Article Vitamin D receptor, STAT3, and TET2 cooperate to establish tolerogenesis Graphical abstract Authors Francesc Català-Moll, Anna G. Ferreté-Bonastre, Gerard Godoy-Tena, ..., Javier Rodrı́guez-Ubreva, Tianlu Li, Esteban Ballestar Correspondence eballestar@carrerasresearch.org In brief Català-Moll et al. show that vitamin D induces DNA demethylation and transcriptional activation at VDR binding sites, as well as IL-6-JAK-STAT3 pathway activation associated with acquisition of tolerogenesis by dendritic cells. VDR, STAT3, and TET2 interact with each other. Pharmacological inhibition of JAK2 reverts vitamin D-induced tolerogenic properties of DCs. Highlights d Vitamin D induces DNA demethylation at VDR binding sites in dendritic cells (DCs) d Differentiation to tolerogenic DCs associates with IL-6-JAK- STAT3 pathway activation d VDR, STAT3, and TET2 interact with each other in tolerogenic DCs d Pharmacological inhibition of JAK2 reverts vitamin D-induced DC tolerogenesis Català-Moll et al., 2022, Cell Reports 38, 110244 January 18, 2022 ª 2021 The Author(s). https://doi.org/10.1016/j.celrep.2021.110244 ll ll OPEN ACCESS Article Vitamin D receptor, STAT3, and TET2 cooperate to establish tolerogenesis Francesc Català-Moll,1,4 Anna G. Ferreté-Bonastre,1,4 Gerard Godoy-Tena,1,4 Octavio Morante-Palacios,1 Laura Ciudad,1 Laura Barberà,1 Federico Fondelli,2,3 Eva M. Martı́nez-Cáceres,2,3 Javier Rodrı́guez-Ubreva,1 Tianlu Li,1,4 and Esteban Ballestar1,5,* 1Epigenetics and Immune Disease Group, Josep Carreras Research Institute (IJC), Badalona, 08916 Barcelona, Spain 2Division of Immunology, Germans Trias i Pujol Hospital, LCMN, Germans Trias i Pujol Research Institute (IGTP), Badalona, 08916 Barcelona, Spain 3Department of Cell Biology, Physiology, Immunology, Autonomous University of Barcelona, Bellaterra, 08193 Barcelona, Spain 4These authors contributed equally 5Lead contact *Correspondence: eballestar@carrerasresearch.org https://doi.org/10.1016/j.celrep.2021.110244 SUMMARY The active form of vitamin D, 1,25-dihydroxyvitamin D3, induces a stable tolerogenic phenotype in dendritic cells (DCs). This process involves the vitamin D receptor (VDR), which translocates to the nucleus, binds its cognate genomic sites, and promotes epigenetic and transcriptional remodeling. In this study, we report the occurrence of vitamin D-specific DNA demethylation and transcriptional activation at VDR binding sites associated with the acquisition of tolerogenesis in vitro. Differentiation to tolerogenic DCs associates with activation of the IL-6-JAK-STAT3 pathway. We show that JAK2-mediated STAT3 phosphorylation is specific to vitamin D stimulation. VDR and the phosphorylated form of STAT3 interact with each other to form a complex with methylcytosine dioxygenase TET2. Most importantly, pharmacological inhibition of JAK2 re- verts vitamin D-induced tolerogenic properties of DCs. This interplay among VDR, STAT3, and TET2 opens up possibilities for modulating DC immunogenic properties in clinics. INTRODUCTION compartment, can interact with a variety of TFs, such as PU.1, C/ EBPa, KLF4, and others, in order to facilitate their recruitment to Dendritic cells (DCs) are a heterogeneous group of innate im- different genomic regions (Costa et al., 2013; Guilhamon et al., mune cells that have a key role in initiating adaptive responses. 2013; de la Rica et al., 2013; Lio et al., 2016; Mendes et al., Also, DCs are not only central for coordinating immune re- 2021; Sardina et al., 2018; Wang et al., 2015; Xiong et al., sponses against a threat but also needed to regulate the immune 2016). Recently, it has been demonstrated that TET2 mutations, system at steady state and for inducing immune tolerance (Mor- which are frequent in myeloid leukemias, lead to DNA hyperme- ante-Palacios et al., 2021). Like in other myeloid cell populations, thylation of enhancer regions and changes in the subsequent the immunological properties of DCs vary with the environment. binding of TFs, particularly members of the basic helix-loop-helix In general, terminal myeloid cell differentiation is highly depen- (bHLH) TF family (Rasmussen et al., 2019). This suggests that dent on the activation of specific signaling pathways in response TET2 recruitment by TFs leads to epigenetic remodeling that fa- to extracellular signals, such as inflammatory cytokines, hor- cilitates the binding of subsequent TFs (Rasmussen et al., 2019). mones, vitamins, and other factors (Álvarez-Errico et al., 2015), Moreover, a reciprocal relationship between DNA methylation which determine the immunogenicity of the resulting myeloid and histone modifications has long been established. TET2 has cells. The activation of signaling pathways leads to the activation been not only described to modulate trimethylation of K4 of his- of specific sets of transcription factors (TFs). Sequence-specific tone H3 (H3K4me3) (Deplus et al., 2013), a mark of active tran- DNA binding of TFs is a pivotal process for establishing gene scription, but also shown to coordinate trimethylation of K27 of expression patterns in concert with the epigenetic machinery histone H3 (H3K27me3), a mark of heterochromatin, in an in- that determines cell identity and function (Monticelli and Natoli, verse manner (Ichiyama et al., 2015). 2017). Recent evidence has shown that several TFs are associ- Calcitriol (1,25-dihydroxyvitamin D3), the active form of vitamin ated with DNA demethylation to increase genomic accessibility D3 (henceforth referred to as vitamin D), is a major modulator of of their binding genomic regions, thus facilitating the binding of the immune system (Barragan et al., 2015; Carlberg, 2019; Mora subsequent TFs (Mahé et al., 2017). In this regard, methylcyto- et al., 2008). DCs are the most susceptible cell type to vitamin D sine dioxygenase ten-eleven translocation (TET2), the most rele- in a mixed immune population (Mora et al., 2008). In these cells, vant enzyme involved in active DNA demethylation in the myeloid vitamin D can generate a stable maturation-resistant tolerogenic Cell Reports 38, 110244, January 18, 2022 ª 2021 The Author(s). 1 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). ll OPEN ACCESS Article A B C D E G F Figure 1. DNA methylation dynamics throughout vitamin D-exposed dendritic-cell differentiation (n = 4, two independent experiments) (A) Schematic overview of the differentiation model from human peripheral blood MOs to DCs and TolDCs. (B) Principal-component analysis of differentially methylated CpGs. (C) DNA methylation heatmap and cluster analysis of four paired samples of MOs and their derived DCs and TolDCs at day 5 of differentiation. The heatmap includes all CpG-containing probes displaying significant methylation changes (differential beta value R0.2 and q value < 0.05) only in the TolDC-DC comparison. The color annotation of the lateral bar represents the membership to cluster 1 in red (DC-specific DNA demethylation), cluster 2 in green (TolDC-specific DNA demethylation), cluster 3 in orange (TolDC-specific DNA hypermethylation), and cluster 4 in blue (DC-specific DNA hypermethylation). (D) Box and violin plots summarizing the distribution of DNA methylation levels per cell type and cluster. (E) Gene ontology (GO) terms associated with CpGs from cluster 1 (red) and cluster 2 (green) as analyzed by GREAT software. Bars represent log-transformed binomial q values of the GO term enrichment. (F) Location proportions of CpGs from each cluster in the context of CpG islands (CGIs) (right) and gene-related regions (left). (legend continued on next page) 2 Cell Reports 38, 110244, January 18, 2022 ll Article OPEN ACCESS phenotype in vitro, with a low level of expression of immunogenic In this study, we studied epigenetic determinants critical for molecules, such as HLA-DR, CD80, and CD86, and increased the acquisition of tolerogenic properties during in vitro human interleukin (IL)-10/IL-12p70 ratios that are maintained even after MO-derived DC differentiation in the presence of vitamin D. We removal of the compound (Van Halteren et al., 2002). After ligand demonstrate an interplay between VDR and the Janus kinase recognition, vitamin D receptor (VDR) translocates to the nucleus (JAK) 2/signal transducer and activator of transcription (STAT) and acts not only as a TF, controlling the expression of a set of im- 3 pathway associated with the generation of a specific TET- mune and metabolic genes (Carlberg, 2019; Ferreira et al., 2013), dependent DNA demethylation signature in TolDCs. It involves but also as a repressor of nuclear factor kappa-light-chain- a direct physical interaction between VDR, STAT3, and TET2 enhancer of activated B cells (NF-kB) at different levels (Carlberg, that leads to the acquisition and stabilization of the tolerogenic 2019; Fetahu et al., 2014). Several studies have shown the capac- properties of DCs in the presence of vitamin D. ity of VDR to interact with a range of TFs, including PU.1 and GABPA, and with chromatin remodeling and histone modification RESULTS enzymes, such as BRD7 and KDM6B (Pereira et al., 2011; Seuter et al., 2017, 2018; Wei et al., 2018). Previous work has shown that Vitamin D induces the acquisition of a specific DNA vitamin D may induce DNA methylation changes in myeloid cells methylation profile associated with tolerogenesis (Ong et al., 2021). However, the molecular mechanism that leads during in vitro DC differentiation to the acquisition of differential methylation patterns remains To investigate the effects of vitamin D in DNA methylation during unexplored. the acquisition of tolerogenic properties by DCs, we first differen- Vitamin D supplementation is generally used as a preventive tiated in vitro peripheral blood MOs from human donors to DCs agent or a co-adjuvant for diseases with underlying autoimmune and TolDCs for 6 days using GM-CSF and IL-4 in the absence or pro-inflammatory states (Bscheider and Butcher, 2016; and presence of vitamin D, respectively (Figure 1A). As previously Dankers et al., 2017). DCs represent an excellent target of described (Penna and Adorini, 2000; Piemonti et al., 2000), vitamin D to dampen autoimmunity and inflammation, not only TolDCs had higher levels of the surface markers CD14 and because these myeloid cells express the whole set of enzymes CD11b and lower levels of HLA-DR, CD1a, and CD86 than did to generate the active form of vitamin D (Mora et al., 2008) but DCs (Figure S1A). To confirm the resemblance between our also because of their unique role as initiators of immune re- in vitro model with in vivo DCs, we integrated the expression pro- sponses. However, the role of DCs in vitamin D-mediated files of MOs, DCs (12 h and 120 h), and TolDCs (12 h and 120 h) immunomodulation is not fully understood. In addition, DCs (Széles et al., 2009) with previously published expression datasets with tolerogenic function (TolDCs) have become a promising (Goudot et al., 2017; Segura et al., 2013) from MOs, in vitro- immunotherapeutic tool for reinstating immune tolerance in derived DCs and macrophages (MACs), and in vivo DCs and autoimmune diseases and in allogeneic bone marrow and solid MACs. According to t-distributed stochastic neighbor embedding organ transplantation (Morante-Palacios et al., 2021). The stabil- (t-SNE) analysis, TolDCs (differentiated in the presence of vitamin ity of the tolerogenic phenotype suggests that regulatory mech- D), among different DC subsets, are the ones nearer different MAC anisms that allow the maintenance of stable changes of gene types with immunosuppressive phenotypes (Figure S1B). expression are involved. In this sense, DNA methylation is a ma- In concordance with previous studies (Piemonti et al., 2000), we jor epigenetic modification closely involved in the acquisition or observed that TolDCs were able to inhibit CD8+ T cell proliferation stabilization of transcriptional states (Luo et al., 2018). Peripheral in vitro, in contrast to DCs, confirming their immunosuppressive blood monocyte (MO)-derived DCs represent a useful model for properties (Figure S1C). Furthermore, we also observed increased studying the properties of DCs. It has been previously described levels of VDR in the nucleus following vitamin D exposure, in that DCs differentiated from isolated MOs by the addition of agreement with previous studies, suggesting that VDR preferen- granulocyte-macrophage colony-stimulating factor (GM-CSF) tially acts in the nucleus (Figure S1D). Altogether, our results and IL-4 in vitro closely resemble CD1c+ DCs at the transcrip- confirmed the validity of this in vitro model to generate and study tional level (Goudot et al., 2017). Exposure of MO-derived DCs TolDCs by the involvement of VDR through vitamin D exposure. to vitamin D results in the inhibition of differentiation and matura- We then obtained and compared the DNA methylation profiles tion into potent antigen-presenting cells and gain in the capacity of MOs, DCs, and TolDCs using BeadChip arrays (see STAR to inhibit T cell proliferation (Piemonti et al., 2000). Similarly, Methods), which interrogate the methylation status of CD1c+ DCs cultured in vitro with vitamin D for 2 days acquire a >850,000 CpG positions across the entire genome, covering typical semi-mature phenotype after exposure to a DC matura- 99% of the reference sequence genes. Principal-component tion cocktail, with low CD83 expression, and a tolerogenic analysis (PCA) showed that most of the variability observed at phenotype, as they suppressed alloimmunity in vivo, in a mouse the DNA methylation level may be explained by events common model (Chu et al., 2012). to the two differentiation processes (principal component 1; (G) Bubble chart depicting the enrichment (red) or depletion (blue) of the CpGs from each cluster in the chromatin states from DCs (Pacis et al., 2015). The circle filling color represents the logarithmic value of the ratio between the percentage of CpGs with the feature in each cluster and the percentage of CpGs with the feature in the background. Circle size indicates the percentage of CpGs from each cluster in the chromatin state, and the circle edge indicates the statistical significance of the enrichment (black: significant; no edge: not significant; q value < 0.01). Statistical tests: paired two-tailed t test (D), Pearson correlation (E), and two-tailed Fisher’s exact test (F and G) (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001, ns = not significant). FDR, false discovery rate. Cell Reports 38, 110244, January 18, 2022 3 ll OPEN ACCESS Article Figure 1B). However, the second principal component is capable and highly dependent on genomic location. CpGs that under- of clustering DCs and TolDCs separately (Figure 1B). Differenti- went TolDC-specific DNA demethylation during differentiation ation mainly resulted in DNA demethylation in which there were were largely situated in open seas corresponding to enhancers; both condition-specific demethylation events and demethylation hence, it is plausible to envision that they control gene expres- events common to both differentiation processes (Figure S1E). A sion, which results in the final tolerogenic phenotype. We small proportion of DNA methylation changes was attributed to therefore integrated our DNA methylation dataset with publicly gains of DNA methylation during differentiation (Figure S1E). available expression data generated in the same in vitro models Hierarchical clustering of differentially methylated CpGs be- (Széles et al., 2009). We observed a significant inverse relation- tween DCs and TolDCs (adjusted p < 0.05 and absolute differen- ship between levels of DNA methylation and mRNA expression tial beta R 0.2) revealed four main groups of CpG sites (Figures at 12 h (r = 0.5926; p = 4.90e-14) and 5 days of differentiation 1C and 1D and Table S1): a group of CpGs that underwent spe- (r = 0.4108; p = 4.57e-11) (Figure 2A). Furthermore, dividing cific demethylation in DCs (cluster 1: 429 CpGs); a second group cluster 1 and 2 CpGs based on their genomic location in relation that was specifically demethylated in TolDCs (cluster 2: 311 to previously identified enhancer regions (Pacis et al., 2015), we CpGs); another group that gained methylation in TolDCs (cluster observed that genes associated with cluster 1 CpGs located at 3: 36 CpGs); and finally a group of CpGs with DC-specific gains active enhancers of DCs displayed higher expression levels in in DNA methylation (cluster 4: 28 CpGs). DCs than in TolDCs (Figure 2B). To confirm these observations in the context of in vivo circu- To explore the dynamics of the relationship between DC- lating DCs, we obtained the DNA methylation profiles of whole (cluster 1) and TolDC-specific demethylation (cluster 2), we per- blood-isolated cDCs (CD1c+ DCs) cultured in the absence and formed bisulfite pyrosequencing and qRT-PCR in a selected presence of vitamin D for 3 days and observed that, similar to group of genes of a set of samples over time. A few genes TolDCs generated in vitro, cDCs exposed to vitamin D under- from each cluster were selected for further analysis based on went DNA demethylation in cluster 2 CpGs (Figure S1F). This the conditions that they had the maximum possible difference confirmed that demethylation observed in cluster 2 CpGs were in DNA methylation during differentiation within their corre- specific to vitamin D exposure. sponding cluster, that they were differentially expressed, and Functional gene ontology (GO) analysis revealed that CpGs in that there were previous reports relating them with relevant im- cluster 1 are associated with immunological categories, such as mune properties. For instance, from cluster 1, we chose IRF4 defense and immune response, whereas those in cluster 2 are and C1QB, which are important for normal DC differentiation more highly enriched in cell activation, positive regulation of im- from MOs (Teh et al., 2011; Murphy et al., 2016), and from cluster mune system process, and wound healing involved in inflamma- 2, CD14 and DPF3 were selected for being specific markers of tory response (Figure 1E). For clusters 3 and 4, GO analysis did TolDCs and being involved in DC chemotaxis, respectively (Liu not show enrichment in any functional categories, probably due et al., 2019; Torres-Aguilar et al., 2010). Bisulfite pyrosequencing to their small size. In all clusters, the majority of changes of these genes showed a high concordance (r = 0.978; p < 2.2 3 occurred in introns and intergenic regions with underrepresenta- 10 16) with the data obtained from the EPIC arrays (Figure 2C). tion of promoter-transcriptional start sites (TSSs). However, DC-specific (cluster 1) genes, such as IRF4 and C1QB, were up- whereas cluster 1 exhibited a marked enrichment of intronic re- regulated in DCs in parallel with their specific DNA demethylation gions with respect to background, the other clusters were en- (Figure 2D). Similarly, for TolDC-specific (cluster 2) genes, such riched in both intronic and intergenic locations (Figure 1F, left). as CD14 and DPF3, transcript upregulation occurred only in Concordantly, CpGs of all clusters were observed to be located TolDCs in parallel with their corresponding DNA demethylation outside of CpG islands, particularly for cluster 2 (Figure 1F, right). (Figure 2E). In agreement with previous reports, stimulus- Next, we mapped the chromatin states of the CpG sites under- induced DNA demethylation occurred succeeding specific going changes in methylation in the four clusters using chromatin gene expression changes (Pacis et al., 2019). In all, our results segmentation data generated in DCs (Pacis et al., 2015) (Fig- suggested that vitamin D-driven DNA demethylation occurred ure 1G). We observed an enrichment in enhancer regions for all in association with upregulation of TolDC-specific genes. clusters and an enrichment for inactive promoters for cluster 4. To further characterize the mechanisms driving DNA demethy- Moreover, cluster 1 (DC-specific demethylation) was enriched lation during MO-to-DC and MO-to-TolDC differentiation, we in weak (H3K27ac + H3K4me1 + H3K4me3) and strong next investigated whether the demethylation was due to active (H3K27ac + H3K4me1) enhancers, while cluster 2 (TolDC-spe- demethylation or replication-mediated passive demethylation. cific demethylation) was more enriched in inactive enhancers Utilizing BrdU proliferation assay, no proliferation was observed (H3K4me1) in DCs, suggesting that these inactive regions in in DCs and TolDCs up to 6 days of differentiation (Figure S2A); DCs are activated in TolDCs. In all, our results indicated that hence, all DNA demethylation events observed were driven by vitamin D-driven demethylation events occurred in regions that active demethylation. In this regard, we and others have previ- may play important roles in regulating gene expression and ously shown that loss of methylation in terminal differentiation establishing the tolerogenic phenotype of TolDCs. from MOs is accompanied by a transient increase in 5-hydroxy- methylcytosine (5hmC) and involves the participation of TET2 DNA demethylation in TolDCs is an active process and is methylcytosine dioxygenase (Garcia-Gomez et al., 2017; Klug associated with changes in gene expression et al., 2013). We then determined the 5hmC levels of CpGs DNA methylation has long been established to influence gene that became demethylated during DC and TolDC differentiation expression (Jones, 2012), although the dynamics are complex and observed that there was indeed a gain of 5hmC in these 4 Cell Reports 38, 110244, January 18, 2022 ll Article OPEN ACCESS A B C D E Figure 2. Integration of gene expression with DNA methylation (A) Scatter plot showing the correlation between DNA methylation differences and gene expression changes between DCs and TolDCs at 12 h (top) and day 5 (bottom) of differentiation. Only differentially methylated CpGs are represented. Dot color indicates gene-related associations. (B) Box and violin plots summarizing the mRNA expression levels per cell type of genes annotated to CpGs from cluster 1 (top) and cluster 2 (bottom) divided by chromatin state annotation of the associated CpG. Adjustment for multiple comparisons was performed with false discovery rate. (C) Scatter plot showing the correlation between methylation array values and bisulfite pyrosequencing DNA methylation values (n = 4, two independent ex- periments). (D) DNA methylation (top) and mRNA expression (bottom) kinetics of two representative examples of cluster 1 genes. CpGs studied include cg10630015 (IRF4) and cg04097715 (C1QB) (n = 3, one single experiment). (E) DNA methylation (top) and mRNA expression (bottom) kinetics of two representative examples of cluster 2 genes (n = 3, two independent experiments). CpGs studied include cg05620710 (CD14) and cg25205844 (DPF3). Statistical tests: Pearson correlation (A) and unpaired two-tailed t test (B, D, and E) (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). Cell Reports 38, 110244, January 18, 2022 5 ll OPEN ACCESS Article CpGs (Figure S2B). Finally, utilizing publicly available DNase- 2010). The dynamics of DNA methylation and gene expression sequencing (seq) datasets from MOs (Feingold et al., 2004), we of these two genes confirmed specific DNA demethylation in observed that more than 75% of cluster 2 CpGs corresponded TolDC, and differential gene expression changes in relation to to closed chromatin in MOs (Figure S2C), which reinforced the DCs (Figures 4E and 4F). hypothesis that DNA demethylation was mediated by an active As indicated in the introduction, TET-mediated demethylation event. Altogether, our results suggested that specific active is associated with histone modifications, such as H3K4me3 (De- DNA demethylation following vitamin D exposure is mediated plus et al., 2013) and H3K27me3 (Ichiyama et al., 2015). Hence, through methylcytosine dioxygenase activity, most likely associ- we speculated that changes in DNA methylation were accompa- ated with TET2. nied by changes in histone modifications, and their dynamics might be associated with VDR recruitment following vitamin D VDR binding is associated with DNA demethylation and exposure. Therefore, we performed ChIP-qPRC of VDR together active chromatin during MO-to-TolDC differentiation with these activating (H3K4me3) and repressive (H3K27me3) In concordance with previous work (Jakob et al., 1992), we histone modifications. We also added an antibody against H3 observed that exposure to vitamin D during DC differentiation acetylation (H3ac), characteristic of active chromatin. To increased the nuclear levels of VDR (Figure S1D). Hence, it is plau- discriminate between the effects of a tolerogenic phenotype ac- sible that liganded VDR plays a direct role in driving DNA demethy- quired through a 6-day differentiation and the effects directly lation following vitamin D exposure during TolDC differentiation. caused by the presence of vitamin D in the medium, we per- Hence, we performed chromatin immunoprecipitation (ChIP)- formed ChIPs in MOs, DCs, TolDCs, and also DCs treated with seq analysis of VDR in DCs and TolDCs. First, we observed that vitamin D for 30 min (DC + vitD). First, we observed a significant exposure to vitamin D during TolDC differentiation led to a sharp increase in VDR binding (Figure 4G) in DCs treated with vitamin D increase in overall VDR genomic binding (Figures 3A and 3B). and in TolDCs. Second, in the aforementioned cluster 2 genes Interestingly, motif discovery analysis revealed promiscuity of GAB2 and HIF1A, we only observed a significant increase, asso- VDR with respect to its genomic binding preferences, with only ciated with VDR binding, for H3ac (Figure 4G). This finding was 37% of regions having the canonical VDR binding motif (Fig- extendable to other cluster 2 genes, such as HOPX, IL6, INHBA, ure 3C), which suggests the cooperation of VDR with other TFs and LYRM1 (Figure S3C). during TolDC differentiation. Second, functional annotation of Hence, altogether, our data suggested the coordination be- VDR-bound genes revealed enrichment of immune- and tween VDR binding, specific DNA demethylation, changes in signaling-related categories, such as myeloid and granulocyte histone H3 acetylation, and gene expression upregulation in activations and cytokine receptor activity (Figure 3D). In fact, TolDC differentiation. several genes previously described to be related to the tolero- genic properties of TolDCs, such as IL10, ANXA1, and CD163 Differentiation to DCs in the presence of vitamin D (Navarro-Barriuso et al., 2018), are direct targets of VDR (Table associates with activation of IL-6-JAK-STAT3 signaling S2). Third, global inspection of VDR genomic occupancy pathway, and both VDR and STAT3 interact with TET2 showed that VDR preferentially binds to promoters and introns Vitamin D, through its receptor VDR, induces changes in cyto- in comparison with background (Figure 3E, left). We also kine production and a profound metabolic reprogramming in observed enrichment of VDR binding in CpG islands, shores, human DC (Ferreira et al., 2015). For this reason, we hypothe- and shelves, which was compatible with the enrichment noted sized that autocrine/paracrine activation of secondary in promoters (Figure 3E, right). Annotation of VDR peaks in rela- signaling pathways during differentiation could lead to the acti- tion to previously published data of DC chromatin states (Pacis vation of a set of TFs downstream to VDR that could be relevant et al., 2015) showed the preference of VDR for binding regions to TolDC differentiation. To explore this possibility, we adapted that correspond to promoters and enhancers in DCs (Figure 3F). a tool initially designed to explore intercellular communication To further characterize the relationship between VDR and DNA in bulk and single-cell expression data to test autocrine/para- methylation, we overlapped our generated DNA methylation crine signal activation (Browaeys et al., 2020). Note that our data with VDR ChIP-seq data and observed a specific enrich- differentiation model does not allow to distinguish between au- ment of VDR binding in TolDCs to CpGs that became demethy- tocrine or paracrine activation. With this approach, and using lated in TolDCs (cluster 2), and this was not observed for the genes associated with both demethylation clusters with signif- other clusters (Figures 4A, 4B, and S3A). In fact, we observed icant expression differences (fold-change <0.5 or >2, and that over 40% of CpG sites in cluster 2 had significant VDR bind- adjusted p < 0.05) as input, we inferred potential ligands that ing (Figure 4C). For instance, cluster 2 CpGs mapped to genes, may regulate these processes (Figure 5A). One of the most such as GAB2 and HIF1A, situated within the binding peaks of interesting ligands due to its role in immune suppression in VDR in TolDCs (Figure 4D) and located in closed chromatin re- the context of tumorigenesis is IL-6 (Park et al., 2017). In fact, gions in MOs (Figure S3B). These genes are of particular interest the IL6 gene is significantly overexpressed in TolDCs compared because GAB2 has been implicated in phosphatidylinositol 3-ki- with DCs (Figure 5B), and its target genes were also observed nase (PI3K) pathway activation (Pratt et al., 2000), a pathway to be overexpressed in TolDCs (Figure 5C). implicated in DC tolerogenesis (Ferreira et al., 2015). Further- We then performed gene set enrichment analysis (GSEA) of more, hypoxia-inducible factor 1-alpha (HIF1A) is a key factor differentially expressed genes between DCs and TolDCs and for the tolerogenic properties of myeloid-derived suppressor visualized that genes differentially overexpressed in TolDCs cells (MDSCs) in the tumor microenvironment (Corzo et al., were enriched in IL-6-JAK-STAT3 signaling pathway (Figure 5D). 6 Cell Reports 38, 110244, January 18, 2022 ll Article OPEN ACCESS A B C D E F Figure 3. Genomic occupancy of vitamin D receptor (ChIP-seq, n = 2, one experiment) (A) Heatmaps showing signal intensity of vitamin D receptor (VDR) ChIP-seq at ± 2.5 Kbp window of significant VDR peaks in MO, DCs and TolDCs (q value < 0.01 and irreproducible discovery rate [IDR] < 0.05). (B) Composite plots of VDR ChIP-seq distribution ±2.5 Kbp around CpGs in MO (gray), DCs (red), and TolDCs (green) for significant VDR peaks. The statistics were computed by comparing the intensity averages of the entire window. (C) Motif discovery analysis using HOMER software showing q values and the percentage of test and background regions with each motif. (D) Results of gene set enrichment analysis using GREAT software. The plot depicts the top enriched terms for biological processes (green), molecular function (orange), and cellular component (purple) categories, based on adjusted p values from the binomial distribution. (E) Location proportions of VDR peaks in the context of CpG islands (CGIs) (right) and gene-related regions (left). (F) Bubble chart depicting the enrichment (red) or depletion (blue) of VDR peaks in the chromatin states of dendritic cells (Pacis et al., 2015). The circle filling represents the logarithm of the ratio between the percentage of VDR peaks with the feature and the percentage of the feature within the background. Circle size indicates the percentage of VDR peaks in the chromatin state, and the circle edge indicates the statistical significance of the enrichment (black: significant; no edge: not significant; q value < 0.01). Statistical tests: two-tailed t test (A and B), cumulative binomial distribution (C and D), and two-tailed Fisher’s exact test (E and F) (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). In fact, VDR binds in several regions upstream of the IL6 gene compared with DCs (Figure 5B). Additionally, significant DNA de- TSS, suggesting that VDR directly regulates its expression (Fig- methylation was observed in 2 CpG sites of the promoter region ure 5E). Furthermore, we detected an increase in IL-6 production of IL6 in TolDCs, and this was coupled with a gain in 5hmC (Fig- and release into the medium in TolDCs (Figure S4A), which was ures S4B and S4C), which suggested the involvement of TET2 in concordant with an upregulation of its gene expression its regulation. Cell Reports 38, 110244, January 18, 2022 7 ll OPEN ACCESS Article A B C D E F G Figure 4. Binding of vitamin D receptor correlates with TolDC-specific DNA demethylation (ChIP-seq, n = 2, one single experiment; DNA methylation, n = 4, two independent experiments) (A) Heatmaps showing signal intensity of vitamin D receptor (VDR) ChIP-seq at ± 2.5 Kbp window from CpGs of cluster 1 (top) and cluster 2 (bottom) in MO, DCs, and TolDCs. (B) Composite plots of VDR ChIP-seq distribution ±2.5 Kbp around CpGs from cluster 1 (top) and cluster 2 (bottom) in MO (gray), DC (red), and TolDC (green). Smooth represents the CIs. (legend continued on next page) 8 Cell Reports 38, 110244, January 18, 2022 ll Article OPEN ACCESS Remarkably, when we blocked IL-6 with an anti-IL-6 antibody ment, we confirmed the inhibition of STAT3 phosphorylation by during TolDC differentiation, we observed the production of western blot (Figure 6A). Given that TG101348 is an inhibitor of decreased levels of IL-10 (Figure S4D), which is involved in tol- JAK2, and therefore can affect upstream signaling of STAT1, erogenesis (Morante-Palacios et al., 2021). This result is STAT3, and STAT5, we checked their phosphorylation and consistent with recent findings in T helper type 1 (Th1) cells observed that the partial inhibition of p-STAT5 and p-STAT1 (Chauss et al., 2022). However, blocking IL-6 during TolDC did not reach statistical significance in TolDCs, unlike differentiation did not result in a reduced ability of TolDCs to p-STAT3 (Figures S5A and S5B). TG101348 treatment also re- suppress CD8+ T cell proliferation (Figure S4E). In contrast, in sulted in a sharp decrease in the production of IL-10 (Figure 6B), proliferation assays performed with TolDCs in the presence of an archetypical anti-inflammatory cytokine that is also a bona anti-IL-6 antibody, we found slightly reduced suppression (Fig- fide target for STAT3 (Schaefer et al., 2009; Ziegler-Heitbrock ure S4E). These results suggest that IL-6 is a contributor to the et al., 2003). In fact, IL-10 secretion by TolDCs is a contributor ability to suppress CD8+ T cell proliferation by TolDCs but not to the suppression of CD8+ T cell proliferation that is halted critical to the acquisition of such properties during TolDC when adding anti-IL-10 to proliferation assays (Figure S4E). differentiation. We also tested the effects of JAK2 inhibition on surface markers In parallel, we utilized DoRothEA (discriminant regulon expres- and observed that JAK2 inhibition resulted in an increase of sion analysis), a manually curated human regulon for estimating CD14 and CD86 protein levels and downregulation of CD1a single-sample TF activities through the expression of their target and CD11b (Figure 6C). In parallel, we investigated the effects genes (Garcia-Alonso et al., 2019) to analyze TF activities of of JAK2 inhibition on the DNA methylation and expression several STATs in genes differentially expressed in TolDCs levels of TolDC-specific demethylated genes. We did not compared with DCs, and observed a specific increase in observe any clear reversion of DNA demethylation (Figure 6D), STAT3 activity that was not observed for other members of the but we did note alterations at the transcriptional level (Fig- STAT family, at 5 days of differentiation (Figure 5F). Furthermore, ure 6E). Changes were observed not only in cluster 2 genes we observed a marked increase in phosphorylation of STAT3 in (TolDC-specific), such as CD14 and DPF3, but also in those TolDCs compared with DCs, which was not observed for STAT5 of cluster 1, such as IRF4 and RASF5 (Figure 6E). These are (Figures 5G and S4E). Although a statistically significant increase likely to be the result of the partial inhibition of phosphorylation in phosphorylation was observed for STAT1, this increase was of STAT1 and STAT5, which might also be involved in activating not to the same extent as STAT3 and may be due to indirect acti- these and other DC and TolDC genes. vation, as previously described (Haan et al., 2005) (Figures 5G Most importantly, JAK2 inhibition by TG101348 treatment dur- and S4F). Thus, our results suggested that vitamin D played a ing differentiation resulted in the loss of the ability to suppress role in STAT3 activation. CD8+ T cell proliferation of DC differentiated in the presence of To explore the possibility that the observed interplay between vitamin D. This reinforces the idea that the activities of VDR VDR and STAT3 involves a physical interaction, we performed and the JAK2-STAT3 pathway coordinate the acquisition of tol- co-immunoprecipitation experiments in TolDCs. Our analysis re- erogenic properties of DCs in the presence of vitamin D vealed a specific interaction between VDR and phosphorylated (Figure 6F). (p)-STAT3 in TolDCs (Figure 5H). We also observed that both VDR and p-STAT3 interacted with TET2 (Figure 5I), which sug- DISCUSSION gests that these two TFs play a role in the targeting of TET2- mediated demethylation to their cognate sites. In this study, we demonstrate that vitamin D is able to induce tol- erogenesis in DCs through a mechanism that involves VDR-spe- Inhibition of JAK2-mediated STAT3 activation affects cific demethylation and activation of key immune genes in a the acquisition of vitamin D-dependent tolerogenesis manner that is coordinated with JAK2-mediated STAT3 activa- We investigated the consequences of inhibiting the JAK2- tion. VDR not only is able to orchestrate a direct response on STAT3 pathway by using TG101348, a pharmacological key immune targets but also associates with activation of the inhibitor of JAK2 (Lasho et al., 2008), during DC and vitamin IL-6-JAK-STAT signaling pathway. We also prove the recruit- D-dependent TolDC differentiation. Following TG101348 treat- ment of TET2 and p-STAT3 by VDR, associated with the (C) Bubble plot representation of significant VDR binding enrichment in each cluster of CpGs. Dots are colored according to their enrichment value, defined as the logarithm of the ratio between the percentage of CpGs with VDR peak within the cluster and the percentage of CpGs with the VDR peak within the background. Bubble size corresponds to the percentage of CpGs in each cluster overlapping with significant VDR peaks. The presence of a black border indicates significant enrichment (q value < 0.01). (D) VDR ChIP-seq signal profiles in the vicinity of the representative genes of CpGs from cluster 2. VDR signals are colored by cell type. At the bottom, the significant VDR binding sites are shown in green and CpG position in red. (E) DNA methylation kinetics of two representative CpGs annotated to GAB2 (cg25310867) and HIF1A (cg14914214) in DCs and TolDCs (n = 3, one experiment). (F) Gene expression kinetics of GAB2 and HIF1a in DCs and TolDCs (n = 3, one experiment). (G) Bar plot representation of ChIP-qPCR results for VDR binding and three histone modifications (H3ac, H3K27me3, and H3K4me4) in the vicinity of VDR peaks close to GAB2 and HIF1A gene sequences (n = 3, one experiment). This analysis was performed in MOs, DCs, and TolDCs and DC + vitD. DC + vitD involves adding vitamin D for 30 min at the end of a 6-day differentiation to DCs. Immunoprecipitation with IgG was used as control. The location of the ChIP primers, the CpG site and the VDR peaks are indicated. Statistical tests: two-tailed Fisher’s exact test (C) and unpaired two-tailed t test (E, F, and G) (*p < 0.05; **p < 0.01; ***p < 0.001). Cell Reports 38, 110244, January 18, 2022 9 ll OPEN ACCESS Article A B C D E F G H I Figure 5. Vitamin D-dependent autocrine/paracrine activation of the IL-6-JAK2-STAT3 pathway (A) Heatmap showing ligand activity prediction based on the Pearson correlation with its target genes. (B) Heatmap displaying average gene expression of ligands for DCs and TolDCs on day 5. (C) Heatmap showing the regulatory potential of each ligand on the target genes based on nichenetr package database (upper panel) and the expression levels of these target genes in each sample (lower panel). (legend continued on next page) 10 Cell Reports 38, 110244, January 18, 2022 ll Article OPEN ACCESS demethylation and activation of target genes. The essential role Limitations of the study of the JAK2-STAT3 pathway in the acquisition of tolerogenesis is One of the limitations of our current study is that we have not fully demonstrated by the functional impact of the pharmacological explored the impact of the VDR and the IL-6-JAK-STAT3 pathway inhibition of this pathway. in vivo, in patients treated or supplemented with vitamin D. It Our results show the direct role of VDR in guiding TET-medi- would have also been relevant to analyze the direct impact of ated DNA demethylation to specific genomic sites during TolDC STAT3 in the epigenetic remodeling in TolDCs, by analyzing their differentiation. We have shown that, in the presence of vitamin D, binding sites and associated expression changes. This partly VDR levels are increased in the nucleus and that interaction with limits our conclusions on the extent and relevance of STAT3 in p-STAT3 and TET2 occurs, thereby promoting TolDC-specific determining the acquisition of the tolerogenic phenotype. demethylation. A recurrent question in the DNA methylation field is whether DNA methylation is causally involved in shaping gene STAR+METHODS expression profiles or if it passively reflects transcriptional states (Schu €beler, 2015). Our own data support both possibilities, and Detailed methods are provided in the online version of this paper some DNA methylation changes appear to be more likely to and include the following: occur after a change in expression than others (Pacis et al., 2019). In our study, we present evidence that TET-mediated de- d KEY RESOURCES TABLE methylation acts as a mechanism facilitating subsequent partic- d RESOURCE AVAILABILITY ipation of other TFs, in this case STAT3. In fact, the absence of B Lead contact interference with DNA demethylation, while activation is B Materials availability impeded following pharmacological inhibition of STAT3 phos- B Data and code availability phorylation, suggests that VDR-dependent demethylation is d EXPERIMENTAL MODEL AND SUBJECT DETAILS necessary and precedes STAT3-mediated gene activation. B Differentiation of TolDCs and DCs from peripheral This proposed mechanism was consistent with the alterations blood monocytes in TF activity reported in TET2 knockout mice (Rasmussen B Isolation and culture of peripheral blood DCs et al., 2019). TET2-associated functions may ensure the binding d METHOD DETAILS + of some TFs, thereby contributing to enhancer-dependent activ- B CD8 cell proliferation assay ity and gene expression. B BrdU proliferation assay Our study identifies a crucial role for the JAK2-STAT3 pathway B Flow cytometry in the acquisition of tolerogenesis in innate immunity. The B Cytokine measurements involvement of STAT3 is also relevant in the context of MDSCs, B Genomic DNA and total RNA extraction which are also characterized by their tolerogenic properties B Bisulfite (BS) and oxidative-bisulfite (oxBS) pyrose- (Corzo et al., 2009; Kumar et al., 2016). We show that the phar- quencing macological impairment of STAT3 phosphorylation, by inhibiting B Real-time quantitative Polymerase Chain Reaction JAK2, directly results in the loss of the tolerogenic properties of (RT-qPCR) TolDCs, which facilitate T-cell proliferation, demonstrating the B Western blot essential role of this pathway for the tolerogenic phenotype. B Co-immunoprecipitation (Co-IP) Our results raise the possibility that tolerogenic properties can B DNA methylation profiling be reverted, not only in the context of vitamin D but also in others. B ChIP-seq analysis These findings could be clinically relevant both in the context of B ChIP-qPCR pathological situations where tolerogenic properties are not B Microarray reanalysis desired, like in the tumor microenvironment or in metastatic pro- B Data analysis cesses (reviewed in DeVito et al., 2019), as well as in those where d QUANTIFICATION AND STATISTICAL ANALYSIS they are intentionally pursued (reviewed in Cauwels and Taver- nier, 2020), including their therapeutic use in the treatment of in- SUPPLEMENTAL INFORMATION flammatory conditions, such as rheumatoid arthritis and multiple Supplemental information can be found online at https://doi.org/10.1016/j. sclerosis (Morante-Palacios et al., 2021). celrep.2021.110244. (D) Gene set enrichment analysis of differentially expressed genes (fold-change < 0.5 or >2 and q-value < 0.05) at 12 h (red) and 120 h (blue). Results for the IL-6- JAK-STAT signaling pathway are shown. (E) VDR ChIP-seq signal profiles in the vicinity of the IL6 gene. VDR signals are colored by cell type. The significant VDR binding sites are shown below in green. (F) Bubble chart depicting the TF activity predicted from mRNA expression of target genes with DoRothEA v2.0. The circle filling represents the normalized enrichment score (NES) (blue: more activity in DCs; red: more activity in TolDCs). Bubble size corresponds to the logarithm of adjusted p values. (G) Representative western blot assays showing the phosphorylated and total protein levels of STAT1, STAT3, and STAT5 on day 3 of differentiation of DCs and TolDCs (n = 4, two independent experiments). (H) Representative western blots showing the results of co-immunoprecipitation assays performed in MOs differentiated to DC and TolDC for 3 days. Protein extracts were immunoprecipitated using anti-VDR or anti-p-STAT3 antibodies (n = 3, two independent experiments). (I) Representative western blots showing the results of co-immunoprecipitation assays performed in MOs differentiated to TolDC for 3 days (n = 3, two inde- pendent experiments). Protein extracts were immunoprecipitated using anti-TET2 antibodies. In both (H and I), IgG was used as a negative control and total protein extract was used as input. Cell Reports 38, 110244, January 18, 2022 11 ll OPEN ACCESS Article A B C D F E Figure 6. Inhibition of STAT3 phosphorylation reverts immunosuppressive properties of vitamin D exposed dendritic cells (A) Representative western blot assays showing the effects of STAT3 at the protein phosphorylation level after pharmacological inhibition of JAK2 with TG101348 (n = 3, two experiments). (B) Bar plot representation of the effect of JAK2 inhibition with TG101348 on IL-10 released by DCs and TolDCs (n = 4, two independent experiments). Protein levels were measured by ELISA. (C) Bar plots showing the impact of JAK2 inhibition with TG101348 on membrane receptor expression (n = 6, two independent experiments). Protein levels were measured with flow cytometry. (D) Dot plot representation of bisulfite pyrosequencing results of four example CpGs, two from cluster 1 (CD14 and DPF3) and two from cluster 2 (C1QB and RASF5), displaying the consequence of JAK2 inhibition with TG101348 in DCs and TolDCs (n = 3, two independent experiments). (E) Dot plot showing mRNA expression of four example genes from cluster 1 and cluster 2 as measured by RT-qPCR, showing the effect of JAK2 inhibition with TG101348 in DCs and TolDCs (n = 4, two independent experiments). Expression was relativized with respect to RPL38 gene expression. (F) Representative example and dot plot showing the effect on CD8+ cell proliferation of DCs and TolDC generated from MO in presence or absence of TG101348 (n = 5, two independent experiments). Statistical tests: two-tailed Wilcoxon rank-sum test (B–F) (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). ACKNOWLEDGMENTS and FEDER. O.M.-P. holds an i-PFIS PhD fellowship (grant number IFI17/ 00034) from Acción Estratégica en Salud 2013–2016 ISCIII, co-financed by We are very grateful to Dr. José Luis Sardina for useful feedback. We thank Fondo Social Europeo. F.F. holds a PhD fellowship from the INsTRuCT Con- CERCA Programme/Generalitat de Catalunya and the Josep Carreras Foun- sortium, which receives. Innovative Training Network subsidy from the EU dation for institutional support. E.B. was funded by the Spanish Ministry of Sci- H2020 program. ence and Innovation (MICINN; grant number PID2020-117212RB-I00/AEI/ 10.13038/501100011033). E.M.-C. is funded with RESTORE project (EU AUTHOR CONTRIBUTIONS H2020 Research and Innovation Programme, number 779316) and Spanish projects PI17/01521 and PI20/01313, integrated in the Plan Nacional de F.C.-M. and E.B. conceived and designed the study; F.C.-M., A.G.F.-B., I+D+I and co-supported by the ISCIII-Subdirección General de Evaluación G.G.-T., O.M.-P., L.C., L.B., F.F., and T.L. performed the differentiation, 12 Cell Reports 38, 110244, January 18, 2022 ll Article OPEN ACCESS chromatin immunoprecipitation, co-immunoprecipitation experiments, and NANOG-dependent function of TET1 and TET2 in establishment of pluripo- immunological assays; F.C.-M. performed the bioinformatic analyses; tency. Nature 495, 370–374. F.C.-M., A.G.F.-B., G.G.-T., E.M.-C., and E.B. analyzed results; J.R.-U. and Dankers, W., Colin, E.M., van Hamburg, J.P., and Lubberts, E. (2017). Vitamin E.B. supervised the study; F.C.-M., T.L., and E.B. wrote the manuscript; all D in autoimmunity: molecular mechanisms and therapeutic potential. Front. authors participated in discussions and interpreting the results. Immunol. 7, 697. de la Rica, L., Rodrı́guez-Ubreva, J., Garcı́a, M., Islam, A.B.M.M.K., Urquiza, DECLARATION OF INTERESTS J.M., Hernando, H., Christensen, J., Helin, K., Gómez-Vaquero, C., and Balles- tar, E. (2013). PU.1 target genes undergo Tet2-coupled demethylation and The authors declare no competing interests. DNMT3b-mediated methylation in monocyte-to-osteoclast differentiation. Genome Biol. 14, R99. Received: June 29, 2021 Deplus, R., Delatte, B., Schwinn, M.K., Defrance, M., Méndez, J., Murphy, N., Revised: September 27, 2021 Dawson, M.A., Volkmar, M., Putmans, P., Calonne, E., et al. (2013). TET2 and Accepted: December 20, 2021 TET3 regulate GlcNAcylation and H3K4 methylation through OGT and SET1/ Published: January 18, 2022 COMPASS. EMBO J. 32, 645–655. DeVito, N.C., Plebanek, M.P., Theivanthiran, B., and Hanks, B.A. (2019). Role REFERENCES of tumor-mediated dendritic cell Tolerization in immune evasion. Front. Immu- nol. 10, 2876. Álvarez-Errico, D., Vento-Tormo, R., Sieweke, M., and Ballestar, E. (2015). Epigenetic control of myeloid cell differentiation, identity and function. Nat. Feingold, E.A., Good, P.J., Guyer, M.S., Kamholz, S., Liefer, L., Wetterstrand, Rev. Immunol. 15, 7–17. K., Collins, F.S., Gingeras, T.R., Kampa, D., Sekinger, E.A., et al. (2004). The ENCODE (ENCyclopedia of DNA elements) project. Science 306, 636–640. Aryee, M.J., Jaffe, A.E., Corrada-Bravo, H., Ladd-Acosta, C., Feinberg, A.P., Hansen, K.D., and Irizarry, R.A. (2014). Minfi: a flexible and comprehensive bio- Ferreira, G.B., Overbergh, L., Verstuyf, A., and Mathieu, C. (2013). 1a,25-Dihy- conductor package for the analysis of infinium DNA methylation microarrays. droxyvitamin D3 and its analogs as modulators of human dendritic cells: a Bioinformatics 30, 1363–1369. comparison dose-titration study. J. Steroid Biochem. Mol. Biol. 136, 160–165. €ller, F., Lutsik, P., Walter, J., Lengauer, T., and Bock, C. (2014). Assenov, Y., Mu Ferreira, G.B., Vanherwegen, A.S., Eelen, G., Gutiérrez, A.C.F., VanLommel, Comprehensive analysis of DNA methylation data with RnBeads. Nat. L., Marchal, K., Verlinden, L., Verstuyf, A., Nogueira, T., Georgiadou, M., Methods 11, 1138–1140. et al. (2015). Vitamin D3 induces tolerance in human dendritic cells by activa- tion of intracellular metabolic pathways. Cell Rep. 10, 711–725. Barragan, M., Good, M., and Kolls, J.K. (2015). Regulation of dendritic cell function by vitamin D. Nutrients 7, 8127–8151. Fetahu, I.S., Höbaus, J., and Kállay, E. (2014). Vitamin D and the epigenome. Front. Physiol. 5, 164. Bibikova, M., Lin, Z., Zhou, L., Chudin, E., Garcia, E.W., Wu, B., Doucet, D., Thomas, N.J., Wang, Y., Vollmer, E., et al. (2006). High-throughput DNA Garcia-Alonso, L., Holland, C.H., Ibrahim, M.M., Turei, D., and Saez-Rodri- methylation profiling using universal bead arrays. Genome Res. 16, 383–393. guez, J. (2019). Benchmark and integration of resources for the estimation of human transcription factor activities. Genome Res. 29, 1363–1375. Browaeys, R., Saelens, W., and Saeys, Y. (2020). NicheNet: modeling intercel- Garcia-Gomez, A., Li, T., Kerick, M., Català-Moll, F., Comet, N.R., Rodrı́guez- lular communication by linking ligands to target genes. Nat. Methods 17, Ubreva, J., De La Rica, L., Branco, M.R., Martı́n, J., and Ballestar, E. (2017). 159–162. TET2- and TDG-mediated changes are required for the acquisition of distinct Bscheider, M., and Butcher, E.C. (2016). Vitamin D immunoregulation through histone modifications in divergent terminal differentiation of myeloid cells. Nu- dendritic cells. Immunology 148, 227–236. cleic Acids Res. 45, 10002–10017. Carlberg, C. (2019). Vitamin D signaling in the context of innate immunity: focus Goudot, C., Coillard, A., Villani, A.C., Gueguen, P., Cros, A., Sarkizova, S., on human monocytes. Front. Immunol. 10, 2211. Tang-Huau, T.L., Bohec, M., Baulande, S., Hacohen, N., et al. (2017). Aryl hy- Carvalho, B.S., and Irizarry, R.A. (2010). A framework for oligonucleotide mi- drocarbon receptor controls monocyte differentiation into dendritic cells croarray preprocessing. Bioinformatics 26, 2363–2367. versus macrophages. Immunity 47, 582–596.e6. Cauwels, A., and Tavernier, J. (2020). Tolerizing strategies for the treatment of Guilhamon, P., Eskandarpour, M., Halai, D., Wilson, G.A., Feber, A., Teschen- autoimmune diseases: from ex vivo to in vivo strategies. Front. Immunol. 11, dorff, A.E., Gomez, V., Hergovich, A., Tirabosco, R., Fernanda Amary, M., et al. 674. (2013). Meta-analysis of IDH-mutant cancers identifies EBF1 as an interaction Chauss, D., Freiwald, T., McGregor, R., Yan, B., Wang, L., Nova-Lamperti, E., partner for TET2. Nat. Commun. 4, 2166. Kumar, D., Zhang, Z., Teague, H., West, E.E., et al. (2022). Autocrine vitamin D Haan, S., Keller, J.F., Behrmann, I., Heinrich, P.C., and Haan, C. (2005). signaling switches off pro-inflammatory programs of T H 1 cells. Nat. Immunol. Multiple reasons for an inefficient STAT1 response upon IL-6-type cytokine 23, 62–74. stimulation. Cell. Signal. 17, 1542–1550. Chu, C.C., Ali, N., Karagiannis, P., Di Meglio, P., Skowera, A., Napolitano, L., Heinz, S., Benner, C., Spann, N., Bertolino, E., Lin, Y.C., Laslo, P., Cheng, J.X., Barinaga, G., Grys, K., Sharif-Paghaleh, E., Karagiannis, S.N., et al. (2012). Murre, C., Singh, H., and Glass, C.K. (2010). Simple combinations of lineage- Resident CD141 (BDCA3) + dendritic cells in human skin produce IL-10 and determining transcription factors prime cis-regulatory elements required for induce regulatory T cells that suppress skin inflammation. J. Exp. Med. 209, macrophage and B cell identities. Mol. Cell 38, 576–589. 935–945. Ichiyama, K., Chen, T., Wang, X., Yan, X., Kim, B.S., Tanaka, S., Ndiaye-Lobry, Corzo, C.A., Cotter, M.J., Cheng, P., Cheng, F., Kusmartsev, S., Sotomayor, D., Deng, Y., Zou, Y., Zheng, P., et al. (2015). The methylcytosine dioxygenase E., Padhya, T., McCaffrey, T.V., McCaffrey, J.C., and Gabrilovich, D.I. Tet2 promotes DNA demethylation and activation of cytokine gene expression (2009). Mechanism regulating reactive oxygen species in tumor-induced in T cells. Immunity 42, 613–626. myeloid-derived suppressor cells. J. Immunol. 182, 5693–5701. Jakob, F., Gieseler, F., Tresch, A., Hammer, S., Seufert, J., and Schneider, D. Corzo, C.A., Condamine, T., Lu, L., Cotter, M.J., Youn, J.I., Cheng, P., Cho, (1992). Kinetics of nuclear translocation and turnover of the vitamin D receptor H.I., Celis, E., Quiceno, D.G., Padhya, T., et al. (2010). HIF-1a regulates in human HL60 leukemia cells and peripheral blood lymphocytes-coincident function and differentiation of myeloid-derived suppressor cells in the tumor rise of DNA-relaxing activity in nuclear extracts. J. Steroid Biochem. Mol. microenvironment. J. Exp. Med. 207, 2439–2453. Biol. 42, 11–16. Costa, Y., Ding, J., Theunissen, T.W., Faiola, F., Hore, T.A., Shliaha, P.V., Jones, P.A. (2012). Functions of DNA methylation: islands, start sites, gene Fidalgo, M., Saunders, A., Lawrence, M., Dietmann, S., et al. (2013). bodies and beyond. Nat. Rev. Genet. 13, 484–492. Cell Reports 38, 110244, January 18, 2022 13 ll OPEN ACCESS Article Klug, M., Schmidhofer, S., Gebhard, C., Andreesen, R., and Rehli, M. (2013). 5- Murphy, T.L., Grajales-Reyes, G.E., Wu, X., Tussiwand, R., Briseño, C.G., Hydroxymethylcytosine is an essential intermediate of active DNA demethyla- Iwata, A., Kretzer, N.M., Durai, V., and Murphy, K.M. (2016). Transcriptional tion processes in primary human monocytes. Genome Biol. 14, R46. control of dendritic cell development. Annu. Rev. Immunol. 34, 93–119. Koressaar, T., and Remm, M. (2007). Enhancements and modifications of Navarro-Barriuso, J., Mansilla, M.J., and Martı́nez-Cáceres, E.M. (2018). primer design program primer3. Bioinformatics 23, 1289–1291. Searching for the transcriptomic signature of immune tolerance induction-bio- Korotkevich, G., Sukhov, V., and Sergushichev, A. (2019). Fast gene set markers of safety and functionality for tolerogenic dendritic cells and regula- enrichment analysis. BioRxiv. https://doi.org/10.1101/060012. tory macrophages. Front. Immunol. 9, 2062. Ong, L.T.C., Schibeci, S.D., Fewings, N.L., Booth, D.R., and Parnell, G.P. Kumar, V., Cheng, P., Condamine, T., Mony, S., Languino, L.R., McCaffrey, (2021). Age-dependent VDR peak DNA methylation as a mechanism for J.C., Hockstein, N., Guarino, M., Masters, G., Penman, E., et al. (2016). latitude-dependent multiple sclerosis risk. Epigenetics Chromatin 14, 9. CD45 phosphatase inhibits STAT3 transcription factor activity in myeloid cells and promotes tumor-associated macrophage differentiation. Immunity 44, Pacis, A., Tailleux, L., Morin, A.M., Lambourne, J., MacIsaac, J.L., Yotova, V., 303–315. Dumaine, A., Danckært, A., Luca, F., Grenier, J.C., et al. (2015). Bacterial infec- tion remodels the DNA methylation landscape of human dendritic cells. Langmead, B., and Salzberg, S.L. (2012). Fast gapped-read alignment with Genome Res. 25, 1801–1811. Bowtie 2. Nat. Methods 9, 357–359. Pacis, A., Mailhot-Léonard, F., Tailleux, L., Randolph, H.E., Yotova, V., Du- Lasho, T.L., Tefferi, A., Hood, J.D., Verstovsek, S., Gilliland, D.G., and Parda- maine, A., Grenier, J.-C., and Barreiro, L.B. (2019). Gene activation precedes nani, A. (2008). TG101348, a JAK2-selective antagonist, inhibits primary he- DNA demethylation in response to infection in human dendritic cells. Proc. matopoietic cells derived from myeloproliferative disorder patients with Natl. Acad. Sci. U S A 116, 6938–6943. JAK2V617F, MPLW515K or JAK2 exon 12 mutations as well as mutation nega- tive patients. Leukemia 22, 1790–1792. Pageaud, Y., Plass, C., and Assenov, Y. (2018). Enrichment analysis with Epi- Annotator. Bioinformatics 32, 1781–1783. Leek, J.T., Johnson, W.E., Parker, H.S., Fertig, E.J., Jaffe, A.E., Zhang, Y., Sto- rey, J.D., Torres, L.C., 2021. Bioconductor. sva: Surrogate Variable Analysis. R Park, J.H., Van Wyk, H., McMillan, D.C., Quinn, J., Clark, J., Roxburgh, C.S.D., package version 3.42.0. Horgan, P.G., and Edwards, J. (2017). Signal transduction and activator of transcription-3 (STAT3) in patients with colorectal cancer: associations with Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., the phenotypic features of the tumor and host. Clin. Cancer Res. 23, 1698– Abecasis, G., and Durbin, R. (2009). The sequence alignment/map format and 1709. SAMtools. Bioinformatics 25, 2078–2079. Penna, G., and Adorini, L. (2000). 1a,25-Dihydroxyvitamin D 3 inhibits differen- Li, T., Garcia-Gomez, A., Morante-Palacios, O., Ciudad, L., Özkaramehmet, tiation, maturation, activation, and survival of dendritic cells leading to S., Van Dijck, E., Rodrı́guez-Ubreva, J., Vaquero, A., and Ballestar, E. (2020). impaired alloreactive T cell activation. J. Immunol. 164, 2405–2411. SIRT1/2 orchestrate acquisition of DNA methylation and loss of histone H3 activating marks to prevent premature activation of inflammatory genes in Pereira, F., Barbáchano, A., Silva, J., Bonilla, F., Campbell, M.J., Muñoz, A., macrophages. Nucleic Acids Res. 48, 665–681. and Larriba, M.J. (2011). KDM6B/JMJD3 histone demethylase is induced by vitamin D and modulates its effects in colon cancer cells. Hum. Mol. Genet. Lio, C.W., Zhang, J., González-Avalos, E., Hogan, P.G., Chang, X., and Rao, A. 20, 4655–4665. (2016). Tet2 and Tet3 cooperate with B-lineage transcription factors to regu- Piemonti, L., Monti, P., Sironi, M., Fraticelli, P., Leone, B.E., Dal Cin, E., Alla- late DNA modification and chromatin accessibility. Elife 5, e18290. vena, P., and Di Carlo, V. (2000). Vitamin D 3 affects differentiation, maturation, Liu, J., Zhang, X., Chen, K., Cheng, Y., Liu, S., Xia, M., Chen, Y., Zhu, H., Li, Z., and function of human monocyte-derived dendritic cells. J. Immunol. 164, and Cao, X. (2019). CCR7 chemokine receptor-inducible lnc-Dpf3 restrains 4443–4451. dendritic cell migration by inhibiting HIF-1a-Mediated glycolysis. Immunity Pratt, J.C., Igras, V.E., Maeda, H., Baksh, S., Gelfand, E.W., Burakoff, S.J., 50, 600–615.e15. Neel, B.G., and Gu, H. (2000). Cutting edge: gab2 mediates an inhibitory phos- Luo, C., Hajkova, P., and Ecker, J.R. (2018). Dynamic DNA methylation: in the phatidylinositol 3’-kinase pathway in T cell antigen receptor signaling. right place at the right time. Science 361, 1336–1340. J. Immunol. 165, 4158–4163. Mahé, E.A., Madigou, T., Sérandour, A.A., Bizot, M., Avner, S., Chalmel, F., €ndar, F., Sarah Diehl, S., Björn A Gru Ramı́rez, F., Du €ning, B.A., and Manke, T. Palierne, G., Métivier, R., and Salbert, G. (2017). Cytosine modifications modu- (2014). deepTools: a flexible platform for exploring deep-sequencing data. late the chromatin architecture of transcriptional enhancers. Genome Res. 27, Nucleic Acids Res. 42, W187–W191. 947–958. Rasmussen, K.D., Berest, I., Kebler, S., Nishimura, K., Simón-Carrasco, L., McLean, C.Y., Bristor, D., Hiller, M., Clarke, S.L., Schaar, B.T., Lowe, C.B., Vassiliou, G.S., Pedersen, M.T., Christensen, J., Zaugg, J.B., and Helin, K. Wenger, A.M., and Bejerano, G. (2010). GREAT improves functional interpre- (2019). TET2 binding to enhancers facilitates transcription factor recruitment tation of cis-regulatory regions. Nat. Biotechnol. 28, 495–501. in hematopoietic cells. Genome Res. 29, 564–575. Mendes, K., Schmidhofer, S., Minderjahn, J., Glatz, D., Kiesewetter, C., Ritchie, M.E., Phipson, B., Wu, D., Hu, Y., Law, C.W., Shi, W., and Smyth, G.K. Raithel, J., Wimmer, J., Gebhard, C., and Rehli, M. (2021). The epigenetic (2015). Limma powers differential expression analyses for RNA-sequencing pioneer EGR2 initiates DNA demethylation in differentiating monocytes at and microarray studies. Nucleic Acids Res. 43, e47. both stable and transient binding sites. Nat. Commun. 12, 1556. Sardina, J.L., Collombet, S., Tian, T.V., Gómez, A., Di Stefano, B., Berenguer, Monticelli, S., and Natoli, G. (2017). Transcriptional determination and C., Brumbaugh, J., Stadhouders, R., Segura-Morales, C., Gut, M., et al. (2018). functional specificity of myeloid cells: making sense of diversity. Nat. Rev. Transcription factors drive Tet2-mediated enhancer demethylation to repro- Immunol. 17, 595–607. gram cell fate. Cell Stem Cell 23, 727–741.e9. Mora, J.R., Iwata, M., and Von Andrian, U.H. (2008). Vitamin effects on the im- Schaefer, A., Unterberger, C., Frankenberger, M., Lohrum, M., Staples, K.J., mune system: vitamins A and D take centre stage. Nat. Rev. Immunol. 8, Werner, T., Stunnenberg, H., and Ziegler-Heitbrock, L. (2009). Mechanism of 685–698. Interferon-gamma mediated down-regulation of Interleukin-10 gene expres- Morante-Palacios, O., Fondelli, F., Ballestar, E., and Martı́nez-Cáceres, E.M. sion. Mol. Immunol. 46, 1351–1359. (2021). Tolerogenic dendritic cells in autoimmunity and inflammatory diseases. €beler, D. (2015). Function and information content of DNA methylation. Schu Trends Immunol. 42, 59–75. Nature 517, 321–326. Mu€ller, F., Scherer, M., Assenov, Y., Lutsik, P., Walter, J., Lengauer, T., and Segura, E., Touzot, M., Bohineust, A., Cappuccio, A., Chiocchia, G., Hosmalin, Bock, C. (2019). RnBeads 2.0: comprehensive analysis of DNA methylation A., Dalod, M., Soumelis, V., and Amigorena, S. (2013). Human inflammatory data. Genome Biol. 20, 55. dendritic cells induce Th17 cell differentiation. Immunity 38, 336–348. 14 Cell Reports 38, 110244, January 18, 2022 ll Article OPEN ACCESS Seuter, S., Neme, A., and Carlberg, C. (2017). Epigenomic PU.1-VDR crosstalk van der Maaten, L. (2014). Accelerating t-SNE using Tree-based algorithms. modulates vitamin D signaling. Biochim. Biophys. Acta Gene Regul. Mech. J. Machine Learn. Res. 15, 3221–3245. 1860, 405–415. Van Halteren, A.G.S., Van Etten, E., De Jong, E.C., Bouillon, R., Roep, B.O., Seuter, S., Neme, A., and Carlberg, C. (2018). ETS transcription factor family and Mathieu, C. (2002). Redirection of human autoreactive T-cells upon inter- member GABPA contributes to vitamin D receptor target gene regulation. action with dendritic cells modulated by TX527, an analog of 1,25 dihydroxy- J. Steroid Biochem. Mol. Biol. 177, 46–52. vitamin D3. Diabetes 51, 2119–2125. Wang, Y., Xiao, M., Chen, X., Chen, L., Xu, Y., Lv, L., Wang, P., Yang, H., Ma, Széles, L., Keresztes, G., Töröcsik, D., Balajthy, Z., Krenács, L., Póliska, S., S., Lin, H., et al. (2015). WT1 recruits TET2 to regulate its target gene expres- Steinmeyer, A., Zuegel, U., Pruenster, M., Rot, A., et al. (2009). 1,25-Dihydrox- sion and suppress leukemia cell proliferation. Mol. Cell 57, 662–673. yvitamin D 3 is an autonomous regulator of the transcriptional changes leading Wei, Z., Yoshihara, E., He, N., Hah, N., Fan, W., Pinto, A.F.M., Huddy, T., to a tolerogenic dendritic cell phenotype. J. Immunol. 182, 2074–2083. Wang, Y., Ross, B., Estepa, G., et al. (2018). Vitamin D switches BAF com- Teh, B.K., Yeo, J.G., Chern, L.M., and Lu, J. (2011). C1q regulation of dendritic plexes to protect b cells. Cell 173, 1135–1149.e15. cell development from monocytes with distinct cytokine production and T cell Xiong, J., Zhang, Z., Chen, J., Huang, H., Xu, Y., Ding, X., Zheng, Y., Nishina- stimulation. Mol. Immunol. 48, 1128–1138. kamura, R., Xu, G.-L., Wang, H., et al. (2016). Cooperative action between Torres-Aguilar, H., Aguilar-Ruiz, S.R., González-Pérez, G., Munguı́a, R., Ba- SALL4A and TET proteins in stepwise oxidation of 5-methylcytosine. Mol. jaña, S., Meraz-Rı́os, M.A., and Sánchez-Torres, C. (2010). Tolerogenic den- Cell 64, 913–925. dritic cells generated with different immunosuppressive cytokines induce an- Ziegler-Heitbrock, L., Lötzerich, M., Schaefer, A., Werner, T., Frankenberger, tigen-specific anergy and regulatory properties in memory CD4+ T cells. M., and Benkhart, E. (2003). IFN-a induces the human IL-10 gene by recruiting J. Immunol. 184, 1765–1775. both IFN regulatory factor 1 and Stat3. J. Immunol. 171, 285–290. Cell Reports 38, 110244, January 18, 2022 15 ll OPEN ACCESS Article STAR+METHODS KEY RESOURCES TABLE REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies Fc Block reagent, human antibody Miltenyi Biotec Cat# 130-059-901; RRID: AB_2892112 Anti-human CD14, FITC conjugated (clone TÜK4) Miltenyi Biotec Cat# 130-080-701;RRID: AB_244303 Anti-human CD80, PE conjugated (clone 2D10) Miltenyi Biotec Cat# 130-097-202; RRID: AB_2659259 Anti-human CD86, APC conjugated (clone FM95) Miltenyi Biotec Cat# 130-113-569; RRID: AB_2726174 Anti-human CD11b, APC conjugated (clone ICRF44) BioLegend Cat# 301310; RRID: AB_314162 Anti-human CD1a, PE conjugated (clone HI149) BioLegend Cat# 300106; RRID: AB_314020 Anti-human HLA-DR, Pe-Cyanine7 conjugated (clone LN3)Thermo Fisher Scientific Cat# 25-9956-42; RRID: AB_1582284 LIVE/DEADTM Fixable Violet Dead Cell Stain Kit Thermo Fisher Scientific Cat# L34964 Anti-acetyl-Histone H3 Antibody Millipore Cat# 06-599; RRID: AB_2115283 Anti-trimethyl-Histone H3 (Lys27) Antibody Millipore Cat# 07-449; RRID: AB_310624 Anti-Trimethyl-Histone H3 (Lys4) Millipore Cat# 17-614; RRID: AB_11212770 Rat IgG1 kappa Isotype Control (clone eBRG1) Thermo Fisher Scientific Cat# 16-4301-85; RRID: AB_470154 IL-10 Monoclonal Antibody (clone JES3-9D7) Thermo Fisher Scientific Cat# 16-7108-85; RRID: AB_469229 IL-6 Monoclonal Antibody (clone MQ2-13A5) Thermo Fisher Scientific Cat# 16-7069-85; RRID: AB_469219 Vitamin D receptor Cell Signaling Technology Cat# 12550; RRID: AB_2637002 Anti-TET2 antibodody Abcam Cat# ab124297; RRID: AB_2722695 Anti-Pstat3 [Y705], (clone 4/P-STAT3) Fluidigm Cat# 3158005A; RRID: AB_2811100 Anti-STAT3, (clone 79D7) Cell Signaling Technology Cat# 4904; RRID: AB_331269 Anti-pSTAT1 [Y701], (clone 58D6) Cell Signaling Technology Cat# 9167; RRID: AB_561284 Anti-STAT1, (clone 42H3) Cell Signaling Technology Cat# 9175; RRID: AB_2197984 Anti-pSTAT5 [Y694], (clone 47) Fluidigm Cat# 3150005A; RRID: AB_2744690 Anti-STAT5 beta, (clone ST5b-10G1) Thermo Fisher Scientific Cat# 13-5300; RRID: AB_2533021 Anti-Lamin B1 Abcam Cat# ab16048; RRID: AB_443298 Anti-alpha-Tubulin, (clone DM1A) Sigma-Aldrich Cat# T6199; RRID: AB_477583 Normal Rabbit IgG Antibody Millipore Cat# 12-370; RRID: AB_145841 Normal Mouse IgG Antibody Millipore Cat# 12-371; RRID: AB_145840 Biological samples Buffy Coats Catalan Blood and Tissue Bank (CBTB)Cat# BB014 Chemicals, peptides, and recombinant proteins Recombinant Human IL-4 Peprotech Cat# 200-04; GenPept: P05112 Recombinant Human GM-CSF Peprotech Ca# 300-03; GenPept: P04141 1a,25-Dihydroxyvitamin D3 Sigma-Aldrich Cat# D1530; CAS: 32222-06-3 TG101348, JAK/STAT pathway inhibitor STEMCELL Cat# 73472; CAS: 936091-26-8 Potassium perruthenate (VII) (KRuO4) Thermo Fisher Scientific Cat# 11877; CAS: 10378-50-4 TritonTM X-100 Sigma-Aldrich Cat# T8787; CAS: 9036-19-5 Benzonase Nuclease Sigma-Aldrich Cat# E1014; CAS: 9025-65-4 DMP (Dimethyl Pimelimidate) Thermo Fisher Scientific Cat# 21666; CAS: 58537-94-3 Critical commercial assays MACS CD14 Microbeads Miltenyi Biotec Cat# 130-050-201 CD1c (BDCA-1)+ Dendritic Cell Isolation Kit Miltenyi Biotec Cat# 130-119-475 DynabeadsTM UntouchedTM Human CD8 T Cells Kit Thermo Fisher Scientific Cat# 11348D Cell Trace CFSE Thermo Fisher Scientific Cat# C34554 (Continued on next page) e1 Cell Reports 38, 110244, January 18, 2022 ll Article OPEN ACCESS Continued REAGENT or RESOURCE SOURCE IDENTIFIER APC BrdU Flow Kit BD Pharmingen Cat# 552598 ELISA MAXTM Deluxe Set Human IL-10 BioLegend Cat# 430604 Access IL-6 reagent kit Beckman-Coulter Cat# A16369 Maxwell RSC Cultured Cells DNA Kit Promega Cat# AS1620 Maxwell RSC simplyRNA cells Kit Promega Cat# AS1390 EZ DNA Methylation-Gold Kit Zymo Research Cat# D5005 PyroMark Q48 Advanced CpG Reagents Qiagen Cat# 974022 Transcriptor First Strand cDNA Synthesis Kit Roche Cat# 04897030001 IMMOLASE DNA polymerase Kit Bioline Cat# BIO-21047 LightCycler 480 SYBR Green I Master Roche Cat# 0487352001 Infinium MethylationEPIC BeadChip Illumina Cat# 20042130 iDeal ChIP-seq kit for Transcription Factors Diagenode Cat# C01010055 Magna ChIPTM Protein A+G Magnetic Beads Sigma-Aldrich Cat# 16-663 cOmpleteTM, EDTA-free Protease Inhibitor Cocktail Sigma-Aldrich Cat# 11873580001 PureProteomeTM Protein G Magnetic Bead System Sigma-Aldrich Cat# LSKMAGG02 Micro Bio-Spin P-6 SSC columns Bio-Rad Cat# 7326200 Deposited data DNA methylation profile of in vitro generated DC This paper GSE145483 and TolDC VDR ChIP-Seq This paper GSE145584 Oligonucleotides Primers for bisulphite pyrosequencing, RT-qPCR and This paper N/A ChIP-qPCR, see Table S2 Software and algorithms PyroMark Assay Design 2.0 software Qiagen Cat# 9019079 Pyromark Q48 Autoprep software Qiagen Cat# 9024325 Minfi (R package) Aryee et al., 2014 https://www.bioconductor.org/packages/ release/bioc/html/minfi.html RnBeads (R package) Assenov et al. (2014) https://bioconductor.org/packages/ release/bioc/html/RnBeads.html Limma (R package) Ritchie et al. (2015) https://bioconductor.org/packages/ release/bioc/html/limma.html Bowtie2 Aligner v2.2.6 Langmead and Salzberg, 2012 http://bowtie-bio.sourceforge.net/ bowtie2/index.shtml MarkDuplicates software v1.126 Broad institute https://broadinstitute.github.io/picard/ Sequence Alignment/Map (SAMtools) v1.2 Li et al. (2009) https://doi.org/10.1093/bioinformatics/ btp352 bamCoverage function (deepTools (v2.0)) Ramı́rez et al. (2014) https://deeptools.readthedocs.io/en/ 2.1.0/content/tools/bamCoverage.html HOMER Motif Analysis Heinz et al. (2010) http://homer.ucsd.edu/homer/motif/ GREAT (version 3.0.0.) McLean et al. (2010) http://great.stanford.edu/public/html EpiAnnotator (R package) Pageaud et al. (2018) http://epigenomics.dkfz.de/EpiAnnotator/ fgsea (R package) Korotkevich et al. (2019) http://bioconductor.org/packages/release/ bioc/html/fgsea.html tSNE van der Maaten, 2014 https://github.com/lvdmaaten/bhtsne/ sva (R package) (Leek et al., 2021) https://bioconductor.org/packages/release/ bioc/html/sva.html DoRothEA (R package) Garcia-Alonso et al. (2019) https://bioconductor.org/packages/release/ data/experiment/html/dorothea.html Nichenetr (R package) Browaeys et al. (2020) https://github.com/saeyslab/nichenetr Cell Reports 38, 110244, January 18, 2022 e2 ll OPEN ACCESS Article RESOURCE AVAILABILITY Lead contact Further information and requests for resources should be directed to the lead contact, Esteban Ballestar (eballestar@ carrerasresearch.org). Materials availability This study did not generate new unique reagents. Data and code availability d DNA methylation and ChIP-seq data for this publication have been deposited in the NCBI Gene Expression Omnibus and are accessible through GEO Series accession numbers GSE145483 and GSE145584. d This paper does not report original code. d Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request EXPERIMENTAL MODEL AND SUBJECT DETAILS Differentiation of TolDCs and DCs from peripheral blood monocytes For in vitro differentiation experiments, we obtained buffy coats from anonymous donors through the Catalan Blood and Tissue Bank (CBTB). The CBTB follows the principles of the World Medical Association (WMA) Declaration of Helsinki. The Committee for Human Subjects of Bellvitge Hospital approved the study (PR275/17). Given the anonymous nature of the volunteers, no information about the gender and age was provided by the CBTB. Before providing the first blood sample, all donors received detailed oral and written information, and signed a consent form at the CBTB. Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Paque gradient centrifugation. MOs were isolated from PBMCs using positive selection with MACS CD14 microbeads (Miltenyi Biotec). Cells were resuspended in RPMI Medium 1640 + GlutaMAXTM-1 (Gibco, Life Technologies) containing 10% fetal bovine serum, 100 units/mL penicillin, and 100 mg/mL streptomycin. For TolDC differentiation, the medium was supplemented with 10 ng/mL human IL-4, 10 ng/mL GM-CSF (PeproTech), and 10 nM of vitamin D3 or calcitriol (Sigma Aldrich). For DCs, the medium did not contain vitamin D. Anti-IL-6 1 mg/mL (Invitrogen) and rat IgG isotype 1 mg/mL (eBioscience) was added during differentiation process when required. In some cases, specified in the text, vitamin D3 was added for 30 min following differentiation to DCs (DC + vitD). In other experiments, differentiation was performed in the presence of a JAK2 inhibitor (TG101348, STEMCELL) at 500 nM. Isolation and culture of peripheral blood DCs For the validation of our differentiation model, peripheral blood CD1c + DCs were isolated and cultured for three days with or without 100 nM vitamin D (Sigma Aldrich). These samples were also obtained from anonymous donors through the Catalan Blood and Tissue Bank (CBTB). The CBTB follows the principles of the World Medical Association (WMA) Declaration of Helsinki. The Committee for Human Subjects of Bellvitge Hospital approved the study (PR275/17). Given the anonymous nature of the volunteers, no information about the gender and age was provided by the CBTB. Before providing the first blood sample, all donors received detailed oral and written information, and signed a consent form at the CBTB. For this, PBMCs were isolated by Ficoll-Paque gradient centrifugation followed by a CD1c + DCs purification with CD1c (BDCA-1)+. Dendritic Cell Isolation Kit (Miltenyi Biotec) according to manufacturer instructions. Obtained cells were cultured at a concentration of 106 cells/mL in RPMI Medium 1640 + GlutaMAXTM-1 (Gibco, Life Technologies) containing 10% fetal bovine serum and 100 units/mL penicillin with or without vitamin D. METHOD DETAILS CD8+ cell proliferation assay Allogenic CD8+ T-cells isolated using negative selection with the human CD8 T Cells Kit (Invitrogen) were labeled with carboxyfluor- escein succinimidyl ester (CFSE) and seeded in 96-well plates at 200,000 cells/well, with TolDCs or DCs at different ratios (TolDC/ DC:CD8+ T-cell ratios: 1:2, 1:4, and 1:6). CD8+ cells were then stimulated with anti-CD3/CD28 Dynabeads 5 mL/mL (Invitrogen) and cultured for 5 days. Anti-IL-6 1 mg/mL (Invitrogen), anti-IL-10 1 mg/mL (eBioscience) and rat IgG isotype 1 mg/mL (eBioscience) was added during co-culture process when required. CD8+ T-cell proliferation was analyzed by FACS and determined by considering the proliferating CD8+ T-cells those where CFSE staining had decreased compared to unstimulated CD8+ T-cells. BrdU proliferation assay MOs were differentiated to DCs and TolDCs as described above and BrdU (APC BrdU Flow kit, BD Pharmingen) pulses were added to a final concentration of 10 mM at days 2 and 4. On days 3, 4 and 5 cells were harvested and 106 cells were prepared for flow cy- tometry as described by the manufacturer. In brief, cells were fixed for 30 minutes on ice, permeabilized for 5 minutes on ice and e3 Cell Reports 38, 110244, January 18, 2022 ll Article OPEN ACCESS treated with DNAse for 1 h at 37 C to expose incorporated BrdU. Cells were then stained with fluorescent anti-BrdU antibody for 20 minutes at room temperature and analyzed in a BD FACSCanto-II flow cytometer. The HAFTL pre-B cell line was used as control for proliferation. Flow cytometry For the study of surface cell markers, cells were harvested after differentiation culture and washed once with PBS. Cell staining was performed in a staining buffer (PBS with 4% fetal bovine serum and 0.4% EDTA) after blocking for non-specific binding with Fc block (BD Pharmingen) for 5 minutes on ice. Cells were stained for 20 minutes on ice. Antibodies used included: CD14-FITC, CD80-PE, CD86-APC (Miltenyi biotec), CD11b-APC, CD1a-PE (Biolegend), HLA-DR-PeCy7 (eBioscience). Cells were also stained with the viability dye LIVE/DEADTM Fixable Violet (Invitrogen) according to manuacturer’s conditions. After staining, cells were fixed with PBS + 4% paraformaldehyde and analyzed in a BD FACSCanto-II flow cytometer in the following 48 h. Cytokine measurements For in vitro experiments, the concentration of IL-10 cytokine was measured from the cell culture supernatants using an enzyme-linked immunosorbent assay (ELISA), according to the manufacturer’s instructions (BioLegend, San Diego, CA, USA). The concentration of IL-6 was measured with Beckman DXI Immunoassay analyzer using the access IL-6 reagent kit (Beckman Coulter) following the man- ufacturer’s instructions. Genomic DNA and total RNA extraction DNA was extracted with a Maxwell RSC Cultured Cells DNA kit (Promega) following manufacturer’s instructions. Similarly, total RNA was extracted with Maxwell RSC simplyRNA cells kit (Promega) following manufacturer’s instructions. Bisulfite (BS) and oxidative-bisulfite (oxBS) pyrosequencing 500 ng of genomic DNA was BS-converted with an EZ DNA Methylation-Gold kit (Zymo Research), following the manufacturer’s in- structions. The oxBS samples were purified via buffer exchange with Micro Bio-Spin P-6 SSC columns (Bio-Rad Laboratories, Inc., USA) and eluted in 22 mL MilliQ-water. After DNA denaturation with 1.25 mL NaOH (1M) for 30 min in a shaking incubator at 37 C, DNA was oxidized with 2 mL KRuO4 (15 mM) (Alfa Aesar, Germany) for 60 min in an ice-water bath (vortexing the reaction twice) and centrifuged at 16000 g for 15 min. Finally, oxidized DNA were BS converted using the EZ DNA MethylationTM kit (Zymo Research, CA, USA). BS- and oxBS-treated DNA was PCR-amplified using IMMOLASE DNA polymerase kit (bioline). Primers were designed with PyroMark Assay Design 2.0 software (Qiagen) (see Table S3 for primer sequences). Finally, PCR amplicons were pyrosequenced with the PyroMark Q24 system and analyzed with PyroMark Q48 Autoprep (Qiagen). 5mC levels were derived from the oxBS data, while 5hmC levels were calculated by subtracting the oxBS values from the BS values (5mC+5hmC) using the same biological replicate, as described in (Garcia-Gomez et al., 2017). Real-time quantitative Polymerase Chain Reaction (RT-qPCR) 250 ng of total RNA were converted to cDNA with Transcriptor First Strand cDNA Synthesis Kit (Roche) following manufacturer’s in- structions. RT-qPCR primers were designed with Primer3 software (Koressaar and Remm, 2007) (see Table S3). RT-qPCR reactions were prepared with LightCycler 480 SYBR Green I Master (Roche) according to manufacturer’s instructions and analyzed with a LightCycler 480 instrument (Roche). Western blot Protein expression and downregulation was visualized by western blotting, performed using standard Western blot. The following antibodies were used for Western blotting and Co-immunoprecipitation: Anti-Vitamin D3 Receptor (Cell Signaling), anti-TET2 (Abcam), anti-pStat3 (Fluidigm), anti-Stat3 (Cell Signaling Technology), anti-pStat1 (Cell Signaling Technology), anti-Stat1 (Cell Signaling Technology), anti-pStat5 (Fluidigm), anti-Stat5 (Thermo Fisher Scientific), anti-Lamin B1 (Abcam), anti-aTubulin (Sigma- Aldrich). anti-rabbit IgGs (Merck Millipore) and anti-mouse IgGs (Merck Millipore). Co-immunoprecipitation (Co-IP) Co-IP assays were performed using TolDCs differentiated from CD14 + monocytes for 3 days. Cell extracts were prepared in lysis buffer [50 mM Tris–HCl, pH 7.5, 1 mM EDTA, 150 mM NaCl, 1% Triton-X-100, protease inhibitor cocktail (cOmpleteTM, Merck)] with corresponding units of Benzonase (Sigma) and incubated at 4 C for 4 h. 100 mL of supernatant was saved as input and diluted with 23 Laemmli sample buffer (5x SDS, 20% glycerol, 1M Tris–HCl (pH 8.1)). Supernatant was first incubated with PureProteomeTM Protein A/G agarose suspension (Merck Millipore) for 1 h to remove background signal. The lysate was then incubated overnight at 4 C with respective crosslinked primary antibody. The cross-linking was performed in 20 mM dimethyl pimelimidate (DMP) (Pierce, Thermo Fisher Scientific, MA, USA) dissolved in 0.2 M sodium borate (pH 9.0). Subsequently, the beats were quenched with 0,2M of etha- nolamine (pH 8.0) and resuspended at 4 C in PBS until use. Beads were then washed three times with lysis buffer at 4 C. Sample elution was done by acidification using a buffer containing 0.2 M glycine (pH 2.3) and diluted with 23 Laemmli. Samples and inputs were denatured at 95 C in the presence of 1% b-mercaptoethanol. Cell Reports 38, 110244, January 18, 2022 e4 ll OPEN ACCESS Article DNA methylation profiling Infinium MethylationEPIC BeadChip (Illumina, Inc., San Diego, CA, USA) arrays were used to analyze DNA methylation. This platform allows >850,000 methylation sites per sample to be interrogated at single-nucleotide resolution, covering 99% of the reference sequence (RefSeq) genes. The samples were bisulfite-converted using EZ DNA Methylation-GoldTM Kit (Zymo Research, Irvine, CA, USA) and were hybridized in the array following the manufacturer’s instructions. Image processing and intensity data extraction software and procedures were as previously described (Bibikova et al., 2006). Each methylation data point was obtained from a com- bination of the Cy3 and Cy5 fluorescent intensities from the methylated and unmethylated alleles. Background intensity computed from a set of negative controls was subtracted from each data point. For representation and further analysis, we used beta and M values. The beta value is the ratio of the methylated probe intensity to the overall intensity (the sum of the methylated and unmethy- lated probe intensities). It can take a value between 0 and 1, and was used to derive heatmaps and to compare DNA methylation percentages from bisulfite-pyrosequencing experiments. The M value is calculated as the log2 ratio of the intensities of the methyl- ated versus unmethylated probes. For the purpose of statistical analysis, M values are more suitable because they are normally distributed. Raw methylation data were preprocessed with the minfi package (Aryee et al., 2014). Data quality was assessed using the minfi and RnBeads packages (Aryee et al., 2014; Assenov et al., 2014; Mu €ller et al., 2019). After Snoob normalization, data were analyzed using aneBayes moderate t test available in the limma package (Ritchie et al., 2015). Several criteria have been proposed as representing significant differences in methylated CpGs, but in this study we considered a probe to be differentially methylated if it had a methyl- ation differential of 20% and if it was significant (q < 0.05). ChIP-seq analysis Chromatin immunoprecipitation was performed using the iDeal ChIP-seq kit for Transcription Factors (Diagenode), according to the manufacturer’s instructions. Briefly, cells on day 3 of differentiation were cross-linked with 1% formaldehyde for 15 min and glycine was added to quench the reaction (final concentration 125 mM, incubated for 5 min at room temperature). Cells were washed once with cold PBS, scraped off the plates, and pelleted. To obtain a soluble chromatin extract, cells were resuspended in 1 mL LB1 (50 mM HEPES, 140 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% NP-40, 0.25% Triton X-100 and 13 complete protease inhibitor) and incubated while rotating at 4 C for 10 min. Samples were centrifuged, resuspended in 1 mL LB2 (10 mM Tris-HCl pH 8.0, 200 mM NaCl, 1 mM EDTA, 0.5 mM EGTA and 13 complete protease inhibitor) and incubated while rotating at 4 C for 10 min. Finally, samples were centrifuged, resuspended in 1 mL LB3 (10 mM Tris-HCl pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 0.1% so- dium deoxycholate, 0.5% N-lauroylsarcosine, 1% Triton X-100 and 13 complete protease inhibitor). Chromatin extracts were son- icated for 12.5 min using a Covaris M220 focused ultrasonicator at a peak power of 75, and a duty factor of 10 and 200 cycles per burst. The lysates were incubated with anti-VDR antibody (12,550, Cell Signaling) bound to 30 mL protein A or protein G Dynabeads and incubated overnight at 4 C, keeping 5% as input DNA. Magnetic beads were sequentially washed with low-salt buffer (150 mM NaCl, 0.1% SDS, 1% Triton X-100, 1 mM EDTA and 50 mM Tris-HCl), high-salt buffer (500 mM NaCl, 0.1% SDS, 1% Triton X-100, 1 mM EDTA and 50 mM Tris-HCl), LiCl buffer (150 mM LiCl, 0.5% sodium deoxycholate, 0.1% SDS, 1% Nonidet P-40, 1 mM EDTA and 50 mM Tris-HCl) and TE buffer (1 mM EDTA and 10 mM Tris-HCl). For ChIP-seq, beads were resuspended in elution buffer (1% SDS, 50 mM Tris-HCl pH 8.0, 10 mM EDTA and 200 mM NaCl) and incubated for 30 min at 65 C. After centrifugation, the eluate was reverse-cross-linked overnight at 65 C. The eluate was then treated with RNaseA for 1 h at 37 C and with Proteinase K (Roche) for 1 h at 55 C and the DNA was recovered using a Qiagen PCR purification kit. Sequencing reads from ChIP-seq experiments were mapped to the hg19 assembly of human reference genome using Bowtie2 Aligner v2.2.6 (Langmead and Salzberg, 2012). After removing reads with MAPQ < 30 with Sequence Alignment/Map (SAMtools) v1.2 (Li et al., 2009), PCR duplicates were eliminated using the Picard function available in MarkDuplicates software v1.126. Peak calling was determined using SPP (with parameters –npeak=300000 –savr –savp -rf). The irreproducible discovery rate (IDR) was used to filter peaks (IDR < 0.05). To visualize individual ChIP-seq data on Integrative Genomics Viewer (IGV), we converted bam output files to normalized bigwig format using the bamCoverage function in deepTools (v2.0). ChIP-qPCR ChIP assays were performed as previously described (Li et al., 2020). Briefly, MOs, DCs and TolDCs were crosslinked with 1% meth- anol-free formaldehyde (Thermo Fisher) for 15 min and subjected to immunoprecipitation after sonication. ChIP experiments were performed using the LowCell# ChIP kitTM protein A (Diagenode, Liège, Belgium). We used antibodies against vitamin D3 Receptor (Cell Signaling), acetylated H3 (H3ac), trimethylated lysine 27 of histone H3 (H3K4me3Millipore) and trimethylated lysine 4 of histone H3. Corresponding rabbit IgG (Diagenode) is used as control. Protein binding was analyzed by real-time quantitative PCR, and data are represented as ratio of the enriched fraction with respect to input. ChIP primers were designed for the areas flanking differentially methylated CpGs and their sequences are shown in Table S3. Microarray reanalysis Affymetrix datasets from human monocytes, and from in vitro- and in-vivo DCs and MACs were obtained from GSE40484 (Segura et al., 2013) and GSE102046 (Goudot et al., 2017). Affymetrix raw data from MOs, DCs and TolDCs were obtained from GSE13762 (Széles et al., 2009). Affymetrix raw data files were normalized by the robust multiarray average (RMA) algorithm and summarized, e5 Cell Reports 38, 110244, January 18, 2022 ll Article OPEN ACCESS after background correction, using the R package oligo (Carvalho and Irizarry, 2010). Normalized expression datasets were then merged and corrected for batch effects using ComBat function of the sva package. Finally, t-Distributed Stochastic Neighbor Embedding (tSNE) of 1000 most variant genes was performed in R using Rtsne package. Data analysis Hierarchical clustering was carried out based on Pearson correlation distances and average linkage criteria. For low-dimensional analysis, we used principal component analysis (PCA). Transcription-factor motifs were enriched for each set using HOMER software v4.10.3. Specifically, we used the findMotifsGenome.pl algorithm (with parameters -size 200 -cpg) to search for significant enrich- ment against a background sequence adjusted to have similar CpG and GC contents. Genomic regions for genetic context location were annotated using the annotatePeaks.pl algorithm in the HOMER v4.10.3 software application (Heinz et al., 2010). To determine the location relative to a CpG island (CGI), we used ‘hg19_cpgs’ annotation in the annotatr v1.8 R package. GREAT software (McLean et al., 2010) was used to enrich downstream pathways and gene ontologies. We used the single nearest gene option to identify as- sociations between genomic regions and genes. Chromatin state analysis for DCs were assessed using the EpiAnnotator R package (Pageaud et al., 2018). Inference of TF activities from expression values were calculated using DoRothEA (Garcia-Alonso et al., 2019). We used the nichenetr package (Browaeys et al., 2020) to predict ligand activity. QUANTIFICATION AND STATISTICAL ANALYSIS All statistical analyses were done in R v3.5.1. Data distributions were tested for normality. Normally distributed data were tested using two-tailed unpaired Student’s t-tests; non-normal data were analyzed with the appropriate non-parametric statistical test. Levels of significance are indicated as: *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. Non-significance (P R 0.05) is indicated as ‘ns’. Cell Reports 38, 110244, January 18, 2022 e6 Cell Reports, Volume 38 Supplemental information Vitamin D receptor, STAT3, and TET2 cooperate to establish tolerogenesis Francesc Català-Moll, Anna G. Ferreté-Bonastre, Gerard Godoy-Tena, Octavio Morante- Palacios, Laura Ciudad, Laura Barberà, Federico Fondelli, Eva M. Martínez- Cáceres, Javier Rodríguez-Ubreva, Tianlu Li, and Esteban Ballestar Figure S1 A B CD14 CD1A CD86 25000 5000 DC 120h ns ns **** ns ns **** GM−CSF/IL4 mo−DC 10000 **** 20000 ** 4000 **** ascites mo−DC **** ** **** 5 CD1c DC 15000 3000 Mean of fluorescence intensity 5000 10000 2000 M−CSF mo−DC CD16 MO 5000 1000 IL34 mo−DC t−SNE 2 0 0 0 0 DC 12h CD11B HLA−DR CD11C TolDC 120h CD14 MO TolDC 12h 60000 ns ** ns ns 4000 ns 6000 * ascites mo−M AC * ns ns MO 0h ** * ns −5 3000 M−CSF mo−M AC 40000 4000 2000 IL34 mo−M AC 20000 2000 1000 −10 −10 −5 0 5 10 0 0 0 t−SNE 1 DC lDC DC olDC DC lDC DC olDC DC lDC DC olDC To T To T To T immature mature immature mature immature mature MO DC MAC C 1:2 1:4 1:6 D ns **** ns ns ns *** 150 ns ns ns VDR % of CD8+ T cells * **** ns 100 Lamin B Tubulin A 50 MO DC TolDC DC TolDC MO DC TolDC DC TolDC day 1 day 3 day 1 day 3 Cytoplasm Nucleus 0 C C C C C C DC TolD DC TolD DC TolD DC TolD DC TolD DC TolD immature mature immature mature immature mature F E cluster 1 cluster 2 cluster 3 cluster 4 1.00 ns ns ns **** DNA methylation level 0.75 MO TolDC DC 0.50 0.25 cluster 2: 311 CpGs cluster 1: 429 CpGs CDc1+ CDc1+ CDc1+ CDc1+ CDc1+ CDc1+ CDc1+ CDc1+ -3 0 3 +VitD +VitD +VitD +VitD normalized Beta value Figure S1. Characterization of DCs and TolDCs, Related to Figure 1 (A) Analysis of DC and TolDC surface markers by flow cytometry with (mature) and without (immature) LPS activation (n = 4, two independent experiments) in differentiated DCs and TolDCs. (B) t-Distributed Stochastic Neighbor Embedding (tSNE) analysis of 1000 most variable genes. Publicly available gene expression data (Goudot et al, 2017; Széles et al, 2009; Segura et al, 2013) of various subpopulations of monocytes (MOs), macrophages (MACs) and dendritic cells (DCs) were integrated as described in Methods. This analysis included mo-MACs and mo-DCs obtained from ascites (ascites-mo-MAC, ascite-mo-DC) (C) In vitro proliferation assay of allogeneic CD8+ T cells co-cultured with DCs and TolDCs at ratios of 1:2, 1:4 and 1:6 before (immature) and after (mature) LPS activation. CD8+ T cell proliferation as analyzed by flow cytometry utilizing CFSE staining (n = 4, two independent experiments). (D) Representative western blot assays showing levels of vitamin D receptor in cytoplasm and nucleus of MO, DCs and TolDCs at different time points of differentiation. Tubulin A and Lamin B were used as cytoplasmic and nuclear markers respectively. (n=3, one experiment) (E) Heatmap showing DNA methylation levels of four paired samples of MOs and their derived DCs and TolDCs. The heatmap includes all CpG-containing probes displaying significant methylation changes (differential of beta value >= 0.2 & q-value < 0.05) in all of three possible comparisons (DC-MO, TolDC-MO and TolDC -DC). The color annotation marks the membership to cluster 1 (DC-specific DNA demethylation) or cluster 2 (TolDC-specific DNA demethylation) as defined in Figure 1C (n=4, two independent experiments). (F) Box and violin plot representation of DNA methylation levels from in vivo circulating DCs (CD1c+ DCs) cultured in vitro for 48h with or without Vitamin D (VitD and DC, respectively). CpGs shown are classified in four clusters as defined in Figure 1C,D. Statistical significance in A and C was calculated using Student’s t-test (* p-value <0.05, ** p-value < 0.01, *** p-value < 0.001, **** p-value < 0.00001). 1 Figure S2 A C Overlap with MO DNase data (%) 100 Proliferating cells (%) 80 ns Open Closed 60 40 50 20 0 Negative Positive DC TolDC 0 control control cluster 1 cluster 2 B IRF4 C1QB CD14 DPF3 * ** 15 30 DNA Hydroxymethylation (%) 10 20 10 10 5 5 0 0 0 0 24 72 0 24 72 0 24 72 0 24 72 Time (h) Time (h) Time (h) Time (h) DC TolDC Figure S2. DNA methylation dynamics during differentiation to TolDCs, Related to Figure 2. (A) Dot plot showing the results of a BrdU proliferation assay for the DCs and TolDCs differentiation model at day 5 (n = 3, one experiment). HAFTL cell line was used as a control. (B) Dot plot representation of 5hmC levels of two representative CpGs from cluster 1 (top) and 2 (bottom) at 24h and 72h of differentiation to DCs and TolDCs. Statistical test: unpaired two-tailed t-test (* p-value < 0.05, ** p-value < 0.01) (n=3, one experiment). (C) Bar plot representation of the chromatin accessibility state of the CpGs from cluster 1 and 2. Chromatin accessibility data was obtained from a DNase dataset from MOs. 1 Figure S3 A B 2400 VDR ChIP-Seq signal MO 2400 DC 2400 TolDC cluster 3 cluster 4 Mean coverage (1bp bin) DNase CpG module 2 (cg25310867) MO GAB2 800 800 400 400 13500 VDR ChIP-Seq signal MO 0 0 13500 -2 -1 0 1 2 -2 -1 0 1 2 DC Distance (Kbp) Distance (Kbp) 13500 TolDC DNase CpG module 2 (cg14914214) MO HIF1A-AS1 HIF1A C VDR H3ac H3K27me3 H3K4me3 * *** 0.20 0.4 0.75 0.3 ** * 0.15 HOPX 0.3 % input 0.50 0.2 * 0.10 0.2 0.25 0.1 0.05 0.1 * 0.00 0.0 0.00 0.0 5 0.6 * **** ** 0.4 0.75 4 ** 0.3 0.4 % input 3 0.50 * * IL6 0.2 2 0.2 ** 0.25 * 1 * 0.1 ** VDR 0 0.0 0.00 0.0 H3ac H3K27me3 0.3 * H3K4me3 * * 0.3 0.6 0.3 IgG * * 0.2 % input 0.2 0.2 0.4 INHBA * 0.1 * 0.2 0.1 0.1 ** 0.0 0.0 0.0 0.0 0.15 * * 0.4 * 0.15 3 * * ** LYRM1 * 0.10 0.3 % input * 0.10 * 2 0.2 0.05 0.05 1 0.1 0.00 0 0.00 0.0 itD itD itD itD C C C C O O O O C C C C lD D lD lD D D D D M M M M +V +V +V +V l To To To To C C C C D D D D Figure S3. Schemes zooming at TolDC-demethylated genes with VDR peaks, Related to Figure 4. (A) Composite plots of VDR ChIP-seq distribution ± 2.5 Kbp around CpGs from cluster 3 (left) and cluster 4 (right) in MO (grey), DC (red) and TolDC (green). Smooth represents the confidence intervals (CIs). (n=2, one experiment) (B) VDR ChIP-seq signal profiles in the vicinity of the representative genes with CpGs of cluster 2. VDR signals are colored by cell type. At the bottom, the significant VDR binding sites are shown in green, CpG position is depicted below in red DNase significant peaks in MOs are colored in blue. (C) Bar plot representation of ChIP-qPCR results for VDR binding and three histone modifications (H3ac, H3K27me3 and H3K4me4) in the vicinity of VDR peaks close to HOPX, IL6, INHB and LYRM1 gene sequences. This analysis was performed in MO, DC and TolDC and DC+VitD, which consist of adding vitamin D for 30 min at the end of a 6-day differentiation to DCs. Immunoprecipitation with IgG was used as control. (n=3, one single experiment) Statistical test: unpaired two-tailed t-test (* p-value < 0.05, ** p-value < 0.01, *** p-value < 0.001, **** p-value < 0.0001). 1 Figure S4 A B C DC TolDC DC TolDC ** Log 2 (IL-6 Concentration (pg/mL)) DNA Hydroxymethylation (%) * 100 100 * * 40 30 DNA Methylation (%) 10 90 90 30 20 80 5 20 70 10 80 10 60 0 0 0 DC TolDC DC TolDC DC TolDC 0 24 72 0 24 72 0 24 72 0 24 72 Time (h) Time (h) Time (h) Time (h) day 1 day 2 day 7 D E 0.039 0.61 0.34 0.028 0.022 100 0.031 0.22 0.046 0.026 50 IL-10 concentration ( pg/ml) 75 Proliferating cells (%) 40 50 30 25 20 0 G -6 0 G -6 0 C C C C -1 -1 Ig Ig IL IL D D D D C C C C IL IL l l α- α- D D D D + + To To G -6 α- α- l l C C Ig To To G -6 IL + + G -6 + + D D Ig α- IL C C Ig G IL l -6 C C To D D α- Ig α- IL D D l To l α- To F pStat1 pStat3 pStat5 * * ns 1.5 1.0 1.0 1.0 Protein quantification (a.u.) 0.5 0.5 0.5 0.0 0.0 0.0 DC TolDC DC TolDC DC TolDC tStat1 tStat3 tStat5 1.5 2.0 ns 1.5 ns ns 1.5 1.0 1.0 1.0 0.5 0.5 0.5 0.0 0.0 0.0 DC TolDC DC TolDC DC TolDC Figure S4. Activation of the IL-6/JAK2/STAT3 pathway in TolDCs, Related to Figure 5. (A) Dot plot representation of IL-6 secretion to the medium by DC and TolDC at day 1, 3 and 7 of differentiation (n = 3, one single experiment). (B-C) Dot plots representing DNA methylation levels (B) and DNA hydroxymethylation levels (C) of upstream regions of IL6 gene on DC and TolDCs at 24h and 72h of differentiation (n=3, one single experiment). CpGs studied are cg02979021 (left) and cg07816106 (right). (D) Dot and box plot representation of the effect of the presence α-IL-6 antibody during the differentiation of DC or TolDC on IL-10 secretion to the medium. Each biological replicate is distinctly represented with a different shape. (n=4, one single experiment) (E) Dot and box plot representation of the effect on CD8+ proliferation of DC and TolDC in the presence of blocking antibodies against IL-6 or IL-10, and isotype IgG was used as a control. IgG DC and IgG TolDC, were differentiated in the presence of either IgG and samples α-IL-6 DC and α-IL-6 TolDC were differentiated in the presence of anti-IL-6. The samples DC + IgG, TolDC + IgG, DC + α-IL-6, TolDC + α-IL-6, DC + α-IL- 10, TolDC + α-IL-10 were differentiated in the absence of any antibody; IgG, α-IL-6 or α-IL10 antibodies were added during the proliferation assay, respectively. Each biological replicate is distinctly represented with a different shape (n=4, two independent experiments) (F) Dot and bar plot representation of STAT1/STAT3/STAT5 protein quantification by Western Blot of DCs and TolDCs (n=3). 1 Figure S5 A B DC TolDC TG101348 - + - + p-STAT1 p-STAT3 p-STAT5 Relative phosphorylation 1.5 2.5 1.5 p-STAT1 * 2.0 1.0 1.0 * t-STAT1 1.5 1.0 p-STAT3 0.5 0.5 0.5 t-STAT3 0.0 0.0 0.0 48 - + - + - + - + - + - + p-STAT5 0 13 T G1 DC TolDC DC TolDC DC TolDC t-STAT5 α-Tubulin Figure S5. Specific inhibition of STAT3 phosphorylation reverts immunosuppressive properties of vitamin D exposed dendritic cells, Related to Figure 6. (A) Representative Western Blot showing the effect of TG101348 at the phosphorylation level of STAT1, STAT3 and STAT5 in DCs and TolDCs. (B) Bar and dot plot showing the proportion of phosphorylated STAT1, STAT3 and STAT5 as measured by Western Blot with or without TG101348 (n = 3, two independent experiments). 1