DNA Demethylation of the TIM-3 Promoter Is Critical for Its Stable Expression on T Cells

ORIGINAL ARTICLE DNA demethylation of the TIM-3 promoter is critical for its stable expression on T cells

F-C Chou1, C-C Kuo1, H-Y Chen2, H-H Chen3 and H-K Sytwu1,2

The T-cell immunoglobulin and mucin domain-containing protein 3 (TIM-3) is selectively expressed on terminally differentiated T helper 1 (Th1) cells and acts as a negative regulator that terminates Th1 responses. The dysregulation of TIM-3 expression on T cells is associated with several autoimmune phenotypes and with chronic viral infections; however, the mechanism of this regulation is unclear. In this study, we investigated the effect of DNA methylation on the expression of TIM-3. By analyzing the sequences of TIM-3 promoter regions in human and mouse, we identiﬁed a CpG island within the TIM-3 promoter and demonstrated that the promoter activity was controlled by DNA methylation. Furthermore, treatment with 5-aza-2′-deoxycytidine enhanced TIM-3 expression on mouse primary CD4+ T cells under Th0-, Th1- or Th2-polarizing conditions. Finally, pyrosequencing analysis revealed that the methylation level of the TIM-3 promoter gradually decreased after each round of T-cell polarization, and this decrease was inversely correlated with TIM-3 expression. These data suggest that the DNA methylation of the TIM-3 promoter cooperates with lineage-speciﬁc transcription factors in the control of Th-cell development. In conclusion, DNA methylation-based regulation of TIM-3 may provide novel insights into understanding the dysregulation of TIM-3 expression under pathogenic conditions.

Genes and Immunity (2016) 17, 179–186; doi:10.1038/gene.2016.6; published online 18 February 2016

INTRODUCTION polymorphism and linkage disequilibrium analysis do not provide The T-cell immunoglobulin and mucin domain-containing protein insight into the possible mechanisms of the dysregulation of TIM-3 3 (TIM-3) molecule is specifically expressed on terminally expression in these diseases, although it has been reported that 22 differentiated T helper 1 (Th1) cells and regulates Th1 cell- TIM-3 is regulated by a Th1 transcription factor, T-bet. However, –/– mediated immune responses.1 The ligand for TIM-3 was identified there is considerable expression of TIM-3 on T cells from T-bet 2 β mice, signal transducer and activator of transcription (STAT)4–/– as galectin-9, a -galactoside-binding lectin that induces T-cell – – – – apoptosis through the calcium–calpain–caspase-1 pathway.3 mice and T-bet / × STAT4 / mice, which suggests that other The roles of the galectin-9–TIM-3 pathway in the regulatory transcription factors and different mechanisms are involved in the functions of T cells have been widely studied in animal models of regulation of TIM-3 expression. Therefore, the regulation of TIM-3 autoimmune diseases, including autoimmune encephalomyelitis,2 expression on Th1 cells in different situations is not understood. autoimmune diabetes,4 rheumatoid arthritis5 and psoriasis.6 Recently, the role of epigenetic regulation in T-cell differentia- In addition, overexpression of galectin-9 has been demonstrated tion has attracted considerable attention. There is growing in several transplantation models to prolong graft survival through evidence that DNA methylation, histone modifications and the downregulation of T-cell responses,7–9 which suggests that the microRNAs are involved in the control of lineage specificity and 23 galectin-9–TIM-3 pathway could be a therapeutic target. In effector function of T-cell subsets (reviewed by Wilson et al. addition, previous reports have demonstrated that the expression and Baumjohann et al.24). During Th-cell differentiation, cell of TIM-3 is dysregulated on T cells from patients with multiple division is accompanied by epigenetic changes that guide the sclerosis,10 autoimmune hepatitis11,12 or during chronic viral production of lineage-specific cytokines and maintain the stability – infections.13–15 These reports indicated that the galectin-9–TIM-3 of the differentiated state.25 27 Because TIM-3 expression is pathway plays a pivotal role in the control of immune responses. restricted to selected cell types, the detailed mechanism of its However, the underlying mechanisms of the regulation of TIM-3 regulation are unclear. We hypothesized that epigenetic regula- expression, especially under pathogenic conditions, are not fully tion may be involved in the control of TIM-3 expression. In this understood. study, we identified a CpG island within the promoter region of Given the importance of TIM-3 in the regulation of T-cell TIM-3 and analyzed the effect of DNA methylation on the immune responses, previous reports have focused on genetic promoter activity. Interestingly, treatment with the DNA methyl- association studies that analyze the single-nucleotide polymorph- transferase (DNMT) inhibitor, 5-aza-2′-deoxycytidine (5-Aza) sig- isms of the TIM-3 promoter in patients with allergic nificantly increased TIM-3 expression on polarized Th1 and Th2 phenotypes16–19 or childhood asthma20 and type 1 diabetes cells, but not on Th17 cells. Finally, using pyrosequencing analysis, families.21 Although researchers have put considerable effort into we demonstrated that the methylation level of the TIM-3 promoter the study of the genetics of TIM-3, the data from single-nucleotide was inversely correlated with TIM-3 expression, indicating that

1Department and Graduate Institute of Microbiology and Immunology, National Defense Medical Center, Taipei, Taiwan; 2Graduate Institutes of Life Sciences, National Defense Medical Center, Taipei, Taiwan and 3Department of Medicine, National Defense Medical Center, Taipei, Taiwan. Correspondence: Professor H-K Sytwu, Department and Graduate Institute of Microbiology and Immunology, National Defense Medical Center, 161, Section 6, MinChuan East Road, Neihu, Taipei 114, Taiwan. E-mail: [email protected] Received 14 September 2015; revised 27 November 2015; accepted 4 January 2016; published online 18 February 2016 DNA methylation controls TIM-3 expression F-C Chou et al 180 DNA methylation was involved in the control of TIM-3 expression. DNA methylation on the activity of the TIM-3 promoter, we cloned These results suggest a novel regulatory mechanism for TIM-3 that different fragments of the DNA sequence from human (Figure 1b) can be applied to investigation of the dysregulation of TIM-3 and mouse (Figure 1c) TIM-3 into the pGL4.21 luciferase reporter expression under pathogenic conditions. or a CpG-free pCpGL luciferase reporter,28 and performed reporter assays. The pCpGL vector contains no CpG site; therefore, it can be used to evaluate the effect of DNA methylation on promoter RESULTS activity. The longest fragments containing the predicted CpG Identification of human and mouse TIM-3 promoter island displayed the strongest promoter activity for both human To examine the transcriptional activity in the 5′-flanking region of (Figure 1d) and mouse (Figure 1e). To test whether the CpG island- TIM-3, we obtained the DNA sequence from the UCSC Genome containing fragments had independent promoter activity or acted Bioinformatics website (http://genome.ucsc.edu/) and analyzed it as enhancers, we first performed the reporter assay and using the rVISTA program. We found that there are three demonstrated that the human − 1702 to − 899 and mouse conserved noncoding sequences near the transcription start site − 3393 to − 3210 constructs had no promoter activity (Figure 1d; of human and mouse TIM-3 (Figure 1a). By using CpG Island data not shown). We then investigated whether DNA methylation Explorer software, we identified a CpG island within the upstream regulated the promoter activity. As shown in Figures 1f–h, DNA sequence of both the human (Figure 1b) and mouse (Figure 1c) methylation greatly reduced the promoter activity in both human TIM-3 genes. To examine the transcriptional activity of the (Figure 1f) and mouse (Figure 1g) promoter constructs. We predicted TIM-3 promoter regions and evaluate the effect of noticed that the reporter activity of the promoter constructs,

Figure 1. Identiﬁcation of the Tim-3 promoter. (a) Comparison of human Tim-3 upstream sequence to that of mouse, rat and cattle. The sequences were obtained from the UCSC database and the interspecies conservation plot was generated using the mVISTA tool on the website. Schematic illustration of the putative promoter fragments of (b) human and (c) mouse Tim-3 used for the reporter assay. The CpG islands indicated in the ﬁgure were predicted using CpGIE software. Determination of (d) human and (e) mouse promoter activity by a luciferase reporter assay. Effect of (f) human and (g and h) mouse promoter methylation on reporter assays. To investigate the effect of CpG methylation on the activity of the Tim-3 promoter, the promoter fragments were cloned into a CpG-free vector for an in vitro methylation and reporter assay. (i) Enhancer activity assay. Putative enhancer fragments were cloned into a CpG-free EF1A-promoter/enhancer cassette and the enhancer activity was evaluated. This construct was further methylated in vitro to evaluate the effect of DNA methylation on enhancer activity. Representative results from at least three independent experiments are shown. CNS, conserved nucleotide sequence. Paired t-test; **Po0.01.

Genes and Immunity (2016) 179 – 186 © 2016 Macmillan Publishers Limited DNA methylation controls TIM-3 expression F-C Chou et al 181 human − 1192 to +42 (Figure 1f) and mouse − 3238 to +67 Effects of DNMT inhibitor on transcription of TIM-3 and (Figure 1g) was reduced by DNA methylation. This phenomenon lineage-specific transcription factors in Th-cell subsets was also observed when we used the mouse minimal promoter To investigate whether DNA methylation regulates TIM-3 expres- construct for an in vitro methylation and reporter assay (Figure 1h). sion and restricts TIM-3 expression on Th1 cells, we analyzed TIM-3 Thus, we propose that the CpG island-containing fragment may expression on different Th subsets. Naive CD4+ T cells were have enhancer activity and that the promoter activity is controlled cultured under Th0-, Th1-, Th2- and Th17-polarizing conditions by other CpG dinucleotides close to the transcriptional start site. and in interleukin (IL)-27 condition (IL-27 promotes early induction 29 Both of these activities could be inhibited by DNA methylation. To of Th1 differentiation ) in the presence of 0.5 μM 5-Aza or 0.001% confirm this hypothesis, we took advantage of a CpG-free pCpGL- dimethyl sulfoxide (DMSO, solvent control; we confirmed that this cytomegalovirus/elongation factor luciferase reporter plasmid that concentration did not alter the T-cell differentiation and TIM-3 contains a cytomegalovirus enhancer and a human elongation expression; see Supplementary Information 2). We found that factor 1 promoter (also CpG free). We replaced the cytomegalovirus transcription of TIM-3 was relatively higher in the cells cultured in enhancer with mouse − 3393 to − 3210, mouse − 3238 to − 2179 or Th1-polarizing and IL-27-stimulated conditions than in cells mouse − 2267 to − 1541 fragments (illustrated in Figure 1c) and cultured in the other conditions, which is consistent with the analyzed the enhancer activity. Our data demonstrated that mouse previous finding that TIM-3 is preferentially expressed on Th1 − 3393 to − 3210 had weak enhancer activity and that its activity cells.1 Interestingly, treatment of T cells with 5-Aza enhanced was controlled by DNA methylation, as demonstrated by the TIM-3 transcription under all stimulation conditions except reduced enhancer activity of the in vitro methylated construct (Me- Th17-polarizing conditions (Figure 2a). Although we cannot rule pCpGL − 3393 to − 3210; Figure 1i). Interestingly, this activity was out that 5-Aza treatment increases TIM-3 expression through the not enhanced by phorbol 12-myristate 13-acetate/ionomycin induction of T-bet, which has been demonstrated to induce TIM-3 stimulation or T-bet expression (Supplementary Information 1). expression,22 the increased TIM-3 transcription in the IL-27- The mouse − 3238 to − 2179 and the mouse − 2267 to − 1541 stimulated condition was independent of increased T-bet expres- fragments showed no promoter activity (data not shown) and were sion (Figure 2b), suggesting a T-bet-independent pathway in the used as a control for enhancer activity (Figure 1i). In summary, we control of TIM-3 expression. Further, the expression of GATA-3 identified the promoter region of TIM-3 and demonstrated that its (Figure 2c) and RORγt (Figure 2d) was not significantly changed by promoter activity is controlled by DNA methylation of the enhancer 5-Aza treatment of Th2 and Th17 cells, respectively. Interestingly, region and proximal promoter. This is analyzed further below. 5-Aza treatment preferentially promoted Th1-related gene

Figure 2. Quantitative PCR analysis of Tim-3, T-bet, Gata-3 and Rorc expression in mouse primary T cells under different polarizing conditions in the presence of 5-Aza. (a) Tim-3,(b) T-bet,(c) Gata-3 and (d) Rorc. DMSO, solvent control (0.001%); DNA methyltransferase inhibitor, 5-Aza (0.5 μM). Data are presented as mean ± s.e.m. Representative results from at least three independent experiments. Paired t-test; signiﬁcance was set at Po0.05 (*).

© 2016 Macmillan Publishers Limited Genes and Immunity (2016) 179 – 186 DNA methylation controls TIM-3 expression F-C Chou et al 182 transcription (Figures 2a and b) and inhibited RORγt expression in demethylation on lineage commitment, we analyzed the Th0 and Th1 cells (Figure 2d). Collectively, our data indicated that cytokine expression of each population. Our data showed that 5-Aza facilitates TIM-3 expression directly through DNA demethy- 5-Aza treatment enhances interferon-γ (IFN-γ;Figure3b)andIL-4 lation of its promoter and/or induction of the Th1-specific (Figure 3c; Supplementary Information 3b) expression in Th1 and transcription factor, T-bet. In addition, because 5-Aza treatment Th2 cells, respectively, and also enhances IFN-γ expression under did not alter the differentiation process of Th2 and Th17 cells, Th0- and Th2-polarizing conditions (Figure 3b), which suggests these two lineage-specific transcription factors may not regulate that DNA methylation regulates the stringency of IFN-γ TIM-3 expression. expression after lineage commitment. Again, 5-Aza treatment did not alter cytokine expressioninTh17cells(Figures3b–d), Effect of DNMT inhibitor on TIM-3 expression and cytokine suggesting that DNA methylation may not regulate IL-17 expression in Th-cell subsets expression or that of the Th17-specific transcription factors that We next analyzed TIM-3 and cytokine expression under different control the stringency of Th17 cells. In summary, we demon- Th-cell-polarizing conditions in the presence of DMSO or 5-Aza. strated that DNA demethylation facilities TIM-3 expression under Our data revealed that TIM-3 was mainly expressed on Th1 cells; Th0-, Th1- and Th2-polarizing conditions, which prompted us to however, Th0 and Th2 cells also express TIM-3 to a lesser extent identify the CpG sites within the TIM-3 promoter that might be after stimulation (Figure 3a). Interestingly, we found that involved in the control of TIM-3 expression. expression of TIM-3 is significantly increased in the presence of 5-Aza under Th0-, Th1- and Th2-polarizing conditions (Figure 3a; Pyrosequencing analysis of the CpG island within the TIM-3 Supplementary Information 3a). Strikingly, TIM-3 is barely promoter expressed on Th17 cells and 5-Aza treatment (Figure 3a), even It is possible that nonspecific demethylation by 5-Aza treatment with 5 μM 5-Aza (Supplementary Information 4), does not induce causes secondary effects that induce the expression of TIM-3 on TIM-3 expression on these cells. To determine the efficiency of the Th cells under polarizing conditions, although the low dose of T-cell polarization experiment and the impact of DNA 5-Aza that we used should minimize the side effects of global

Figure 3. Flow cytometric analysis of TIM-3 and cytokine expression in CD4+ T cells under Th0-, Th1-, Th2- and Th17-polarizing conditions in the presence of 5-Aza. Purified CD4+ T cells were stimulated in the indicated conditions for 3 days and the expression of (a) TIM-3, (b) IFN-γ, (c) IL-4 and (d) IL-17 was analyzed by flow cytometry. Data are presented as mean ± s.e.m. Unpaired t-test; significance was set at Po0.05 (*).

Genes and Immunity (2016) 179 – 186 © 2016 Macmillan Publishers Limited DNA methylation controls TIM-3 expression F-C Chou et al 183 demethylation. To analyze the correlation between DNA methyla- analyzed the methylation level of the CpG sites in region 2 of tion and TIM-3 expression directly, we analyzed the methylation the minimal promoter (Figure 4a). CD4+ T cells after one round status of the CpG sites in the TIM-3 promoter by bisulfite of stimulation showed similar methylation levels to naive pyrosequencing. Because we have identified that the enhancer CD4+ T cells, even when treated with 5-Aza (Supplementary region contains a CpG island and that the enhancer activity is Information 5c). Interestingly, we found that after four rounds of controlled by DNA methylation, we first analyzed the 17 CpG sites stimulation, Th1 cells displayed reduced methylation in region 2 within the CpG island-containing region 1 (Figure 4a) of the TIM-3 and expressed high levels of TIM-3, which suggests that TIM-3 promoter in Th1-polarized cells. The methylation status of region 1 expression induced by repeated stimulation correlates with DNA was only slightly decreased in Th1-polarized cells, even those demethylation (Supplementary Information 5c). Thus, we stimu- treated with 5-Aza, compared with that in naive CD4+ T cells lated cells for four rounds in Th1-, Th2- and Th17-polarizing (Figure 4b). We also analyzed the methylation status of region 1 of conditions, and analyzed TIM-3 expression and the methylation Th0-, Th2- and Th17-polarized cells, and found that the methyla- status of region 2. We found that TIM-3 expression increased tion status was also slightly decreased in 5-Aza-treated Th0 and after each round of stimulation (Figure 4c), and that this was Th2 cells, whereas Th17 cells, which express low levels of TIM-3, inversely correlated with the methylation status of region 2 showed high levels of DNA methylation (Supplementary (Figures 4d–f). These data suggested that DNA demethylation in Information 5a). To determine whether the demethylation of region 2 is crucial for TIM-3 expression, and that the enhancer CpG sites in region 1 (Figure 4a) are crucial for the regulation of region plays only a minor role in the promotion of TIM-3 TIM-3 expression, we stimulated Th1 cells for several rounds in expression. Interestingly, under certain conditions, DNA methyla- polarizing conditions to induce stable TIM-3 expression as tion may be higher in the hierarchy of control of TIM-3 described previously,1 and analyzed the methylation level of expression than are lineage-specific transcription factors. region 1. We found that the expression of TIM-3 on the cell surface markedly increased after the second round of stimulation (Figure 4c). Unexpectedly, the methylation status of region 1 DISCUSSION was not significantly decreased (Supplementary Information 5b). Epigenetic mechanisms for regulation of gene expression have Although we did not identify a CpG island within the mouse been widely studied in many fields, including embryogenesis, minimal promoter region, we have demonstrated that DNA stem cell biology, oncology and immunology. There is growing methylation markedly reduces its activity (Figure 1h). Thus, we evidence that histone modification, DNA methylation and

Figure 4. Pyrosequencing analysis of the methylation level of CpG motifs within the Tim-3 promoter. (a) CpG sites within the enhancer (region 1) and minimal promoter (region 2). (b) Methylation level of 17 CpG sites within region 1 of the Tim-3 promoter. DNA from naive CD4+ T cells and DMSO-treated or 5-Aza-treated Th1 cells was extracted for determination of the methylation level. (c) Expression of Tim-3 in Th cells after repetitive stimulation. Naive CD4+ T cells were subjected to four rounds of polarization in the indicated condition, and the expression of Tim-3 was determined after each round of stimulation. Methylation levels of the four CpG sites within the Tim-3 minimal promoter under (d) Th1-, (e) Th2- and (f) Th17-polarizing conditions were determined by pyrosequencing. Viable cells were collected from each round of stimulation and DNA was extracted for pyrosequencing analysis.

© 2016 Macmillan Publishers Limited Genes and Immunity (2016) 179 – 186 DNA methylation controls TIM-3 expression F-C Chou et al 184 microRNAs control the lineage stringency and plasticity of Th thymic regulatory T cells, but heavily methylated in conventional cells.30 TIM-3 plays a crucial role in the control of T-cell responses T cells and peripheral-induced regulatory T cells.45,46 These results in autoimmunity,31 transplant acceptance,7,8,32 chronic viral suggested that, in coordination with lineage-specific transcription infections13,14,33 and immune surveillance of tumors.34 In this factors, DNA methylation controls accessibility to the promoter or study, we identified a CpG island within the TIM-3 enhancer region gene in the regulation of gene expression in Th cells. and demonstrated that the enhancer activity was controlled by It has been demonstrated that under pathogenic conditions, DNA methylation. More importantly, methylation level of the four such as during chronic viral infections in mice, demethylation of CpG sites within the minimal promoter region was highly the promoter region of the pdcd-1 gene results in a high level of correlated with the level of TIM-3 expression on T cells. In expression of the inhibitory molecule PD-1 on T cells, which is addition to the role of cytokine signaling and lineage-specific associated with clonal exhaustion.47 In clinical studies, aberrant transcription factors in the control of the stringency of TIM-3 signal transduction leads to hypomethylation of the promoter expression on Th1 cells, it is noteworthy that the DNA methylation region of methylation-sensitive genes (LFA-1, CD70, CD40L and status controls the stable levels of TIM-3 expression in terminally perforin) in T cells, which causes T-cell overactivation and leads to differentiated cells. These findings may help to clarify the autoimmune diseases such as lupus erythematosus.48–50 Thus, it is mechanism of the dysregulation of TIM-3 expression in auto- possible that the dysregulation of TIM-3 expression in chronic viral immune disorders and chronic infections. These findings also infections13 and autoimmune diseases10 may be the result of provide a new perspective by indicating that epigenetic mechan- abnormal regulation of DNA methylation of the TIM-3 promoter. isms regulate not only T-cell differentiation and effector function We are now studying the regulation of TIM-3 in an autoimmune (expression of transcription factors and effector cytokines), but diabetes mouse model. In our preliminary results, we found that also T-cell homeostasis (TIM-3-expressing T cells are susceptible to CD4+ T cells from the inflammatory pancreas expressed pre- galectin-9-mediated apoptosis). dominant IFN-γ but less TIM-3. Besides, by pyrosequencing Although it is well documented that TIM-3 is specifically analysis, these CD4+ T cells maintained high methylation level expressed on Th1 cells1 and dendritic cells,35 Th17 cells also on the TIM-3 promoter. Our data suggested that in autoimmune express low levels of TIM-3 in a T-bet-dependent manner after diabetes, dysregulated demethylation of TIM-3 promoter reduced immunization and restimulation.36 Thus, 5-Aza-induced TIM-3 TIM-3 expression and eventually caused uncontrolled Th1 expression under Th0-, Th1- and Th2-polarizing conditions may be responses (our unpublished data). However, the detailed mechan- partially the result of slightly increased T-bet expression. However, isms of, for example, the regulation of DNMT expression in T cells we did not detect a significant level of TIM-3 expression on one under T-cell receptor and/or cytokine stimulation, or the process round-polarized Th17 cells even when treated with 5-Aza, which of active DNA demethylation by Tet proteins51 during T-cell suggests that IL-6 and/or transforming growth factor-β may activation, need to be further investigated. suppress the induction of TIM-3 at an early stage. It was unexpected that Th2 and Th17 cells expressed TIM-3 after several rounds of polarization (Figure 4c). In this situation, it is possible MATERIALS AND METHODS that T-cell receptor/CD28-induced signal transduction and/or Prediction of promoters and CpG island analysis T-cell proliferation accompanied by passive DNA demethylation The upstream sequences of TIM-3 were obtained from the UCSC Genome induce TIM-3 expression. Previous studies have demonstrated that database (4000 base pairs) and were subjected to comparative analysis TIM-3 is expressed on Th1 cells after the third round of using the rVISTA program.52 The putative promoter region and CpG islands polarization in in vitro cultures, even in the absence of the Th1- within these sequences were analyzed using website software and the specific transcription factors T-bet and STAT4.1,22 Besides, previous CpGIE program, respectively. reports also indicated that TIM-3 is not exclusively expressed on Th1 cells under certain stimulation conditions. Sanchez-Fueyo Construction of luciferase reporter plasmids and in vitro DNA et al.31 have demonstrated that Th2 cells also express TIM-3 after methylation repeated stimulation, albeit the expression level is lower than The promoter fragments were cloned into a pGL4.21 luciferase vector Th1 cells. Besides, Th17 cells also express low levels of TIM-3 in a (Promega, Mannheim, Germany) and CpG-free pCpGL or pCpGL-cytomega- T-bet-dependent manner after immunization and restimulation.36 lovirus/elongation factor 1 luciferase vectors,28 depending on the experi- The discrepancy of these results may be due to different ment. Primer sets for cloning the promoter fragment are listed in Table 1. To experimental settings (for example, stimulation condition, mouse determine the effects of DNA methylation on promoter/enhancer activity, strains and antibody clones used for staining). Altogether, these results suggested that there may be methylation-dependent regulatory mechanisms that are more efficient than, or synergistic Table 1. Primer sequences for promoter cloning with, Th1-specific transcription factors for the control of TIM-3 expression. Moreover, a decreased methylation status of the four 5′–3′ First base location Sequence CpG sites within the minimal promoter may facilitate T-bet-mediated transcription of TIM-3, even with the low level of Human 5′–3′ − 1702 GCCTTGACCAAGTTCATGCT T-bet expression in Th0, Th2 and Th17 cells. Human 5′–3′ − 1192 CCTGGACAAAAAGAGCGAAA DNA methylation has been implicated in the regulation of the Human 5′–3′ − 787 AGTGTGGCTGGAACTCAACA γ 37 (ref. 38) Human 5′–3′ − 451 GAGGCTTATGCTGGGAGTTG expression in T cells of cytokines, including IFN- , IL-2 and ′– ′ − IL-3,39 and of T-cell lineage markers including perforin, CD4 and Human 5 3 241 GGACATGCTCCATTTCAGGT Human 3′–5′ − 894 CTCACAGGCTGAGTGGTTTCTG CD8, suggesting that DNA methylation may help to establish and/ ′– ′ + Human 3 5 +42 AGCAGTAGCCGCCCCCAC or maintain the different effector responses exhibited by CD4 ′– ′ − + 40,41 Mouse 5 3 3393 CCCTTCACAGAAGGCAAGAA and CD8 T cells. Moreover, genetic ablation of DNMT in Mouse 5′–3′ − 3238 CCCCTATTGTGATGCCAGTC T cells has demonstrated that DNMT1 is crucial for the Mouse 5′–3′ − 1570 TAGAGCAAGCCTAGGGCTCA development, function and survival of T cells,42,43 whereas Mouse 5′–3′ − 1261 CCTCTGTGGGAAAACTTTGG DNMT3a controls lineage stability after differentiation.44 The Mouse 5′–3′ − 492 AAAGGTATACACCGCCATGC expression of Foxp3 in naturally occurring regulatory T cells is Mouse 3′–5′ − 3210 TCTCATGGAGACTGGCATCA Mouse 3′–5′ − 2179 CAATAAACATACGAGGGGGAAA also controlled by DNA methylation. Analysis of the promoter ′– ′ − region of the Foxp3 gene showed that this noncoding region Mouse 3 5 1541 ACAAGTCCCCTGAGCCCTA Mouse 3′–5′ +67 TGAGTACTTGGCAGGGGAAA contains CpG motifs, which are completely demethylated in

Genes and Immunity (2016) 179 – 186 © 2016 Macmillan Publishers Limited DNA methylation controls TIM-3 expression F-C Chou et al 185 the constructed pCpGL and pCpGL-elongation factor vectors were subjected Table 2. Primers used for real-time PCR to in vitro methylation with SssI methylase in the presence of S-adenosylmethionine (New England Biolabs, Ipswich, MA, USA). The Primer Sequence (5′–3′) efficiency of methylation was evaluated using restriction enzymes HpaII (methylation sensitive) and MspI (methylation insensitive), respectively. Tim3-F GCTAAAGGGCGATCTCAACAA Tim3-R GGAGAAGCTGTAGTAGAGTCCC Electroporation of T cells and reporter assay Tbx21-F TGCCCGAACTACAGTCACGAAC Tbx21-R AGTGACCTCGCCTGGTGAAATG Human and mouse promoter constructs were transfected into Jurkat and EL4 Gata3-F AGAACCGGCCCCTTATGAA cell lines, respectively, by electroporation. A total of 107 cells were μ μ Gata3-R AGTTCGCGCAGGATGTCC cotransfected with 20 g of promoter constructs and 0.4 g of phRL-TK Rorc-F TCTACGCTATGAGGAAGGAAGGC Renilla plasmid (Promega, Mannheim, Germany) using an ECM830 electro- Rorc-R GACTATGGAGGAGAAACAGGTCCC poration system (BTX Harvard Apparatus, Holliston, MA, USA; Jurkat cells: Rps29-F ACGGTCTGATCCGCAAATAC 290 V, 20 ms; EL4 cells: 225 V, 35 ms). For the in vitro methylation of Rps29-R AGCATGATCGGTTCCACTTG plasmids, 2.5 × 106 cells were prepared in Ingenio Electroporation Solution (Mirus Bio LLC, Madison, WI, USA) and cotransfected with 2 μg of promoter constructs and 0.04 μg of phRL-TK Renilla plasmid. After 24 h, cell lysates were assayed for fireflyandRenilla luciferase activity using the Dual- Table 3. Primer sequences for pyrosequencing Luciferase Reporter Assay System (Promega) on a Packard TopCount Microplate Scintillation Counter (PerkinElmer, Shelton, CT, USA). The firefly Primer Sequence (5′–3′) Base pair luciferase activity of individual transfections was normalized against their Renilla luciferase activity. Data are presented as relative luciferase units Region 1 (luciferase/Renilla). PCR-forward AGTGTTGTTTTTTTATAGAAGGTAAGA 529 PCR-reverse ACAACATCCAAAAAAACTACCTATTCA Sequencing 1 TTTTATAGAAGGTAAGAAATTGT T-cell polarization and flow cytometry (CpG1–4) CD4+CD25– T cells and CD4+CD62L+ T cells (naive CD4 T cells) were Sequencing 2 GGTTTTGGTATAGTGTAGATT isolated from BALB/c mice using the BD IMag Cell Separation System (CpG5–8) (BD Pharmingen, San Jose, CA, USA). A total of 2 × 106 cells were stimulated Sequencing 3 GGTGGTTGTTATTATTATTGTTA − – with IL-2 (5 ng ml 1), and plate-bound anti-CD3 (1 μg per well) and anti- (CpG 9 16) CD28 (1 μgml− 1) monoclonal antibodies for 3 days in 24-well plates under conditions for polarizing Th0 (anti-IFN-γ,5μgml− 1; anti-IL-4, 5 μgml− 1), Region 2 Th1 (IL-12, 10 μgml− 1; anti-IL-4, 5 μgml− 1 or IL-27, 10 μgml− 1; anti-IL-4, PCR-forward GATAATTGGTTTGGTTTGTATATGGA 160 − − − PCR-reverse ACACCCTACAAAACACTCTAAAAT 5 μgml 1), Th2 (IL-4, 10 μgml 1; anti-IFN-γ,5μgml 1) and Th17 − − Sequencing GTTTGGTTTGTATATGGAT (transforming growth factor-β, 5 ng ml 1; IL-6, 10 ng ml 1; anti-IFN-γ, − 1 − 1 (CpG1) 5 μgml ; anti-IL-4, 5 μgml ). 5-Aza-2′-deoxycytidine (0.5 μM) was added PCR-forward TTGTAGGGTGTATTTAGTGTGT 222 at the beginning of culture. For intracellular cytokine staining, the T cells PCR-reverse CCAAAAACAACTCTATATAAAAATAACTT were stimulated for 4 h with phorbol 12-myristate 13-acetate and Sequencing TTTTGATGTGATTAGATTAATAGTA ionomycin in the presence of monensin. All chemicals were from Sigma- (CpG3–5) Aldrich (St Louis, MO, USA). Cells were stained with anti-CD4 (RM4-5), anti- IFN-γ (XMG1.2), anti-IL-4, anti-IL-17 (TC11-18H10.1), anti-TIM-3 (B8.2C12) and isotype antibodies (BioLegend, San Diego, CA, USA). Cells were CONFLICT OF INTEREST analyzed using a FACSCalibur flow cytometer and CellQuest software (BD The authors declare no conflict of interest. Pharmingen). For the analysis of the DNA methylation level of the TIM-3 promoter after each round of polarization, cells were polarized for 3 days − 1 and rested for 2 days in the presence of IL-2 (5 ng ml ) before the next ACKNOWLEDGEMENTS round of polarization. Viable cells from each round of polarization were isolated using Histopaque-1077 (Sigma-Aldrich) density gradient This work was supported by the Ministry of Science and Technology, Taiwan, ROC sedimentation. (MOST 103-2320-B-016-017-MY3 and MOST 104-2320-B-016-014-MY3 to H-KS; NSC 102-2321-B-016-005-MY3 to F-CC), and Tri-service General Hospital foundation (TSGH-C103-005-007-009-S01 and TSGH-C104-008-S02 to H-KS). We thank Drs Maja Quantitative reverse transcription PCR analysis Klug and Michael Rehli for kindly providing us the CpG-free plasmids. After Th cells were polarized for 3 days, RNA was extracted using a NucleoSpin RNA kit (Macherey-Nagel GmbH, Duren, Germany) and converted to cDNA using a SuperScript III synthesis kit (Invitrogen, REFERENCES Carlsbad, CA, USA) according to the manufacturer’s instructions. The 1 Monney L, Sabatos CA, Gaglia JL, Ryu A, Waldner H, Chernova T et al. Th1-specific primers used for quantitative reverse transcription PCR using the SYBR cell surface protein TIM-3 regulates macrophage activation and severity of an Green method are listed in Table 2. autoimmune disease. Nature 2002; 415:536–541. 2 Zhu C, Anderson AC, Schubart A, Xiong H, Imitola J, Khoury SJ et al. The TIM-3 DNA extraction, bisulfite conversion and pyrosequencing analysis ligand galectin-9 negatively regulates T helper type 1 immunity. Nat Immunol 2005; 6:1245–1252. Genomic DNA was extracted using a QIAamp DNA Mini kit (Qiagen, Hilden, 3 Kashio Y, Nakamura K, Abedin MJ, Seki M, Nishi N, Yoshida N et al. Galectin-9 Germany). One microgram of genomic DNA was subjected to sodium induces apoptosis through the calcium-calpain-caspase-1 pathway. J Immunol fi bisul te treatment using an EZ DNA Methylation kit (Zymo Research, 2003; 170: 3631–3636. Orange, CA, USA) and the treated genomic DNA was used for 4 Chou FC, Shieh SJ, Sytwu HK. Attenuation of Th1 response through galectin-9 and pyrosequencing. The CpG sites within the CpG island or the minimal T-cell Ig mucin 3 interaction inhibits autoimmune diabetes in NOD mice. Eur J fi fi promoter region were rst ampli ed by PCR using various primer sets and Immunol 2009; 39: 2403–2411. the PCR product was used for pyrosequencing analysis. PCR was 5 Seki M, Oomizu S, Sakata KM, Sakata A, Arikawa T, Watanabe K et al. Galectin-9 performed in a 20-μl reaction mixture containing 1-μl-bisulfite-converted suppresses the generation of Th17, promotes the induction of regulatory T cells, DNA, 2 × RBC SensiZyme HotStart Taq Mastermix (RBC Bioscience Corp., and regulates experimental autoimmune arthritis. Clin Immunol 2008; 127:78–88. Taipei, Taiwan) and primers, using the following program: 95 °C for 15 min, 6 Niwa H, Satoh T, Matsushima Y, Hosoya K, Saeki K, Niki T et al. Stable form of then 49 cycles of 95 °C for 30 s, followed by 62 °C for 30 s and 72 °C for galectin-9, a TIM-3 ligand, inhibits contact hypersensitivity and psoriatic reactions: 30 s, with a final extension at 72 °C for 10 min. The primer sequences for a potent therapeutic tool for Th1- and/or Th17-mediated skin inflammation. Clin pyrosequencing are listed in Table 3. Immunol 2009; 132: 184–194.

© 2016 Macmillan Publishers Limited Genes and Immunity (2016) 179 – 186 DNA methylation controls TIM-3 expression F-C Chou et al 186 7 Chou FC, Kuo CC, Wang YL, Lin MH, Linju Yen B, Chang DM et al. Overexpression 31 Sanchez-Fueyo A, Tian J, Picarella D, Domenig C, Zheng XX, Sabatos CA et al. of galectin-9 in islets prolongs grafts survival via downregulation of Th1 TIM-3 inhibits T helper type 1-mediated auto- and alloimmune responses and responses. Cell Transplant 2013; 22: 2135–2145. promotes immunological tolerance. Nat Immunol 2003; 4: 1093–1101. 8 He W, Fang Z, Wang F, Wu K, Xu Y, Zhou H et al. Galectin-9 significantly prolongs 32 Sabatos CA, Chakravarti S, Cha E, Schubart A, Sanchez-Fueyo A, Zheng XX et al. the survival of fully mismatched cardiac allografts in mice. Transplantation 2009; Interaction of TIM-3 and TIM-3 ligand regulates T helper type 1 responses and 88:782–790. induction of peripheral tolerance. Nat Immunol 2003; 4:1102–1110. 9 Wang F, He W, Yuan J, Wu K, Zhou H, Zhang W et al. Activation of TIM-3-galectin-9 33 Vali B, Jones RB, Sakhdari A, Sheth PM, Clayton K, Yue FY et al. HCV-specific pathway improves survival of fully allogeneic skin grafts. Transpl Immunol 2008; T cells in HCV/HIV co-infection show elevated frequencies of dual TIM-3/PD-1 19:12–19. expression that correlate with liver disease progression. Eur J Immunol 2010; 40: 10 Koguchi K, Anderson DE, Yang L, O'Connor KC, Kuchroo VK, Hafler DA. Dysre- 2493–2505. gulated T cell expression of TIM3 in multiple sclerosis. J Exp Med 2006; 203: 34 Klibi J, Niki T, Riedel A, Pioche-Durieu C, Souquere S, Rubinstein E et al. Blood 1413–1418. diffusion and Th1-suppressive effects of galectin-9-containing exosomes released 11 Li H, Wu K, Tao K, Chen L, Zheng Q, Lu X et al. TIM-3/galectin-9 signaling pathway by Epstein-Barr virus-infected nasopharyngeal carcinoma cells. Blood 2009; 113: mediates T-cell dysfunction and predicts poor prognosis in patients with hepatitis 1957–1966. B virus-associated hepatocellular carcinoma. Hepatology 2012; 56:1342–1351. 35 Anderson AC, Anderson DE, Bregoli L, Hastings WD, Kassam N, Lei C et al. Pro- 12 Liberal R, Grant CR, Holder BS, Ma Y, Mieli-Vergani G, Vergani D et al. The impaired motion of tissue inflammation by the immune receptor TIM-3 expressed on immune regulation of autoimmune hepatitis is linked to a defective galectin-9/ innate immune cells. Science 2007; 318:1141–1143. – tim-3 pathway. Hepatology 2012; 56:677 686. 36 Chen Y, Langrish CL, McKenzie B, Joyce-Shaikh B, Stumhofer JS, McClanahan 13 Jones RB, Ndhlovu LC, Barbour JD, Sheth PM, Jha AR, Long BR et al. TIM-3 T et al. Anti-IL-23 therapy inhibits multiple inflammatory pathways and amelio- fi expression de nes a novel population of dysfunctional T cells with highly ele- rates autoimmune encephalomyelitis. J Clin Invest 2006; 116: 1317–1326. – vated frequencies in progressive HIV-1 infection. J Exp Med 2008; 205: 2763 2779. 37 Penix LA, Sweetser MT, Weaver WM, Hoeffler JP, Kerppola TK, Wilson CB. The 14 McMahan RH, Golden-Mason L, Nishimura MI, McMahon BJ, Kemper M, Allen TM proximal regulatory element of the interferon-gamma promoter mediates fi et al. TIM-3 expression on PD-1+ HCV-speci c human CTLs is associated with viral selective expression in T cells. J Biol Chem 1996; 271: 31964–31972. persistence, and its blockade restores hepatocyte-directed in vitro cytotoxicity. 38 Bruniquel D, Schwartz RH. Selective, stable demethylation of the interleukin-2 – J Clin Invest 2010; 120: 4546 4557. gene enhances transcription by an active process. Nat Immunol 2003; 4: 15 Golden-Mason L, Palmer BE, Kassam N, Townshend-Bulson L, Livingston S, 235–240. McMahon BJ et al. Negative immune regulator TIM-3 is overexpressed on T cells 39 Fitzpatrick DR, Shirley KM, McDonald LE, Bielefeldt-Ohmann H, Kay GF, Kelso A. in hepatitis C virus infection and its blockade rescues dysfunctional CD4+ and Distinct methylation of the interferon gamma (IFN-gamma) and interleukin 3 CD8+ T cells. J Virol 2009; 83: 9122–9130. (IL-3) genes in newly activated primary CD8+ T lymphocytes: regional IFN-gamma 16 Zhang J, Daley D, Akhabir L, Stefanowicz D, Chan-Yeung M, Becker AB et al. Lack promoter demethylation and mRNA expression are heritable in CD44(high)CD8+ of association of TIM3 polymorphisms and allergic phenotypes. BMC Med Genet T cells. J Exp Med 1998; 188:103–117. 2009; 10: 62. 40 Makar KW, Wilson CB. DNA methylation is a nonredundant repressor of the Th2 17 Chae SC, Park YR, Lee YC, Lee JH, Chung HT. The association of TIM-3 gene effector program. J Immunol 2004; 173: 4402–4406. polymorphism with atopic disease in Korean population. Hum Immunol 2004; 65: 41 Sellars M, Huh JR, Day K, Issuree PD, Galan C, Gobeil S et al. Regulation of DNA 1427–1431. methylation dictates Cd4 expression during the development of helper and 18 Chae SC, Park YR, Shim SC, Yoon KS, Chung HT. The polymorphisms of Th1 cell cytotoxic T cell lineages. Nat Immunol 2015; 16:746–754. surface gene TIM-3 are associated in a Korean population with rheumatoid 42 Lee PP, Fitzpatrick DR, Beard C, Jessup HK, Lehar S, Makar KW et al. A critical role arthritis. Immunol Lett 2004; 95:91–95. for Dnmt1 and DNA methylation in T cell development, function, and survival. 19 Gao PS, Mathias RA, Plunkett B, Togias A, Barnes KC, Beaty TH et al. Genetic Immunity 2001; 15: 763–774. variants of the T-cell immunoglobulin mucin 1 but not the T-cell immunoglobulin 43 Makar KW, Perez-Melgosa M, Shnyreva M, Weaver WM, Fitzpatrick DR, Wilson CB. mucin 3 gene are associated with asthma in an African American population. Active recruitment of DNA methyltransferases regulates interleukin 4 in thymo- J Allergy Clin Immunol 2005; 115:982–988. – 20 Zhang CC, Wu JM, Cui TP, Wang P, Pan SX. [Study on relationship between cytes and T cells. Nat Immunol 2003; 4: 1183 1190. fi polymorphism sites of TIM-3 and allergic asthma in a population of adult Hans 44 Gamper CJ, Agoston AT, Nelson WG, Powell JD. Identi cation of DNA methyl- from Hubei province of China]. Zhonghua Yi Xue Yi Chuan Xue Za Zhi 2006; 23: transferase 3a as a T cell receptor-induced regulator of Th1 and Th2 differentia- – 74–77. tion. J Immunol 2009; 183: 2267 2276. 21 Bruck P, Ramos-Lopez E, Bartsch W, Bohme A, Badenhoop K. TIM-3 polymorph- 45 Baron U, Floess S, Wieczorek G, Baumann K, Grutzkau A, Dong J et al. DNA isms in type 1 diabetes families. J Hum Genet 2008; 53:559–564. demethylation in the human FOXP3 locus discriminates regulatory T cells from – 22 Anderson AC, Lord GM, Dardalhon V, Lee DH, Sabatos-Peyton CA, Glimcher LH activated FOXP3(+) conventional T cells. Eur J Immunol 2007; 37: 2378 2389. et al. T-bet, a Th1 transcription factor regulates the expression of TIM-3. 46 Ohkura N, Hamaguchi M, Morikawa H, Sugimura K, Tanaka A, Ito Y et al. T cell Eur J Immunol 2010; 40: 859–866. receptor stimulation-induced epigenetic changes and Foxp3 expression are 23 Wilson CB, Rowell E, Sekimata M. Epigenetic control of T-helper-cell differentia- independent and complementary events required for Treg cell development. – tion. Nat Rev Immunol 2009; 9:91–105. Immunity 2012; 37: 785 799. 24 Baumjohann D, Ansel KM. MicroRNA-mediated regulation of T helper cell differ- 47 Youngblood B, Oestreich KJ, Ha SJ, Duraiswamy J, Akondy RS, West EE et al. entiation and plasticity. Nat Rev Immunol 2013; 13:666–678. Chronic virus infection enforces demethylation of the locus that encodes PD-1 in 25 Bird JJ, Brown DR, Mullen AC, Moskowitz NH, Mahowald MA, Sider JR et al. Helper antigen-specific CD8(+) T cells. Immunity 2011; 35:400–412. T cell differentiation is controlled by the cell cycle. Immunity 1998; 9:229–237. 48 Gorelik G, Richardson B. Aberrant T cell ERK pathway signaling and chromatin 26 Young HA, Ghosh P, Ye J, Lederer J, Lichtman A, Gerard JR et al. Differentiation of structure in lupus. Autoimmun Rev 2009; 8:196–198. the T helper phenotypes by analysis of the methylation state of the IFN- 49 Gorelik G, Fang JY, Wu A, Sawalha AH, Richardson B. Impaired T cell protein kinase gamma gene. J Immunol 1994; 153: 3603–3610. C delta activation decreases ERK pathway signaling in idiopathic and hydralazine- 27 Valapour M, Guo J, Schroeder JT, Keen J, Cianferoni A, Casolaro V et al. Histone induced lupus. J Immunol 2007; 179: 5553–5563. deacetylation inhibits IL4 gene expression in T cells. J Allergy Clin Immunol 2002; 50 Deng C, Kaplan MJ, Yang J, Ray D, Zhang Z, McCune WJ et al. Decreased Ras- 109: 238–245. mitogen-activated protein kinase signaling may cause DNA hypomethylation in T 28 Klug M, Rehli M. Functional analysis of promoter CpG methylation using a lymphocytes from lupus patients. Arthritis Rheum 2001; 44:397–407. CpG-free luciferase reporter vector. Epigenetics 2006; 1:127–130. 51 Ichiyama K, Chen T, Wang X, Yan X, Kim BS, Tanaka S et al. The methylcytosine 29 Owaki T, Asakawa M, Morishima N, Hata K, Fukai F, Matsui M et al. A role for IL-27 dioxygenase tet2 promotes DNA demethylation and activation of cytokine gene in early regulation of Th1 differentiation. J Immunol 2005; 175: 2191–2200. expression in T cells. Immunity 2015; 42:613–626. 30 Kanno Y, Vahedi G, Hirahara K, Singleton K, O'Shea JJ. Transcriptional and 52 Loots GG, Ovcharenko I, Pachter L, Dubchak I, Rubin EM. rVista for comparative epigenetic control of T helper cell specification: molecular mechanisms under- sequence-based discovery of functional transcription factor binding sites. Genome lying commitment and plasticity. Annu Rev Immunol 2012; 30: 707–731. Res 2002; 12:832–839.

Supplementary Information accompanies this paper on Genes and Immunity website (http://www.nature.com/gene)