Ash1l and lnc-Smad3 coordinate Smad3 locus accessibility to modulate iTreg polarization and T cell autoimmunity

Regulatory T (Treg) cells are important for the maintenance of immune homoeostasis and prevention of autoimmune diseases. Epigenetic modifications have been reported to modulate autoimmunity by altering Treg cell fate. Here we show that the H3K4 methyltransferase Ash1l facilitates TGF-β-induced Treg cell polarization in vitro and protects mice from T cell-mediated colitis in vivo. Ash1l upregulates Smad3 expression by directly targeting Smad3 promoter to increase local H3K4 trimethylation. Furthermore, we identify an lncRNA, namely lnc-Smad3, which interacts with the histone deacetylase HDAC1 and silences Smad3 transcription. After TGF-β stimulation, activated Smad3 suppresses lnc-Smad3 transcription, thereby recovering the Smad3 promoter accessibility to Ash1l. By revealing the opposite regulatory functions of Ash1l and lnc-Smad3 in Smad3 expression, our data provide insights for the epigenetic control of Treg cell fate to potentially aid in the development of therapeutic intervention for autoimmune diseases.


Supplementary Figure 2. Ash1l affects the induced differentiation of T cells under polarization conditions.
(a, b) Flow cytometry profiles the pooled spleens from WT and Ash1l-silenced mice (n = 3 mice per group), showing expression of IFN-γ, IL-17, and Foxp3 in the naï ve CD4 + T cells (Th0) or in the CD4 + T cells stimulated under indicated polarizing conditions (Th1, Th17 or iTreg cell skewing conditions respectively) for 3 days. Numbers beside outlined areas indicate percentage of cells in the CD4 + gate. (c, d) Flow cytometry profiles the mLN (mesenteric lymph nodes) and thymus from WT and Ash1l-silenced mice (n = 3 mice per group), showing expression of Foxp3 in the CD4 + T cells stimulated under iTreg cell-skewing conditions (with TGF-β) for 3 days. Numbers beside outlined areas indicate percent Foxp3 + cells in the CD4 + gate. Error bars represent SD. Student's t test. **p < 0.01. Data are representative of three independent experiments (a, c) or are from three independent experiments (b, d; mean ± SD of technical triplicates).

Supplementary Figure 3. Silencing of Ash1l does not affect the expression of other H3K4 methyltransferases.
(a) The mRNA level of Ash1l in the CD4 + T cells stimulated under iTreg cell-skewing conditions (with TGF-β) for indicated times. Results are relative to the baseline Ash1l expression in unstimulated WT CD4 + T cells, set as 1. (b) The mRNA level of Ash1l in the naï ve CD4 + T cells (Th0), T cells stimulated under indicated polarizing conditions (Th1, Th17 or iTreg cell skewing conditions respectively) for 3 days, and in macrophages. Results are relative to the baseline Ash1l expression in Th0 cells, set as 1. (c) The mRNA levels of 8 H3K4 methyltransferases in WT and Ash1l-silenced naï ve CD4 + T cells. Results are relative to those in WT naï ve CD4 + T cells, set as 1. Error bars represent SD. Student's t test. **p < 0.01. NS, not significant. All data are from three independent experiments (mean ± SD of technical triplicates).

Supplementary Figure 4. Ash1l facilitates TGF-β-mediated Treg cell function.
(a) Flow cytometry profiles the expression of indicated costimulatory molecules on WT and Ash1l-silenced iTreg cells. Data are mean fluorescence intensity (MFI) ± SD, × 1000; n = 3. (b, c) Proliferation of CD45.1 + WT CFSE-labeled CD4 + T cells incubated with BMDCs and WT or Ash1l-silenced iTreg cells, assessed by flow cytometry (b) and quantified (c) as dilution of CFSE at 72h. Numbers above plots indicates percent of CFSE low cells in the CD45.1 + gate. MFI, mean fluorescence intensity. Error bars represent SD. Student's t test. *p < 0.05, **p < 0.01. Data are from three independent experiments (a, c; mean ± SD of technical triplicates) or representative of three independent experiments (b). Figure 5. In vivo silencing of Ash1l renders mice more susceptible to T cell-mediated colitis. (a) TNBS-colitis was induced as in Fig. 2a. Histological sections of colons from WT and Ash1l-silenced mice (n = 5 mice per group) were examined at day3 after TNBS-induction. Outlined areas are shown at higher magnification at right. Scale bars represent 50μm. (b) Percentages of Foxp3 + Tregs in splenocytes and mLN (mesenteric lymph nodes) cells were determined at day3 after TNBS-induction. (c) T cell contransfer colitis was induced as in Fig. 2d. Histological sections of colons from Rag1 -/mice transferred with WT CD4 + CD25 -T cells with or without WT or Ash1l-silenced iTreg cells were examined at week 6 after T cell transfer. Outlined areas are shown at higher magnification at right. Scale bars represent 50μm. (d, e) Representative flow cytometry (d) and quantification (e) of the percentages of Foxp3 + Tregs in splenocytes of Rag1 -/mice 4 weeks following adoptive transfer of naï ve CD4 + T cells from WT and Ash1l-silenced mice. Numbers indicate percentage of cells in the CD4 + gate. Error bars represent SD. Student's t test. **p < 0.01. Data are representative of three independent experiments (a, c, d) or from three independent experiments (b, e; mean ± SD of five mice). Figure 6. ASH1L, FOXP3 and SMAD3 are significantly downregulated in CD4 + T cells from peripheral blood mononuclear cells (PBMCs) of patients with rheumatoid arthritis mRNA expression of ASH1L, FOXP3 and SMAD3 in CD4 + T cells sorted from the peripheral blood mononuclear cells (PBMCs) of 10 individuals with rheumatoid arthritis (RA) or 9 healthy controls (HC). Error bars represent SD. Student's t test. *p < 0.05, **p < 0.01. Data are shown as mean ± SD.

Supplementary Figure 7. Less iTreg cell generation in Smad3-KO mice.
(a-d) The proportion of CD4 and CD8 in the thymus and spleen (a, b), and the expression of CD44 and CD62L in peripheral CD4 + T cells (c, d) from WT and Smad3-KO mice (n = 3 mice per group), assayed by flow cytometry (a, c) and quantified (b, d).
(e, f) The percentage of Foxp3 + CD25 + in the CD4 + T cells from WT and Smad3-KO splenocytes (left of e and f) or thymocytes (right of e and f) stimulated under iTreg cell-skewing conditions (with TGF-β) for 3 days, assayed by flow cytometry (e) and quantified (f). Error bars represent SD. Student's t test. NS, not significant, *p < 0.05. Data are representative of three independent experiments (a, c, e) or from three independent experiments (b, d, f; mean ± SD of three mice). (a) Quantitative PCR detection of the lnc-Smad3 retrieved by Smad3 or Ash1l specific antibody compared with immunoglobulin G (IgG) in the RIP assay within CD4 + T cells. Normalized data are shown as percentage of input control (% Inp). IgG serves as a RIP control. (b) Chromatin accessibility of the Foxp3 promoter region by quantitative PCR with DNase I pretreated nucleus of CD4 + T cells transduced with a control lentivirus (Lenti-CTR) or lnc-Smad3-expressing lentivirus (Lenti-lnc-Smad3) and cultured under iTreg cell-skewing conditions (with TGF-β) for 2 days. Changed fold are concluded using 2 ΔCt with respect to CD4 + T cells transduced with Lenti-CTR, set as 1.

Supplementary
(c) CHIP analysis of the accumulation of Ash1l and H3K4me3 modification at Foxp3 promoter regions in CD4 + T cells transduced and cultured as in b. Normalized data are shown as percentage of input control (% Inp). IgG serves as a ChIP control. Error bars represent SD. Student's t test. NS, not significant. Data are from three independent experiments (mean ± SD of technical triplicates).

Supplementary Figure 12. Ash1l competes with lnc-Smad3 to promote iTreg cell polarization via upregulating TGF-β-mediated Smad3 expression.
In naï ve T cells, lnc-Smad3 turns off the transcription activation of the nearby gene locus Smad3 by recruiting HDAC1. Upon TGF-β stimulation, activated Smad3 suppresses lnc-Smad3 expression by binding to its promoter region. TGF-β-mediated reduction of lnc-Smad3 relieves its suppression on Smad3 promoter region, restoring the accessibility of Smad3 promoter to Ash1l. Hence Ash1l is allowed to accumulate at Smad3 promoter, increase the H3K4me3 modification and activate Smad3 transcription. Consequently, Ash1l-upregulated Smad3 facilities Foxp3 expression as a transcription factor and finally enhances TGF-β-mediated Treg cell generation. Figure 13. The gating and sorting strategies of indicated cell subsets.

Supplementary
(a, b) The representative FACS plots show the gating strategies of CD4 + and CD8 + T cell subsets from the thymocytes or splenocytes of mice, corresponding to Supplementary Fig. 1a, b and Supplementary Fig. 7a, b. (c) The representative FASC plots show the gating strategies of CD62L + and CD44 + T cell subsets from splenocytes and mLN (mesenteric lymph nodes) of mice, corresponding to Supplementary  Fig. 1c, d and Supplementary Fig. 7c, d. (d) The representative FASC plots show the gating strategies of Foxp3 + CD25 + T cell subset from splenocytes, mLN, thymocytes and purified CD4 + T cells of mice, corresponding to Fig. 1a