G9a-mediated H3K9me2 orchestrates intestinal epithelial regeneration through epigenetic silencing of cell cycle-related genes

Abstract

Histone modifications play an important role in intestinal homeostasis and regeneration. Here, we identify histone H3 lysine 9 di-methylation (H3K9me2) as an epigenetic regulator of intestinal epithelial repair through mass spectrometry-based screening of histone modifications. We then find that H3K9me2 and its methyltransferase G9a levels are reduced during acute injury and progressively increase during regeneration in both mouse models and human clinical samples. Genetic ablation of G9a in intestinal epithelial cells or pharmacological inhibition of its enzymatic activity substantially impairs intestinal regeneration and reduces survival following irradiation. Mechanistically, integrative genomic analyses reveal that G9a-mediated H3K9me2 suppresses chromatin accessibility and transcriptional activity of cell cycle arrest genes, including Rb1cc1, Rb1, Cdkn1a, and Pten, thereby promoting intestinal stem cell proliferation. Furthermore, we elucidate that IL-4-STAT6 signaling controls G9a expression during regeneration, i.e., IL-4 upregulation leads to STAT6 phosphorylation and subsequent transcriptional activation of G9a. These findings establish the IL-4-STAT6-G9a-H3K9me2 regulatory axis as a critical epigenetic mechanism controlling intestinal regeneration with therapeutic potential for gastrointestinal disorders.

Here they show that G9a-mediated H3K9me2 regulates intestinal homeostasis and injury repair through repression of cell cycle arrest genes in both mouse and human.

Introduction

The intestinal epithelium is a meticulously arranged system, composed of differentiated villi and actively proliferating crypts¹. Continuous renewal of intestinal tissue is supported by intestinal stem cells (ISCs) positioned at the bottom of the crypts². These ISCs, as master architects of intestinal regeneration, continually divide to generate both stem cells and a population of transit amplifying (TA) daughter cells^3,4. The TA cells serve as progenitors for a suite of terminally differentiated cell types⁵. Balanced proliferation state of crypt cells, including ISCs and TA cells, is essential not only for the physiological integrity of the intestinal system but also for intestinal repair triggered by epithelial injury, caused by either irradiation or inflammation^1,6. Hundreds of well-orchestrated genes guide crypt cell proliferation, thus ensuring the maintenance of tissue homeostasis and facilitating a timely response to injury. Although the genome is largely stable, the epigenome allows for dynamic control of gene expression and cellular identity⁷. Accordingly, varying levels of epigenetic regulation contribute to intestinal health through transcriptional activation or repression of specific sets of essential genes that control crypt cell proliferation and differentiation^7–9.

Histone modifications, as one of the core epigenetic mechanisms, dynamically regulate chromatin accessibility and transcriptional activity in response to intracellular or extracellular signals^10,11. For example, histone acetylation, primarily catalyzed by histone acetyltransferases, generally promotes chromatin opening and gene activation, while histone deacetylases remove these marks to facilitate gene silencing¹². Histone methylation presents a more complex regulatory landscape, where methylation of different lysine residues (e.g., H3K4, H3K9) and varying degrees of methylation (mono-, di-, tri-methylation) can lead to distinct outcomes, such as gene activation or repression¹³. The regulatory enzymes for histone methylation are particularly diverse, encompassing a large family of methyltransferases and demethylases with distinct substrate specificities^14,15. For example, SET1/MLL complexes add activating H3K4 methylation marks¹⁶, G9a complex primarily deposits repressive H3K9me2 marks^17,18, and SETDB1 mediates H3K9me3 deposition¹⁹. Histone demethylases such as LSD1 and JMJD2 family members provide additional regulatory control by removing specific methylation marks²⁰. This intricate enzymatic network enables precise temporal and spatial control of gene expression through dynamic histone methylation states, allowing cells to rapidly respond to environmental stimuli.

In the intestinal epithelium, several histone modifications have been implicated in maintaining epithelial homeostasis, stem cell function, and cellular differentiation. Histone acetyltransferases such as p300 and CBP regulate intestinal stem cell proliferation and differentiation by modulating Wnt signaling target genes²¹. SETDB1-mediated H3K9me3 is necessary for intestinal epithelial differentiation and the prevention of intestinal inflammation by silencing endogenous viral elements in the genome^22,23. Chronic stress has been observed to induce alterations in H3K9me1 at the promoters of tight junction genes in rat colon epithelial cells²⁴. While these studies have provided valuable insights into the roles of specific histone modifications on steady-state conditions or chronic disease models, our understanding of how histone modifications dynamically regulate gene expression programs during intestinal injury and repair remains limited.

In this study, we systematically screened for differential histone modifications associated with intestinal epithelial regeneration and identified H3K9me2 as exhibiting significant changes. We then evaluated the correlation between the levels of H3K9me2 and its modifying enzyme G9a and the extent of intestinal injury repair. Subsequently, we reduced the H3K9me2 levels by either knocking out the G9a gene in intestinal epithelial cells (IECs) and ISCs or inhibiting its enzymatic activity to analyze its roles in intestinal homeostasis and injury repair in mice. Next, we explored the underlying mechanism of H3K9me2 in this process. Finally, we investigated the upstream regulatory pathways controlling this modification. Our findings reveal a significant epigenetic modification that controls ISCs and TA cells during intestinal injury and the repair process, and highlight H3K9me2 as a potential therapeutic target for interventions in gastrointestinal epithelium-related diseases.

Results

H3K9me2 and G9a levels correlate with intestinal epithelium injury and regeneration

To investigate the relationship between histone modifications and intestinal injury regeneration, we employed an irradiation-induced injury model and performed histone mass spectrometry analysis on chromatin fractions isolated from intestinal crypt cells (Fig. 1a). Among detected 14 histone H3 tail and 5 histone H4 tail modifications, H3K9me2 was significantly upregulated at 96 h of post-irradiation, while H3K27me3 and H3K14ac were significantly downregulated (Fig. 1b, c). To validate these findings, we performed immunoblotting to examine the temporal dynamics of these three modifications along with other H3K9 methylation states (H3K9me1 and H3K9me3) throughout the injury-repair process. H3K9me2 exhibited a biphasic pattern, with initial reduction followed by subsequent upregulation during regeneration (Fig. 1d). G9a, the methyltransferase responsible for H3K9me2 modification, showed temporal changes that closely paralleled H3K9me2 levels (Fig. 1e). IHC analysis of H3K9me2 and G9a levels confirmed these observations (Supplementary Fig. 1a, b). To corroborate these findings, we examined a DSS-induced injury and regeneration model and observed similar dynamics (Supplementary Fig. 2a–d). These data suggested that H3K9me2 and its methyltransferase G9a are closely associated with intestinal injury and repair processes.

Fig. 1. H3K9me2 and G9a levels correlate with intestinal epithelium injury and regeneration.

a Schematic illustration of the irradiation-induced intestinal injury model and experimental workflow. b, c LC-MS quantification of histone modifications in intestinal crypt cells. Relative normalized intensity of 14 histone H3 tail modifications (b) and 5 histone H4 tail modifications (c) at baseline (0 h, blue, n = 3 mice) and 96 h post-irradiation (96 h, red, n = 3 mice were calculated. Immunoblotting analyses of H3K9me1, H3K9me2, H3K9me3, H3K27me3, and H3K14ac (d) and G9a (e) at the indicated time points during the injury-regeneration process. Total histone H3 or ACTB (β-actin) was used as a loading control. Numbers below each blot represent relative band intensities normalized to 0 h. Each experiment was repeated independently three times with similar results. f Representative IHC images of H3K9me2 and G9a staining in human rectal tissue samples from non-radiotherapy patients (normal region) and radiotherapy patients (injury region and repair region). Lower panels show magnified views of boxed areas. Scale bar, 100 μm. Quantification of H3K9me2 (g) and G9a (h) staining scores in normal region (non-radiotherapy patients, n = 15), injury region (radiotherapy patients, n = 20), and repair region (radiotherapy patients, n = 20). Each dot represents one patient. i Correlation analysis between H3K9me2 and G9a staining scores in human rectal tissue samples. Each dot represents one sample. Data are presented as mean ± SEM. Statistical analysis was performed using one-way ANOVA test in (b and c) and by the Kruskal–Wallis test in (g and h). Correlation analysis was performed using Pearson correlation coefficient (two-sided) in (i) with correlation strength (R²) and statistical significance (P-value) indicated. Source data are provided as a Source Data file. LC-MS liquid chromatography-mass spectrometry.

To further validate these findings in clinical settings, we collected large intestinal tissue samples from 20 patients with rectal cancer who had undergone radiotherapy and 15 patients who had not received radiotherapy. Within the same patients, the staining intensity for H3K9me2 and the corresponding G9a in the nuclei of IECs at the repair region was significantly higher than that in the injured region, and also exceeded that of IECs in the normal region of non-radiotherapy patients (Fig. 1f–h). Moreover, there was a positive correlation between the levels of H3K9me2 and G9a (Fig. 1i). We also analyzed these in human intestinal tissue samples from patients with CD at various stages, including moderate, severe, and remission phases. H3K9me2 level in IECs decreased during disease progression and negatively correlated with the Crohn’s Disease Activity Index (Supplementary Fig. 3a–c). Interestingly, H3K9me2 level was restored and increased during the remission phase (Supplementary Fig. 3b). A similar correlation was observed between G9a expression and CD severity, i.e., the more severe the disease, the lower the level of G9a in the nuclei of IECs (Supplementary Fig. 3a, d, e). A positive correlation was found between the levels of H3K9me2 and G9a within these tissues (Supplementary Fig. 3f). Immunoblotting confirmed the expression patterns of H3K9me2 and G9a, whereas no significant changes were observed for H3K9me1 and H3K9me3 (Supplementary Fig. 3g).

Collectively, these findings demonstrate a positive association between the nuclear levels of H3K9me2 and G9a and the degree of intestinal regeneration, with the concordance between mouse and human data suggesting an evolutionarily conserved epigenetic mechanism underlying intestinal repair.

H3K9me2 is required for intestinal regeneration

We next employed IEC-specific G9a knockout (G9a^ΔIEC, Supplementary Fig. 4a–d) and a G9a methyltransferase competitive inhibitor²⁵, UNC0642, to reduce H3K9me2 levels in the intestinal epithelium, and combined these approaches with the two aforementioned models to corroborate the role of H3K9me2 in intestinal regeneration. The results showed that G9a deletion in the intestinal epithelium or treatment with UNC0642 significantly reduced H3K9me2 level without affecting H3K9me1 and H3K9me3 expression, confirming that both methods effectively decrease H3K9me2 modification (Supplementary Fig. 4e, f).

In the 8 Gy irradiation-induced model, we found that G9a^ΔIEC mice exhibited significantly lower body weight and small intestine length than the controls (Fig. 2a–c). Histological staining of the small intestine revealed that the number of regenerating crypts, as well as the corresponding ISCs (OLFM4 positive) and proliferative cells (Ki67 positive), were significantly reduced in G9a^ΔIEC mice compared to controls (Fig. 2d–f), indicating that the IEC-specific knockout of G9a exacerbates post-irradiation symptoms. Similarly, treatment with UNC0642 significantly slowed down the regeneration process after irradiation, as evidenced by a more rapid decline in body weight, a greater reduction in small intestine length, and a lower number of regenerating crypts (Fig. 2g–l). When the irradiation was increased to 10 Gy, both G9a^ΔIEC mice and UNC0642-treated mice exhibited significantly reduced survival time compared to controls (Fig. 2m).

Fig. 2. Role of G9a-mediated H3K9me2 in irradiation-induced intestinal injury and regeneration processes.

a and g Relative body weight changes in mice following irradiation. n = 5 mice/group. b and h Representative images of small intestines from mice at 96 h post-irradiation. c and i Small intestine length. n = 5 mice/group. d and j Representative images of H&E staining (left column), and Ki67 and OLFM4 IHC staining (middle and right columns, respectively) of small intestine tissues. Scale bar: 100 μm. e and k Quantification of viable crypts per field in small intestine tissues; at least 20 fields counted per mouse. n = 5 mice/group. f and l Quantification of Ki67- and OLFM4-positive cells per crypt in small intestine tissues; at least 30 crypts counted per mouse. n = 5 mice/group. m Survival curve of mice following 10 Gy irradiation. n = 8 mice/group. Data are presented as mean ± SEM. Statistical analysis was performed using a two-tailed unpaired t-test in (c, e, f, i, k, and l) and a two-way ANOVA test in (a and g). Survival analysis was performed using Kaplan–Meier survival curve analysis with the log-rank test in (m). Source data are provided as a Source Data file.

In the DSS-induced model, G9a^ΔIEC mice also displayed increased susceptibility compared to G9a^fl/fl mice, marked by more significant weight loss, a higher disease activity index, and enhanced intestinal permeability (Supplementary Fig. 5a–d). Furthermore, analyses of small intestinal tissues collected on day 14 showed that G9a knockout resulted in a significant decrease in the number of intestinal crypts, stem cells, and proliferating cells (Supplementary Fig. 5e–g).

Taken together, these results indicate that G9a-mediated H3K9me2 plays a crucial role in the repair of intestinal injury.

H3K9me2 depletion causes villi-crypt structural abnormalities under homeostatic conditions

As components of the most vigorously self-renewing tissue, crypt ISCs and TA cells play key roles in maintaining intestinal homeostasis and regeneration²⁶. Therefore, we evaluated the influence of H3K9me2 on the morphology and cell composition of the villi-crypt structure after G9a knockout or treatment with the UNC0642 inhibitor under homeostatic conditions without any injury treatment. In G9a^ΔIEC mice, we observed significantly reduced crypt depth and villus height compared to G9a^fl/fl mice (Supplementary Fig. 6a, b). IHC staining revealed a reduction of OLFM4-positive cells (indicating ISCs) and Ki67-positive cells (mainly representing TA cells) in G9a^ΔIEC mice, with corresponding downregulated expression of both genes (Fig. 3a–c). ISCs differentiate into progenitors and eventually form various cell types in the villi, including Paneth, goblet, and enteroendocrine cells (Supplementary Fig. 6c)²⁷. Therefore, we compared these cell types in both mice and found significant reductions in their numbers and associated marker gene expressions in the G9a knockout mice (Supplementary Fig. 6d–f). Consistent with the results from G9a knockout mice, UNC0642 treatment also altered villi-crypt structure, including reduced crypt depth and villus height, and decreased numbers and marker gene expressions of ISCs and TA cells and their differentiated downstream cells (Supplementary Fig. 7a–h). Collectively, these results suggest G9a-mediated H3K9me2 in crypt cells governs ISC and TA cell numbers, thus maintaining crypt and villus structure at steady state.

Fig. 3. Control of crypt cell proliferation state by G9a-mediated H3K9me2 under homeostatic conditions.

a Representative IHC staining images for OLFM4 and Ki67 in small intestine tissues. Red boxes indicate magnified regions. Scale bar: 100 μm. b Quantification of OLFM4-positive and Ki67-positive cells per crypt; at least 50 crypts counted per mouse. n = 4 mice/group. c Expression levels of Olfm4 and Ki67 in crypt tissues. n = 4 mice/group. d Representative immunofluorescence images for EdU and OLFM4 co-staining. Nuclei are stained with Hoechst (blue). Scale bar: 25 μm. e Quantification of OLFM4 and EdU double-positive cells per crypt and proportion of double-positive cells relative to OLFM4 single-positive cells; at least 60 crypts were counted per mouse. n = 3 mice/group. f Levels of cell cycle-related proteins in crypt epithelial cells, including Cyclin A2, Cyclin E2, Cyclin D1, and Cyclin H, with ACTB serving as loading control. Numbers below each blot represent relative band intensities normalized to mouse #1 of the G9a^fl/fl group. g Representative images of intestinal organoid growth on days 3 and 5. Scale bar: 100 μm. h Quantification of intestinal organoid area on days 3 and 5; at least 60 organoids assessed per mouse. n = 4 mice/group. i Representative images of individual organoid growth tracking. Scale bar: 50 μm. j Quantification of organoid bud numbers on days 3, 4, and 5; at least 50 organoids assessed per mouse. n = 4 mice/group. All samples and organoids in this figure were obtained from G9a^fl/fl and G9a^ΔIEC mice at homeostasis without injury treatment. Data are presented as mean ± SEM. Statistical analysis was performed using a two-tailed unpaired t-test in (b, c, e, h, and j). Source data are provided as a Source Data file.

H3K9me2 controls the proliferation state of ISCs under both homeostatic and injury conditions

The maintenance of crypt cell numbers relies on their proliferation capacity²⁸. Intriguingly, we also observed a predominant localization of G9a in the nuclei of ISCs and TA cells in both human and mouse small intestine tissues (Supplementary Fig. 8a–f), indicating that G9a and its catalyzed H3K9me2 may play a role in these cells. To investigate whether G9a-mediated H3K9me2 regulates ISCs and TA cells proliferation, we employed the EdU pulse-chase assay to trace the proliferative cells under homeostatic conditions. The results showed that the labeled cells were primarily located within crypts, and their numbers were significantly reduced in G9a^ΔIEC mice after 1.5 h of EdU incorporation (Supplementary Fig. 9a, b). At later time points, these EdU-labeled cells were positioned lower in the crypts of G9a^ΔIEC mice compared to G9a^fl/fl mice (Supplementary Fig. 9a, b), suggesting a slower proliferation rate in the presence of diminished H3K9me2 level. Immunofluorescence analysis further confirmed the notable decrease in the ISCs' proliferation after G9a deletion (Fig. 3d, e). Correspondingly, the expression levels of cell cycle-related proteins in crypt cells of G9a^ΔIEC mice, including Cyclin A2, Cyclin E2, Cyclin D1, and Cyclin H, were significantly reduced compared to G9a^fl/fl mice (Fig. 3f). When we treated these mice with UNC0642, we observed the same phenomenon as in G9a^ΔIEC mice, further confirming the role of H3K9me2 in crypt proliferation (Supplementary Fig. 10a–e).

To further evaluate the impact of H3K9me2 on ISCs' function, we generated mice with ISC-specific G9a deletion (Supplementary Fig. 11a) and assessed its influence on crypt homeostasis and intestinal injury repair separately. Under homeostatic conditions, loss of G9a in ISCs resulted in a substantial reduction of H3K9me2 levels, leading to decreased stem cell proliferation, reduced cell numbers, and abnormal crypt architecture (Supplementary Fig. 11b–i). Furthermore, under radiation-induced injury conditions, ISC-specific G9a deficiency impaired regenerative capacity (Supplementary Fig. S12a–e), corroborating the critical role of H3K9me2 in maintaining ISC function during both homeostasis and regeneration.

To complement our in vivo findings, we established small intestinal organoids from isolated crypts and observed that G9a deficiency significantly reduced organoid area, indicating an impact on their growth (Fig. 3g, h). Individual organoid tracking revealed that those derived from G9a^ΔIEC crypts had significantly fewer buds compared to the ones from G9a^fl/fl, illustrating that H3K9me2 depletion leads to a reduction in the number of differentiated cells downstream of ISCs (Fig. 3i, j). We also treated organoids with UNC0642, which phenocopied the G9a deletion results (Supplementary Fig. 13a–d).

Overall, these results indicate that H3K9me2, regulated by G9a, controls the proliferation state of intestinal stem cells in crypts, thereby influencing their numbers; its dysfunction would damage intestinal homeostasis and regeneration capacity.

H3K9me2 suppressed cell cycle arrest genes

To elucidate the molecular events following H3K9me2 modification, we performed comprehensive genomic analyses, including transposase-accessible chromatin using sequencing (ATAC-seq), H3K9me2 and G9a chromatin immunoprecipitation with sequencing (ChIP-seq), and RNA sequencing (RNA-seq), to identify its responsive genes in isolated intestinal crypts under homeostatic conditions. ATAC-seq analysis revealed a significant increase in global chromatin accessibility in the crypt cells from G9a^ΔIEC mice, especially at transcription start sites (Fig. 4a). We identified 26,271 chromatin accessibility sites in G9a^ΔIEC crypt cells compared to 20,815 sites in G9a^fl/fl cells. Gene ontology enrichment analysis of these differential accessibility sites indicated primary association with regulation of cyclin-dependent protein serine/threonine kinase activity and cyclin-dependent protein kinase activity (Fig. 4b).

Fig. 4. Analysis of H3K9me2-regulated genes in intestinal crypt cells under homeostatic conditions.

a Heatmap of ATAC-seq analysis. Scatter plot displays average read counts within ±3 kb of the transcription start site (TSS). Heatmaps are organized based on G9a^fl/fl samples, and the color bar represents counts per million (CPM) mapped reads. b Pathway analysis of genes associated with differentially accessible sites identified by ATAC-seq. c Heatmap and average intensity plots (top row) illustrating H3K9me2 occupancy within ±3 kb of the TSS for all H3K9me2-positive genes. Analysis includes spike-in normalization controls for accurate quantification. d Pathway analysis of genes associated with differential H3K9me2 occupancy at TSS. e Heatmap and average intensity plots (top row) illustrating G9a occupancy within ± 3 kb of the TSS for all G9a-positive genes. f Pathway analysis of genes associated with differential G9a occupancy at TSS. g Gene expression heatmap of cell cycle-promoting and -inhibiting genes. h Pathway analysis of differentially expressed genes. i Venn diagram showing overlap of differentially regulated genes identified in ATAC-seq, ChIP-seq, and RNA-seq analysis. j Pathway analysis of overlapping genes identified in (i). All sequencing data in this figure were generated from intestinal crypt cells isolated from G9a^fl/fl and G9a^ΔIEC mice at homeostasis without injury treatment. Source data are provided as a Source Data file.

H3K9me2 ChIP-seq analysis showed that chromatin occupancy of H3K9me2 was significantly reduced in G9a^ΔIEC mice (30,861 peaks) compared to G9a^fl/fl mice (59,355 peaks), suggesting a global reduction in H3K9me2 levels following G9a deletion (Fig. 4c). Functional enrichment analysis of genes corresponding to significantly downregulated H3K9me2 peaks revealed enrichment in pathways related to negative regulation of cell proliferation (Fig. 4d). Similarly, G9a ChIP-seq analysis identified 20,851 G9a binding sites and their corresponding target genes, with functional annotation showing enrichment in negative regulation of cell proliferation and cell cycle pathways (Fig. 4e, f).

RNA-seq analysis revealed significant downregulation of cell cycle-promoting genes and upregulation of cell cycle-inhibiting genes in G9a^ΔIEC mice; differential gene enrichment analysis also pointed to cell cycle processes (Fig. 4g, h). Given that G9a-mediated H3K9me2 inhibits chromatin accessibility and thereby gene expression, we integrated the data from all three analyses to identify common differentially expressed genes, resulting in a list of 691 genes (Fig. 4i). Enrichment analysis of these differential genes primarily centered on pathways negatively regulation of cell proliferation, negatively regulation of cell cycle G1/S phase transition and regulation of cell cycle (Fig. 4j). These high-throughput data analysis confirms that in crypt cells, G9a-catalyzed H3K9me2 reduces chromatin accessibility at cell cycle-related loci and suppresses expression of genes that negatively regulate cell proliferation.

We then selected 4 genes closely related to the cell cycle from the aforementioned 691 genes, including Rb1cc1, Rb1, Cdkn1a, and Pten, for further validation. Genomic browser tracks showed that G9a knockout led to a significant reduction in H3K9me2 and G9a level at these gene loci, resulting in increased chromatin accessibility and upregulation of gene expression (Fig. 5a). These phenomena were confirmed by ChIP-qPCR. Our data showed significant binding of H3K9me2 and G9a to Rb1cc1, Rb1, Cdkn1α, and Pten in the crypt cells of G9a^fl/fl mice, whereas these bindings were markedly reduced in the crypt cells of G9a^ΔIEC mice (Fig. 5b). Correspondingly, these genes were significantly upregulated following G9a knockout (Fig. 5c). Similarly, treatment with UNC0642 under homeostatic conditions also resulted in decreased H3K9me2 levels at these gene regions and increases in their expression levels (Supplementary Fig. 14a, b).

Fig. 5. Suppression of cell cycle arrest gene expression by G9a-mediated H3K9me2.

a Genomic browser views showing ATAC-seq, ChIP-seq, and RNA-seq data at four representative target genes (Rb1cc1, Rb1, Cdkn1α, and Pten). Views highlight changes in chromatin accessibility, H3K9me2 and G9a binding, and gene expression between genotypes. Data were obtained from intestinal crypt cells isolated from mice at homeostasis. b H3K9me2 and G9a occupancy at the promoters of four genes at homeostasis. n = 3 mice/group. c Expression levels of four target genes at homeostasis. n = 3 mice/group. d Schematic of the experimental workflow for analyzing four target gene expression during irradiation and the regeneration process. e H3K9me2 occupancy at the promoters of four genes at the indicated time points post-irradiation. n = 3 mice/group. f Expression levels of the four target genes at the indicated time points post-irradiation. n = 3 mice/group. Data are presented as mean ± SEM. Statistical analysis was performed using a one-way ANOVA test in (b) and a two-tailed unpaired t-test in (c). Source data are provided as a Source Data file.

To establish a direct relationship between H3K9me2 and cell cycle arrest genes during intestinal regeneration, we detected H3K9me2 levels and these gene expressions in the crypts after irradiation (Fig. 5d). Indeed, the levels of H3K9me2 at these gene loci were significantly increased during the repair process. However, following G9a deletion, the H3K9me2 levels in these regions were almost vanished (Fig. 5e). Concurrently, G9a deficiency led to a slower decrease in the expression of these genes, resulting in delayed repair (Fig. 5f). These data further clarify that H3K9me2, modified by G9a, suppresses the expression of cell cycle-inhibiting genes, thereby participating in both intestinal homeostasis maintenance and injured tissue regeneration.

IL-4-STAT6 signaling regulates G9a expression during intestinal regeneration

To investigate the upstream regulatory pathway of G9a expression during intestinal epithelial regeneration initiation, we first employed a bioinformatic approach to identify potential transcription factors that could bind to the G9a promoter region. Using the JASPAR database, a collection of transcription factor DNA-binding profiles, we identified potential transcription factor binding sites in the G9a promoter (Supplementary Data 1). PANTHER signaling pathway enrichment analysis revealed that these transcription factors were associated with the JAK/STAT signaling pathway (Fig. 6a). Single-cell transcriptomic analysis of murine and human intestinal epithelium²⁹ demonstrated high expression of Stat6 and its receptor-encoding gene Jak1 in ISCs and TA cells, which aligned with the expression pattern of G9a in the intestinal epithelium (Fig. 6b). Meanwhile, we identified specific STAT6-binding motifs in the G9a promoter region (Fig. 6c), and ChIP-qPCR analysis confirmed that phosphorylated STAT6 directly binds to the G9a promoter, with binding significantly increased at 48 and 96 h post-irradiation (Fig. 6d). To determine which interleukin mediates this response, we analyzed the expression of various IL family members in isolated intestinal crypts during intestinal repair. Notably, only IL-4 showed significant upregulation during the repair process at both mRNA and protein levels, with peak expression occurring at 48–96 h post-irradiation (Fig. 6e, f). Correspondingly, STAT6 phosphorylation levels increased during repair, consistent with G9a expression patterns (Fig. 6g). These findings demonstrate that G9a expression is regulated by the IL-4-STAT6 signaling pathway during intestinal injury repair (Fig. 6h).

Fig. 6. IL-4-STAT6 signaling regulates G9a expression during intestinal regeneration.

a PANTHER signaling pathway enrichment analysis of transcription factors predicted to bind to the G9a promoter region. b Expression levels of Stat and Jak family members in ISCs and TA cells from mouse and human single-cell RNA sequencing data. Circle size represents expression level, and color indicates cell type. c STAT6 consensus binding motif identified in the G9a promoter region. d Phosphorylated STAT6 (pSTAT6) binding to G9a promoter at different time points post-irradiation. n = 3 mice/group. e Heatmap showing the expression of IL family members at different time points post-irradiation. f Quantitative analysis of IL-4 expression at different time points post-irradiation. n = 6 mice/group. g Immunoblotting of phosphorylated STAT6 and total STAT6 levels at different time points post-irradiation. Numbers below each blot represent relative band intensities normalized to 0 h. Each experiment was repeated independently three times with similar results. h Schematic model illustrating IL-4-STAT6-G9a signaling pathway during intestinal regeneration. Data are presented as mean ± SEM. Statistical analysis was performed using two-way ANOVA in (d) and one-way ANOVA in (f). Source data are provided as a Source Data file.

Discussion

Histone modifications play a critical role in controlling gene expression to maintain intestinal homeostasis. In this study, we identified that H3K9me2 and its methyltransferase G9a levels in IECs are downregulated at the intestinal injury stage and increased during the regeneration process. Intestinal repair in mice was significantly hindered when G9a was genetically deleted or its methyltransferase activity was inhibited, both of which led to a decrease in H3K9me2 levels. Moreover, G9a deficiency also impaired intestinal homeostasis under steady-state conditions, as evidenced by reduced crypt depth, villus height, and decreased numbers of ISCs and differentiated cell types. Integrative genomic analyses revealed that G9a exerted its effects by elevating H3K9me2 levels at the promoter regions of cell cycle arrest genes, including Rb1cc1, Rb1, Cdkn1α, and Pten, consequently suppressing their expression. This, in turn, promoted the proliferative capacity of IECs, a crucial step for both maintaining intestinal homeostasis and facilitating the regeneration of the injured epithelium. Furthermore, we found that IL-4 upregulation leads to STAT6 phosphorylation and subsequent transcriptional activation of G9a during intestinal regeneration. These findings may have significant therapeutic implications for diseases related to intestinal injury and/or regeneration, highlighting the potential of targeting this epigenetic mechanism in developing treatment strategies.

Previous studies have revealed that histone modifications control IECs' self-renewal and differentiation processes. For example, MLL1-mediated H3K4me3 modification is essential for intestinal stem cell maintenance by promoting the expression of key transcription factors, with its loss resulting in ISC depletion and secretory lineage bias³⁰. PRC2-mediated H3K27me3 modification controls progenitor cell proliferation and lineage balance, being particularly critical for radiation-induced regeneration through Cdkn2a repression^31,32. SETDB1-mediated H3K9me3 modification contributes to genomic stability by silencing endogenous retroviral elements^22,23. H3K9me2, a highly conserved histone methylation modification, has been shown to inhibit gene expression by promoting heterochromatin formation and hindering the assembly of transcription initiation complexes with tissue and cell specificity³³. In this study, we found that H3K9me2 modification is present in intestinal crypt ISCs and TA cells, and is selectively enriched in the promoter regions of cell cycle arrest-related genes. By inhibiting the expression of these genes, H3K9me2 modification maintains a high proliferative activity of crypt cells under both homeostatic and regenerative conditions. Therefore, this study not only expands current understanding of the histone modification landscape in maintaining intestinal homeostasis but also provides insights into the biological functions of H3K9me2 modification.

Importantly, our study unveils an unexpected role of G9a-mediated H3K9me2 modification in maintaining intestinal epithelial renewal by suppressing cell cycle arrest gene expression, contrasting with the traditional view of H3K9me2 as a repressive mark that inhibits cell behavior^33,34. The enrichment of G9a-mediated H3K9me2 modification in intestinal crypt cells and its selective enrichment in promoter regions containing C2H2 zinc finger motifs (Supplementary Fig. 15) may contribute to this unique negative regulatory mechanism. This mechanism represents a specialized modulatory function distinct from master epigenetic regulators. The relatively mild phenotype observed with constitutive G9a loss, compared to the rapid and severe intestinal defects seen with inducible loss of master epigenetic regulators like EZH2 or RING1B^35,36, reflects G9a’s specialized role as a modulatory rather than master regulator. While Polycomb complexes control fundamental developmental programs and core stem cell identity maintenance³⁶, G9a functions as a “fine-tuning” mechanism that provides regulatory precision without disrupting essential transcriptional networks. This modulatory approach complements the well-established positive regulatory mechanisms, such as Wnt and Notch pathways, in fine-tuning intestinal stem cell proliferation and self-renewal^37,38. Under physiological conditions, moderate negative regulation can finely control the level of stem cell proliferation, avoiding tissue homeostasis disruption caused by excessive self-renewal. Under pathological conditions such as intestinal injury, the attenuation of negative regulation (decreased expression of cell cycle arrest genes) can remove the “brake” on cell proliferation, thereby promoting tissue repair and regeneration. Therefore, this negative regulatory approach may play a critical “checks and balances” role in both maintaining intestinal homeostasis and enabling timely repair of intestinal injury.

Dysfunction in intestinal injury repair is closely associated with various diseases, such as radiation enteritis and inflammatory bowel disease, which severely impact patients’ quality of life and can even be life-threatening^39,40. Epigenetic reprogramming has emerged as a promising approach for drug development and clinical treatment, as it allows for the regulation of the genome’s transcriptional state³³. This approach has already been applied to diseases such as cancer, primarily by modulating the enzymatic activities involved in establishing and maintaining epigenetic modifications, thereby reactivating epigenetically silenced tumor suppressors and DNA repair genes. For instance, DNA methylation inhibitors are used to treat hematological malignancies by reactivating the expression of tumor suppressor genes and endogenous retroviruses³⁴. In this study, we discovered that G9a-mediated H3K9me2 modification plays a crucial role in the intestinal regeneration process, suggesting that G9a may be a potential therapeutic target. Our findings provide a foundation for developing strategies to promote intestinal repair by targeting epigenetic regulators. One potential approach could be to employ gene delivery systems, such as adenoviruses, to specifically overexpress G9a in injured intestinal tissues, thereby accelerating the intestinal epithelial repair process. However, further drug development and translational research based on our findings will be necessary to validate the efficacy and safety of such interventions.

In conclusion, our study reveals that G9a-mediated H3K9me2 modification suppresses cell cycle arrest genes to maintain intestinal stem cell proliferation under homeostasis and is dynamically upregulated during regeneration to promote tissue regeneration. Disruption of H3K9me2 impairs proliferation dynamics, highlighting its importance in both intestinal homeostasis and regeneration. These findings could lead to more precise and effective interventions in a range of intestinal disorders.

Methods

Ethics declarations

This study complied with all relevant ethical regulations regarding the use of human study participants and was conducted in accordance with the Declaration of Helsinki. Ethical approval for the use of human subjects was obtained from the ethics committees of Second Affiliated Hospital of Zhejiang University School of Medicine (#2017-072) and Sir Run Run Shaw Hospital of Zhejiang University School of Medicine (#20210622-31). All animal studies were performed in compliance with the Guide for the Care and Use of Laboratory Animals, and adopted the protocol approved by the Medical Experimental Animal Care Commission of ZJU (#ZJU20220219 updated).

Human subjects

Twenty rectal cancer patients who had received long-course neoadjuvant radiotherapy were enrolled in this study. The radiotherapy regimen consisted of 1.8 Gy per fraction administered 5 times per week for a total dose of 50.4 Gy. Tissue sections from both radiotherapy-induced injury regions and adjacent repair regions were obtained during radical resection at the Pathology Department of the Second Affiliated Hospital of Zhejiang University School of Medicine (Supplementary Table 1). Meanwhile, 15 non-radiotherapy patients were enrolled to serve as controls. Crohn’s disease (CD) patient samples used in this study were collected from the Inflammatory Bowel Disease Center at the Sir Run Run Shaw Hospital, affiliated with the Zhejiang University School of Medicine (Supplementary Table 2)^41,42. The diagnosis of CD was performed according to a standard combination of clinical, endoscopic, histological, and radiological criteria. Patient characteristics including age at diagnosis (A1, <16 years; A2, 16–40 years; A3, >40 years), disease localization (L1, ileal; L2, colonic; L3, ileocolonic; L4, upper gastrointestinal tract), and disease behavior (B1, non-stricturing and non-penetrating; B2, stricturing; B3, penetrating) were documented according to the Montreal classification. Disease activity was assessed using the Crohn’s Disease Activity Index (CDAI), a standardized scoring system that incorporates abdominal pain, general well-being, stool frequency, and other clinical parameters to provide a numerical score ranging from 0 to 600, with higher scores indicating more active disease⁴³. Sex-based analysis was not performed in this study for the following reasons: (1) the primary objective was to investigate the epigenetic mechanisms of intestinal epithelial regeneration, which are fundamental cellular processes not expected to differ substantially between sexes; (2) the sample size was not powered to detect potential sex differences; and (3) patients were enrolled based on clinical diagnosis and tissue availability rather than sex-stratified recruitment. Both sexes were included to enhance the generalizability of findings. Future studies with larger, sex-balanced cohorts will be needed to explore potential sex-specific differences in intestinal regeneration.

Mice

The G9a conditional knockout mice (G9a^fl/fl) and intestinal epithelial-specific Cre recombinase-expressing mice (Villin-cre) were obtained from the Shanghai Model Organisms Center. The C57BL/6 mice and Lgr5-EGFP-creERT2 knock-in mice were obtained from Shanghai Slack Company and Jackson Laboratory, respectively. Intestinal epithelial cell-specific G9a knockout mice (G9a^ΔIEC) and their littermate controls (G9a^fl/fl) were generated by crossing Villin-cre mice with G9a^fl/fl mice. Intestinal stem cell-specific G9a knockout mice were generated by crossing Lgr5-EGFP-creERT2 mice with G9a^fl/fl mice, followed by tamoxifen induction (75 mg/kg body weight, intraperitoneally, once daily for 5 consecutive days). Genotypes were confirmed by PCR amplification using primers listed in Supplementary Table 3. Littermates were randomly assigned to experimental groups. All mice were maintained under pathogen-free conditions at the Laboratory Animal Centre of Zhejiang University with a controlled 12-h light/dark cycle, ambient temperature of 22 ± 1 °C, and relative humidity of 50 ± 10%. Mice were provided with water and a standard laboratory diet ad libitum unless otherwise noted. Male mice aged 8–10 weeks were used in all experiments unless otherwise specified. Sex was considered in the study design; male mice were selected to minimize potential confounding effects of estrous cycle-related hormonal fluctuations on intestinal epithelial dynamics and to ensure experimental consistency. Data are reported for male mice, and findings may not be generalizable to female mice without further investigation.

Intestinal injury and regeneration mouse model

For the irradiation-induced model, eight-week-old C57BL/6 male mice were subjected to 8 Gy or 10 Gy whole-body IR with an RS2000 X-source irradiator at a dose rate of 60 or 85 cGy/min. Intestinal tissue samples were collected at 0, 1, 6, 12, 24, 48, and 96 h post-irradiation for subsequent analyses. In this model, IECs typically undergo an apoptotic phase within hours post-irradiation, followed by a proliferation and repair phase starting around 24 h later, during which surviving cells proliferate and restore the intestinal lining⁴⁴.

For the dextran sulfate sodium (DSS) induced model, eight-week-old C57BL/6 male mice were treated with DSS (2.5% w/vol, molecular weight: 36–50 kDa, #160110; MP Biomedicals) in drinking water for 7 days, followed by regular drinking water. Intestinal tissue samples were collected at 0, 7, 10, and 14 days after DSS treatment initiation for subsequent analyses. In this model, mice reached peak injury severity on day 10 of DSS treatment before gradually returning to baseline levels⁴⁵.

G9a pharmacological inhibition

The G9a inhibitor UNC0642 (CAS: 1481677-78-4; #A15487; AdooQ Bioscience) was dissolved in a vehicle solution containing 10% DMSO, 20% β-cyclodextrin, and normal saline. The stock solution was prepared at a 5 mg/mL concentration and stored at −20 °C until use. For treatment, the stock solution was diluted 10-fold with sterile saline to achieve a working concentration of 0.5 mg/mL. Eight-week-old male C57BL/6 J mice were administered with UNC0642 via intraperitoneal injection at a dose of 5 mg/kg body weight every other day for four consecutive weeks. Control mice received an equivalent volume of vehicle solution following the same injection schedule. Body weight was monitored throughout the treatment period to assess potential toxicity. Following the treatment period, mice were used for subsequent experimental analyses, including tissue collection, organoid culture, and phenotypic assessments.

Histone modification screening and identification

For comprehensive histone modification analysis in intestinal regeneration, we performed liquid chromatography-mass spectrometry (LC-MS) analysis⁴⁶. Intestinal crypts were isolated from mice before and after irradiation, then lysed in buffer containing 2 mM EDTA, 3 μM trichostatin A (TSA; #HY-15144, MedChemExpress), 50 mM nicotinamide (NAM; #72340, Sigma-Aldrich), 1% Triton X-100, and protease inhibitor cocktail (#04693132001, Roche) at 4 °C for 30 min. Following centrifugation at 20,000 × g for 10 min, the pellet was resuspended in 0.2 M H₂SO₄, sonicated on ice, and centrifuged again to obtain the histone-enriched supernatant. Protein concentrations were determined using a BCA assay kit.

Extracted histones were propionylated in 1 M triethylammonium bicarbonate (TEAB; #T7408, Sigma-Aldrich) buffer at room temperature, followed by overnight trypsin digestion at 37 °C at an enzyme-to-histone ratio of 1:20 (w/w). Samples were subsequently treated with hydroxylamine (final concentration 0.25%) and ammonia water (final concentration 4%) to remove nonspecific chemical modifications, then quenched with formic acid (final concentration 1%). The resulting peptides were analyzed by LC-MS/MS using an EASY-nLC 1200 UPLC system coupled to an Orbitrap Exploris 480 mass spectrometer (both Thermo Fisher Scientific) operating with FAIMS compensation voltage set at −45 V. LC-MS/MS analysis was performed by PTM BIO (Hangzhou, China). Data were processed using Spectronaut software v16 (Biognosys) with a comprehensive variable modification search including acetylation, mono-, di-, and tri-methylation of lysine residues, enabling identification and quantification of histone post-translational modifications across experimental conditions.

Hematoxylin and eosin (H&E), immunohistochemical (IHC), and Periodic Acid-Schiff (PAS) staining

The intestinal tissue samples were washed with 70% ethanol and then fixed in 10% formalin solution overnight. The fixed tissues were embedded in paraffin and sectioned. Sections were deparaffinized in xylene for 30 min and rehydrated through a graded ethanol series (100%, 95%, and 80%) for 5 min each, followed by a 5-min incubation in distilled water. For H&E staining, the sections were immersed in hematoxylin for 5 min and then in phosphate-buffered saline (PBS) for 3 min to prevent background staining, followed by staining in eosin for 2 min and distilled water for 5 min. For IHC staining, the sections were subjected to high-temperature antigen retrieval in 0.01 M citrate buffer (pH 6.0) for 5 min, blocked in 1.5% normal goat serum (#ZLI-9056, ZSGB-BIO) for 30 min at room temperature, and then incubated with mouse anti-Ki67 antibody (#ab15580, Abcam), rabbit anti-OLFM4 antibody (#39141, Cell Signaling Technology), rabbit anti-LYZ antibody (#ab108508, Abcam) and rabbit anti-Chromogranin A antibody (#ab237979, Abcam) overnight at 4 °C. The primary antibody was detected using an avidin/biotin-based peroxidase system (#ab64261 or #ab64264, Abcam) according to the manufacturer’s instructions. Briefly, the sections were incubated with a biotinylated goat anti-rabbit IgG or anti-mouse IgG secondary antibody for 60 min. After three washes in PBS, the sections were treated with a solution of avidin-biotin-peroxidase complex for 60 min and then incubated with 3,3′-diaminobenzadine (DAB) and H₂O₂ in PBS for 45–90 s to develop the insoluble brown chromogen. Hematoxylin was used as a counterstain for 5 min to visualize the nuclei. For PAS staining, the deparaffinized and rehydrated sections were immersed in a 0.5% periodic acid solution (#DG0005, LEAGENE) for 5 min, followed by a wash in distilled water. The sections were then submerged in Schiff’s reagent (#DG0005, LEAGENE) for 15 min and subsequently washed in distilled water. Counterstaining was performed using Hematoxylin solution for 1 min. After H&E, IHC, or PAS staining, the sections were dehydrated through a graded ethanol series (80%, 90%, and 100%) for 5 min each, followed by a 15-min incubation in xylene. The stained sections were mounted using neutral resin and a coverslip, and observed and digitally photographed using a Leica DM2000 LED.

Histological analysis

Histological assessment of stained sections was performed by a board-certified pathologist who was blinded to the experimental groups. The scoring criteria for staining intensity were as follows: 0 = no staining; 1 = weak staining (light yellow); 2 = moderate staining (yellow-brown); and 3 = strong staining (brown). For each sample, five randomly selected high-power fields were examined. Any discrepancies in scoring between the pathologist and a second observer were resolved by consensus. For the evaluation of colitis severity in mouse samples, the extent and severity of inflammation and ulceration of the mucosa were assessed using the following criteria. The severity score for inflammation was as follows: 0, normal (within the normal limit); 1, mild (small, focal, or widely separated, limited to the lamina propria); 2, moderate (multifocal or locally extensive, extending to the submucosa); and 3, severe (transmural inflammation with ulcers covering >20 crypts). The score for ulceration was as follows: 0, normal (no ulcers); 1, mild (1–2 ulcers involving up to a total of 20 crypts); 2, moderate (3–4 ulcers involving a total of 20-40 crypts); and 3, severe (>4 ulcers or >40 crypts). The overall colitis score was calculated as the sum of the inflammation severity score and the ulceration score, resulting in a total score ranging from 0 to 6. Higher scores indicate more severe colitis.

Immunofluorescence staining

Intestinal tissues were fixed with 4% paraformaldehyde (PFA) at 4 °C for 2 h. After washing with cold PBS, the tissues were dehydrated in 30% sucrose at 4 °C for 12 h and then embedded in optimal cutting temperature (OCT) compound for cryosectioning. The frozen sections were hydrated in PBS for 5 min and blocked with 1.5% normal goat serum for 1 h at room temperature. The sections were then incubated with rabbit anti-EHMT2/G9a antibody (#ab185050, Abcam) overnight at 4 °C. The primary antibody was detected using a fluorescent-conjugated species-specific secondary antibody (Goat anti-Rabbit IgG antibody/Alexa Fluor 555, #A31572, Invitrogen). Nuclei were counterstained with Hoechst 33258 dye (#H3570, Life Technologies) for 15 min at room temperature. After three washes in PBS, the stained sections were mounted using neutral resin and coverslips. The sections were then examined and digitally photographed using a Nikon A1R Confocal Microscope.

EdU pulse-chase assay

To assess cell proliferation and track the fate of EdU-labeled cells, mice were administered a single injection of 8 mg/kg EdU (#A10044, Thermo Fisher Scientific) and analyzed at the indicated time points (1.5 h, 24 h, and 48 h post-injection). EdU staining was performed using the BeyoClickTM EdU Cell Proliferation Kit with Alexa Fluor 594 (#C0078S, Beyotime Biotechnology) according to the manufacturer’s protocol. Briefly, intestinal tissue sections were fixed with 3.7% formaldehyde and permeabilized with 0.5% Triton X-100 in PBS for 20 min. EdU labeling was visualized by incubating the sections with freshly prepared Click-iT reaction buffer containing azide-conjugated Alexa Fluor 594 for 30 min at room temperature. To delineate the cell boundaries, EdU-stained sections were washed with 3% bovine serum albumin (BSA, #ST025, Beyotime Biotechnology) in PBS and incubated with a primary rabbit anti-E-cadherin antibody (#3195, Cell Signaling Technology) overnight at 4 °C. The sections were then incubated with a secondary goat anti-rabbit IgG antibody conjugated to Alexa Fluor 647 (#A21446, Thermo Fisher Scientific) for 1 h at room temperature. Finally, the sections were stained with Hoechst 33342 for 15 min to visualize the nuclei. Images were acquired using a Nikon A1 confocal microscope system.

Crypt isolation

To isolate intestinal crypts, the mouse small intestine was dissected and opened longitudinally. The tissue was briefly washed with ice-cold Dulbecco’s phosphate-buffered saline (DPBS, #14190144, Thermo Fisher Scientific) and further dissected into small pieces. Using a glass coverslip, the villi were gently scraped off, and the tissue fragments were washed with cold DPBS to remove unattached epithelial fragments. The intestinal fragments were then incubated in DPBS containing 3 mM EDTA for 10 min on ice. After removing the EDTA solution, the fragments were washed once with DPBS and vigorously suspended in DPBS using a 10 mL pipette. The supernatant containing the villous fraction was discarded, and the sediment was resuspended in PBS. To enrich for crypts, the resuspended sediment was vigorously suspended once more, and the resulting supernatant was collected. This crypt-enriched fraction was passed through a 70 μm cell strainer to remove residual villous material and then centrifuged at 70 g for 3 min to separate the crypts from single cells. The purified crypts were counted and subsequently used for organoid culture or fluorescence-activated cell sorting (FACS), depending on the experimental requirements.

Organoid culture

For organoid culture, 100–300 isolated crypts were embedded in 30 μL of Matrigel (#356231, BD Biosciences) and cultured in IntestiCult Organoid Growth Medium (#06005, STEM CELL Technologies) supplemented with 100 U/mL penicillin and 100 μg/mL streptomycin (#15070063, Thermo Fisher Scientific). The culture medium was replaced every 2–3 days, and the organoids were maintained under standard tissue culture conditions (37 °C, 5% CO₂). To assess organoid growth and morphology, bright-field images of the organoids were acquired using an inverted microscope (ECLIPSE Ti2, Nikon) at various time points during culture. The images were analyzed using ImageJ software v1.53 (National Institutes of Health). Quantitative analysis of organoid growth was performed by determining the number of de novo buds and measuring the area of the organoids between day 3 and day 5 of culture.

Fluorescence-activated cell sorting (FACS)

To isolate specific cell populations from the intestinal crypt, a single-cell suspension was prepared by filtering the crypt cells through a 40 μm cell strainer. The cells were then simultaneously labeled with the following fluorescence-conjugated antibodies in staining buffer (2 mM EDTA and 3% fetal bovine serum (FBS, #10100147, Thermo Fisher Scientific) in PBS): CD31-BV510 (#102407, BD Biosciences), CD45-BV510 (#103105, BD Biosciences), FVD450-BV450 (to distinguish live cells from dead/dying cells) (#65-0863-14, eBioscience), EpCAM-APC (#17-5791, Thermo Fisher Scientific) and CD24-PerCP-Cyanine 5.5 (#562360, BD Biosciences). The cells were gently shaken for 30 min on ice to allow for antibody binding. After labeling, the cells were washed three times with staining buffer and resuspended at a concentration of 5 × 10⁶ cells/mL in Advanced DMEM/F12 medium (#12634028, Gibco). The labeled cells were then analyzed and sorted using a BD FACS Aria II flow cytometer. The following markers were used to identify specific cell populations: LGR5-EGFP^hi; EpCAM⁺; CD24^med/−; CD31⁻; FVD450⁻; CD45⁻; TA cells were LGR5-EGFP^low; EpCAM⁺; CD24^med/−; CD31⁻; FVD450⁻; CD45⁻; and Paneth cells were LGR5-EGFP^neg; EpCAM⁺; CD24^hi; CD31^-; FVD450^-; CD45⁻; Side scatter^hi. The sorted cell populations were collected for further analysis or downstream applications. Cell population analysis was performed using FlowJo software v10.4.

RNA isolation and real-time quantitative PCR (RT-qPCR)

For mRNA analysis, total RNA was extracted from the samples using TRIzol reagent (#15596026, Thermo Fisher Scientific) following the manufacturer’s instructions. The quality and quantity of the isolated RNA were assessed using a Nanodrop 2000 spectrophotometer. Reverse transcription of mRNA to cDNA was performed using Moloney murine leukemia virus (MMLV) reverse transcriptase (#639524, Takara Bio). The resulting cDNA was then subjected to RT-qPCR using SYBR Green master mix (#RR420A, Takara Bio) on a LightCycler 480 real-time PCR system (Roche). The PCR primers used were designed using the NCBI Primer-BLAST tool and listed in Supplementary Table 4. The relative quantities were calculated using the ΔΔCt method. To ensure the integrity of RNA samples, all equipment used for RNA manipulations was sterilized according to standard laboratory protocols, and diethylpyrocarbonate-treated water was used throughout the process.

RNA sequencing (RNA-Seq) and data analysis

The mRNA was enriched from 500 ng of total RNA by poly-dT enrichment using the NEBNext Poly(A) mRNA Magnetic Isolation Module (#E7490L, NEB) according to the manufacturer’s instructions. The enriched samples were then directly subjected to the workflow for strand-specific RNA-Seq library preparation using the Ultra II Directional RNA Library Prep kit (#E7760L, NEB). The libraries were quantified using the Agilent 2100 Bioanalyzer (Agilent Technologies) and equimolarly pooled and sequenced with 75 bp single-end reads on a NextSeq 500 (Illumina). After sequencing, RNA-SeQC (1.1.8) was used to perform basic quality control, including the assessment of exonic, intronic, and intergenic distribution of the reads and the rRNA rate within each sample. Alignment of the reads to the mouse reference genome (mm10) was performed using GSNAP (v2018-07-04), and Ensembl gene annotation V.92 was used to detect splice sites. Uniquely aligned reads were counted with featureCounts (V.1.6.3) using the same Ensembl annotation. Normalization of the raw read counts based on library size and testing for differential expression between the two conditions was performed using the DESeq2 R package (V.1.24) and IHW (V.1.12).

Chromatin immunoprecipitation sequencing (ChIP-Seq) and data analysis

At least 2 × 10⁵ intestinal crypts were lysed in 1 mL single-cell dissociation buffer (TrypLE Express Enzyme (#12604-021, Gibco), 1% DNase I (260 U/μL, #18047019, Invitrogen)). After incubation for 2 min at 32 °C, the sample was filtered through a 40 μm cell strainer. Following centrifugation at 1300 × g for 3 min at 4 °C, the sample was resuspended in 1 mL ice-cold phosphate-buffered saline. To enable spike-in normalization for accurate quantification of histone modifications across experimental conditions, the ChIP DNA from HEK293T cells, equivalent to 1% of the murine crypt cell ChIP DNA amount, was added to the murine ChIP samples, following the established protocol for spike-in control of epigenomic analysis⁴⁷. The cells were crosslinked in 1% formaldehyde, quenched in 0.125 M glycine, and lysed in cell lysis buffer (50 mM Tris pH 8.1, 10 mM EDTA, 1% SDS). Lysates were sonicated with Bioruptor Pico (Diagenode) and diluted in dilution buffer (20 mM Tris pH 8.1, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100) with a pre-incubated mixture of rabbit anti-H3K9me2 antibody (#ab1220, Abcam), rabbit anti-G9a/EHMT2 antibody (#68851, Cell Signaling Technology), rabbit anti-Phospho-Stat6 antibody (#56554, Cell Signaling Technology), or species-matched IgG (#2729, Cell Signaling Technology) and Dynabeads Protein A and Protein G (#80106 G, Invitrogen). After overnight incubation, samples were washed five times with 1 mL low salt washing buffer (75 mM NaCl, 50 mM Tris-HCl pH 7.5, 10 mM EDTA, and 0.01% NP40), 1 mL high salt washing buffer (100 mM NaCl, 50 mM Tris-HCl pH 7.5, 10 mM EDTA, and 0.01% NP40), and 1 mL LiCl washing buffer (50 mM HEPES pH 7.6, 0.5 M LiCl, 1 mM EDTA, 0.7% sodium deoxycholate, 1% NP40), followed by two washes with Tris-EDTA Buffer (10 mM Tris pH 8, 1 mM EDTA). DNA was eluted overnight in decrosslinking Buffer (0.1 M NaHCO₃, 1% SDS) and purified using the QIAquick PCR Purification Kit (#28106, QIAGEN). The ChIP DNA was further treated with RNase A (#EN0531, Invitrogen) and Protease K (#25530049, Invitrogen), followed by DNA purification using a PCR purification kit (#28104, QIAGEN). The DNA samples were quantified using Qubit fluorometric quantification (Thermo Fisher Scientific). The libraries were constructed through end repair, A-tailing, adapter ligation, size selection, and PCR amplification, followed by fragment size analysis using the Agilent 5400 system and quantification to 1.5 nM via qPCR. and sequenced on the Illumina platform, generating paired-end reads of 150 bp to achieve an average of 20 M reads per sample at the Novogene Genomics Center (China).

For analysis, reads were trimmed using fastp (v 0.20.0) and aligned to the mouse reference genome (mm10) using BWA (v 0.7.12). Spike-in reads were aligned to the human reference genome (hg38) to calculate spike-in normalization factors for each sample. Normalization was performed by scaling ChIP-seq signal intensities based on the ratio of spike-in reads between samples, following established guidelines to correct for technical variations in chromatin preparation and immunoprecipitation efficiency. After mapping reads to the reference genome, the MACS2 (v2.2.7.1) peak calling software was used to identify regions of IP enrichment over background. A q-value threshold of 0.05 was used for all data sets. DiffBind was used to predict and annotate differential peaks with the following parameters. Peak calling and differential analysis incorporated spike-in normalization factors to ensure accurate detection of biologically relevant changes in histone modifications. Differential peaks with a false discovery rate (FDR) P < 0.05 were considered statistically significant.

For qPCR validation of ChIP-seq data, the DNA samples were diluted appropriately and analyzed using SYBR Green-based real-time PCR on a LightCycler 480 real-time PCR system. The PCR primers used were designed using the NCBI Primer-BLAST tool and listed in Supplementary Table 5. The relative enrichment of each target region was calculated using the percent input method, where the amount of immunoprecipitated DNA is normalized to the input DNA for each sample.

Assay for transposase-accessible chromatin using sequencing (ATAC-Seq) and data analysis

The single cells from intestinal crypts were lysed in lysis buffer (10 mM Tris-HCl pH 7.4, 10 mM NaCl, 3 mM MgCl₂, and 0.15% NP40) for 10 min on ice to prepare the nuclei. Immediately after lysis, nuclei were centrifuged at 500 × g for 5 min to remove the supernatant, and incubated with Tn5 transposome (#53150, Active Motif) and tagmentation buffer (#53150, Active Motif) at 37 °C for 30 min. Following tagmentation, the stop buffer (#53150, Active Motif) was directly added to the reaction. PCR amplification of the library was performed for 15 cycles using the following conditions: 72 °C for 3 min; 98 °C for 30 s; and thermocycling at 98 °C for 15 s, 60 °C for 30 s, and 72 °C for 3 min; followed by a final extension at 72 °C for 5 min. After PCR, libraries were purified with 1.2× AMPure beads (A63880, Beckman) before proceeding to mitochondrial DNA depletion. The libraries were quantified using the Qubit dsDNA HS Assay Kit (#Q32854, Thermo Fisher Scientific), and the size distribution was assessed using the Agilent 2100 Bioanalyzer. The libraries were then diluted to a final concentration of 2 nM and pooled in equimolar ratios. 2 × 150 bp paired-end (PE150) sequencing was performed on an Illumina NovaSeq 6000 platform at a depth of approximately 50 million reads per sample. Raw reads were trimmed using Trim Galore (v0.4.4) and aligned to the mouse reference genome (mm10) using Bowtie 2 (v2.2.6). Peak detection was performed using MACS2 software to identify open chromatin regions across the genome for each sample. HOMER (v4.9) was used to predict and annotate differential peaks. Differential peaks with a fold change ≥1.5 and an FDR < 0.001 were considered statistically significant. Gene Ontology (GO) analysis of differentially expressed genes (DEGs) was performed using the ShinyGO (http://bioinformatics.sdstate.edu/go/).

Immunoblotting

Intestinal tissues, organoids, or cells were lysed in radioimmunoprecipitation assay buffer (RIPA, #P0013C, Beyotime Biotechnology) supplemented with protease inhibitor cocktail (#04693132001, Roche). Protein concentrations were quantified using a BCA protein assay kit (#P0012, Beyotime Biotechnology). Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and then transferred to a 0.45 µm polyvinylidene fluoride (PVDF) membrane (#FFP39, Beyotime Biotechnology). The membrane was blocked with 5% BSA prepared in Tris-buffered saline containing 0.05% Tween 20 (TBST) for 1 h, washed with TBST for 5 min, and incubated with primary antibodies diluted in 5% BSA overnight at 4 °C on a rocker. After washing with TBST for 5 min, the membrane was incubated with HRP-conjugated secondary antibodies (anti-mouse IgG, #SA00001-1; anti-rabbit IgG, #SA00001-2; Proteintech) diluted in 5% BSA for 1 h at room temperature. Chemiluminescence signals were detected using SuperSignal West Pico Chemiluminescent Substrate (#34580, Thermo Fisher Scientific), and images were acquired using a chemiluminescence imager, Amersham Imager 600 (GE Healthcare Life Sciences). The following primary antibodies were used in this study: rabbit anti-EHMT2/G9a (#ab185050, Abcam), rabbit anti-Mono-Methyl-Histone H3K9 (#A2358, ABclonal), rabbit anti-Di-Methyl-Histone H3K9 (#4658, Cell Signaling Technology), rabbit anti-Tri-Methyl-Histone H3K9 (#A2360, ABclonal), rabbit anti- Acetyl-Histone H3K14 (#7627, Cell Signaling Technology), rabbit anti-Tri-Methyl-Histone H3K27 (#9733, Cell Signaling Technology), rabbit anti-Phospho-Stat6 antibody (#56554, Cell Signaling Technology), rabbit anti-STAT6 antibody (#ab32520, Abcam), rabbit anti-Cyclin A2 (#A19036, ABclonal), rabbit anti-Cyclin H (#A4076, ABclonal), rabbit anti-Cyclin E2 (#A9305, ABclonal), rabbit anti-Cyclin D1 (#A19038, ABclonal), mouse anti-Histone H3 (#sc-517576, Santa Cruz), and mouse anti-ACTB (#60004-1-Ig, Proteintech).

Sample size and sample collection

No statistical methods were used to predetermine sample size for in vivo and in vitro experiments, but at least three samples were used per experimental group and condition. The number of samples used in each experiment is indicated in the Source Data. Samples and experimental animals were randomly assigned to experimental groups. Animal procedures (that is, genotyping and treatments) were performed by investigators unaware of the experimental design.

Statistical analysis

All data in this paper are presented as mean ± standard error of the mean (SEM). Normality of the data was assessed using the Shapiro–Wilk test. For comparisons between two groups, a two-tailed unpaired t-test (Student’s t-test) was used for normally distributed data, and the Mann-Whitney U-test was used for non-normally distributed data. Differences among variables were evaluated using one-way analysis of variance (ANOVA) and two-way ANOVA for normally distributed data, and the Kruskal–Wallis test for non-normally distributed data. All statistical analyses were performed using GraphPad Prism 10. Statistical significance was defined as P < 0.05 unless otherwise specified, and exact P values are reported where applicable.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

J.C., X.S., X.Z., and J.H. share co-first authorship. J.C., J.S., and Z.X. conceived of the project. J.C., X.S., L.X., Z.H., J.G., and X.S. performed the mouse experiments. J.C. and X.F. performed high sequencing analysis. J.C., Z.H., J.G., and J.H. were involved in the molecular biology experiments. X.G., X.Z., Q.X., and W.Z. provided essential materials and analyzed the clinical samples. J.C. and J.S. wrote the original draft. J.S., R.B., and Z.X. reviewed and revised the final version of the text. R.B., J.S., and Z.X. supervised the study.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Data availability

The RNA-seq, ATAC-seq, and ChIP-seq data generated in this study have been deposited in the NCBI’s Sequence Read Archive (SRA) database under accession code PRJNA1135937, PRJNA1136160, and PRJNA1136210, respectively. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-68626-7.