGF-β1 concentration increased, whereas KO cells showed increased TH17 cells only up to TGF-β1 (0.05 ng/ml) and plateaued thereafter (fig. S6, C and D). In addition, IL-1β and IL-23, known to enhance TH17 (8, 37), displayed better effect on TH17 cell production in WT than KO cells (fig. S7, A and B). STAT3 activation, a canonical downstream of IL-6 signal, is crucial for TH17 differentiation. However, we did not observe a significant difference in pSTAT3 levels between WT and KO cells under the TH17-polarizing condition (fig. S8). On the other hand, TGF-β1–induced expression of Il17a and Rorc is primarily dependent on the phosphorylation of SMAD2/3. However, Dhx9 deficiency did not affect the levels of pSMAD2 and pSMAD3 in both Treg and TH17 cells (fig. S8). Together, these data indicate that Dhx9 is required for IL-6 and TGF-β inducing TH17 cell differentiation without directly affecting the activation of their downstream pathways. The helicase activity of Dhx9 is required for Il17a and Rorc transcription Dhx9 is known as a member of the DExH family of adenosine triphosphatase (ATPase)–dependent helicase. To investigate whether Dhx9 regulates the transcription of Il17a and Rorc through its helicase activity, we performed assay for transposase-accessible chromatin with high-throughput sequencing (ATAC-seq) on WT and Dhx9-deficient TH17 cells. Clearly, Dhx9-deficient cells exhibited weaker ATAC-seq peaks at the Il17a and Rorc loci than WT cells (Fig. 3A), indicating reduced chromatin accessibility and transcriptional activity at these two loci. The helicase activity of Dhx9 is required for Il17a and Rorc transcription. (A) ATAC-seq peaks at the Il17a and Rorc loci in WT and KO cells under TH17-polarizing conditions. (B) Schematic diagram of the WT and mutant forms of Dhx9 protein, as indicated. Amino acid (aa) positions are indicated by numbers. Domain abbreviations: DsRBD1/2, dsRNA binding domain 1/2; HA2, helicase-associated domain 2; OB fold, oligonucleotide/oligosaccharide-binding fold; RGG, glycine-rich domain; ATPB, ATP-binding site (413-421aa). (C to F) Naïve CD4+ T cells from WT and KO mice were activated with plate-coated anti-CD3 and soluble anti-CD28 antibodies for 16 hours and then infected with retroviruses overexpressing (OE) Mock, Dhx9 FL, ∆RBD, ∆Helicase, or ∆ATPB constructs, followed by culture under TH17-polarizing conditions for 48 hours. Representative FACS plots (C) and percentages (D) of IL-17A+ cells in CD4+ T cells (n = 3). (E) Representative histogram plots of RORγt expression in CD4+ T cells overexpressing indicated Dhx9 protein domains. (F) Statistic bar graphs for RORγt MFI in CD4+ T cells of indicated groups (n = 3). Data are shown as the means ± SEM and are representative of three independent experiments. Statistical analysis: one-way ANOVA. ***P ≤ 0.01. To further validate the helicase function of Dhx9, we generated three mutant constructs: an RNA/DNA binding domain mutant (∆RBD; amino acids 255 to 1383), an ATPase/helicase mutant (∆Helicase; amino acid 1 to 254 and 834 to 1383), and an ATP-binding site mutant (∆ATPB; amino acids 1 to 412 and 422 to 1383) (Fig. 3B). These three constructs were overexpressed in WT and Dhx9-deficient naïve CD4+ T cells, followed by the subsequent culture under TH17-polarizing conditions for 48 hours. Unlike the Dhx9 full-length (FL) construct, all three mutants (∆RBD, ∆Helicase, and ∆ATPB) failed to rescue IL-17A expression in the Dhx9-deficient cells (Fig. 3, C and D). Similar results were observed for RORγt expression (Fig. 3, E and F). Therefore, these results demonstrated that the helicase activity of Dhx9 is essential for the transcriptional regulation of Il17a and Rorc. Dhx9 establishes chromatin accessibility for key factors regulating TH17 differentiation On the basis of the abovementioned findings, we hypothesized that Dhx9 facilitates chromatin accessibility for canonical TFs to regulate Il17a and Rorc expression, thereby controlling TH17 lineage differentiation. To test this, we first performed chromatin immunoprecipitation (ChIP) analysis in WT CD4+ T cells cultured under TH17-polarizing conditions, which revealed marked Dhx9 binding at the regulatory regions of Il17a and Rorc genes (Fig. 4C), whereas no binding was detected at the Foxp3 regulatory region (fig. S9), which is consistent with the fact that Dhx9 deletion had no effect on Treg cell differentiation (Fig. 2H). This further supports a specific role of Dhx9 in TH17 cell, rather than in Treg cell, lineage differentiation. Moreover, Dhx9 exhibited similar binding patterns with SMAD2, SMAD3, STAT3, and IRF4 in TH17-polarizing cells (Fig. 4C). Overexpression of these TFs significantly enhanced TH17 cell differentiation in WT cells but not in KO cells, except for BATF, indicating that SMAD2, SMAD3, STAT3, and IRF4 regulate TH17 cell differentiation in a Dhx9-dependent manner (Fig. 4, A and B). Dhx9 establishes chromatin accessibility for key factors regulating TH17 differentiation. (A and B) WT and KO naïve CD4+ T cells were activated for 16 hours and then infected with either Mock-, Smad2-, Smad3-, Stat3-, Batf-, or Irf4-overexpressing retrovirus, followed by TH17-polarizing conditions for 48 hours. Representative FACS plots (A) and percentages (B) of IL-17A+ cells in CD4+ T cells (n = 3). (C) ChIP-seq peaks at the Il17a and Rorc loci bound by either Dhx9, SMAD2, SMAD3, STAT3, or IRF4 in WT and KO TH17 cells. (D and E) Representative histogram plots of RORγt expression (D) and statistic MFI of RORγt (E) in Mock- or Smad3-overexpressing WT and KO CD4+ T cells cultured under TH17-polarizing conditions for 48 hours (n = 3). (F and G) WT naïve CD4+ T cells were stimulated with anti-CD3 and anti-CD28 antibodies for 16 hours and then infected with either Mock or Nono–short hairpin RNA (shRNA) retrovirus. Cells were subsequently cultured under TH17-polarizing conditions for 48 hours. Representative FACS plots (F) and frequencies (G) of IL-17A+ cells in CD4+ T cells (n = 3). (H and I) Representative FACS plots (H) and frequencies (I) of Mock- or Nono-overexpressing WT and KO CD4+ T cells cultured under TH17-polarizing conditions for 48 hours (n = 3). (J) Luciferase activity in 293T cells transfected with the Il17a or Rorc luciferase reporter constructs along with either Mock or Nono-overexpressing plasmid (n = 3). (K) Luciferase activity in WT and Dhx9 KO Jurkat T cells transfected with the Il17a luciferase reporter construct along with either Mock or Nono-overexpressing plasmid then activated for 36 hours with PDBu and ionomycin (n = 3). Data are shown as the means ± SEM and are representative of three independent experiments. Statistical analysis: two-way ANOVA [(B), (E), and (I)] or two-tailed Student’s t test [(G) and (J)] or one-way ANOVA (K). *P ≤ 0.05, ***P ≤ 0.001, and ****P ≤ 0.0001. To precisely validate the role of Dhx9 in facilitating canonical TFs, we performed ChIP-seq analysis using antibodies against SMAD2, SMAD3, STAT3, and IRF4 in WT and Dhx9-deficient CD4+ T cells cultured under TH17-polarizing conditions. Binding of these TFs to the Il17a promoter and enhancer regions was markedly reduced in Dhx9-deficient TH17 cells compared to WT controls (Fig. 4C). At the Rorc loci, SMAD3 binding to the promoter was decreased in Dhx9-deficient cells, whereas bindings of SMAD2, STAT3, and IRF4 remained largely unchanged. Flow cytometry confirmed that Smad3 overexpression increased RORγt expression in WT cells but not in Dhx9-deficient CD4+ T cells (Fig. 4, D and E). Together, these findings support a critical role of Dhx9 in TH17 differentiation by facilitating the binding of canonical TFs to the Il17a and Rorc loci, thereby regulating their expression. Dhx9 is required for Nono regulating TH17 cell differentiation To further investigate whether other proteins also regulating the transcription of Il17a and Rorc in a Dhx9-dependent manner, we conducted immunoprecipitation (IP) coupled with mass spectrometry to identify Dhx9-associating proteins in TH17 cells. First, naïve CD4+ T cells were cultured under TH17-polarizing conditions for 48 hours, and cell lysates were incubated with Dhx9 antibody-conjugated agarose beads, with isotype control immunoglobulin G (IgG)–conjugated agarose beads as a negative control. Specific bands from the Dhx9 antibody–IP samples were then subjected to mass spectrometry. The top 10 proteins with the highest matching scores for Dhx9 interaction are listed in fig. S10A. Next, we assessed the role of these proteins in TH17 cell differentiation and found that knockdown of Nono significantly diminished TH17 cell differentiation (Fig. 4, F and G, and fig. S10B). Nono, known as a non-POU domain-containing octamer-binding protein, belongs to the Drosophila behavior/human splicing (DBHS) family and is involved in transcriptional regulation (38). Because mass spectrometry analysis showed that Nono was associated with Dhx9 in TH17 cells (fig. S10A), we next explored whether Nono regulates TH17 cell differentiation in a manner dependent on its interaction with Dhx9. Our data found that Nono overexpression increased TH17 cell proportion in WT group, but this effect was abolished in the Dhx9-deficient group (Fig. 4, H and I). However, Nono overexpression had no effect on Rorc expression (fig. S10C). Moreover, dual-luciferase reporter assays confirmed that Nono regulates Il17a promoter activity but does not affect the Rorc promoter (Fig. 4J). To further investigate whether Nono modulates the Il17a transcriptional activity in conjunction with Dhx9, we performed luciferase reporter assays in Dhx9 KO Jurkat cells. We found that Nono enhances Il17a promoter activity in a Dhx9-dependent manner (Fig. 4K). Therefore, our findings identify Nono as a previously unidentified regulator of TH17 cell differentiation that acts in a manner dependent on its interaction with Dhx9. IL-6–STAT3 signaling enhances Dhx9 expression To understand the upstream signal governing TH17 cell-specific expression of Dhx9, we evaluated Dhx9 expression in CD4+ T cells treated with IL-6 or TGF-β1. IL-6 significantly up-regulated Dhx9 expression at both mRNA and protein levels (Fig. 5, A to D), whereas Dhx9 expression remained unchanged regardless of TGF-β1 treatment (Fig. 5, A to D). Given that STAT3 is a major downstream effector of IL-6 signaling, we decided to assess whether STAT3 regulates Dhx9 expression. STAT3 overexpression significantly increased Dhx9 expression at both the transcription and protein levels (Fig. 5, E to G), whereas C188-9, a STAT3 inhibitor, markedly suppressed Dhx9 expression in dosage-dependent manner (Fig. 5, H to J). ChIP-seq analysis revealed STAT3 binding at the Dhx9 promoter region (Fig. 5K). To validate this regulatory interaction, we first computed the STAT3-binding motifs within the mouse Dhx9 promoter using the JASPAR database (Fig. 5L). Two putative STAT3-binding sites were identified at −179 and −152 base pairs (bp) relative to the transcription start site (TSS) of the Dhx9 gene (Fig. 5, L and M). We then construed plasmids containing either WT or mutant STAT3-binding fragments and assessed the effect of STAT3 on the transactivation of the Dhx9 promoter via luciferase reporter assays. Dhx9 promoter activity was enhanced by CA-STAT3 overexpression in the presence of a WT-binding fragment, whereas mutant ones (−179~−170 and −152~−143 bp) abolished this enhancement (Fig. 5N). These findings indicate that IL-6–STAT3 signaling directly regulates Dhx9 expression during TH17 cell differentiation. IL-6–STAT3 signaling enhances Dhx9 expression. (A) qPCR analysis of Dhx9 mRNA expression in WT CD4+ T cells stimulated with anti-CD3 and anti-CD28 antibodies for 24 hours, in the presence or absence of IL-6 or TGF-β1 (n = 4). (B to D) WT naïve CD4+ T cells were stimulated for 48 hours in the presence or absence of IL-6 or TGF-β1. Dhx9 expression was measured by flow cytometry. Representative FACS plots (B), percentages (C), and MFI (D) of Dhx9 expression in each group (n = 3). (E and F) Representative histograms (E) and MFI (F) of Dhx9 expression in CD4+ T cells overexpressing either Mock or Stat3 retrovirus for 48 hours (n = 3). (G) qPCR analysis of Dhx9 mRNA expression in WT CD4+ T cells overexpressing either Mock or Stat3 retrovirus for 36 hours (n = 3). (H to J) Naïve CD4+ T cells sorted from WT mice were activated in the presence or absence of IL-6, together with 1.25, 2.5, or 5 μM C188-9, STAT3 inhibitor, for 48 hours. Dhx9 expression was measured by flow cytometry. Representative FACS plots (H), percentages (I), and MFI (J) of Dhx9 expression in each group (n = 3). (K) ChIP-seq analysis of STAT3-binding peaks at the Dhx9 locus in WT TH17 cells (GSE40918). (L) Predicted DNA binding motifs bound by STAT3 at the Dhx9 locus. (M) STAT3-binding site (orange) at the Dhx9 promoter region. (N) The Dhx9 WT, S1 mutant, S2 mutant, or S1+S2 mutant promoter reporter plasmid was cotransfected with a construct encoding constitutively activated STAT3 (CA-STAT3) or Mock into 293T cells, and luciferase activity was detected (n = 3). Data are shown as the means ± SEM and are representative of three independent experiments. Statistical analysis: two-way ANOVA [(A), (C), and (D)] or two-tailed Student’s t test [(F) and (G)] or one-way ANOVA [(I), (J), and (N)]. *P ≤ 0.05, ***P ≤ 0.001, and ****P ≤ 0.0001. Punicalagin, a potential Dhx9 inhibitor, prevents TH17 lineage differentiation and EAE development On the basis of the regulatory role of Dhx9 in TH17 cell differentiation, we have explored whether Dhx9 could serve as a therapeutic target for treating EAE. The Dhx9 gene is conserved across multiple species, including Homo sapiens, Mus musculus, Xenopus tropicalis, and Danio rerio, particularly within its ATP-binding domain (fig. S11A). According to the binding affinity, we carried out computer virtual screening by docking 14,800 compounds from the MCE Bioactive Compound Library Plus. The top 40 candidates were selected for further experimental validation (Fig. 6A and table S1). WT CD4+ T cells cultured under TH17-polarizing conditions in the presence of the candidate drugs revealed that punicalagin markedly inhibited TH17 cell differentiation (fig. S11B). The chemical structure of punicalagin, along with its predicted docking modes and binding scores with Dhx9, were shown in Fig. 6B. We further demonstrated that punicalagin inhibited TH17 cell differentiation in a dose-dependent manner (Fig. 6, C and D) and suppressed RORγt expression in TH17 cells (Fig. 6, E and F). Consistent with observations in Dhx9-deficient cells, punicalagin treatment did not alter the phosphorylation levels of SMAD2 and SMAD3 upon TGF-β1 stimulation (fig. S11C). Punicalagin, a potential Dhx9 inhibitor, prevents TH17 lineage differentiation and EAE development. (A) Schematic diagram screening small-molecule drugs that target the helicase domain of Dhx9. (B) Top: Chemical structure of punicalagin. Bottom: Molecular docking mode and score between Dhx9 and punicalagin. (C and D) Representative FACS plots (C) and percentages (D) of IL-17A+ cells in CD4+ T cells cultured under TH17-polarizing conditions in the presence of DMSO or punicalagin (2.5 and 5 μM) for 72 hours (n = 3). (E and F) Histogram plots (E) and MFI (F) of RORγt expression in DMSO- or punicalagin-treated TH17 cells (n = 3). (G and H) Naïve CD4+ T cells (CD4+CD25−CD45RA+CD45RO−) from human donors were cultured under TH17-polarizing conditions in the presence of DMSO or punicalagin for 6 days. Representative FACS plots (G) and percentages (H) of IL-17A+ cells in CD4+ T cells (n = 4). (I) ChIP-seq peaks at the Il17a and Rorc loci bound by either IgG or Dhx9 in DMSO- or punicalagin-treated TH17 cells. (J) Experimental schematic of the EAE model treated with small-molecule drugs. WT mice were immunized with MOG35–55 to induce EAE and then injected intraperitoneally (ip) with saline control or punicalagin (4 mg/kg) on days 4, 7, and 10 postimmunization. PTx, pertussis toxin; i.h., subcutaneous injection. (K) Clinical scores of mice recorded from day 7 postimmunization [control (Ctr), n = 5; punicalagin, n = 6]. (L) H&E staining of spinal cord sections from EAE mice. (M to O) Representative FACS plots (M), percentages (N), and numbers (O) of IL-17A+, IFN-γ+, or IFN-γ+IL-17A+ CD4+ T cells in the spinal cord from the EAE mice treated with saline or punicalagin (control, n = 3; punicalagin, n = 4). Data are shown as the means ± SEM and are representative of three independent experiments. Statistical analysis: one-way ANOVA (D) or two-tailed Student’s t test [(F), (H), (N), and (O)]. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. Furthermore, we investigated whether punicalagin affects human TH17 differentiation. Human naïve CD4+ T cells (CD4+CD25−CD45RA+CD45RO−) were sorted from healthy donors and cultured under TH17-polarizing conditions in the presence of dimethyl sulfoxide (DMSO) or punicalagin for 6 days. We found that punicalagin significantly restricted human TH17 cell differentiation, consistent with our observations in mice (Fig. 6, G and H). Because Dhx9 binds to the loci of Il17a and Rorc genes to facilitate their transcription (Fig. 4C), examining the binding changes of Dhx9 at the Il17a and Rorc regions in TH17 cells treated with punicalagin provides an approach to demonstrate that punicalagin may target the Dhx9 protein. To test this, we performed cleavage under targets and tagmentation sequencing (CUT&Tag-seq) analysis using antibodies against Dhx9 in DMSO-treated or punicalagin-treated CD4+ T cells cultured under TH17-polarizing conditions. Bindings of Dhx9 to the Il17a promoter and enhancer regions, as well as the Rorc promoter, were markedly reduced in punicalagin-treated TH17 cells compared with the DMSO-treated group (Fig. 6I). These results indicate that punicalagin may affect the binding of Dhx9 to Il17a and Rorc loci, which is consistent with the observation in Dhx9-deficient TH17 cells. Last, we examined the therapeutic potential of punicalagin in the EAE mouse model. EAE was induced by subcutaneously injecting WT mice with a MOG35–55 peptide (emulsified in FCA) along with pertussis toxin on day 0 and an additional dose on day 2. Mice were subsequently administered saline or punicalagin (4 mg/kg) on days 4, 7, and 10 postimmunization (Fig. 6J). Punicalagin treatment effectively ameliorated the symptoms of EAE (Fig. 6K). The numbers of central nervous system–infiltrating inflammatory cells were noticeably reduced in punicalagin-treated mice compared to controls (Fig. 6L). The frequencies and absolute numbers of IL-17A+ and IFN-γ+ IL-17A+ CD4+ T cells in the spinal cord were markedly decreased in punicalagin-treated mice (Fig. 6, M to O). Collectively, these findings demonstrated that punicalagin may act as a potential Dhx9 inhibitor, effectively suppressing TH17 cell differentiation and attenuating EAE progression. DISCUSSION TH17 cells represent a predominant subset of T cells involved in the development and progression of autoimmune diseases. TGF-β and IL-6 signaling, along with their downstream molecules (SMAD2, SMAD3, STAT3, and IRF4), regulate TH17 lineage differentiation through modulating the transcription of Il17a and Rorc. However, it remains largely unknown how these canonical TFs are coordinated within the transcriptional program. Here, we demonstrate that elevated expression of Dhx9 is positively correlated to TH17 lineage commitment and autoimmune disease progression. T cell–specific deletion of Dhx9 impaired TH17 differentiation and ameliorated autoimmune disease symptoms, including EAE and RA. Mechanistically, Dhx9 enhanced chromatin accessibility at the Il17a and Rorc loci via its helicase activity, thereby facilitating the recruitment of SMAD2, SMAD3, STAT3, and IRF4 to promote Il17a and Rorc transcription and, consequently, TH17 cell differentiation (Fig. 7). Notably, we identified punicalagin, a potential small-molecule inhibitor of Dhx9, which effectively suppressed TH17 cell differentiation in both mice and humans, as well as ameliorated EAE progression in mice (Fig. 7). Schematic model illustrating that Dhx9 orchestrates TH17 transcriptional program and promotes the development of autoimmune diseases. In conjunction with TCR stimulation, TGF-β and IL-6 signals activate canonical TFs such as SMAD2, SMAD3, and STAT3. The basal level of Dhx9 modulates double-stranded DNA unwinding via its helicase activity and facilitates the binding of SMAD2, SMAD3, STAT3, and/or IRF4 to the Il17a and Rorc loci, thereby promoting their transcription. Meanwhile, Nono, as a binding partner and coactivator of Dhx9, directly binds to the Il17a promoter and regulates Il17a gene expression. Furthermore, Dhx9 is up-regulated by IL-6–STAT3 signaling, thereby enhancing Il17a and Rorc transcription. These collective transcription events promote TH17 cell differentiation, thereby contributing to the onset and progression of autoimmune diseases. Conversely, genetic depletion or pharmacological inhibition of Dhx9 by punicalagin effectively inhibits TH17 cell differentiation and ameliorates the progression and symptoms of autoimmune diseases. In the current study, we demonstrated that Dhx9 selectively promoted TH17 lineage differentiation, whereas there is no significant effect on TH1, TH2, and Treg cell differentiation. Consistent with this observation, Dhx9 expression was higher under TH17-polarizing conditions compared to other TH cell subtypes and was required for both IL-6 and TGF-β1 to induce TH17 differentiation. Although TGF-β1 is also a lineage-specific cytokine essential for inducible Treg cell differentiation, Dhx9 deficiency did not affect Treg cell differentiation. As we know, activation of SMAD2/3, downstream of TGF-β signaling, is required for the expression of Foxp3, Rorc, and Il17a (9, 39). However, T cell–specific deletion of Dhx9 had no effect on the phosphorylation of SMAD2/3 in either Treg or TH17 cells, indicating that Dhx9 does not directly participate in the regulation of TGF-β downstream signaling. Our ChIP-seq analysis revealed that Dhx9 was capable to bind to the Il17a and Rorc loci but not to the Foxp3 locus. Furthermore, we approved that Dhx9 deficiency reduced chromatin accessibility and decreased the binding of SMAD2, SMAD3, STAT3, and IRF4 at the Il17a and Rorc loci. Accordingly, overexpression of these canonical TFs failed to fully restore Rorc expression and TH17 cell differentiation in the absence of Dhx9. In contrast, SMAD2/3 could still be recruited to the Foxp3 locus independently of Dhx9, suggesting a locus-specific role for Dhx9. We hypothesize that one or more lineage-specific TFs may be responsible for recruiting Dhx9 to the Il17a and Rorc loci but not to the Foxp3 locus. However, mass spectrometry data did not reveal any known TH17-related TFs associated with Dhx9. Notably, Nono, a top Dhx9-interating protein identified by mass spectrometry, also participates in TH17 cell differentiation. However, it remains unclear whether Nono is required for the bindings of SMAD2, SMAD3, STAT3, and IRF4 to the Il17a and Rorc loci. Future studies should aim to identify the TFs responsible for recruiting Dhx9 specifically to these loci. Dhx9 regulates the transcription of both Rorc and IL17a at two distinct stages during TH17 cell differentiation. Within the first 24 hours, Dhx9 contributes to up-regulating Rorc expression, as evidenced by the reduced Rorc levels observed in Dhx9-deficient TH17 cells. Notably, Dhx9 expression itself does not increase during this initial period, suggesting that the basal level of Dhx9 is sufficient to mediate Rorc up-regulation. After 24 hours, IL-6–activated pSTAT3 causes the up-regulation of Dhx9, which subsequently promotes Il17a transcription. Our RNA-seq and qPCR data revealed that Rorc expression decreases by 48 hours, consistent with previous finding by Yosef et al. (40). Moreover, there is no significant difference in Rorc expression between WT and Dhx9-deficient cells at this later stage, indicating the impotency of Dhx9 in regulating Rorc expression during the early stage of TH17 cell differentiation. Therefore, basal Dhx9 levels regulate Rorc transcription initially, whereas IL-6–STAT3–driven Dhx9 up-regulation predominantly enhances ll17a transcription during the later stage. We also observed a reduction in the number of TH1 cells in the spinal cord of Dhx9-deficient mice, which may contribute to the attenuated development of EAE. Notably, Dhx9 deletion led to a decreased TH1 population, a defect that was reversed upon treatment with a pan-caspase inhibitor. A previous study demonstrated that, in CD8+ T cells, Dhx9 interacted with the CD3ε-LCK-ZAP70 complex to promote ZAP70 phosphorylation and subsequently regulated Bim and Bcl-2 expression via ERK phosphorylation, thereby supporting CD8+ T cell survival (33). In this study, we also observed a reduction in TCR signaling and resultant Bcl-2 expression in the absence of Dhx9, suggesting that Dhx9 similarly regulates TH1 cell survival. However, Dhx9 expression in CD8+ T cells was markedly up-regulated following TCR stimulation, whereas no substantial change in Dhx9 mRNA levels was observed in activated CD4+ T cells, indicating that basal Dhx9 expression may be sufficient to affect the survival of CD4+ T cells. In CD8+ T cells, up-regulated Dhx9 regulates Id2 transcription to drive effector differentiation, whereas in CD4+ T cells, IL-6–STAT3 signaling-induced Dhx9 facilitates Il17a transcription. These distinct regulatory mechanisms by which TCR or cytokine-cytokine receptor signaling pathways mediate Dhx9 expression in CD4+ and CD8+ T cells warrant further investigation. As we know, T cells are one of the major contributors to the development and progression of autoimmune diseases, such as EAE and RA. A recent study has demonstrated that CD8+ T cells exert diverse effector and regulatory functions in the pathogenesis of autoimmune diseases (41). However, IL-17+ and/or IFN-γ+ CD4+ T cells remain the primary drivers and key mediators in most autoimmune diseases due to the preferential presentation of autoantigens by MHC (major histocompatibility complex) class II molecules and the robust effector functions (42–44). In our study, we have provided a coherent and compelling evidence supporting that Dhx9 plays a critical role in autoimmune diseases by regulating CD4+ T cell survival and TH17 cell differentiation. We also found impaired CD8+ T cells in the mouse models of EAE and RA. Therefore, Dhx9 seems to influence the development of EAE primarily through regulating TH17 fate, together with CD8+ T cell fate. In the future, adaptive transfer approach investigation will be required to specifically delineate the contribution of CD4+ T cells to the ameliorated EAE and RA phenotypes observed in Dhx9-deficient mice. Dhx9 is involved in regulating both innate and adaptive immunity through diverse mechanisms. As a nuclear sensor, Dhx9 can sense RNA and subsequently activate type I IFNs via NF-κB and IRF3 signaling pathways (31) and induces gasdermin D–mediated pyroptosis by facilitating inflammasome assembly (32). Beyond its role as a sensor, Dhx9 can also function as a protein partner by interacting with STAT1 to promote ISG transcription (34), binding to LCK and ZAP70 to mediate ZAP70 phosphorylation and ensure CD8+ T cell survival (33) and associating with NF-κB and RNAPII to activate inflammatory cytokine gene expression (35, 36). Furthermore, Dhx9 acts as a TF by directly binding to the Id2 promoter, thereby enhancing its transcription and promoting effector CD8+ T cell differentiation (33). In the current study, we reveal a previously unidentified helicase function of Dhx9 that enhances chromatin accessibility at the Rorc and Il17a loci, thereby promoting TH17 cell differentiation. Collectively, these findings highlight the multifaceted roles of Dhx9 in regulating innate and adaptive immunity through distinct mechanisms. Of note, further investigation into the functions and mechanisms of Dhx9 across a broader spectrum of immune and nonimmune cell types is warranted. To explore the potential of Dhx9 as a therapeutic target for autoimmune diseases, we identified punicalagin, a small-molecule drug that effectively suppresses TH17 differentiation and ameliorates EAE symptoms. Punicalagin is the main active ingredient of pomegranate peel and exhibits broad pharmacological effects including anti-inflammatory, antioxidant, antitumor, and antimicrobial activities (45, 46). Notably, punicalagin displays anti-inflammatory effects in autoimmune diseases such as inflammatory bowel disease (47), ankylosing spondylitis (48), psoriasis (49, 50), and RA (51, 52), primarily by inhibiting NF-κB, mitogen-activated protein kinase (MAPK), and/or Janus kinase (JAK)–STAT signaling pathways in innate immune and tissue cells, thereby reducing the production of inflammatory cytokines such as IL-1β, TNF-α, and IL-6. However, our study reveals a previously unrecognized role for punicalagin in regulating TH17 cell differentiation. Besides the fact that NF-κB binds to the Il17a and Rorc promoters with the assistance of NFAT and STAT3 (53–56), MAPK signaling contributes to Rorc transcriptional activation and TH17 differentiation through STAT3 phosphorylation or other TFs (57–61). Thus, in addition to targeting Dhx9, punicalagin may also inhibit TH17 differentiation by targeting the NF-κB and/or MAPK pathways. Future investigations will help clarify this possibility and its underlying contribution. In conclusion, punicalagin may suppress inflammatory responses through both adaptive and innate immune mechanisms, exhibiting potential anti-inflammatory effects in multiple mouse models of autoimmune diseases. Of note, in the present study, it cannot be excluded that punicalagin confers protection in the EAE model by targeting additional cell types, such as innate immune or neuronal cells, rather than acting exclusively on TH17 cells. Further studies are required to determine whether the amelioration of EAE is primarily attributable to punicalagin-mediated targeting of Dhx9 and the consequent reduction in TH17 cell differentiation. Collectively, our findings, along with reports from others, support the potential application of punicalagin in treating various autoimmune diseases in mouse models (47–52), providing a rationale for advancing this compound to clinical trials, albeit through distinct mechanisms. Punicalagin, identified as a potential Dhx9 inhibitor, recapitulates the phenotype observed in Dhx9-deficient cells. We acknowledge, however, that additional biochemical studies are warranted to validate the binding affinity between Dhx9 and punicalagin, as well as its direct inhibitory effect on Dhx9 activity. In summary, we identified a previously unrecognized role of Dhx9 in promoting TH17 lineage differentiation by enhancing chromatin accessibility at the Il17a and Rorc loci via its helicase activity. T cell–specific deletion of Dhx9 markedly protected mice from the onset and progression of EAE and RA. Notably, we identified punicalagin, a potential small-molecule inhibitor, which effectively suppressed TH17 cell differentiation and ameliorated EAE progression. Together, these findings reveal a previously unknown mechanism by which Dhx9 regulates TH17 cell differentiation and highlight its potential as a therapeutic target for autoimmune diseases. MATERIALS AND METHODS Mice C57BL/6J mice were obtained from the Jackson Laboratory, and CD4Cre transgenic mice were purchased from Taconic Biosciences. The generation of Dhx9f/f strain has been described previously (33). T cell–specific Dhx9 KO mice (CD4Cre+Dhx9f/f) were generated by crossing CD4Cre+ mice with Dhx9f/f mice. All mice were maintained under specific pathogen–free conditions at the Laboratory Animal Center of Xi’an Jiaotong University. All procedures were approved by the Biomedical Ethics Committee of Health Science Center of Xi’an Jiaotong University (no. 2018-284). Antibodies and flow cytometry The following antibodies and reagents used for flow cytometry were purchased from BioLegend (San Diego, CA, USA): APC/Cy7 anti-mouse CD4 (clone GK1.5), PE/Cy5 anti-mouse CD4 (clone GK1.5), PE/Cy7 anti-mouse CD4 (clone GK1.5), FITC anti-mouse CD4 (clone GK1.5), PB anti-mouse CD8α (clone 53-6.7), PE anti-mouse CD25 (clone 3C7), FITC anti-mouse/human CD44 (clone IM7), PE/Cy7 anti-mouse/human CD44 (clone IM7), APC/Cy7 anti-mouse CD62L (clone MEL-14), PE anti-mouse IL-17A (clone TC11-18H10.1), PE/Cy7 anti-mouse IFN-γ (clone XMG1.2), PB anti-mouse IL-4 (clone 11B11), APC anti-mouse TNF-α (clone 11B11), Alexa Fluor 647 anti-mouse FOXP3 (clone MF-14), PE Annexin V (catalog no. 640947), 7-AAD Viability Staining Solution (catalog no. 420404), PE donkey anti-rabbit IgG (clone Poly4064), APC/Cy7 anti-human CD4 (clone RPA-T4), PE/Cy7 anti-human CD25 (clone M-A251), PE/Cy5 anti-human CD45RA (clone HI100), FITC anti-human CD45RO (clone UCHL1), FITC anti-human IL-17A (clone BL168), Fixation Buffer (catalog no. 420801), and Intracellular Staining Perm Wash Buffer (catalog no. 421002). The PE/Cy7 anti-mouse RORγt (clone B2D) antibody and Foxp3/Transcription Factor Staining Buffer Kit were purchased from eBioscience (Thermo Fisher Scientific). For surface marker analysis, cells were stained in phosphate-buffered saline (PBS) containing 2% fetal bovine serum (FBS) on ice for 30 min. For intracellular cytokine analysis, cells were stimulated in vitro with phorbol 12,13-dibutyrate (PDBu; catalog no. P1269, Sigma-Aldrich) and ionomycin (catalog no. I8800, Solarbio) in the presence of brefeldin A (catalog no. 420601, BioLegend) and monensin (catalog no. 420701, BioLegend). Stimulated cells were fixed and permeabilized using a Fixation/Permeabilization Solution Kit (BioLegend). For TF analysis, cells were fixed and permeabilized using the Foxp3/Transcription Factor Staining Buffer Kit (eBioscience). All samples were analyzed on a CytoFLEX flow cytometer (Beckman Coulter). Flow cytometry data were analyzed using FlowJo (Tree Star) and CytExpert (Beckman Coulter). Cells were sorted on a FACSAria III cell sorter (BD Biosciences). Induction and assessment of EAE EAE was induced by subcutaneously immunizing mice at the flanks with 200 μg of the myelin oligodendrocyte glycoprotein (MOG35–55) peptide emulsified in FCA supplemented with Mycobacterium tuberculosis (400 μg/ml; catalog no. F5881, Sigma-Aldrich). On days 0 and 2 after immunization, mice were intraperitoneally injected with 200 ng of pertussis toxin (catalog no. B7273, APExBIO). C57BL/6 mice were intraperitoneally injected with either saline (control) or punicalagin (4 mg/kg; catalog no. HY-N0063, MCE) on days 4, 7, and 10 postimmunization. EAE clinical symptoms were monitored daily on the basis of a standardized 0 to 5 scoring system as follows (16): 0, no symptoms; 0.5, partial limp tail; 1, completely paralyzed tail; 1.5, mild ataxia; 2, hindlimb paresis; 2.5, one hindlimb paralyzed; 3, both hindlimbs paralyzed; 3.5, both hindlimbs paralyzed with forelimb weakness; 4, one forelimb paralyzed; 4.5, both forelimbs paralyzed; and 5, moribund status or death. Construction and assessment of the CIA model WT and KO mice were initially immunized at the base of the tail with 100 μg of chick type II collagen (CII; catalog no. 20012, Chondrex) emulsified in an equal volume of FCA (catalog no. F5881, Sigma-Aldrich). After 21 days, a booster immunization was administered using the same emulsion as in the initial injection. Arthritis progression was monitored daily starting 2 weeks after the second immunization, using a previously described clinical scoring system: 0, normal; 1, slight swelling and/or erythema; 2, pronounced swelling; and 3, ankylosis. Each of the four paws was scored individually, and the total arthritis score was calculated as the sum of all four scores (62). H&E staining and histopathology Spinal cords from WT and KO EAE mice, as well as from small-molecule drug-treated WT EAE mice, and ankle joints from WT and KO CIA mice were harvested. Samples were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned at a thickness of 4 μm. Sections were stained with H&E and used to evaluate inflammatory cell infiltration. Photomicrographs were captured at 2× and 20× magnifications. Cytokine measurement Serum samples were collected from WT and KO mice with EAE. The concentrations of TNF-α, IFN-γ, IL-6, MCP-1 (monocyte chemoattractant protein-1), G-CSF (granulocyte colony-stimulating factor), IL-10, IL-17A, and GM-CSF (granulocyte-macrophage colony-stimulating factor) in the serum were measured using a Luminex assay kit (LX-MultiDTM-23, Labex) according to the manufacturer’s instructions. Human samples All patients and healthy donors in this study were enrolled from the Department of Rheumatology and Immunology at the First Affiliated Hospital of Xi’an Jiaotong University. A total of 11 patients with RA (all females; mean age: 50 ± 15.5 years), 8 patients with systemic lupus erythematosus (SLE; one male and seven females; mean age: 35.8 ± 8.2 years), and 15 healthy donors (nine males and six females; mean age: 35.1 ± 14.5 years) were enrolled in the study. SLE was diagnosed according to the 2012 Systemic Lupus International Collaborating Clinics (SLICC) criteria (63). RA was diagnosed on the basis of the 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) classification criteria (64). The study protocol was approved by the Ethics Committee of the First Affiliated Hospital of Xi’an Jiaotong University (no. KYLLSL-2023-092 for SLE; no. KYLLSL-2023-288 for RA). All participants provided informed consent for the use of their peripheral blood mononuclear cell (PBMC) samples. Mouse TH cell differentiation Naïve CD4+ T cells (CD4+CD25−CD44−CD62L+) were isolated from spleens and lymph nodes of WT and Dhx9 KO mice using a FACSAria III cell sorter (BD Biosciences). Purified cells were stimulated with plate-bound anti-mouse CD3 antibody (2.5 μg/ml; clone 145-2c11, BioLegend) and soluble anti-mouse CD28 antibody (0.5 μg/ml; clone 37.51, BioLegend) in the presence of cytokine cocktails for lineage-specific differentiation as follows: TH0: mIL-2 (200 U/ml; catalog no. 575406, BioLegend); TH1: mIL-12 (20 ng/ml; catalog no. 577002, BioLegend), mIL-2 (200 U/ml), and anti-mouse IL-4 antibody (10 μg/ml; clone 11B11, BioLegend); TH2: mIL-4 (100 ng/ml; catalog no. 715004, BioLegend), mIL-2 (50 U/ml), anti-mouse IL-12/IL-23 p40 antibody (10 μg/ml; clone C17.8, BioLegend), and anti-mouse IFN-γ antibody (10 μg/ml; clone XMG1.2, BioLegend); TH17: mTGF-β1 (0.5 ng/ml; catalog no. HY-P7117, MCE) and mIL-6 (50 ng/ml; catalog no. 575704, BioLegend); iTreg: mTGF-β1 (1 ng/ml) and mIL-2 (50 U/ml). In some experiments, small-molecule drugs (MCE) were added at the culture. Cells were harvested on day 1 or 2 for qPCR, RNA-seq, CUT&Tag-seq, and ATAC-seq analyses and on day 3 for fluorescence-activated cell sorting (FACS) analysis. Human TH17 differentiation Human PBMCs were isolated from the peripheral blood of healthy donors by density gradient centrifugation using Ficoll-Paque PLUS (catalog no. 7111011, DAKEWE). Naïve CD4+ T cells (CD4+CD25−CD45RA+CD45RO−) were sorted using a FACSAria III cell sorter (BD Biosciences) and stimulated with plate-bound anti-human CD3 antibody (5 μg/ml; clone OKT3, BioLegend) and soluble anti-human CD28 antibody (2 μg/ml; clone CD28.2, BioLegend) under TH17-polarizing conditions, including hTGF-β1 (10 ng/ml; catalog no. 781802, BioLegend), hIL-6 (30 ng/ml; catalog no. 570802, BioLegend), hIL-1β (100 ng/ml; catalog no. 579402, BioLegend), hIL-23 (20 ng/ml; catalog no. 574102, BioLegend), anti-human IFN-γ antibody (10 μg/ml; clone B27, BioLegend), and anti-human IL-4 antibody (10 μg/ml; clone MP4-25D2, BioLegend). DMSO or punicalagin (10 μM, MCE) were added at the start of the culture. After 6 days of culture, cells were subjected to intracellular staining and FACS analysis. RNA isolation and qPCR Total RNA was extracted from purified T cells using the Quick-RNA Microprep Kit (catalog no. R1051, Zymo Research) according to the manufacturer’s instructions. cDNA synthesis was performed using a cDNA synthesis kit (catalog no. FSQ-301, TOYOBO). qPCR was carried out with the SYBR Green RT-qPCR Master Mix (GenStar) on a StepOnePlus Real-Time PCR System (Thermo Fisher Scientific). Relative gene expression levels were normalized to β-actin. Primer sequences used in this study were as follows: Actb forward primer, 5′-CATTGCTGACAGGATGCAGAAGG-3′; Actb reverse primer, 5′-TGCTGGAAGGTGGACAGTGAGG-3′; Dhx9 forward primer, 5′-AAACTCCCCATTGAACCTCG-3′; Dhx9 reverse primer, 5′-TGTATCCCAGGCGTTTTCC-3′; Il17a forward primer, 5′-CAGACTACCTCAACCGTTCCAC-3′; Il17a reverse primer, 5′-TCCAGCTTTCCCTCCGCATTGA-3′; Rorc forward primer, 5′-GTG-GAGTTTGCCAAGCGGCTTT-3′; Rorc reverse primer, 5′-CCTGCACATTCTGACTAGGACG-3′. Western blotting In vitro–cultured TH1, Treg, or TH17 cells were lysed in radioimmunoprecipitation assay (RIPA) lysis buffer (catalog no. P0013B, Beyotime Biotechnology) containing phenylmethylsulfonyl fluoride (PMSF; catalog no. P0100, Solarbio) and phosphatase inhibitor cocktail A (catalog no. P1081, Beyotime Biotechnology) for total protein extraction. Proteins were separated by SDS–polyacrylamide gel electrophoresis (PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore). After blocking with 5% skimmed milk, the membranes were incubated overnight at 4°C with the indicated primary antibodies, including anti-Caspase-3 antibody (catalog no. 9664, Cell Signaling Technology), anti-Bcl-2 antibody (catalog no. sc-7382, Santa Cruz Biotechnology), anti-p-ERK1/2 antibody (catalog no. 4370, Cell Signaling Technology), anti-ERK1/2 antibody (catalog no. sc-135900, Santa Cruz Biotechnology), anti-p-Zap70 antibody (catalog no. 2717, Cell Signaling Technology), anti-Zap70 antibody (catalog no. 3165, Cell Signaling Technology), anti-p-SMAD2 antibody (catalog no. 3108, Cell Signaling Technology), anti-SMAD2 antibody (catalog no. ab33875, Abcam), anti-p-SMAD3 antibody (catalog no. ab52903, Abcam), anti-SMAD3 antibody (catalog no. sc-101154, Santa Cruz Biotechnology), anti-p-STAT3 antibody (catalog no. 9134, Cell Signaling Technology), anti-Dhx9 antibody (catalog no. ab26271, Abcam), and anti-β-Actin antibody (catalog no. 60008, Proteintech). After washing with Tris-buffered saline containing Tween 20, the membranes were incubated with the according secondary antibodies at room temperature for 1 hour. Protein expression was detected using the Tanon 5200 chemiluminescent imaging system (Tanon Science). Retroviral transduction For retroviral transduction, 2.4 μg of each retroviral vector along with 0.6 μg of the pCL-Eco retrovirus packaging vector were cotransfected into BOSC cells using jetPRIME reagent (Polyplus). After 48 and 72 hours, retroviral supernatants were collected through 0.45-μm filters and supplemented with polybrene (8 μg/ml). Naive CD4+ T cells (CD4+CD25−CD44−CD62L+) were activated with plate-bound anti-CD3 (2.5 μg/ml) and soluble anti-CD28 (0.15 μg/ml) antibodies for 16 hours. Activated cells were then spin infected with the retroviral supernatant at 2500 rpm for 90 min at 37°C. After centrifugation, supernatants were replaced with culture medium containing 10% FBS supplemented with TGF-β1 (0.5 ng/ml) and IL-6 (5 ng/ml). GFP+ cells were analyzed by FACS 48 hours after transfection. IP and mass spectrometry The cells from TH17 culture (1 × 107) were lysed in RIPA lysis buffer (catalog no. P0013D, Beyotime Biotechnology) supplemented with 1 mM PMSF on ice for 30 min. Anti-Dhx9 antibody or normal mouse IgG (Millipore) was added to the lysates and incubated with rotation overnight at 4°C. The samples were then incubated with magnetic protein A beads (catalog no. 1001D, Thermo Fisher Scientific) at 4°C for 4 hours. After several washes with Tris-NaCl buffer [20 mM Tris (pH 7.4) and 120 mM NaCl], the immunoprecipitated proteins were eluted from the beads using lysis buffer containing 4× loading buffer at 95°C for 10 min. The proteins were separated by SDS-PAGE and visualized using Coomassie staining and subsequently subjected to mass spectrometry analysis. Bulk RNA-seq library preparation and sequencing Naive CD4+ T cells (CD4+CD25−CD44−CD62L+) were sorted from the spleens and lymph nodes of WT and Dhx9 KO mice using a FACSAria III cell sorter (BD Biosciences) and subsequently cultured in the presence of IL-6 and TGF-β1 for 24 hours. Total RNA was extracted using the Quick-RNA Microprep Kit (catalog no. R1051, Zymo Research) and quantified with a Qubit Fluorometer (Invitrogen). After quality assessment by RNase-free agarose gel electrophoresis and the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA), RNA-seq libraries were prepared from 200 ng of total RNA using the TruSeq RNA Sample Preparation Kit (Illumina). Poly(A)+ mRNA was isolated using oligo(dT) beads and fragmented into smaller fragments. First-strand cDNA synthesis was performed using random hexamer primers, followed by second-strand synthesis to generate double-stranded cDNA. After a series of terminal repairs steps, the double-stranded cDNA library was generated by PCR enrichment and size selection. The prepared cDNA libraries were then sequenced on the Illumina NovaSeq 6000 platform after quality control. Three biological replicates were included in the RNA-seq analysis. Raw reads were aligned to the mouse genome (GRCm38) using the Tophat2 alignment tool. Unique reads were retained to quantify gene expression counts based on the Tophat2 alignment files. Expression analysis of TH17 lineage-associated genes was performed using the R package DESeq2, and statistical significance was determined by the Wald test. Scatterplots were generated using the R package. ScRNA-seq library preparation and sequencing Spinal cords from WT or Dhx9 KO EAE mice (n = 3) were dissociated into single-cell suspensions as follows: Tissues were collected, minced with a sharp razor blade, and digested with collagenase IV (1 mg/ml; Worthington Biochemical) at 37°C for 45 min. Single cells were obtained by filtering the dissociated tissue through a 70-μm cell strainer, followed by density gradient centrifugation using a 40/80% Percoll solution (Solarbio). Isolated cells were stained with anti-CD45.2 antibody at 4°C for 30 min before flow cytometry sorting using a FACSAria III cell sorter (BD Biosciences). Cells from each group were pooled to generate one representative WT sample and one representative KO sample. Sorted CD45+ cells with >90% viability were used for scRNA-seq using 10x Genomics platform. scRNA-seq libraries were prepared following the standard protocol of the 10x Chromium Single Cell 3′ v3 Reagent Kits. Raw sequencing data were processed using Cell Ranger (v5.0.1) to generate gene expression matrices. Specifically, sequencing reads were aligned to the mouse reference genome (mm10), and unique molecular identifier (UMI) matrices were generated for each cell using the Cell Ranger count pipeline. ScRNA-seq data analysis Gene expression matrices from all samples were imported into Seurat (version 4.0.0) (65) and merged for subsequent analyses. Low-quality cells were excluded on the basis of the following criteria: Cells with fewer than 200 or more than 6000 detected genes, and those with mitochondrial gene content greater than 10%, were removed. In total, 19,781 cells (WT = 10,687; KO = 9084), with a median of 1868 genes per cell, were included in the analysis. To visualize clustering, nonlinear dimensionality reduction was performed using the Uniform Manifold Approximation and Projection (UMAP), and cluster-specific marker genes were identified using the FindAllMarkers function in the Seurat package. To further dissect the heterogeneity within the T cell compartment, all T cells were extracted and reclustered into nine subtypes using the standard procedure. As a result, a total of 2177 WT T cells and 1157 KO T cells were included in the analysis. In downstream analyses, each cell was considered an individual observational unit (i.e., one cell = one sample). Differentially expressed genes were assessed using the Wilcoxon test implemented in the Seurat package. ATAC-seq library preparation and sequencing ATAC-seq was performed as previously described (66). Briefly, naive CD4+ T cells (CD4+CD25−CD44−CD62L+), sorted from WT and Dhx9 KO mice, were cultured with IL-6 and TGF-β1 for 48 hours. Purified cells were washed with ice-cold PBS and resuspended in cold ATAC-Resuspension Buffer (RSB) [10 mM tris-HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl2, 0.1% NP-40, 0.1% Tween 20, and 0.01% digitonin] for nuclear isolation. Cell lysis was performed on ice for 3 min, followed by immediate quenching with 2 volumes of wash buffer (RSB without detergents). The nuclei suspensions were incubated in a transposition reaction mixture containing 2.5 μl of Tn5 transposase at 37°C for 30 min. Immediately after transposition, DNA fragments were purified, amplified, and sequenced using the DNBSEQ-T7 platform. After removing adapter sequences and low-quality bases, cleaned paired-end ATAC-seq reads were aligned to the mm10 reference genome (GRCm38) using Bowtie2 (v2.2.9). Poorly mapped, nonuniquely mapped, and mitochondrial reads were removed using Samtools. PCR duplicates were removed using picard MarkDuplicates. Only concordantly mapped reads were retained for downstream analysis. Two biological replicates were included in the ATAC-seq analysis, and differential chromatin accessibility analysis was performed using DESeq2, with statistical significance determined via the Wald test. High-throughput CUT&Tag assay CUT&Tag libraries were prepared using the Hyperactive Universal CUT&Tag Assay Kit (catalog no. TD904, Vazyme) following the manufacturer’s instructions with minor modifications. Nuclei, purified from WT and Dhx9 KO CD4+ T cells cultured for 48 hours in the presence of IL-6 and TGF-β1, as previously described (66), were incubated with concanavalin A (Con A)–coated magnetic beads at room temperature for 10 min. Bead-bound nuclei were then incubated overnight at 4°C on a rotator with the indicated primary antibodies, including Dhx9 (catalog no. ab26271, Abcam), SMAD2 (catalog no. ab33875, Abcam), SMAD3 (catalog no. F0364, Selleck Chemicals), p-STAT3 (catalog no. 9134, Cell Signaling Technology), IRF4 (catalog no. 15106, Cell Signaling Technology), or control IgG (catalog no. AC005, Abclonal). The beads were subsequently incubated with secondary antibodies (1:100 dilution) at room temperature for 60 min. Following incubation, the beads were washed three times with Dig-wash buffer [20 mM Hepes (pH 7.5), 150 mM NaCl, 0.5 mM spermidine, and 0.05% digitonin] and subsequently incubated with protein A–Tn5 transposome at room temperature for 1 hour. Afterward, the beads were washed with Dig-med buffer [20 mM Hepes (pH 7.5), 300 mM NaCl, 0.5 mM spermidine, 0.05% digitonin, and 1× protease inhibitor cocktail] and resuspended in TTBL (Tagment Buffer L) at 37°C for 30 min to facilitate DNA fragmentation. The fragmented DNA was then purified, amplified, and sequenced using the DNBSEQ-T7 platform. Two biological replicates were included per experimental group in CUT&Tag analysis. Peaks were called using the MACS2 package with the “callpeak” function, using a Poisson distribution–based statistical model to assess significance and identify enriched regions relative to background. Luciferase assay DNA fragments were cloned into the pGL4.16 (luc2CP/Hygro) vector (Promega) containing the luciferase reporter gene luc2CP. A mixture of the pGL4.16 plasmid, the control vector pGL4.74 (hRluc/TK) encoding Renilla luciferase reporter gene hRluc, and candidate gene-expressing plasmids was transfected into 293T or Jurkat cells. Forty-eight hours posttransfection, cells were lysed, and the activities of both firefly and Renilla luciferases were measured using the Dual-Luciferase Reporter Assay Kit (Promega) on a Synergy Neo2 multimode reader (BioTek). High-throughput virtual screening To identify potential Dhx9 inhibitors, structure-based high-throughput virtual screening (HTVS) was performed using Schrödinger Maestro (version 11.4). The crystal structure of human Dhx9 [Protein Data Bank (PDB) ID: 3LLM] was retrieved from the RCSB Protein Data Bank and prepared using the Protein Preparation Wizard module of Schrödinger. Energy optimization of the MCE Bioactive Compound Library Plus, containing 14,800 compounds, was performed using the LigPrep module of Schrödinger, and the generated three-dimensional (3D) structures were for virtual screening. These optimized compounds were subsequently subjected to a multistep virtual screening workflow for molecular docking. Initially, all compounds were docked using Glide HTVS mode, and the top 30% highest-scoring compounds were further screened using Glide Standard Precision (SP) mode. The top 30% of compounds from Glide SP were subjected to Glide Extra Precision (XP). The top 40 hit compounds, ranked by absolute docking scores, were selected as candidates for subsequent experimental validation. Statistics Statistical analyses were performed using two-tailed Student’s t test, one-way analysis of variance (ANOVA), or two-way ANOVA in GraphPad Prism 9 software (GraphPad Software, San Diego, CA, USA). All bar graphs represent means and error bars to indicate data variability. Statistical significance is denoted as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P ≤ 0.0001.