Abstract
FOXP3-expressing regulatory T (Treg) cells play a pivotal role in maintaining immune homeostasis and tolerance, with their activation being crucial for preventing various inflammatory responses. However, the mechanisms governing the epigenetic program in Treg cells during their dynamic activation remain unclear. In this study, we demonstrate that CXXC finger protein 1 (CXXC1) interacts with the transcription factor FOXP3 and facilitates the regulation of target genes by modulating H3K4me3 deposition. Cxxc1 deletion in Treg cells leads to severe inflammatory disease and spontaneous T-cell activation, with impaired immunosuppressive function. As a transcriptional regulator, CXXC1 promotes the expression of key Treg functional markers under steady-state conditions, which are essential for the maintenance of Treg cell homeostasis and their suppressive functions. Epigenetically, CXXC1 binds to the genomic regulatory regions of Treg program genes in mouse Treg cells, overlapping with FOXP3 binding sites. Given its critical role in Treg cell homeostasis, CXXC1 presents itself as a promising therapeutic target for autoimmune diseases.
Introduction
Regulatory T (Treg) cells are a distinct subset of CD4+ T cells that play a critical role in maintaining immune homeostasis and self-tolerance by suppressing excessive or aberrant immune responses to foreign or self-antigens 1–3. These cells can be further categorized into thymus-derived regulatory T cells (tTreg cells), periphery-derived Treg cells (pTreg cells), and induced Treg cells (iTreg cells) 4. They uniquely express the transcription factor FOXP3, a member of the forkhead winged-helix family, which is essential for Treg cell lineage commitment and suppressive function 5,6. Deletion or mutation of the Foxp3 gene leads to a range of immunological disorders, including allergies, immunopathology, and autoimmune diseases in both mice and humans 7,8.
It is well established that FOXP3 recruits various cofactors to form complexes that either promote or repress the expression of downstream genes, with histone and DNA modifications playing pivotal roles in this process. FOXP3 can activate or repress the transcription of key regulators of Treg cell activation and function by recruiting the histone acetyltransferases (HATs) or histone deacetylases (HDACs)9,10. Notably, FOXP3-bound sites exhibit significant enrichment of H3K27me3, a modification essential for FOXP3-mediated repressive chromatin remodeling under inflammatory conditions11. However, the direct role of FOXP3 as a transcriptional activator through interactions with epigenetic regulators, particularly via modulation of H3K4 trimethylation, remains poorly documented.
In mammals, six proteins have been identified that catalyze H3K4 methylation. These proteins contain the SET domain and include MLL1 (KMT2A), MLL2 (KMT2B), MLL3 (KMT2C), MLL4 (KMT2D), SETD1A, and SETD1B 12,13. For example, MLL1 plays a critical role as an epigenetic regulator in Treg cell activation and functional specialization14. Additionally, Placek et al. demonstrated that MLL4 is essential for Treg cell development by catalyzing H3K4me1 at distant unbound enhancers through chromatin looping 15. H3K4me3, which is enriched at the transcription start site (TSS) and the CpG island (CGI), converts chromatin into active euchromatin by recruiting activating factors 16. CXXC finger protein 1(CXXC1, also known as CFP1), which contains a SET1 interaction domain (SID), is required for binding to the histone H3K4 methyltransferases SETD1A and SETD1B 17. Previous studies have shown that CXXC1 regulates promoter patterns during T cell maturation, GM-CSF-derived macrophage phagocytosis, TH17 cell differentiation, and the function of ILC3 cells with aging by modulating H3K4me3 modification18–21. Despite the well-documented role of CXXC1 in various immune effector cells, its role in Treg cells remains unclear.
Here, we demonstrate that CXXC1 interacts with FOXP3 and enhances the expression of FOXP3 target genes. Treg cell-specific deletion of Cxxc1 triggers systemic autoimmunity, accompanied by multiorgan inflammation, ultimately resulting in early-onset fatal inflammatory disease in mice. Cxxc1-deficient Treg cells exhibit a significant disadvantage in proliferation and homeostasis, even in non-inflammatory mice where coexisting wild-type (WT) Treg cells were present. Moreover, Cxxc1-deficient Treg cells display intrinsic defects in the expression of key suppression molecules, including CTLA-4, CD25, ICOS, and GITR. Consistent with these findings, mechanistic studies strongly suggest that CXXC1 functions as an essential cofactor of FOXP3, playing a crucial role in sustaining H3K4me3 modifications at key Treg cell genes.
Materials and methods
Mice
All mice used in this study were bred for a minimum of seven generations on a C57BL/6 background. Mouse experiments, or cells from mice of the same genotype, compared littermates or age-matched control animals. The Cxxc1fl/fl mouse strain has been previously described 22. The Foxp3cre mice (JAX,016959) were generously provided by Bin Li (Shanghai Jiao Tong University School of Medicine, Shanghai, China). CD45.1 (NM-KI-210226) mice were purchased from the Nanjing Biomedical Research Institute of Nanjing University. Rag1-/- mice (stock#T004753) were purchased from GemPharmatech. 2D2 (MOG35-55-specific TCR transgenic) mice were graciously supplied by Prof. Linrong Lu (Zhejiang University School of Medicine, Hangzhou, Zhejiang, China). Foxp3cre (WT) and Foxp3cre Cxxc1fl/fl (cKO) were utilized at 3 weeks old unless otherwise stated. All mice were housed in the Zhejiang University Laboratory Animal Center under specific pathogen-free conditions, and all animal experimental procedures were approved by the Zhejiang University Animal Care and Use Committee (approval no.ZJU20230246).
Cell culture
HEK 293T cells were obtained from ATCC, and Plat E cells were kindly provided by Prof. Xiaolong Liu (Shanghai Institutes for Biological Sciences). Both cell lines were cultured in Dulbecco-modified Eagle medium (DMEM) containing 10% (v/v) fetal bovine serum (FBS), supplemented with 1% penicillin/streptomycin.
Immunofluorescence (IF) Staining
As previously described 23, coverslips were treated with a 0.01% poly-L-lysine solution (P4707; Sigma) for 10 minutes, air-dried, and then coated with CD4+ YFP+ Treg cells. The cells were then fixed in 4% formaldehyde for 15 minutes at room temperature, permeabilized with 0.2% Triton X-100, and blocked with 1% BSA. Antibodies against CXXC1 (ab198977; Abcam) and FOXP3 (17-5773-82; Invitrogen) were diluted in Image iT FX signal enhancer (I3693; Invitrogen) and incubated with cells overnight at 4 °C. After washing with phosphate-buffered saline (PBS), the cells were incubated with a goat anti-rabbit antibody Alexa Fluor 594 (1:250;10015289; Invitrogen) secondary antibody and stained with DAPI (200 ng/mL; D523; Dojindo). Slides were washed with PBS and sealed with an antifade solution(P36934; Invitrogen) before imaging with an Olympus FV3000 fluorescence microscope. The images were visualized using the FV31-SW software.
Co-immunoprecipitation and western blot
Harvest the appropriate transfected cell lines and primary cells from culture and wash them with ice-cold PBS. Lyse the cells in NETN300 buffer (300 mM NaCl, 0.5 mM EDTA, 0.5% (v/v) NP-40, 20 mM Tris-HCl pH 8.0) supplemented with a protease inhibitor (1:100, P8340, Sigma-Aldrich) and PMSF (1 mM) on ice for 10 minutes. Take a small portion of the whole-cell lysate as input, and incubate the remaining lysate with either Anti-FLAG M2 Beads (M8823; Sigma) or Anti-HA Beads (HY-K0201; MCE) on a rotator at 4°C overnight. For the endogenous Co-IP assay targeting CXXC1 and FOXP3, incubate the cell lysate with protein G magnetic beads along with anti-CXXC1 (ab198977; Abcam) or anti-FOXP3 (14-4774-82; Invitrogen) antibodies on a rotator at 4°C overnight. Wash the beads three times with IP buffer (100 mM NaCl, 0.5 mM EDTA, 0.5% (v/v) NP-40, 20 mM Tris-HCl pH 8.0) to remove non-specific binding. Boil the washed beads with 1× Laemmli sample buffer (1610747; Bio-Rad) to elute the bound proteins. Separate the denatured proteins by SDS-PAGE. Transfer the separated proteins onto PVDF membranes (Millipore) for immunoblotting.Immunoblot the PVDF membranes (IPVH00010) with the following antibodies: anti-CXXC1 (1:1000; ab198977; Abcam), anti-FOXP3 (1:500; 14-7979-80; Invitrogen), anti-FlAG (1:1000; 14793; Cell Signaling Technology), anti-HA (1:1000; 3724S; Cell Signaling Technology).Detect the immunoblotted proteins using a secondary HRP-conjugated goat anti-rabbit antibody (1:1000; HA1001-100; Huabio) and visualize the bands using an appropriate detection method.
Flow cytometry
Spleen and lymph nodes were harvested from 3-week-old mice. Tissues were placed in a dish containing PBS with 2% fetal bovine serum (FBS). The tissues were minced into small pieces and then filtered through a 70 µM nylon filter to obtain a single-cell suspension. For surface marker analysis, cells were incubated for 15 minutes with purified anti-mouse CD16/32 antibody (101320; BioLegend) to block Fc receptors. After blocking, cells were stained with the indicated antibodies for surface markers. To determine cytokine expression, cells were stimulated for 4 hours at 37℃ with phorbol12-myristate13-acetate (50ng/mL; S1819; Beyotime), ionomycin (1mg/ml; S1672; Beyotime), and brefeldin A (BFA;00-4506-51; Invitrogen). After stimulation, cells were labeled with a fixable viability dye and surface markers. Cells were then fixed and permeabilized according to the manufacturer’s instructions (00-8222-49; Invitrogen). For transcription factor staining, samples were fixed using the FOXP3/Transcription Factor Staining Buffer Set (00-5523-00; Invitrogen). Flow cytometry was performed using a BD Fortessa (BD Biosciences). Flow cytometry data were acquired and analyzed using FlowJo software.
The following antibodies were purchased from Invitrogen or BioLegend: Zombie Violet f-ixable viability(423113), Zombie NIR fixable(423105), CD25(PC61), CD8α(53-6.7), CD62L(MEL-14), PD-1(J43), CD44(IM7), IL-17A(TC1118H10), KL-RG1(2F1), CD4 (GK1.5), TCRβ(H57-597), IFN-γ(XMG1.2), FOXP3(FJK-16s), CTLA-4(UC10-4B9), ICOS(15F9), GITR(DTA-1), LAG3(C9B7W), TCRVβ11(RR3-15), CD45RB(C363-16A) CD69 (H1.2F3), MHCII(M5/114.15.2), CD80(16-10A1), T-bet(eBio4B10), IKZF4(ESB7C2), IL-4(11B11), CCR7(4B12), CD73(TY/11.8).
Real-time PCR
Total RNA was extracted from Treg cells using the RNAAiso Plus(9109; Takara) reagent according to the manufacturer’s instructions, and cDNA synthesis was performed using the Prime Script RT Reagent Kit (Takara). TB Green Premix Ex Taq (RR420A; Takara) was used for quantitative real-time PCR (qPCR). The expression levels of target mRNA were normalized to the level of β-actin expression. The primers for qPCR are as follows:
Cxxc1 qPCR Forward: ATCCGGGAATGGTACTGTCG
Cxxc1 qPCR Reverse: CTGTGGAGAAGATTTGTGGG
β-actin qPCR Forward: CTGTCCCTGTATGCCTCTG
β-actin qPCR Reverse: ATGTCACGCACGATTTCC
CD4+T and YFP+ Treg cells adoptive transfer in Experimental autoimmune encephalomyelitis (EAE)
CD4+YFP+ Treg cells from Foxp3cre and Foxp3cre Cxxc1fl/fl mice were enriched using the Mouse CD4 T Cell Isolation Kit (480005; Biolegend) and then sorted using the BD Aria II flow cytometer. Naïve CD4+T cells from 2D2 (MOG35-55-specificTCR transgenic) mice were isolated using the Mouse CD4 Naïve T cell Isolation Kit480039 (480039; Biolegend). As previously described24, 2D2 naïve CD4+ T cells alone (5 x 105 per mouse), or 2D2 naïve CD4+ T cells (5x 105 per mouse) together with WT or Foxp3creCxxc1fl/fl Treg cells (2x105 per mouse), were transferred into Rag1−/− mice via the tail vein. One day after cell transfer, the recipient mice were inoculated subcutaneously (s.c.) with 200μg MOG35-55 peptide (MEVGWYRSPFSRVVHLYRNGK; GenemeSynthesis) emulsified in complete Freund’s adjuvant (CFA) (F5506; Sigma). Intravenous administration of 200 ng of Pertussis toxin (181; List Biological Laboratories) was performed on days 0 and 2 after peptide inoculation. The severity of EAE was monitored and blindly graded using a clinical score from 0 to 5: 0, no clinical signs; 1, limp tail; 2, paraparesis (weakness, incomplete paralysis of one or two hind limbs); 3, paraplegia (complete paralysis of two hind limbs); 4, paraplegia with forelimb weakness or paralysis; 5, dying or death.
Isolation lymphocytes from the central nervous system (CNS)
On day 14 after EAE induction, mice were perfused with transcardially administered PBS to eliminate contaminating blood cells in the central nervous system (CNS). The forebrain and cerebellum were dissected to expose the spinal cord, which was then carefully removed from the spinal canal. The fresh spinal cord was harvested and cut into 2 mm pieces. The CNS tissue pieces were homogenized using a syringe and passed through a 70 µM cell strainer to obtain a single-cell suspension. The single-cell suspension was digested with collagenase D (2μg/ml;11088858001; Roche) and deoxyribonuclease I (DNase I; 1 μg/ml; DN25; Sigma-Aldrich) at 37°C for 20 minutes under rotation. After digestion, the cell suspension was centrifuged to pellet the cells. The cell pellets were resuspended in 40% Percoll and layered onto a discontinuous Percoll gradient. Centrifugation at 80% Percoll allowed for the separation of cells at the 40–80% Percoll interface, which were collected as CNS mononuclear cells. The collected CNS mononuclear cells were washed with PBS to remove any remaining Percoll. CNS mononuclear cells were stimulated for 4 hours with PMA and ionomycin in the presence of Brefeldin A to induce cytokine production. After stimulation, cells were fixed, rendered permeable, and stained with appropriate antibodies for intracellular cytokine detection.
Histological analyses
The lungs, skin, liver, and colon were excised from three-week-old mice. Prior to histological analysis, the samples were fixed in formalin, embedded in paraffin, and stained with hematoxylin and eosin. For CNS histology, spinal cords were fixed in 4% paraformaldehyde, paraffin-embedded, sectioned, and stained with Luxol Fast Blue and hematoxylin and eosin (H&E). To examine colon histology, colons from Rag1−/− hosts were similarly processed and stained with H&E.
Adoptive transfer colitis model
Colitis was induced following the protocol described25. In brief, CD4+ YFP+ Treg cells were isolated from 3-week-old CD45.2+Foxp3creCxxc1fl/fl and CD45.2+ Foxp3cre mice. A total of 2x105 Treg cells from each group were mixed with 4x105 Teff cells (CD45.1+CD4+CD45RBhi) sorted from CD45.1+ mice and transferred into the Rag1-/- mice via intraperitoneal injection. Teff cells alone were transferred as a control group. Mouse body weight was measured weekly post-adoptive transfer. The percentage change in body weight was calculated by comparing the current weight with the initial weight on day 0. Mice were euthanized when any had reached 80% of their initial body weight. The large intestines were sectioned into 4μm thick slices and stained with hematoxylin.
In vitro Treg suppression assay
Naïve CD4+ T cells isolated from WT mice were labeled with CFSE (C34554; Invitrogen). CD4+YFP+ Treg cells from Foxp3cre and Foxp3creCxxc1fl/fl mice were cultured with naïve CD4+ T cells (1ⅹ105 cells) at various ratios in the presence of 2μg/mL anti-CD3(16-0031-85; Invitrogen) and 3μg/mL anti-CD28(16-0281-85; Invitrogen). On day 3, cells were analyzed by flow cytometry.
CUT&Tag
CUT&Tag assays of CD4+YFP+ Treg cells were conducted as previously described26. Briefly, approximately 1x105 single cells were carefully pipetted into wash buffer twice. The pelleted cells were resuspended in wash buffer, activated concanavalin (BP531; Bangs Laboratories), and incubated for 15 minutes at room temperature. Cells bound to the beads were resuspended in Dig-Wash Buffer and incubated with a 1:50 dilution of primary antibodies (rabbit anti-H3K4me3, Active Motif,39016; rabbit anti-CXXC1, abcam, ab198977; normal IgG, Cell Signaling, 2729) at 4 ℃ overnight. The beads were incubated with a secondary antibody (goat anti-rabbit IgG; SAB3700883; Sigma-Aldrish) diluted 1:100 in Dig-Wash buffer for 60 minutes at room temperature. Cells were treated with Hyperactive pG-Tn5 Transposase (S602; Vazyme) diluted in Dig-300 Buffer for 1 hour at room temperature. The cells were subsequently resuspended in Tagmentation buffer (10mM MgCl2 in Dig-300 Buffer) and incubated at 37℃ for 1 hour. To halt tagmentation, 10μl was spiked with 0.5M EDTA, 3μl with 10%SDS, and 3μl with 20mg/ml Proteinase K and incubated at 55℃ for 1h. DNA library amplification was performed according to the manufacturer’s instructions and purified using VAHTS DNA Clean Beads (N411; Vazyme). Libraries were sequenced on the Illumina NovaSeq platform (Annoroad Gene Technology).
CUT&Tag and ChIP-seq data analysis
FOXP3 ChIP-seq data was obtained from GSE121279. H3K27me3 ChIP-seq data was obtained from GSE14254. CUT&Tag and ChIP-seq reads were trimmed to 50 bp and aligned against the mouse genome build mm9 using Bowtie2(v2.3.4.1) with default parameters. All PCR duplicates and unmapped reads were removed. Peak calling was performed using MACS2(v2.1.1.20160309) and signal tracks for each sample were generated using the ‘wigToBigWig’ utility of UCSC. We classified the H3k4me3 peaks around TSSs into three groups: broad (>5 kb), medium (1-5kb), and narrow (<1 kb). The top 5% of the widest peaks were considered as broad peaks. The average intensity profiles were generated using deepTools (v2.5.4). Motif analysis was performed using the ‘findmotifsGenome.pl’ command inHomer2 package. Epigenetic factors were identified using the Epigenetic Factor Database(https://epifactors.autosome.org/) and then screened for those that exclusively regulate the expression of their target genes by modulating the deposition of H3K4me3.
Genomic distribution was analyzed using the “genomation” R package. GO pathway analysis was performed using the “clusterProfiler” R package. The sequencing information of CUT&Tag data used in this study is summarized in Supplementary Table S1.
Clustering analysis
Promoters were defined as ±2 kb regions flanking the annotated TSS. Reads in promoters were counted using the “coverage” command in bedtools (v2.26.0) and further normalized to RPKM. The k-means clustering of H3K4me3 and H3K27me3 enrichment at promoters was conducted using the “kmeans” function in R.
RNA-seq and data analysis
Total RNA was extracted from sorted CD4⁺YFP⁺ Treg cells using the RNeasy Plus Mini Kit (Qiagen, #74134), following the manufacturer’s protocol. RNA-Seq libraries were constructed and sequenced by Haplox (Nanchang, China), using an Illumina platform with paired-end reads of 150 bp. RNA-seq data was obtained from GSE82076. Raw reads were trimmed to 50 bp and mapped to the mouse genome (mm9) using TopHat (v2.1.1) with default parameters. Only uniquely mapped reads were kept for downstream analysis. The RNA abundance of each gene was quantified using Cufflinks (v2.2.1).
Single-Cell RNA-Sequencing
A total of 300,000 sort-purified CD4+YFP+ Treg cells from Foxp3cre and Foxp3cre Cxxc1fl/fl were resuspended in BD Pharmingen Stain Buffer (FBS) (554656; BD). Single cells were isolated using a chromium controller (BD platform, BD Bioscience) according to the manufacturer’s instructions, as previously described27. The single cells were labeled with sample tags using the BD Mouse Immune Single-Cell Multiplexing Kit (633793; BD). Following standard protocols, cDNA amplification and library construction were performed to generate scRNA-seq libraries.
Targeted scRNA-seq data processing
The raw FASTQ files were processed by BD Rhapsody using the Targeted analysis pipeline. After alignment and filtering, the distribution-based error correction(DBEC)-adjusted molecules were loaded into R studio(version 4.3.2). All subsequent analyses were performed using the package Seurat (version 4.4.0) with default parameters. Specifically, the scRNA-seq data counts were log-normalized. All targeted genes were scaled and then were used for Principal components analysis (PCA). The batch effects were removed by the HarmonyMatrix function in the Harmony package (version 1.1.0). The first 20 principal components were used to calculate nonlinear dimensionality reduction using RunUMAP. Differential gene expression (DEGs) between clusters was assessed using the FindAllMarkers function. The clusters were then annotated based on DEGs. Barplots were generated using ggplot2 (version 3.4.4). Heatmaps were generated using pheatmap (version 1.0.12).
Analysis of the single-cell TCR-seq repertoire
Raw V(D)J fastq reads were processed using BD Rhapsody Pipeline and then were analyzed using scRepertoire (version 1.12.0). The TCR clonotype was called using the nucleotide sequence of the CDR3 region for both TCR alpha and beta chains. For cells with multiple chains, the top two clonotypes with the highest expression were selected for downstream analysis. Clonal overlap between different cell types was calculated using the clonalOverlap function of scRepertoire. A clonotype was defined as expansion if it could be detected in at least two cells.
scRNA-seq trajectory analysis
UMAP embeddings obtained from the Seurat package were projected into the Slingshot(version 2.10.0) package to construct pseudotime Trajectories for Treg cells. Naïve subsets were set as the root state.
Whole genome bisulfite sequencing (WGBS) and data analysis
Sorted CD4+YFP+ Treg cells(3ⅹ106) were lysed in cell lysis buffer to release DNA. The bisulfite-treated DNA was used to prepare the sequencing library. DNA libraries were transferred to the Illumina Platform for sequencing using 150 bp paired-end reads. Raw reads were trimmed using TrimGalore (v0.4.4) with default parameters. Subsequently, the reads were mapped against the mm9 reference genome using Bismark v0.19.0 with parameters ‘--bowtie2’. PCR duplicates were removed and the methylation levels were calculated using ‘bismark_methylation_extractor’. We calculated the mean CpG methylation levels of various genome elements: promoter, 5′-UTR, exon, intron, 3′-UTR, genebody, intergenic, CGIs, and repeats using in-house scripts. The sequencing information of WGBS data used in this study is summarized in Supplementary Table S2.
Statistical analysis
The statistical significance analysis was performed using Prism 8.0 (GraphPad). Error bars are presented as mean ± SD. P values of < 0.05 were deemed statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.000). Statistical analyses were performed with unpaired Student’s t-test or two-way ANOVA and Holm–Sidak post hoc test.
Results
FOXP3 binds regulatory loci primed for activation and repression in Treg cells
FOXP3-mediated gene expression is well-recognized, with several studies highlighting its dual role as both a transcriptional activator and repressor 28–32. However, the regulation of Treg-specific gene transcription by FOXP3 through intermediary mechanisms, such as epigenetic factors, remains incompletely understood. To address this, we employed CUT&Tag to generate global H3K4me3 maps in Treg cells and compared these with an H3K27me3 ChIP-Seq dataset from Wei et al.33 at previously identified FOXP3-bound loci(Konopacki et al. 2019)34. This comparison aimed to identify FOXP3-dependent genes associated with H3K4me3 and H3K27me3 deposition. As expected, H3K4me3 was enriched at gene promoters (fig. S1, A and B). A Venn diagram revealed a significant overlap between FOXP3 binding sites and H3K4me3 peaks, with minimal overlap with H3K27me3 peaks (Fig. 1, A and B). The overlapping regions between FOXP3 binding sites and H3K4me3 or H3K27me3 peaks were predominantly located at promoters (Fig. 1, A and B). We then clustered the promoters into four groups based on the enrichment of H3K4me3 and H3K27me3. Clusters 1 and 3 showed strong enrichment of H3K4me3; cluster 2 was enriched with H3K27me3; and cluster 4 showed weak enrichment of both modifications. FOXP3 enrichment analysis revealed that FOXP3 preferentially binds to the promoters of clusters 1 and 3, which display high levels of H3K4me3 (Fig. 1C). Correspondingly, genes in these clusters exhibited high transcription levels, as demonstrated by our reanalysis of previously published RNA-seq data (Oh et al. 2017) shown in the right panels of Figure 1D35. In contrast, genes with H3K27me3 enrichment at their promoters were transcribed at low levels. Gene Ontology (GO) analysis of these four clusters revealed that cluster 1 was significantly enriched in genes involved in mRNA processing, covalent chromatin modification, and histone modification, while cluster 3 was enriched in genes related to DNA repair and mitochondrion organization (Fig. 1E). Cluster 2, on the other hand, was enriched in the pattern specification process, whereas cluster 4 showed no significant correlation with Treg cells. As expected, signature Treg cell genes such as Tnfrsf18, Nrp1, Stat5a, Lag3, Icos, and Pdcd1 were enriched in clusters 1 and 3, which exhibited strong H3K4me3 marks (fig. S1C). Conversely, FOXP3-bound sites in cluster 2 were specifically enriched for H3K27me3, including genes like Hic1, Trp73, and Rnf157, which are associated with inflammatory responses(fig. S1C). These observations support the hypothesis that FOXP3 primarily functions as an activator by promoting H3K4me3 deposition in Treg cells.
CXXC1 interacts with FOXP3 and binds H3K4me3-enriched sites in Treg cells
We conducted an enrichment analysis of known motifs at the overlapping peaks of FOXP3 ChIP-seq and H3K4me3 CUT&Tag in Treg cells to identify epigenetic factors that directly interact with FOXP3 to mediate chromatin remodeling and transcriptional reprogramming. Motif analysis of the overlapping peaks between FOXP3 binding sites and regions enriched in H3K4me3 revealed that, in addition to transcription factors, the most abundant motif associated with H3K4me3 was the epigenetic factor CXXC1(fig. S2A). To investigate this further, we performed CUT&Tag for endogenous CXXC1 in Treg cells to examine the genome-wide co-occupancy of CXXC1 and FOXP3. Over half of these CXXC1 binding sites were located at promoter regions (fig. S2B). Additionally, CXXC1 exhibited strong binding at transcription start site (TSS) and CpG islands (CGIs) (Fig. 2A and fig. S2C). As illustrated by the Venn diagram (Fig. 2B), more than half of the FOXP3-bound genes and H3K4me3-enriched genes were also bound by CXXC1. Similarly, more than half of CXXC1 peaks were overlapped with FOXP3 peaks (fig. S2D). Furthermore, the CXXC1-specific and FOXP3-specific binding sites also demonstrated modest binding of FOXP3 and CXXC1, respectively (Fig. 2C). These findings indicate that FOXP3 and CXXC1 share a substantial number of target genes in Treg cells. To confirm this interaction, we further validated the reciprocal immunoprecipitation of both endogenous and exogenous CXXC1 and FOXP3 (Fig. 2D and fig. S2E). An immunofluorescence assay revealed predominant colocalization of CXXC1 with FOXP3 in the nucleus (Fig. 2E). Overall, these results suggest that CXXC1 primarily functions as a coactivator of FOXP3-driven transcription in Treg cells.
Complete ablation of Cxxc1 in Treg cells leads to a fatal autoimmune disease
To investigate the role of CXXC1 in Treg-cell homeostasis and function, we generated Foxp3CreCxxc1fl/fl mice (conditional knockout [cKO] mice) by crossing Cxxc1fl/fl with Foxp3YFP-Cre 36 mice, thereby specifically deleting Cxxc1 in Treg cells. The effective depletion of Cxxc1 in Treg cells was confirmed through quantitative PCR (qPCR) and Western blotting (fig. S3A). Notably, cKO mice appeared normal at birth but later exhibited spontaneous mortality starting around three weeks of age (Fig. 3A). Deletion of Cxxc1 in Treg cells led to the development of severe inflammatory disease, characterized by reduced body size, stooped posture, crusting of the eyelids, ears, and tail, and skin ulceration, particularly on the head and upper back (Fig. 3, B and C). Additionally, cKO mice developed extensive splenomegaly and lymphadenopathy (Fig. 3D). Histopathological analysis revealed massive lymphocyte and myeloid cell infiltration in the skin, lungs, liver sinusoids, and colon mucosa (Fig. 3E). Moreover, cKO mice displayed a significant increase in CD8+ T cell percentages (fig. S3B), along with a marked rise in cells exhibiting an effector/memory phenotype (CD44hi CD62Llo) (Fig. 3F). Furthermore, T cells from cKO mice produced elevated levels of IFN-γ, IL-17, and IL-4 in CD4+ T cells, as well as increased IFN-γ production in CD8+ T cells (Fig. 3G and fig. S3, C and D). These phenotypes closely resembled those observed in Foxp3-deficient mice 37 or mice with depleted Treg cells38, suggesting a deficiency in immune suppression.
CXXC1 is necessary for the maintenance of Treg-cell suppressive activity
Despite the development of severe autoimmune disease, we observed an increase in both the absolute number and percentage of FOXP3+ Treg cells in the lymph nodes (fig. S4A). The expression level of the FOXP3 protein was only slightly altered in Cxxc1-deficient Treg cells (fig. S4B). In an in vitro suppression assay, Treg cells from Foxp3CreCxxc1fl/fl and WT mice exhibited similar suppressive effects on naïve T (Tn) cell proliferation (fig. S4C). The expression of the hallmark Treg-cell marker CTLA-4 showed a modest increase in Cxxc1-deficient Treg cells compared to WT Treg cells, while the expression of GITR remained unchanged (fig. S4D). To further assess the suppressive capacity of Cxxc1-deficient Treg cells in vivo, we employed the experimental autoimmune encephalomyelitis (EAE) model. Naïve CD4+ T cells from 2D2 mice were co-transferred with Treg cells from either Foxp3Cre or Foxp3Cre Cxxc1fl/fl mice into Rag1-/- recipients, and EAE was induced in these recipient mice. Mice that received only naïve CD4+ T cells from 2D2 mice developed more severe EAE symptoms (Fig. 4A). The addition of WT Treg cells from Foxp3Cre mice slightly mitigated EAE progression and reduced Th17 cells in the spinal cord (Fig. 4, A to D). In contrast, Foxp3Cre Cxxc1fl/fl Treg cells failed to suppress EAE (Fig. 4, A to D), and the cKO mice showed a significant reduction in Treg cells frequency in CNS tissues(Fig. 4E). Finally, we examined the role of CXXC1 in Treg cell-mediated suppression using T cell transfer-induced colitis, in which naïve T cells were transferred to Rag1-/- recipients either alone or together with WT or Foxp3Cre Cxxc1fl/fl Treg cells. The transfer of naïve T cells led to weight loss and intestinal pathology in recipient mice(Fig. 4, F and G). Mice receiving WT Treg cells continued to gain weight (Fig. 4F), whereas those that received Treg cells from cKO mice were unable to prevent colitis and exhibited a reduced percentage of Treg cells (Fig. 4, F to H). These findings underscore the critical role of CXXC1 in maintaining Treg cell function in vivo.
Treg cell lineage stability and proliferation depend upon CXXC1
Treg cells harbor a diverse T-cell receptor (TCR) repertoire, which likely play a critical role in their immune suppression function 39–41. To explore the role of CXXC1 in Treg-mediated suppression, we performed single-cell RNA sequencing (scRNA-seq) combined with TCR sequencing (TCR-seq) on CD4+YFP+ Treg cells isolated from mouse lymph nodes. After quality control and removal of doublets, 18,577 cells were retained for further analysis. Through unsupervised clustering and uniform manifold approximation and projection (UMAP) analysis, we identified eight distinct Treg cell clusters based on the expression of well-characterized markers, with a particular focus on two clusters of activated Treg cells that exhibited higher expression of markers and gene sets relative to naïve Treg cells (Fig. 5A and fig. S5, A to C). A comparison between Cxxc1-deficient and WT Treg cells within each cluster revealed a reduction in Cxxc1-deficient cells in the naïve subsets, while an increase was observed in the Gzmb+ and H2-Eb1+ subsets (Fig. 5B). To further elucidate the transition of Treg cells along a dynamic biological timeline, we constructed pseudo-time trajectories using Slingshot 42. The pseudo-time gradient depicted a progression from quiescent to activated Treg cells, ultimately encompassing the Gzmb+ and H2-Eb1+ subsets (Fig. 5C). Given the antigen-specific suppression capabilities of Treg cells 43,44, we examined their clonal expansion. The analysis revealed that expanded WT TCR clonotypes (n ≥ 2) were predominantly distributed among the Nt5e+ subsets, while Cxxc1-deficient Treg cells showed expanded clonotypes primarily within the Gzmb+ and H2-Eb1+ subsets(Fig. 5D and Fig. S5D). TCR sharing analysis indicated significant clonotype sharing among various clusters of WT Treg cells, suggesting a degree of homogeneity. However, the reduced TCR sharing in Cxxc1-deficient cells implies that decreased TCR diversity may impair the suppressive activity of Treg cells (Fig. 5E) 41. Furthermore, the Cxxc1-deficient group exhibited lower expression of several Treg-specific genes associated with suppressive functions, such asNt5e, Il10, Pdcd1, Klrg1, as well as genes that inhibit effector T cells differentiation, including Sell and Tcf7 (Fig. 5F and fig. S5E). Conversely, Cxxc1-deficient Treg cells demonstrated elevated expression of Gzmb, Il2ra, and Cd69 compared to WT, reflecting a profile indicative of increased activation (Fig. 5, F and G). Additionally, we observed increased expression of genes linked to Th1-type inflammation, such as Ifng, Tbx21, and Hif1a, in Cxxc1-deficient Treg cells, likely due to extreme inflammatory conditions (Fig. 5, F to H). The proportion of FOXP3+Ki67+ Treg cells was lower in cKO mice compared to WT mice (Fig. 5I). These findings underscore the crucial role of CXXC1 in maintaining Treg cell stability.
Cxxc1 deletion leads to a reduction in essential Treg cell markers
To confirm that the deficiency in Treg cell function in Treg-specific Cxxc1-deficient animals is due to intrinsic defects caused by Cxxc1 deficiency, rather than severe autoimmune inflammation in Foxp3Cre Cxxc1fl/fl mice, we examined Cxxc1-sufficient and Cxxc1-deficient Treg cell subsets in heterozygous Foxp3Cre/+ Cxxc1fl/fl (designated as “het-KO”) and littermate Foxp3Cre/+ Cxxc1fl/+ (designated as “het-WT”) female mice (Fig. 6A). Notably, het-KO female mice did not exhibit overt signs of autoimmunity, as random X-chromosome inactivation led to the coexistence of both Cxxc1-cKO and Cxxc1-WT Treg cells. However, both the frequency and absolute numbers of FOXP3+YFP+ Treg cells within the total Treg population were reduced in het-KO mice compared to their counterparts in het-WT littermates, indicating that Cxxc1 deficiency imposes a competitive disadvantage on Treg cells (Fig. 6B). Additionally, Cxxc1-deficient YFP+ Treg cells failed to upregulate the proliferation marker Ki-67 (Fig. 6C). Moreover, YFP+ Treg cells in het-KO female mice showed reduced expression of key genes essential for suppressive function, including Icos, Il2ra, Ctla4, and Tnfrsf18, compared to YFP-Treg cells from the same mice (Fig. 6D). To investigate the molecular program affected by the deletion of Cxxc1 in Treg cells, we performed RNA-sequencing (RNA-seq) analysis on CD4+YFP+ Treg cells isolated from het-WT and het-KO mice. We then conducted a differential gene expression (DGE) analysis based on the RNA-seq data. Among all expressed genes, 865 were upregulated (“Up” genes) and 761 were downregulated (“Down” genes) by ≥1.5-fold in CD4+YFP+ Treg cells from het-KO mice compared to het-WT mice, with a false discovery rate (FDR)–adjusted P-value cutoff of 0.05 (fig. S6A). Gene Ontology (GO) enrichment analysis revealed that the downregulated genes in Cxxc1-deficient Treg cells were predominantly enriched in pathways related to the negative regulation of immune system process and regulation of cell−cell adhesion (fig. S6B). The Cxxc1-deficient Treg cells also showed reduced expression of several genes associated with Treg cell suppressive function, including Il10, Tigit, Lag3, Icos, Nt5e(encoding CD73) and Itgae (encoding CD103) (fig. S6C). Thus, although YFP− WT Treg cells can prevent autoimmunity in het-KO mice, the absence of Cxxc1 in YFP+ Treg cells leads to the dysregulation of key Treg cell markers under steady-state conditions.
The FOXP3-CXXC1 complex regulates the expression of key factors in Treg cells that are associated with the breadth of H3K4me3
CXXC1 binds to unmethylated CpG DNA via its N-terminal CXXC finger domain, facilitating its interaction with DNA methyltransferase 1 (DNMT1). This binding stabilizes the DNMT1 protein, thereby regulating DNA methylation 45,46. To investigate whether CXXC1 depletion affects DNA methylation in Treg cells, we performed whole genome bisulfite sequencing (WGBS) on Treg cells isolated from both WT and cKO mice. On average, Cxxc1-deficient Treg cells exhibited no significant changes in DNA methylation at gene loci or across genome-wide CpG sites, irrespective of chromosomal region (fig. S7, A to C). Furthermore, Cxxc1 knockout Treg cells did not show a significant increase in DNA methylation at key Treg signature gene loci (fig. S7D). Given the pivotal role of MLL4-mediated H3K4me1 in establishing the enhancer landscape and facilitating long-range chromatin interactions during Treg cell development15, we performed CUT&Tag to assess changes in H3K4me1 levels in Cxxc1-deficient Treg cells. This analysis revealed that H3K4me1 levels were similar in both WT and Cxxc1-deficient Treg cells (fig. S7, E to G).
While H3K4me3 modifications typically form sharp 1-to 2-kb peaks around promoters, some genes exhibit broader H3K4me3 regions, referred to as broad H3K4me3 domains (H3K4me3-BDs), which can extend to cover part or all of the gene’s coding sequences (up to 20 kb) 47,48. Broad H3K4me3 domains are preferentially associated with genes essential for the identity or function of specific cell types 47,49 and have been implicated in enhancing transcriptional elongation and increasing enhancer activity 49. To further explore the connection between broad H3K4me3 domains and the expression of immune-regulatory genes, we analyzed genes containing broad H3K4me3 regions. We classified the H3K4me3 domains surrounding transcription start sites (TSSs) into three categories: broad (more than 5 kb), medium (between 1 and 5 kb), and narrow (less than 1 kb) (Fig. 7A). Notably, Cxxc1-deficient Treg cells exhibited weaker H3K4me3 signals compared to WT cells within the broad H3K4me3 domains where CXXC1 binding is prominent (Fig. 7, A and B). Using the criteria established by Benayoun et al.47, which defines the top 5% of the widest H3K4me3 domains as BDs, we observed similar enrichment results (fig. S7, H and I). We then compared three groups of genes: BD-associated genes with reduced H3K4me3 levels following Cxxc1 deletion, genes with direct CXXC1 binding, and genes with direct FOXP3 binding. The Venn diagram revealed that the majority of genes (283 out of 294, 96%) with CXXC1 binding and reduced H3K4me3 levels overlap with FOXP3-bound genes, suggesting that CXXC1 presence enhances FOXP3 binding within the broad H3K4me3 domain (Fig. 7C). Furthermore, GO term analysis indicated that BD-associated genes are enriched in biological processes related to the negative regulation of immune system processes (Fig. 7D). Genome browser views displayed the enrichments of FOXP3, CXXC1, and H3K4me3 at key signature genes in Treg cells, such as Ctla4, Il2ra, Icos, and Tnfrsf18, with lower H3K4me3 densities observed at these loci in Cxxc1-deficient Treg cells (Fig. 7E). Similar patterns were observed at core genes involved in Treg homeostasis and suppressive function (e.g., Lag3, Nt5e, Ikzf4, and Cd28) (fig. S7, J)50,51. These findings suggest that CXXC1 and FOXP3 collaboratively promote sustained Treg cell homeostasis and function by preserving the H3K4me3 modification at key Treg cell genes.
Discussion
Treg cells specifically express the transcription factor FOXP3, which is essential for maintaining Treg lineage stability and suppressive function52. However, FOXP3 alone is insufficient to fully regulate the transcriptional signature and functionality of Treg cells; its interaction with protein partners is crucial for this regulation. In this study, we identify CXXC1 as a key transcription factor that, by acting as a critical cofactor of FOXP3, orchestrates the Treg transcriptional program, which in turn plays a vital role in Treg cell homeostasis and function.
FOXP3 is known to promote histone H3 acetylation at the promoters and enhancers of its target genes, such as Il2ra, Ctla4, and Tnfrsf18, following Treg cell activation, thereby functioning as a transcriptional activator53,54. Conversely, FOXP3 can act as a repressor by silencing target genes like Il2 and Ifng through the induction of histone H3 deacetylation, mediated by the recruitment of histone deacetylases and transcriptional co-repressors10,55. Additionally, FOXP3 exerts this repression by recruiting the Ezh2-containing polycomb repressive complex to target genes during activation, as FOXP3-repressed genes are associated with H3K27me3 deposition and reduced chromatin accessibility11,56. Recent studies have begun to explore the biological significance of H3K4me3 breadth, revealing a positive correlation between H3K4me3 breadth and gene expression 49,57,58. These studies also suggest that H3K4me3 breadth contributes to defining specific cell identities during development and disease, including systemic autoimmune diseases like systemic lupus erythematosus and various cancers 47,59–62. Our study demonstrates that the loss of Cxxc1 leads to reduced H3K4me3 levels, predominantly in genes with broader peaks, such as Il2ra, Tnfrsf18, Ctla4 and Icos, directly impairing their suppressive function in Treg cells. Although our research primarily focused on the role of CXXC1 in Treg cells, it is plausible that similar mechanisms may be operative in other cell types.
The Treg-specific deletion of Cxxc1 leads to a rapid and fatal autoimmune disorder, characterized by systemic inflammation and tissue damage, underscoring the essential role of CXXC1 in maintaining immune self-tolerance within Treg cells. Interestingly, despite this severe phenotype, our findings show that H3K4me1 levels were comparable between WT and Cxxc1-deficient Treg cells. In contrast, MLL4 is critical for Treg cell development in the thymus, primarily by regulating H3K4me1, though it is not required for peripheral Treg cell function15. Numerous studies investigating the effectors of Treg cell-mediated suppression have identified a wide range of molecules and mechanisms. These include the upregulation of the inhibitory co-stimulatory receptor CTLA-4, which initiates inhibitory signaling; the sequestration of the T-cell growth factor IL-2 via CD25; the secretion of inhibitory cytokines, such as interleukin (IL)-10, IL-35 and TGF-β and the activity of ectoenzymes CD39 and CD73 on the Treg cells surface, which convert extracellular ATP, a potent pro-inflammatory mediator, into its anti-inflammatory counterpart, adenosine41,63. Mechanistically, we demonstrated that CXXC1 interacts with FOXP3 to regulate Treg cell function by trimethylating H3K4 at broad H3K4me3 domains of multiple genes involved in suppressive functions such as Il2ra, Nt5e, and Ctla4. Additionally, MLL1, another KMT, controls Treg cell activation and function by specifically regulating H3K4 trimethylation at genes encoding key Treg-related molecules such as Tigit, Klrg1, Tbx21, Cxcr3, and serves as a crucial epigenetic regulator in establishing a stable Th1-Treg lineage14.
Our results offer provide novel insights into the suppressive functions, heterogeneity, and regulatory mechanisms of Treg cells. Maintaining Treg cell homeostasis and function remains a significant challenge in harnessing Treg cells for the treatment of autoimmune diseases and the prevention of graft rejection.
Acknowledgements
We thank B. Li (Shanghai Jiao Tong University School of Medicine, Shanghai, China) for providing Foxp3YFP-Cre mice, and X. L. Liu (Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences) for his generous gifts of cell lines. We thank Y. Huang, Y. Li, J. Wan, C.Guo, and Z. Lin from the Core Facilities, Zhejiang University School of Medicine for their technical support and S. Hong, Y. Ding, H. Jin, Q. Wang, and X. Zhang from Animal Facilities, Zhejiang University, for feeding the mice.
Additional information
Competing interests
The authors declare that they have no competing interests.
Footnotes
This work was supported by grants from the National Natural Science Foundation of China (32341002, 32030035, 32321002, 32100693 and 32270839), the National Key R & D Program of China (2023YFA1800202), the Zhejiang Provincial Natural Science Foundation of China (LZ21C080001, LZ23C070003), Science and Technology Innovation 2030-Major Project (2021ZD0200405), Key project of the Experimental Technology Program of Zhejiang University (SZD202203).
Author contributions
L.W. and L.S.: supervision, conceptualization, project administration, and writing—review and editing. L.W., L.S. and X.M: funding acquisition. X.M: investigation, methodology, project administration, and writing— original draft. Y.Z.and K.L.: bioinformatics analysis. Y. W., X.L., J.C., C.L., Y.Z., S.W., S.T., Q.X., L.D., X.S., and X.G.: investigation and methodology.
References
- 1Regulatory T cells and immune toleranceCell 133:775–787https://doi.org/10.1016/j.cell.2008.05.009
- 2Annual Review of ImmunologyAnnual Review of Immunology 30:531–564
- 3Regulatory T cells as immunotherapyFrontiers in Immunology 5https://doi.org/10.3389/fimmu.2014.00046
- 4Transcriptional and epigenetic basis of Treg cell development and function: its genetic anomalies or variations in autoimmune diseasesCell Research 30:465–474https://doi.org/10.1038/s41422-020-0324-7
- 5Th17 and Regulatory T Cells in Mediating and Restraining InflammationCell 140:845–858https://doi.org/10.1016/j.cell.2010.02.021
- 6The plasticity and stability of regulatory T cellsNature Reviews Immunology 13:461–467https://doi.org/10.1038/nri3464
- 7Foxp3 programs the development and function of CD4+CD25+ regulatory T cellsNat Immunol 4:330–336https://doi.org/10.1038/ni904
- 8A Mutation in the Transcription Factor Foxp3 Drives T Helper 2 Effector Function in Regulatory T CellsImmunity 50https://doi.org/10.1016/j.immuni.2018.12.016
- 9FOXP3 Orchestrates H4K16 Acetylation and H3K4 Trimethylation for Activation of Multiple Genes by Recruiting MOF and Causing Displacement of PLU-1Molecular Cell 44:770–784https://doi.org/10.1016/j.molcel.2011.10.012
- 10FOXP3(+) regulatory T cell development and function require histone/protein deacetylase 3Journal of Clinical Investigation 125:1111–1123https://doi.org/10.1172/jci77088
- 11Inflammation-induced repression of chromatin bound by the transcription factor Foxp3 in regulatory T cellsNature Immunology 15:580–587https://doi.org/10.1038/ni.2868
- 12Methylation of lysine 4 on histone H3: Intricacy of writing and reading a single epigenetic markMolecular Cell 25:15–30https://doi.org/10.1016/j.molcel.2006.12.014
- 13Structural analysis of the core COMPASS family of histone H3K4 methylases from yeast to humanProceedings of the National Academy of Sciences of the United States of America 108:20526–20531https://doi.org/10.1073/pnas.1109360108
- 14The histone lysine methyltransferase MLL1 regulates the activation and functional specialization of regulatory T cellsCell Reports 43https://doi.org/10.1016/j.celrep.2024.114222
- 15MLL4 prepares the enhancer landscape for Foxp3 induction via chromatin loopingNature Immunology 18https://doi.org/10.1038/ni.3812
- 16Analyzing the impact of CFP1 mutational landscape on epigenetic signalingFaseb Journal 35https://doi.org/10.1096/fj.202100427R
- 17CXXC finger protein 1 restricts the Setd1A histone H3K4 methyltransferase complex to euchromatinFebs Journal 277:210–223https://doi.org/10.1111/j.1742-4658.2009.07475.x
- 18CXXC finger protein 1 is critical for T-cell intrathymic development through regulating H3K4 trimethylationNat Commun 7https://doi.org/10.1038/ncomms11687
- 19Cxxc Finger Protein 1 Positively Regulates GM-CSF-Derived Macrophage Phagocytosis Through Csf2rα-Mediated SignalingFrontiers in Immunology 9https://doi.org/10.3389/fimmu.2018.01885
- 20Epigenetic initiation of the T(H)17 differentiation program is promoted by Cxxc finger protein 1Science Advances 5https://doi.org/10.1126/sciadv.aax1608
- 21Cxxc finger protein 1 maintains homeostasis and function of intestinal group 3 innate lymphoid cells with agingNature Aging 3https://doi.org/10.1038/s43587-023-00453-7
- 22CXXC finger protein 1 is critical for T-cell intrathymic development through regulating H3K4 trimethylationNature Communications 7https://doi.org/10.1038/ncomms11687
- 23T cell proliferation requires ribosomal maturation in nucleolar condensates dependent on DCAF13Journal of Cell Biology 222https://doi.org/10.1083/jcb.202201096
- 24AIM2 in regulatory T cells restrains autoimmune diseasesNature 591https://doi.org/10.1038/s41586-021-03231-w
- 25mTORC1 couples immune signals and metabolic programming to establish T(reg)-cell functionNature 499:485–490https://doi.org/10.1038/nature12297
- 26The phosphatase PAC1 acts as a T cell suppressor and attenuates host antitumor immunityNature Immunology 21https://doi.org/10.1038/s41590-019-0577-9
- 27Single-Cell Profiling of Tumor Immune Microenvironment Reveals Immune Irresponsiveness in Gastric Signet-Ring Cell CarcinomaGastroenterology 165:88–103https://doi.org/10.1053/j.gastro.2023.03.008
- 28Foxp3 occupancy and regulation of key target genes during T-cell stimulationNature 445:931–935https://doi.org/10.1038/nature05478
- 29Genome-wide analysis of Foxp3 target genes in developing and mature regulatory T cellsNature 445:936–940https://doi.org/10.1038/nature05563
- 30Foxp3 controls regulatory T-cell function by interacting with AML1/Runx1Nature 446:685–689https://doi.org/10.1038/nature05673
- 31A multiply redundant genetic switch ‘locks in’ the transcriptional signature of regulatory T cellsNature Immunology 13:972–980https://doi.org/10.1038/ni.2420
- 32T Cell Receptor Stimulation-Induced Epigenetic Changes and Foxp3 Expression Are Independent and Complementary Events Required for Treg Cell DevelopmentImmunity 37:785–799https://doi.org/10.1016/j.immuni.2012.09.010
- 33Global Mapping of H3K4me3 and H3K27me3 Reveals Specificity and Plasticity in Lineage Fate Determination of Differentiating CD4(+) T CellsImmunity 30:155–167https://doi.org/10.1016/j.immuni.2008.12.009
- 34Transcription factor Foxp1 regulates Foxp3 chromatin binding and coordinates regulatory T cell functionNature Immunology 20https://doi.org/10.1038/s41590-018-0291-z
- 35An NF-kappaB Transcription-Factor-Dependent Lineage-Specific Transcriptional Program Promotes Regulatory T Cell Identity and FunctionImmunity 47:450–465https://doi.org/10.1016/j.immuni.2017.08.010
- 36Regulatory T cell-derived interleukin-10 limits inflammation at environmental interfacesImmunity 28:546–558https://doi.org/10.1016/j.immuni.2008.02.017
- 37Foxp3 programs the development and function of CD4(+)CD25(+) regulatory T cellsNature Immunology 4:330–336https://doi.org/10.1038/ni904
- 38Regulatory T cells prevent catastrophic autoimmunity throughout the lifespan of miceNature Immunology 8:191–197https://doi.org/10.1038/ni1428
- 39Treg Heterogeneity, Function, and HomeostasisFrontiers in Immunology 10https://doi.org/10.3389/fimmu.2019.03100
- 40Tissue-specific abundance of interferon-gamma drives regulatory T cells to restrain DC1-mediated priming of cytotoxic T cells against lung cancerImmunity 56https://doi.org/10.1016/j.immuni.2023.01.010
- 41Principles of regulatory T cell functionImmunity 56:240–255https://doi.org/10.1016/j.immuni.2023.01.004
- 42Slingshot: cell lineage and pseudotime inference for single-cell transcriptomicsBmc Genomics 19https://doi.org/10.1186/s12864-018-4772-0
- 43CD25+ CD4+ T cells, expanded with dendritic cells presenting a single autoantigenic peptide, suppress autoimmune diabetesJournal of Experimental Medicine 199:1467–1477https://doi.org/10.1084/jem.20040180
- 44Specificity requirements for selection and effector functions of CD25+4+ regulatory T cells in anti-myelin basic protein T cell receptor transgenic miceProceedings of the National Academy of Sciences of the United States of America 99:8213–8218https://doi.org/10.1073/pnas.122224799
- 45CFP1 Interacts with DNMT1 Independently of Association with the Setd1 Histone H3K4 Methyltransferase ComplexesDNA and Cell Biology 27:533–543https://doi.org/10.1089/dna.2007.0714
- 46DNA Methyltransferase Protein Synthesis Is Reduced in CXXC Finger Protein 1-Deficient Embryonic Stem CellsDNA and Cell Biology 28:223–231https://doi.org/10.1089/dna.2009.0854
- 47H3K4me3 Breadth Is Linked to Cell Identity and Transcriptional ConsistencyCell 158:673–688https://doi.org/10.1016/j.cell.2014.06.027
- 48Transcription-Dependent Generation of a Specialized Chromatin Structure at the TCRβ LocusJournal of Immunology 194:3432–3443https://doi.org/10.4049/jimmunol.1400789
- 49Broad H3K4me3 is associated with increased transcription elongation and enhancer activity at tumor-suppressor genesNature Genetics 47https://doi.org/10.1038/ng.3385
- 50An obligate cell-intrinsic function for CD28 in TregsJournal of Clinical Investigation 123:580–593https://doi.org/10.1172/jci65013
- 51Selective deletion of Eos (Ikzf4) in T-regulatory cells leads to loss of suppressive function and development of systemic autoimmunityJournal of Autoimmunity 105https://doi.org/10.1016/j.jaut.2019.06.011
- 52Regulatory T-cell functions are subverted and converted owing to attenuated Foxp3 expressionNature 445:766–770https://doi.org/10.1038/nature05479
- 53Transcriptional regulation by Foxp3 is associated with direct promoter occupancy and modulation of histone acetylationJournal of Biological Chemistry 281:36828–36834https://doi.org/10.1074/jbc.M608848200
- 54Differential roles of epigenetic changes and Foxp3 expression in regulatory T cell-specific transcriptional regulationProceedings of the National Academy of Sciences of the United States of America 111:5289–5294https://doi.org/10.1073/pnas.1312717110
- 55Eos Mediates Foxp3-Dependent Gene Silencing in CD4+ Regulatory T CellsScience 325:1142–1146https://doi.org/10.1126/science.1176077
- 56The Chromatin-Modifying Enzyme Ezh2 Is Critical for the Maintenance of Regulatory T Cell Identity after ActivationImmunity 42:227–238https://doi.org/10.1016/j.immuni.2015.01.007
- 57Distinct features of H3K4me3 and H3K27me3 chromatin domains in pre-implantation embryosNature 537https://doi.org/10.1038/nature19362
- 58Coordinated regulation of cellular identity-associated H3K4me3 breadth by the COMPASS familyScience Advances 6https://doi.org/10.1126/sciadv.aaz4764
- 59Broad histone H3K4me3 domains in mouse oocytes modulate maternal-to-zygotic transitionNature 537https://doi.org/10.1038/nature19360
- 60Allelic reprogramming of the histone modification H3K4me3 in early mammalian developmentNature 537https://doi.org/10.1038/nature19361
- 61H3K4 tri-methylation breadth at transcription start sites impacts the transcriptome of systemic lupus erythematosusClinical Epigenetics 8https://doi.org/10.1186/s13148-016-0179-4
- 62The H3K4-Methyl Epigenome Regulates Leukemia Stem Cell Oncogenic PotentialCancer Cell 28:198–209https://doi.org/10.1016/j.ccell.2015.06.003
- 63Adenosine generation catalyzed by CD39 and CD73 expressed on regulatory T cells mediates immune suppressionJournal of Experimental Medicine 204:1257–1265https://doi.org/10.1084/jem.20062512
Article and author information
Author information
Version history
- Sent for peer review:
- Preprint posted:
- Reviewed Preprint version 1:
Copyright
© 2024, Meng et al.
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
Metrics
- views
- 83
- download
- 1
- citations
- 0
Views, downloads and citations are aggregated across all versions of this paper published by eLife.