Intestinal microbiota-derived short-chain fatty acids regulation of immune cell IL-22 production and gut immunity

Innate lymphoid cells (ILCs) and CD4+ T cells produce IL-22, which is critical for intestinal immunity. The microbiota is central to IL-22 production in the intestines; however, the factors that regulate IL-22 production by CD4+ T cells and ILCs are not clear. Here, we show that microbiota-derived short-chain fatty acids (SCFAs) promote IL-22 production by CD4+ T cells and ILCs through G-protein receptor 41 (GPR41) and inhibiting histone deacetylase (HDAC). SCFAs upregulate IL-22 production by promoting aryl hydrocarbon receptor (AhR) and hypoxia-inducible factor 1α (HIF1α) expression, which are differentially regulated by mTOR and Stat3. HIF1α binds directly to the Il22 promoter, and SCFAs increase HIF1α binding to the Il22 promoter through histone modification. SCFA supplementation enhances IL-22 production, which protects intestines from inflammation. SCFAs promote human CD4+ T cell IL-22 production. These findings establish the roles of SCFAs in inducing IL-22 production in CD4+ T cells and ILCs to maintain intestinal homeostasis.

I nterleukin 22 , a member of the IL-10 family, was initially characterized as a Th1 cytokine 1 , and later as a Th17 as well as Th22 cytokine 2,3 , which is central to host protection against inflammatory insult in the intestine by inducing antimicrobial peptides and promoting epithelial barrier function 4,5 . Both innate lymphoid cells (ILCs) and CD4 + T cells produce IL-22. Although IL-22-producing ILCs express the transcription factors aryl hydrocarbon receptor (AhR) and the retinoid-related orphan receptor gamma t (RORγt), and require IL-23 stimulation to produce IL-22, a recent study demonstrated that CD4 + T cells IL-22 production depends on AhR and T-bet, but not primarily IL-23, for IL-22 production 6 . IL-22 receptor complex, a heterodimer of IL-10R2 and IL-22R1, is confined to non-hematopoietic cells, and thus the IL-22-IL-22R axis provides a critical link in the integration of immune responses with barrier function at the mucosal surface 5,7 . Accumulating evidence indicates an important role of IL-22 in inflammatory bowel disease (IBD), in that the majority of IL-22-associated molecules are encoded by IBD susceptibility genes 8 . IL-22 produced by both innate and adaptive lymphocytes is indispensable for maintaining intestine homeostasis 9 . Although ILCs provide a rapid source of IL-22 that is essential for early protection of epithelial barrier function upon inflammatory insult, CD4 + T cells become the dominant source of IL-22 in the intestinal lamina propria (LP) during chronic intestinal inflammation. However, the factors that regulate IL-22 production in CD4 + T cells and ILCs in the intestines and the mechanisms involved are still not completely understood.
In this report, we demonstrate that SCFAs promote IL-22 production in CD4 + T cells and ILCs through histone deacetylase (HDAC) inhibition and GPR41, but not GPR43 and GPR109a. SCFAs upregulate IL-22 production through promoting AhR and hypoxia-inducible factor (HIF)1α expression, which are differentially regulated by mTOR and Stat3. HIF1α directly binds to the Il22 promoter, and SCFAs increase the accessibility of HIF1αbinding sites in the Il22 promoter through histone modification. Furthermore, SCFA supplementation in vivo protects mice from intestinal inflammation upon Citrobacter rodentium infection and inflammatory insult, which is mediated by enhanced IL-22 production.

SCFAs promote IL-22 in CD4 + T cells and ILCs in vitro.
SCFAs have been shown to promote regulatory T cell (Treg) development as well as CD4 + T cell IL-10 production 20,26,28 . To explore how SCFAs regulate CD4 + T cells more comprehensively, splenic CD4 + T cells were isolated from wild-type (WT) C57BL/ 6J (B6) mice and activated with anti-CD3 mAb and anti-CD28 mAb in the presence or absence of butyrate, one of the major SCFAs in the intestines, for 2 days. The RNA transcriptome was analyzed by RNA sequencing (RNA-seq). Principal component analysis (PCA) and volcano plot analysis demonstrated butyrate treatment led to a different transcriptional profile ( Supplementary  Fig. 1a, b). Consistent with previous studies 20, 28 , butyrate promoted expression of Il10 and Ifng by CD4 + T cells (Fig. 1a). Interestingly, Il22 was significantly increased in butyrate-treated CD4 + T cells (Fig. 1a). To confirm SCFAs induction of IL-22 in gut microbiota antigen-specific T cells, we cultured splenic CD4 + T cells of CBir1 TCR transgenic (CBir1 Tg) mice, which are specific for an immunodominant microbiota antigen CBir1 flagellin 29 , with antigen-presenting cells (APCs) and CBir1 peptide in the presence or absence of acetate, propionate, or butyrate, the three major SCFAs, for 2 days. Acetate, propionate, and butyrate all increased IL-22 production at both mRNA and protein (Fig. 1b, c). We also confirmed that acetate, propionate, and butyrate increased Il22 expression in WT B6 CD4 + T cells activated with anti-CD3 mAb and anti-CD28 mAb ( Supplementary Fig. 1c, d).
IL-22 was initially characterized as a Th1 cytokine, and later as a Th17 as well as Th22 cytokine 30 . To determine whether SCFAs promote IL-22 production in different T cell subtypes, CBir1 CD4 + T cells were cultured under neutral, Th1, Th17, and Treg polarization conditions, in the presence or absence of butyrate, for 2 days. Butyrate induced Il22 expression in CD4 + T cells under all conditions (Fig. 1d), although Treg cells expressed the lowest Il22 levels without butyrate stimulation among all CD4 + T cell subtypes tested, which is consistent with the previous reports 31 . Interestingly, Il22 expression in Th1 cells was higher even than those under Th17 conditions (Fig. 1d), which is likely due to TGFβ inhibition of IL-22 expression under Th17 conditions 32,33 , and higher Tbx21 expression in Th1 cells ( Supplementary Fig. 1e), which is critical for inducing IL-22 in CD4 + T cells 6,34 . We then performed RNA-seq to determine transcriptional profiles in CD4 + T cells under Th1 conditions with or without butyrate treatment. Treatment with butyrate changed the transcription landscape of CD4 + T cells under Th1 conditions ( Supplementary Fig. 1f, g). Consistently, Il22, in addition to Il10 and Ifng, was increased after the treatment of butyrate (Fig. 1e). We then measured Il22 expression at different time points in CD4 + T cell cultures with or without butyrate under Th1 conditions. Butyrate induced Il22 expression as early as at 24 h and reached peak mRNA level at 60 h (Fig. 1f), indicating that butyrate promotes Il22 expression in a timedependent manner. Consistently, butyrate increased IL-22 protein in CD4 + T cells under Th1 conditions (Fig. 1g, h). However, butyrate did not affect IFN-γ production in CD4 + T cells ( Fig. 1h and Supplementary Fig. 2). We obtained similar results in CD4 + T cells cultured under Th17 conditions, in that butyrate promoted CD4 + T cell production of IL-22 (Supplementary Fig. 1h, i). Butyrate also promoted Treg but inhibited Th17 cell differentiation ( Supplementary Fig. 2). We then investigated whether SCFAs regulate IL-22 production in CD4 + T cells from spleen and MLN differently. Similar to splenic CD4 + T cells, butyrate treatment of MLN CD4 + T cells enhanced IL-22 production ( Supplementary Fig. 3a). Furthermore, splenic and MLN CD4 + T cells proliferated (Supplementary Fig. 3b) and polarized at similar levels under Th1, Th17, and Treg conditions ( Supplementary Fig. 3c).
Given that gut microbiota is required for IL-22 production in ILCs 12 , we next investigated whether SCFAs regulate IL-22 in ILCs. CD4 + T cell-depleted splenic cells were stimulated with IL-23 in the presence or absence of acetate, propionate, and butyrate overnight. Acetate, propionate, and butyrate upregulated IL-22 production, but not IL-17, in ILCs when cells were treated with IL-23 (Fig. 1i). However, SCFA treatment alone did not affect IL-22 production in ILCs without IL-23 stimulation ( Supplementary  Fig. 4). Consistently, butyrate upregulated ILC production of IL-22 in LP ( Supplementary Fig. 3d). Given IL-22 can be produced by NKT cells 35 , we also determined whether butyrate affect NKT cell production of IL-22. However, butyrate did not affect IL-22 in NKT cells (Supplementary Fig. 5a).
Butyrate promotes IL-22 production by CD4 + T cells and ILCs in vivo. To investigate whether SCFAs upregulate IL-22 production in vivo, WT B6 mice were administrated with or without 200 mM butyrate in drinking water for 3 weeks. Butyrate supplementation did not affect weight gain (Fig. 2a). Butyrate levels in the colon were increased after administration of butyrate (Fig. 2b). When mice were killed on day 21, butyrate upregulated IL-22 production in the serum and the colonic organ culture (Fig. 2c, d), and increased IL-22 production in CD4 + T cells in the spleen, MLN, and LP compared with control mice (Fig. 2e). Some CD4 − cells   Fig. 6). Butyrate also promoted ILC production of IL-22 in the spleen, MLN, and LP in vivo ( Fig. 2g and Supplementary Fig. 7a To investigate whether inhibition of HDAC also contributes to SCFA induction of CD4 + T cell IL-22 production, we first determined whether butyrate at the dose of 0.5 mM could suppress HDAC activity. We treated CD4 + T cells cultured under Th1 condition with or without butyrate for 24 h. Butyrate indeed inhibited HDAC activity in CD4 + T cells (Fig. 3d). We then treated CD4 + T cells cultured under Th1 or Th17 conditions with the HDAC inhibitor, Trichostatin A (TSA), to determine whether it mimics the effect of butyrate on the induction of IL-22. Treatment with TSA increased IL-22 production at both mRNA and protein levels in CD4 + T cells under Th1 (Fig. 3a-c) and Th17 ( Supplementary Fig. 8i-j) conditions. Furthermore, treatment with HDAC inhibitor and GPR41 agonist together further promoted IL-22 production in CD4 + T cells to levels similar to that induced by butyrate (Fig. 3a-c). Butyrate induced similar levels of IL-22 in WT, Gpr43 −/− , and Gpr109 −/− ILCs ( Supplementary Fig. 9a, b). Similar to our findings in CD4 + T cells, treatment with AR420626 or TSA promoted IL-22 production in ILCs ( Supplementary Fig. 9c). Taken together, these data indicated that butyrate promotes IL-22 production through GPR41 and inhibiting HDAC in CD4 + T cells and ILCs.
Next, we investigated whether butyrate induced HIF1α and AhR through GPR41 or HDAC inhibition. Treatment with GPR41 agonist, but not HDAC inhibitor, upregulated Hif1a and Ahr expression in CD4 + T cells (Fig. 4l, m), indicating GPR41 mediates butyrate induction of HIF1α and AhR expression in CD4 + T cells.
We have previously shown that Blimp1-mediated SCFAinduced IL-10 production in CD4 + T cells 20,23 . We then explored whether Blimp1 also regulates butyrate-induced IL-22 production in CD4 + T cells. Butyrate increased Il22 expression in Blimp1deficient CD4 + T cells from Cd4 cre Prdm1 fl/fl mice to the levels  Stat3 and mTOR are involved in butyrate induction of IL-22. Stat3 and mTOR have been implicated in the regulation of HIF1α/AhR expression [43][44][45][46] , and SCFAs activated Stat3 and mTOR 20,21 . To explore whether butyrate increases HIF1α and AhR expression in CD4 + T cells through activation of Stat3 and mTOR, we first analyzed phosphorylation levels of Stat3 and  (Fig. 5a, b), whereas phosphorylated mTOR was increased 24 h after butyrate treatment (Fig. 5c, d). mTOR activation was further confirmed by increased phosphorylated S6 ribosomal protein levels (Fig. 5e) (Fig. 5f, g). Interestingly, while inhibition of Stat3 decreased both Hif1a and Ahr expression induced by butyrate, mTOR inhibitor only downregulated butyrate-induced expression of Ahr but not Hif1a (Fig. 5h, i), suggesting Stat3 and mTOR activation differentially regulates HIF1α and AhR to promote IL-22 production in CD4 + T cells. Furthermore, while HJC0152 did not affect mTOR activation ( Supplementary  Fig. 13a), rapamycin moderately suppressed phosphorylated Stat3 induction by butyrate ( Supplementary Fig. 13b) in CD4 + T cells. Addition of mTOR inhibitor further reduced butyrate-induced IL-22 production suppressed by Stat3 inhibitor in CD4 + T cells ( Supplementary Fig. 13c). The role of Stat3 in butyrate-induced IL-22 production under Th1 conditions was further confirmed in Stat3-deficient CD4 + T cells from Cd4 Cre Stat3 fl/fl mice, in that butyrate-induced IL-22 production was decreased in Stat3deficient CD4 + T cells compared to WT CD4 + T cells ( Fig. 5j and Supplementary Fig. 11e, f). In addition, we obtained similar results for the roles of Stat3 and mTOR in butyrate-induced IL-22 production and expression of HIF1α and AhR in CD4 + T cells cultured under Th17 conditions ( Supplementary Fig. 12g-l). Similarly, butyrate activated Stat3 and mTOR ( Supplementary  Fig. 9f, g), and inhibition of Stat3 and mTOR suppressed their ability to produce IL-22 in ILCs ( Supplementary Fig. 9h).
Butyrate promotes HIF1α binding to the Il22 promoter. Upon hypoxia, HIF1α dimerizes with HIF1β, together with co-activators, to translocate into the nucleus to regulate target gene expression by binding to the hypoxia response element (HRE), NCGTG, in the promoters of the target genes. By retrieving genomics data in PubMed and Ensembl and using the consensus core (NCGTG), we found the putative HRE on Il22 promoter region (−2000 bp) (Fig. 6a). To confirm the direct binding of HIF1α to the Il22 promoter in CD4 + T cells, we performed a CHIP assay in CD4 + T cells under Th1 conditions. Compared to the control with anti-IgG antibody, qRT-PCR amplification of DNA that was immunoprecipitated with anti-HIF1α antibody resulted in specific enrichment of the HRE region in the Il22 promoter (Fig. 6b), suggesting HIF1α directly binds to the Il22 promoter. Next, we asked whether butyrate enhances the binding of HIF1α to the Il22 promoter. We treated CD4 + T cells with or without butyrate under Th1 conditions for 48 h. Butyrate enhanced HIF1α binding to the HRE region of the Il22 promoter (Fig. 6c).
Butyrate induces histone acetylation of HRE on Il22 promoter.
Histone modifications, such as acetylation and methylation, are associated with gene transcription through regulation of accessibility to the target DNA. Lysine 9 (K9) acetylation on histone H3 (H3K9ac) and K4 trimethylation on histone H3 (H3K4me3), which are co-localized in gene promoters, are the active and repressive markers of transcription, respectively 48 . Given that butyrate is an effective HDAC inhibitor, we investigated whether butyrate suppresses HDAC to facilitate HIF1α binding to the Il22 promoter. We performed a CHIP assay using antibodies against H3K9ac and H3K9me3 to determine the accessibility of HIF1αbinding sites in the Il22 promoter in CD4 + T cells cultured with or without butyrate under Th1 conditions. Butyrate increased the H3K9 acetylation, but suppressed the trimethylation of H3K9 in HRE sites of the Il22 promoter (Fig. 6d, e), indicating butyrate increases the accessibility of HIF1α-binding sites in the Il22 promoter through histone modification.
Butyrate inhibits colitis through promoting IL-22. IL-22 is a crucial cytokine in host defense against enteric infection of Citrobacter rodentium (C. rodentium), which is similar to human enteropathogenic Escherichia coli (EPEC) associated with IBD 49 . We then investigated whether butyrate protects intestines from C. rodentium infection and the role of IL-22 invovled. We infected WT mice orally with C. rodentium (5 × 10 8 CFU/mice) on day 0 and administrated with or without butyrate in drinking water for 10 days. The mice treated with butyrate exhibited less weight loss (Fig. 7a), and decreased intestinal inflammation (Fig. 7b), with higher IL-22 and IL-10 production in intestinal LP CD4 + T cells (Fig. 7c). IFN-γ + CD4 + T cells, but not IL-17 + CD4 + T cells, were decreased in the intestinal LP in mice treated with butyrate (Fig. 7c). In addition, butyrate supplementation increased IL-22 production, but did not affect IL-17 production, in the intestinal ILCs (Fig. 7d). Administration of butyrate decreased the fecal C. rodentium CFU (Fig. 7e) and C. rodentium CFU in liver (Fig. 7f), indicating butyrate promotes C. rodentium clearance in the colon and decreases C. rodentium dissemination to the liver.
To determine whether ILCs, CD4 + T cells, or both ILCs and CD4 + T cells are required for the effects of butyrate on intestinal inflammation, we infected mice with C. rodentium (5 × 10 8 CFU/ mice) orally on day 0 followed by treatment with or without Fig. 4 HIF1α and AhR mediate butyrate induction of IL-22 in CD4 + T cells. a WT CD4 + T cells were activated with anti-CD3/CD28 mAbs under Th1 conditions ± butyrate (0.5 mM) for 2 days (n = 3 biologically independent samples per group). RNA sequencing was performed. Hif1α, Ahr, and Prdm1 were shown in heatmap. b-f CD4 + T cells were activated with anti-CD3/CD28 mAbs ± butyrate (0.5 mM) under Th1 conditions (n = 3/group). Hif1a (b) and Ahr (c) were analyzed by qRT-PCR. HIF1α (d) and AhR (e) protein was analyzed by western blot on day 2. HIF1α activity was measured using HIF1α Transcription Factor Assay Kit (f). g Raw 264.7 cells were transduced with XRE/AhR Luciferase Reporter Gene Lentivirus, and treated ± butyrate (0.5 mM) 3 days post transduction. AhR activity was assessed by luciferase. h-j Cbir1 Tg CD4 + T cells were activated with APCs and Cbir1 peptide under Th1 conditions with butyrate (0.5 mM) ± YC-1 (5 µM) or/and CH-223191 (3 µM) for 60 h (n = 3/group). IL-22 mRNA (h) and protein (i) were measured by qRT-PCR and ELISA. j IL-22 was measured by flow cytometry on day 5. k WT and HIF1α −/− CD4 + T cells were activated with anti-CD3/CD28 mAbs ± butyrate (0.5 mM) for 5 days (n = 3/group). IL-22 was assessed by flow cytometry. l, m CD4 + T cells were activated with anti-CD3/CD28 mAbs under Th1 conditions with or without butyrate (0.5 mM), AR420626 (5 µM), or TSA (10 nM) for 60 h (n = 3/group). Hif1a (l) and Ahr (m) were measured by qRT-PCR. One representative of three independent experiments was shown (b-m). Data were expressed as mean ± SD. Statistical significance was tested by two-tailed unpaired Student t-test (b-g) or two-tailed one-way ANOVA (h-m). b **p = 0.0033 (24 h butyrate in drinking water for 10 days. Groups of the mice received either anti-CD4 mAb to deplete CD4 + T cells, anti-Thy1 mAb to deplete CD4 + T cells and ILCs, or control IgG. Depletion of CD4 + T cells led to more severe colitis, and depletion of both CD4 + T cells and ILCs further aggravated the severity of colitis (Fig. 7g), indicating that both CD4 + T cells and ILCs are critical in protecting the intestines against C. rodentium infection. Feeding butyrate provided partial protection against C. rodentium infection in CD4 + T cell-depleted mice (Fig. 7g). However, depletion of both CD4 + T cells and ILCs completely abrogated the protective effects of butyrate (Fig. 7g). Consistent with the severity of the disease, administration of butyrate decreased fecal CFU in CD4 + T cell-depleted mice to a lesser degree than in control mice, but was unable to affect fecal CFU in mice depleted of both CD4 + T cells and ILCs (Fig. 7h). This indicates that both ILCs and CD4 + T cells are important in butyrate protection of the intestines from C. rodentium infection.
To investigate whether increased IL-22 mediates butyrate protection against C. rodentium infection, we infected WT and Il22 −/− mice orally with C. rodentium and treated with or without butyrate in drinking water. Il22 −/− mice suffered more weight loss compared with WT mice, and butyrate administration decreased weight loss in WT mice, but not in Il22 −/− mice (Fig. 8a). Il22 −/− mice showed more severe colitis compared with WT control mice. Feeding butyrate decreased colitis severity in WT but not in Il22 −/− mice (Fig. 8b). In addition, butyrate treatment decreased colonic IL-6 and TNF levels in WT mice but not in Il22 −/− mice (Fig. 8c, d). Fecal C. rodentium CFU was decreased in WT but not Il22 −/− mice treated with butyrate (Fig. 8e), indicating that butyrate promotes C. rodentium clearance in the gut in an IL-22-dependent manner. We also checked the C. rodentium dissemination in the liver. C. rodentium CFU in liver was significantly higher in Il22 −/− mice compared with WT mice. Butyrate treatment decreased CFU levels in the liver of WT mice but not Il22 −/− mice (Fig. 8f), suggesting butyrate limits C. rodentium dissemination from the intestine to the liver through upregulating IL-22 production.
To investigate whether butyrate-induced IL-22 regulates intestinal inflammation upon inflammatory insult, we assessed the role of IL-22 in butyrate inhibition of Dextran sulfate sodium (DSS)-induced colitis. Similar to the C. rodentium model, Il22 −/− control mice showed more weight loss and developed more severe colitis compared with WT control mice (Supplementary Fig. 14a, b). Butyrate-treated WT mice showed less weight loss compared with control WT mice, while there was no difference in weight loss between butyrate-treated IL-22 −/− mice and control Il22 −/− mice ( Supplementary Fig. 14a). Administration of butyrate alleviated intestinal inflammation in WT but not in Il22 −/− mice, characterized by less inflammation and lower histopathological scores ( Supplementary Fig. 14b), and decreased levels of IL-6 and TNFα in colonic organ culture ( Supplementary Fig. 14c, d). In addition, butyrate increased IL-22 + CD4 + T cells, but did not affect the percentage of IFN-γ + CD4 + T cells and IL-17 + CD4 + T cells, in the intestinal LP of DSS-treated WT mice (Supplementary Fig. 14e). Consistent with our previous study, butyrate promoted IL-10 production in LP CD4 + T cells of WT mice (Supplementary Fig. 14e). Furthermore, butyrate promoted IL-22, but not IL-17, production in LP ILCs (Supplementary Fig. 14f). Taken together, these results demonstrated that butyrate protects the intestines from inflammation induced by both enteric infection and intestinal injury through the upregulation of IL-22 production. Relative enrichment (Fold) Relative enrichment (Fold) Fig. 6 Butyrate promotes HIF1α binding to Il22 promoter in CD4 + T cells. a Schematic diagram of HIF1α binding to Il22 promoter. b WT CD4 + T cells were activated with anti-CD3/CD28 mAbs under Th1 conditions for 2 days (n = 3/group). HIF1α binding to Il22 promoter was analyzed by CHIP assay. c-e WT CD4 + T cells were cultured under Th1 conditions with or without butyrate (0.5 mM) for 2 days (n = 3/group). HIF1α binding to Il22 promoter was analyzed by CHIP assay (c). The H3K9 acetylation (d) and trimethylation (e) levels in HIF1α-binding site on Il22 promoter were assessed by CHIP assay. One representative of three independent experiments (b, c), or two independent experiments (d, e) was shown. Data were expressed as mean ± SD. Statistical significance was tested by two-tailed unpaired Student t-test. b **p = 0.0047; c *p = 0.0278; d *p = 0.0105; e **p = 0.0094.
Butyrate promotes human CD4 + T cell IL-22 production. To assess whether butyrate induces CD4 + T cell IL-22 production in IBD patients for translational potential, we treated peripheral blood CD4 + T cells, isolated from healthy volunteers, and patients with active Crohn's disease (CD) and ulcerative colitis (UC), with anti-CD3 mAb and anti-CD28 mAb with or without butyrate. Butyrate promoted Il22 mRNA levels in human CD4 + T cells, including healthy volunteers, CD, and UC patients (Fig. 9a). The percentage of IL-22 + CD4 + T cells and IL-22 production were increased after treatment with butyrate (Fig. 9b, c). Similar to the results from the mouse study, butyrate increased Hif1a and Ahr expression in human CD4 + T cells (Fig. 9d, e). Furthermore, inhibition of HIF1α and AhR using HIF1α inhibitor YC-1 and AhR inhibitor CH-223191 suppressed butyrate-induced    (Fig. 9f), indicating that butyrate promotes IL-22 production in human CD4 + T cells, including IBD patients, through regulation of HIF1α and AhR.

Discussion
Emerging evidence indicates that interaction between microbiota and IL-22 is central at barrier sites in the regulation of intestinal homeostasis. IL-22 regulates the gut microbiota composition through promoting intestinal barrier function by inducing epithelial cell production of antimicrobial peptides, mucins, and other beneficial effects. On the other hand, gut microbiota also regulates intestinal IL-22 production, in which the mechanisms are still not well-established. Many functions of gut microbiota in modulating health and disease are through their metabolites 50,51 .
Our current study demonstrated that SCFAs, the major microbiota metabolites from high fiber diet, promote CD4 + T cells and ILC IL-22 production through upregulating AhR and HIF1α, and, thus, provided novel insights into how gut microbiota, through their metabolites SCFAs, regulates IL-22 production to maintain the intestinal homeostasis. As IL-22 is critical in the regulation of intestinal barrier function and intestinal homeostasis, how its production is regulated has been intensively investigated. Interestingly, although both ILC3 and CD4 + T cells produce IL-22 under steady conditions, most insights into how IL-22 production is regulated in both cells are achieved either upon enteric infection or intestinal inflammation. Several proinflammatory cytokines have been identified critical in mediating IL-22 production, including IL-23, IL-6, and IL-1β 6,52,53 . As IL-22 production in the intestines depends on gut microbiota, this begs the questions on what products of gut microbiota function as inducers of IL-22 production. It has been shown that AhR ligands produced by specific gut bacteria species are able to stimulate ILC3 and CD4 + T cells Fig. 7 Butyrate protects the intestines from Citrobacter rodentium infection. a-f WT mice (n = 4 mice/group) were orally infected with Citrobacter rodentium (C. rodentium, 5 × 10 8 CFU/mice), and treated with or without butyrate (200 mM) in drinking water for 10 days. Mice were weighed daily (a), and killed on day 10. Colonic histopathology (b), LP IL-22 + , IL-10 + , IFN-γ + , and IL-17 + CD4 + T cells (c), and IL-22 + and IL-17 + ILCs (d) were measured. CFU in feces (e) and liver (f) were measured. g, h WT mice (n = 4 mice/group) were orally infected with C. rodentium (5 × 10 8 CFU/mice) on day 0, and with or without butyrate (200 mM) in drinking water for 10 days. Mice were administrated with anti-IgG antibody (25 mg/kg), anti-CD4 antibody (25 mg/kg), or anti-Thy1 (25 mg/kg) i.p. every other day. Mice were killed on day 10. Colonic histopathology was assessed (g), and CFU in feces was measured (h). One representative of three independent experiments (a-f) or two independent experiments (g, h) was shown. Scale bar, 300 µm (b, g). Data were expressed as mean ± SD. Statistical significance was tested by two-tailed unpaired Student t-test (a, c-f, h), the non-parametric twotailed Mann- Whitney U test (b, g) to produce IL-22 14,15 . Colonization of Clostridia, which produces SCFAs 18,19 , in antibiotic-treated neonatal mice induces IL-22 production in ILCs and CD4 + T cells 17 , which raised the question whether Clostridia promote IL-22 production through upregulating SCFAs production. SCFAs as major microbiota metabolites from high fiber diets are present in the intestinal lumen at high concentrations, which make them the potential major players in the induction of IL-22 production in the intestines. Our study demonstrated that SCFAs induce IL-22 production in CD4 + T cells and ILCs. Consistently, a recent report showed GPR43, one of the major receptors for SCFAs, promotes ILC3 development and function 27 , further supporting a role of the SCFAs in promoting IL-22 production in both ILCs and CD4 + T cells in the intestines to maintain the intestinal homeostasis. Our previous study showed butyrate promotes differentiated Th1 cell IL-10 production via Blimp1 and GPR43 pathways. Although we confirmed butyrate increased Blimp1 expression in CD4 + T cells, loss of Blimp1 and GPR43 did not affect the IL-22 expression induced by butyrate. Furthermore, we found GPR41 mediates butyrate induction of IL-22 production in CD4 + T cells and ILCs. Thus, butyrate regulates CD4 + T cells production of IL-10 and IL-22 through different mechanisms. Indeed, we found that butyrate-treated CD4 + T cells expressed higher levels of AhR expression, a master regulator of IL-22, and inhibition of AhR suppressed butyrate-induced IL-22. HIF1α, a subunit of HIF1 54 , has been shown to regulate the functions of different CD4 + T cells. Butyrate was reported to increase oxygen consumption to activate HIF1 in intestinal epithelial cells 55 . Interestingly, HIF1α was also increased in butyrate-treated CD4 + T cells and ILCs, and blockade of HIF1α both pharmacologically and genetically suppressed butyrate-induced IL-22, indicating a crucial role of HIF1α in mediating butyrate induction of IL-22. Furthermore, hypoxia promoted CD4 + T cell IL-22 production, and treatment with butyrate also promoted CD4 + T cell IL-22 expression under hypoxia condition. This result has a particular significance for SCFAs induction of IL-22 in the intestines in vivo. It has been shown that physiological hypoxia predominates in the normal intestinal mucosa, especially in the colon, and the inflammatory lesions in the inflamed intestines of the experimental colitis are profoundly hypoxic or even anoxic 56,57 . Our data that SCFAs promote CD4 + T cell IL-22 production under hypoxia condition suggest SCFAs as major inducers of CD4 + T cell IL-22 production in the intestines under both physiological and inflammatory conditions through induction HIF1α. HIF1α has been shown to promote Th17 cell IL-17 production through association with RORγt at the Il17 promoter 58 . Although butyrate treatment increases HIF1α in Th17 cells, it inhibits their IL-17 production, which is likely due to the decreased expression of Rorα and Rorγt as we recently reported 23 . Our study, thus, identifies HIF1α as a transcription factor for IL-22 production in CD4 + T cells and ILCs.
HIF1α modifies several target genes by binding to the HRE in their promoters 59 . In this study, we demonstrated that HIF1α directly binds to the Il22 promoter. More interestingly, butyrate promoted hypoxia binding to the Il22 promoter through increasing the H3K9 acetylation and suppressing the H3K9 trimethylation in HRE sites of IL-22 promoter. SCFAs have been shown as potent HDAC inhibitor, which modifies chromatin 37 . We found TSA, a HDAC inhibitor, upregulated IL-22 production in CD4 + T cells and ILCs, which mimics butyrate induction of IL-22 production, suggesting butyrate promotes IL-22 production at least partially through inhibition of HDAC.
Consistent with in vitro data, butyrate supplementation increased IL-22 production in intestinal LP CD4 + T cells and ILCs under both steady conditions and inflammatory conditions. Butyrate administration ameliorated C. rodentium infection severity, promoted C. rodentium clearance, and decreased C. rodentium dissemination from colon to liver in WT mice, but not Il22 −/− mice, suggesting an indispensable role of butyrateinduced IL-22 production in butyrate protective role for host defense. Consistent with the previous study 20 , butyrate administration protected mice from colitis in WT mice. However, butyrate supplementation did not protect Il22 −/− mice from DSSinduced colitis, although butyrate treatment in vivo increased both IL-22 and IL-10 production in mice, which suggests that butyrate induction of IL-10 alone is insufficient to decrease colitis severity in this model. Butyrate can modulate oxygen availability in the intestines, and the gut microbiota can shift drastically based on oxygen availability 60,61 . Previously, we showed that butyrate affects gut microbiota composition in mice 21 . In the current study, we found that the administration of butyrate increased butyrate levels in the colon. Thus, butyrate induction of IL-22 in vivo is likely due to the combination of altered gut microbiota and increased butyrate in the intestines. Interestingly, mice administrated with butyrate showed goblet cell hyperplasia, which might be due to higher IL-22 production that has been found to mediate goblet cell hyperplasia 62 .
Overall, we demonstrate gut microbiota-derived metabolites SCFAs promote CD4 + T cell and ILC production of IL-22 through GPR41 and HDAC inhibition. HIF1α and AhR are mediated in the butyrate induction of IL-22, which is differentially regulated by mTOR and Stat3 (Fig. 10), thus, providing a novel function of gut microbiota-derived metabolites in the regulation of intestinal homeostasis. Interestingly, SCFAs also promote human CD4 + T cell IL-22 production, even from IBD patients, our study provides SCFAs as a new potential therapeutic target for suppressing intestine infection and inflammation, and eventually treatment of IBD patients.

Methods
Mice. Wild-type (WT) C57BL/6J mice were purchased from Jackson Laboratory. Prdm1 fl/fl mice and Stat3 fl/fl mice were purchased from Jackson Laboratory, and bred to B6.Cd4 cre mice from Jackson Laboratory. Gpr43 −/− (Ffar2 tm1Lex ) mice were obtained from Bristol-Myers Squibb. Il22 −/− mice on the B6 background were obtained from Amegen (Thousand Oaks, CA). Gpr109a −/− mice were obtained from Dr. Nagendra Singh of the Augusta University. Cd4 cre Hif1α fl/fl mice were obtained from Dr. Fan Pan of Sidney Kimmel Comprehensive Cancer Center. CBir1 TCR transgenic (CBir1 Tg) mice were bred in the Animal Resource Center of University of the Texas Medical Branch (UTMB). All the mice were maintained on a 12 h-light/dark cycle and at the temperature of 20-26°C with 30-70% humidity in the specific pathogen-free animal facilities. The animal use and care were in accordance with institutional guidelines of UTMB, and all experiments were approved by the Institutional Animal Care and Use Committee of UTMB.
Written informed consent was obtained from all participants, and all the human studies were approved by the Institutional Review Board for Clinical Research of Shanghai Tenth People's Hospital, Tongji University. The characteristics of patients and healthy volunteers are described in Supplementary Table 1. CD4 + T cells isolation and culture. Mouse CD4 + T cells were isolated from spleen or mesenteric lymph node (MLN) using anti-mouse CD4 magnetic particles (Cat#551539, BD Biosciences). CD4 + T cells were seeded in the 24-well plates, and activated with 5 µg/ml anti-CD3 mAb (Clone#145-2C11, Cat#BE0001-1, Bio X Cell) and 2 µg/ml anti-CD28 mAb (Clone#37.51, Cat#BE0015-1, Bio X Cell), or 0.2 million/ml irradiated APCs and CBir1 peptide (ThermoFisher Scientific) in the presence or absence of acetate (10 mM, Sigma Aldrich), propionate (0.5 mM, Sigma Aldrich), or butyrate (0.5 mM, Sigma Aldrich), under neutral (without exogeneous cytokines), Th1 (10 ng/ml IL-12), Th17 (15 ng/ml TGFβ, 30 ng/ml IL-6, 10 µg/ml anti-IFNγ mAb, 5 µg/ml anti-IL-4 mAb), or Treg (5 ng/ml TGFβ and 10 µg/ml anti-IFNγ mAb) polarization conditions. Cells were cultured at 37°C with 5% CO 2 . On day 5, the cell yield was about 2 million/ml. RNA sequencing. Mouse splenic CD4 + T cells were activated with anti-CD3 (Clone#145-2C11, Cat#BE0001-1, Bio X Cell) and anti-CD28 mAb (Clone#37.51, Cat#BE0015-1, Bio X Cell) in the presence or absence of butyrate (0.5 mM) under neutral or Th1 conditions for 48 h. Cellular RNA was extracted, qualified, and followed by library construction at Novogene using an NEBNext @ Ultra RNA Library Prep Kit for Illumia @ . Briefly, mRNA was enriched, purified, and randomly fragmented. After synthesis of the first-strand cDNA using random hexamers primers, the second-strand cDNA was generated by using a custom second-strand synthesis buffer (Illumina), dNTPs, RNase H, and DNA polymerase I. After completing the double-stranded cDNA library through several steps, including terminal repair, poly-adenylation, sequencing adaptor ligation, size selection, and PCR enrichment, the 250-350 bp insert libraries were quantified by quantitative PCR. Qualified libraries were sequenced on an Illumina Novaseq Platform using a paired-end 150 run. Data were analyzed using Novosmart.
Quantitative real-time PCR. Total RNA was extracted from CD4 + T cells by using Trizol, and reverse-transcribed to cDNA by using qScript cDNA Synthesis Kit (P/N#84035, Quantabio). Quantitative real-time PCR was performed for analysis of gene expression by using SYBR Green Gene Expression Assays (Cat#1725124, Bio-Rad). All the primers were ordered from Integrated DNA Technologies, and listed in Supplementary Table 2. ELISA. IL-22, IL-6, and TNFα production were measured using ELISA MAX TM Deluxe Sets from Biolegend (IL-22, Cat#436304; IL-6, Cat#431304; TNF-α, Cat#430904). Microplate wells were incubated with capture antibodies overnight at 4°C, and blocked with 1% BSA for 1 h at room temperature. Samples were incubated in the wells for 2 h, and followed by the addition of detection antibodies. Horseradish peroxidase-labeled streptavidin was then incubated in the wells for 30 min. After adding TMB substrate, cytokines concentrations were measured at 450 nm using BioTek Gene5 instrument.
All the events were collected by BD FACS Diva software and analyzed using Flowjo. All the gating strategies were included in Supplementary Fig. 15.
All the events were collected by BD FACS Diva software. IL-22 and IL-17 production in ILCs (Thy1 + Lineage − cells), or in ILC3 (Thy1 + Lineage − RORγt + cells) was analyzed using FlowJo. Gating strategies are included in Supplementary  Fig. 15.
Statistical analysis. Student's unpaired or paired t-test was used to compare two groups when data were normally distributed. The nonparametric Mann-Whitney U test was used to measure the difference between two groups in which data were not normally distributed. For comparing more than two groups, one-way ANOVA was performed. All the tests were two-sided. All the analysis was performed by using Graphpad Prism 8.0 software. All the data were presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
RNA-seq data have been deposited in GEO database under the accession number GSE139631. All other data supporting the findings of this study are available within the article and its supplementary information files or from the corresponding author upon reasonable request. The source data underlying Figs. 1a-i, 2a-h, 3a-d, 4a-m, 5a-j, 6b-e, 7a-h, 8a-f, 9a-f, and Supplementary Figs. 1c-e, 1h-j, 2, 3a, d, 4, 5a, b, 8a-j, 9a-h, 10a-f, 11a-f, 12a-l, 13a-c, 14a-f. Source data were provided as a Source Data file. Source data are provided with this paper.