Fasting primes small intestinal regeneration after damage via a microbiome–metabolite–chromatin axis

Significance

Fasting changes the gut microbiome, but how these changes help the body recover from damage is not well understood. We found that fasting increases a helpful bacterium, Akkermansia muciniphila, which produces propionate, which drives epigenetic changes by modifying histones and regulating gene activity. These changes promote the expansion of primed persister cells that help the intestine recover after radiation. This study shows how fasting and gut bacteria work together to protect healthy tissue and suggests that diet or microbial treatments could help reduce side effects of cancer radiotherapy.

Abstract

Fasting enhances small intestinal regeneration after radiation, but the contribution of the gut microbiome to this process remains uncharacterized. We identify Akkermansia muciniphila (AKK) as a key mediator of this response. AKK was enriched in fasted mice and its antibiotic depletion abrogated radioprotection, whereas reintroduction restored both organismal survival and intestinal integrity. Fasting elevated propionic acid, consistent with AKK’s metabolic output. AKK-conditioned medium and propionate induced histone H3 acetylation in intestinal stem cell cultures while in vivo fasting induced AKK-dependent H3K27ac and H3K9ac, remodeling promoter–enhancer landscapes in crypt epithelial cells. Epigenetic profiling revealed a rewired core regulatory program enriched for pioneer transcription factors (Foxa, Gata, Klf), architectural organizers (Ctcf, Boris), and lineage-defining and metabolic regulators (Cdx2, Hnf4). This program supports expansion of a population of primed persister cells characterized by open chromatin accessibility at key stem and regenerative-associated loci including Clu, Olfm4, Lgr5, Ascl2, Lrig1, Sox9, Rnf43, and Axin2. These findings define a fasting-induced microbiome–metabolite–chromatin axis that epigenetically primes highly plastic persister cells for rapid regeneration of the intestinal epithelium following radiation-induced injury.

Fasting induces widespread changes in metabolism, chromatin architecture, and gene expression across multiple tissues (1, 2) and reshapes the gut microbiota (3). We previously demonstrated that fasting protects mice from small intestinal injury caused by high-dose etoposide (4) and ionizing radiation (5). Fasted mice maintain superior intestinal architecture compared with fed mice shortly after chemotherapy or irradiation, with deeper crypts, taller villi, and reduced epithelial cell damage (4, 5), supporting the conclusion that fasting improves tissue structure early after injury. The protection arises, in part, from the preservation of stem cells in the small intestines (SI), which promotes tissue homeostasis and organismal survival (4, 5). However, the molecular mechanisms by which fasting primes the SI for regeneration after injury remain poorly understood.

The continuous turnover of the intestinal epithelium is driven by multipotent LGR5+ crypt-base columnar cells (CBCs) located at the base of intestinal crypts (6). However, CBCs are highly sensitive to injury and are rapidly lost following insults such as irradiation (7). Despite this loss, the intestinal epithelium exhibits a remarkable capacity for regeneration (8), suggesting the involvement of additional stem or progenitor populations.

A second population of quiescent “+4” cells, also referred to as reserve stem cells (RSCs), was previously proposed to mediate intestinal regeneration following damage (9–12). Although CBCs and RSCs were initially considered mutually exclusive populations (12, 13), subsequent studies demonstrated that LGR5+ CBCs express markers associated with RSCs (14). Moreover, functional studies revealed that RSCs are dispensable for intestinal repair, whereas LGR5+ cells are essential for regeneration following damage (8).

More recently, a rare damage-induced stem cell population, termed revival stem cells (revSCs), has been identified as a key contributor to intestinal regeneration. revSCs are characterized by high CLU expression and are extremely rare under homeostatic conditions (15). These cells are multipotent and can give rise to all major intestinal epithelial lineages, including LGR5+ CBCs. Following irradiation-induced injury, revSCs undergo a transient YAP1-dependent expansion, reconstitute the LGR5+ CBC compartment, and are required for restoration of a functional intestinal epithelium.

In addition to revSCs, multiple progenitor populations, including absorptive enterocyte progenitors (16), secretory lineage progenitors (17–21), and slow-cycling LGR5+ cells (22), have been shown to contribute to epithelial regeneration. The transcriptional regulator YAP1, a critical mediator of intestinal regeneration, has been proposed to promote a prosurvival phenotype in LGR5+ cells (23). Collectively, these findings raise an unresolved question: whether intestinal regeneration is driven primarily by specialized injury-responsive stem cell populations or by broader cellular plasticity across multiple epithelial lineages and whether the regenerative response of the SI can be preconditioned by stress-associated states such as fasting.

Akkermansia muciniphila (AKK), a mucin-degrading commensal bacterium, is enriched in the intestinal microbiome of fasted mice (24). Here, we show that AKK is required for fasting-mediated SI radioprotection. Depletion of AKK abrogated radioprotection, whereas reintroduction of AKK restored epithelial integrity and improved survival. Mechanistically, AKK produces the short chain fatty acid propionate, which promotes histone H3 acetylation at lysine 9 and 27 in SI crypt epithelial cells, thereby remodeling enhancer–promoter landscapes. Using integrated single-cell ATAC-seq and H3K27ac CUT&Tag profiling, we identify an AKK-dependent transcription factor (TF) regulatory program enriched in Hnf, Foxa, Gata, and Klf activities that likely drives expansion of a Clu+Olfm4+ population which we name “primed persister cells”. This cellular state persists across injury-regeneration phases and is associated with transcriptional programs that accelerate small intestinal regeneration. Our work provides evidence for revitalization of epigenetic memory of developmental chromatin regulatory programs by regenerative cells postradiation damage and identifies a fasting–AKK–metabolite–chromatin axis that primes the intestinal epithelium for injury resilience. Together, these findings establish a microbiome-based framework for protecting normal tissues during radiation therapy.

Results

Fasting Induces Accumulation of AKK in the SI of Mice.

Given that fasting is known to impact the gut flora (24), we compared the diversity and composition of bacteria in the SI of fed and fasted animals (Fig. 1). Mice were allowed to feed freely or were fasted for 24 h according to the schema (Fig. 1A). Mice received total abdominal irradiation (TA-XRT) as previously described (25). Consistent with prior findings (5), animals fed ad libitum before high-dose radiation died between 7 and 9 d after treatment, whereas all fasted mice survived to the study endpoint (Fig. 1B).

Fig. 1.

AKK accumulates in the SI of fasted mice. (A) Schematic of the experimental design. C57BL/6J mice were allowed to feed freely or were fasted for 24 h. On day 0, small intestinal ileum samples were collected from one cohort of mice for 16S rRNA sequencing to analyze the SI microbiome composition. Another cohort of mice was exposed to total abdominal radiation (TA-XRT) at a dose of 11.5 Gy. After radiation exposure, mice were returned to single-housed cages with food and survival was monitored daily. (B) Survival of fed and fasted mice following TA-XRT was monitored daily for 30 d. P <0.0001 (Log-rank (Mantel–Cox) test). (C) Diversity of the SI microbiome in ileum samples collected on D0 was quantified using Simpson’s reciprocal index. P = 1 (using the Mann–Whitney U test). (D) Absolute abundance of total 16S copies per gram of ileum sample collected from fed and fasted mice were measured by the standard spike-in method. Representative data shown from five mice. P = 0.44. Data are presented as mean ± SEM, analyzed by Student’s t test, unpaired. (E) Principal coordinates analysis (PCoA) based on weighted UniFrac distances was performed to evaluate the differences in microbiome composition between fed and fasted mice. Statistical significance was determined by permutational MANOVA testing. P = 0.001. (F) Relative abundance of AKK in ileum samples of fed and fasted mice. P = 4e-05 (Mann–Whitney U test). (G) Relative abundance of bacteria at the genus level in ileum samples collected on day 0. Each column represents an individual mouse.

Assessment of microbiome using 16S rRNA based gene sequencing of fecal material collected from the ilea of fed and fasted mice revealed minimal changes in the diversity (Fig. 1C) or absolute abundance (Fig. 1D) of bacterial flora between the two groups. However, significant compositional differences between fed and fasted cohorts were readily detected (Fig. 1E), with AKK abundance significantly enriched in the SI of fasted mice compared with their fed counterparts (Fig. 1 F and G). Other changes in bacterial flora associated with fasting included a significant increase in the relative abundance of Erysipelatoclostridium, Clostridium sensu stricto 1, Coprococcus, and Marvinbryantia and a significant decrease in the relative abundance of Lachnospiraceae, Lactobacillus, Streptococcus, and the Clostridia vadin BB60 group (SI Appendix, Fig. S1).

AKK Contributes to Organismal and Small Intestinal Radioprotection.

To determine if AKK contributes to organismal and SI radioprotection, mice were exposed (or not) to broad spectrum antibiotics (Ampicillin and Enrofloxacin) in their drinking water for 7 d according to the schema (Fig. 2A) (26). Microbiome analysis revealed significant reductions in both the diversity (SI Appendix, Fig. S2A) and composition (SI Appendix, Fig. S2 B and C) of bacteria in the SI of antibiotic treated mice. Antibiotics were allowed to clear from the remaining cohorts of mice and either AKK or vehicle (PBS) were introduced by gavage. 16S rRNA sequencing of ileal fecal material demonstrated that AKK successfully colonized the gut of antibiotic treated mice (Fig. 2B). Other cohorts of mice were then mock-irradiated or exposed to TA-XRT at 11.5 Gy and SIs were isolated 5 d later for histology, while remaining mice were followed for 30 d to monitor organismal survival.

Fig. 2.

Fasting and AKK are both required to protect mice and preserve small intestinal architecture from high-dose irradiation. (A) Schema of experimental design. C57BL/6J mice were administered antibiotics (Ampicillin 0.5 g/L and Enrofloxacin 0.25 g/L) in drinking water for 7 d followed by 3 d of antibiotic clearance. AKK (1 × 10⁸ CFU) or vehicle (PBS) were introduced by gavage on day −7 and day −4. On day −1, mice were fed or fasted for 24 h and then exposed to total abdominal radiation (11.5 Gy; D0). Mice were then housed singly with food and on day 5 postradiation a cohort of mice was humanely killed, and ileal samples were collected for analysis. Remaining mice were monitored for survival until day 30 postradiation. Ileal samples were also isolated from mice on D-10 and D0 for 16S rRNA sequencing. (B) Relative abundances of bacteria at the genus level in ileum samples collected on D0. Each column represents an individual mouse. (C) Survival of mice following TA-XRT was monitored daily for 30 d. P = 0.0023 for FED (−AKK) vs. FAST (+AKK), P = 0.8488 for FED (−AKK) vs. FED (+AKK), P = 0.2698 for FED (−AKK) vs. FAST (−AKK), P = 0.026 for FAST (−AKK) vs. FAST (+AKK) [using Log-rank (Mantel–Cox) test]. (D) Representative images of H&E-stained ileum collected on Day 5. (Scale bars, 200 µm). Magnification, 20×. Representative images shown from four mice. (E) Number of crypts per 2 mm of SI (ileum) and their depth were quantitated for each day 5 sample. The average number of crypts per mm of ileum length and the average crypt depth were plotted (n = 10 fields per mouse). ns = nonsignificant; *P < 0.05; ****P < 0.0001 by one-way ANOVA. Error bars are ± SEM. (F) Representative images of ileum stained for Olfm4 (day 5 samples). (Scale bars, 200 µm). Magnification, 20×. Representative images shown for four mice. (G) Percentage of crypt epithelial cells staining positive for Olfm4 is shown. ns, nonsignificant; ****P < 0.0001 by one-way ANOVA. Error bars are ± SEM. 10 fields per mouse from n = 4 mice.

Disruption of the SI microbiota through antibiotic treatment impaired the ability of fasting to protect mice from high dose radiation as only 50% of antibiotic treated fasted mice survived to study endpoint (Fig. 2C). Importantly, reconstitution with AKK fully restored the ability of fasting to provide radioprotection to antibiotic treated mice as 100% of these mice survived to study endpoint. Although AKK levels varied among fasted mice (Fig. 2B), all fasted animals survived irradiation (Fig. 2C) suggesting that a threshold level of colonization is sufficient to confer radioprotection. Interestingly, reconstitution with AKK did not significantly improve the survival of antibiotic treated fed mice (Fig. 2C). Normally 100% of fed mice die within 6 to 10 d after exposure to high dose radiation. Interestingly, this was not observed in fed mice that received broad-spectrum antibiotics prior to radiation exposure, suggesting that antibiotics provide partial protection from the loss of the barrier function which is compromised by high-dose radiation in fed mice. Importantly, the SIs of fasted antibiotic-treated mice reconstituted with AKK exhibited a significant increase in crypt depth (Fig. 2 D and E) as well as significantly more Olfm4+ stem cells (Fig. 2 F and G) relative to antibiotic treated fasted cohorts not reconstituted with AKK. Conversely, these increases in crypt depth and Olfm4+ stem cells were not observed in fed antibiotic-treated mice reconstituted with AKK. Taken together, these results suggest that AKK contributes to fasting-induced radioprotection at both the organismal and tissue levels.

Specific Depletion of AKK from the Gut Microbiota Disrupts Fasting-Induced Radioprotection.

Whereas broad spectrum antibiotics significantly disrupted the gut microbiome (Fig. 2), a 3-wk course of tetracycline has been shown to deplete AKK without specific depletion of other bacterial species (27, 28). Therefore, we treated mice with tetracycline to determine if loss of AKK followed by its reconstitution impacted host responses to high dose radiation (Fig. 3A). A 3-wk course of tetracycline exposure in drinking water was sufficient to deplete AKK from the gut microbiome of mice (SI Appendix, Fig. S3 A and B). 16S rRNA gene sequencing revealed changes in the composition of bacteria in the SI of tetracycline treated mice (SI Appendix, Fig. S3C), but not to the same extent as observed in mice treated with broad spectrum antibiotics (SI Appendix, Fig. S2C). Importantly, tetracycline pretreatment impaired the ability of fasting to radioprotect and reconstitution with AKK was sufficient to fully restore radioprotection to fasted animals (Fig. 3 B and C). By contrast, reconstitution with AKK did not improve survival of similarly treated fed animals. Analysis of small intestinal integrity revealed a significant increase in crypt number, crypt depth, and Olfm4+ stem cells in fasted mice reconstituted with AKK compared with all other groups (Fig. 3 D–G). Taken together, these data suggest an essential role for AKK in fasting-mediated radioprotection.

Fig. 3.

Specific elimination of AKK hampers fasting mediated radiation protection in mice. (A) Schema of experimental design. C57BL/6J mice were administered vehicle or antibiotics (Tetracycline 3 g/L and 10% sucrose) in drinking water for 21 d followed by 3 d of antibiotic clearance. AKK (1 × 10⁸ CFU) or vehicle (PBS) was introduced by gavage on day −7 and day −4. On day −1, mice were fed or fasted for 24 h and then exposed to total abdominal radiation (11.5 Gy; D0). Mice were then singly housed with food and on day 5 postradiation, a cohort of mice was humanely killed, and ileal samples were collected for analysis. Remaining mice were monitored for survival until day 30 postradiation. Ileal samples were also isolated from mice on D-10 and D0 for 16S rRNA sequencing. (B) Relative abundances of bacteria at the genus level in ileum samples collected at D0. Each column represents an individual mouse. (C) Survival of mice following TA-XRT was monitored daily for 30 d. P = 0.0038 for FED (−AKK) vs. FAST (+AKK), P = 0.1036 for FED (−AKK) vs. FED (+AKK), P = 0.4313 for FED (−AKK) vs. FAST (−AKK), P = 0.0226 for FAST (−AKK) vs. FAST (+AKK) [using Log-rank (Mantel–Cox) test]. (D) Representative images of H & E-stained ileum collected on Day 5. (Scale bars, 200 µm). Magnification, 20×. Representative images shown from four mice. (E) Number of crypts per 2 mm of SI (ileum) and their depth was quantitated for each day 5 sample. The average number of crypts per mm of ileum length and the average crypt depth was plotted (n = 10 fields per mouse). ns, nonsignificant; ***P < 0.0003; ****P < 0.0001 by one-way ANOVA. Error bars are ± SEM. (F) Representative images of ileum stained for Olfm4 (day 5 samples). (Scale bars, 200 µm). Magnification, 20×. Representative images shown for four mice. (G) Percentage of crypt epithelial cells staining positive for Olfm4 is shown. ns, nonsignificant, ****P < 0.0001 by one-way ANOVA. Error bars are ± SEM. 10 fields per mouse from n = 4 mice.

AKK Metabolite Promotes Histone H3 Acetylation in Small Intestinal Epithelial Cells.

AKK is a gram-negative anaerobic bacterium that degrades and utilizes mucus produced by intestinal goblet cells (29). Single cell RNA sequencing demonstrated that goblet cells became significantly enriched in the SI of fasted animals relative to their fed cohorts (Fig. 4 A–C and SI Appendix, Fig. S4) and this likely accounts for the enrichment of AKK in the gut microbiome of fasted animals. The mucin degradation activity of AKK leads to the production of short chain fatty acids (SCFAs) including propionate and acetate, which can be used by some non-mucus-degrading bacteria to produce butyrate (30).

Fig. 4.

Fasting and AKK induces epigenetic changes in SI epithelial cells. (A) C57BL/6J mice were allowed to feed freely or were fasted for 24 h. Small intestinal cells were isolated and subjected to single cell RNA-Sequencing. Percentage of Goblet cells, identified by Muc2 expression, was quantified. Total 4,034 cells form fasted and 3,724 cells from fed group. (B) Ileum samples from fed and 24 h fasted mice were stained with Alcian Blue to identify Goblet cells. (Scale bars, 100 µm). Magnification, 20×. Representative images shown from n = 4 mice. (C) The number of Goblet cells in ileum samples from fed and fasted mice was determined by counting Alcian Blue positive cells in 10 fields from four mice. Data are presented as mean ± SEM, analyzed by Student’s t test, unpaired. (D) Murine AKK (MDAJAX AM001) was cultured in brain heart infusion media containing 5 mg/mL of mucin under strict anaerobic conditions. Relative abundance of short-chain fatty acids measured in AKK conditioned- and control-media was determined by mass spectrometry and is presented as relative fold change. (E) Ileum samples were isolated from fed (n = 10) and 24 h fasted (n = 10) mice and absolute concentrations of beta-hydroxybutyrate and propionic acid were determined using ultra-high-resolution IC-MS. Data are presented as mean ± SEM, analyzed by Student’s t test, unpaired. *P = 0.0161 and ****P = < 0.0001. (F) Relative abundance of ileum SCFA concentrations in fed (n = 10) and 24 h fasted mice (n = 10). Data are presented as mean ± SEM, analyzed by Student’s t test, unpaired. *P < 0.02. (G) Representative western blot analysis of H3K9bhb, H3K9ac, and H3K27ac of acid-extracted histones from stem cell–enriched epithelial spheroids generated from small intestinal crypts. Spheroids were treated with (R)-(-)-3-hydroxybutyric acid (R-bHB) for 24 h, or for 3 h with other listed components. 2 mg of histones were loaded per well, and histone H3 was used as the loading control.

Metabolic profiling identified propionic acid as the most significantly enriched metabolite in AKK- conditioned media compared with control media (Fig. 4D). In addition, metabolic profiling of ileal contents from fed and fasted mice revealed elevated levels of propionic acid and beta-hydroxybutyrate (β-OHB) in fasted animals relative to fed controls (Fig. 4E) (31). In addition, pentanoic acid, a microbial fermentation product related to propionic acid, and glyceric acid, an intermediate in carbohydrate and glycerol metabolism, were significantly increased in the ilea of fasted animals (Fig. 4F). The accumulation of these microbial- and host derived- metabolites points to a coordinated fasting-induced reprogramming of gut microbial fermentation and intestinal energy metabolism.

SCFAs regulate diverse cellular functions including fueling SI goblet cells to enhance mucin production and thereby promote AKK accumulation. They also modulate epigenetic and transcriptional programs either directly, through histone acetylation or propionylation (32), or by serving as substrates for butyrate-producing bacteria to inhibit histone deacetylases (HDACs) (33). β-OHB also functions as a chromatin regulator both via lysine β-hydroxybutyrylation (Kbhb) of histones H3 and H4 (31, 34) and by inhibiting class I HDACs (35). Consistent with this, AKK-conditioned medium and its major metabolite propionate induced histone H3 acetylation in stem cell–enriched SI epithelial spheroids cultured in vitro (Fig. 4G). Moreover, Kbhb enrichment was observed in these cultures following treatment with the bioactive β-OHB derivative R-bHB.

AKK Induces Epigenetic Changes in Small Intestinal Epithelial Cells, Contributing to Protection Against Radiation.

We previously reported that fasting alters specific chromatin states of SI epithelial cells in vivo leading to upregulation of specific pathways involved in adaptation to nutrient deprivation (31). To dissect the contribution of AKK-derived metabolites to this process, we mapped active enhancers and promoters by CUT&Tag profiling of the histone acetylation marks H3K27ac and H3K9ac in SI crypt epithelial cells under fasting and AKK-depleted conditions. Mice were divided into fed (Fed) and fasted (Fast) groups and each group was treated with or without tetracycline (−/+ Tetra) to selectively deplete AKK (Fig. 5A). Fasted mice harbored significantly increased average intensity and number of H3K27ac and H3K9ac peaks (Fast −Tetra vs. Fed −Tetra) (Fig. 5 B and C and SI Appendix, Fig. S5 A and B). This effect was significantly reduced when AKK was depleted by tetracycline in the fasted group (Fast +Tetra). However, AKK depletion showed minimal effect on acetylation peak enrichment in fed mice (Fed +Tetra). Overlap analysis identified 37,601 H3K27ac peaks and 88,040 H3K9ac peaks (P value = 10⁻⁵) that were uniquely present under the Fast (−Tetra) condition, indicating an AKK-dependent effect on many regulatory elements. The most pronounced changes were observed in H3K27ac, followed by H3K9ac (Fig. 5C and SI Appendix, Fig. S5B). Together, these data suggest a key role for AKK in regulating fasting-induced histone acetylation patterns in SI crypt epithelial cells.

Fig. 5.

AKK differentially enriches H3K27ac at proximal promoters of DNA repair and cell proliferation program in SI crypts. (A) Schematic of whole-crypt isolation for CUT&Tag and bulk RNA seq analyses. (B) Enrichment plot for H3K27ac peaks in Fed (−Tetra), Fed (+Tetra), Fast (−Tetra), and Fast (+Tetra). (C) Upset plots showing overlap of CUT&Tag seq H3K27ac peaks among Fed (−Tetra), Fed (+Tetra), Fast (−Tetra), and Fast (+Tetra). The CUT&Tag seq peaks unique to Fast (−Tetra) were analyzed separately. (D) Top 10 MSigDB GSEA pathways based on H3K27ac proximal promoter peaks overlapping with publicly available HiChIP and inhouse ChIP seq data in Fast (−Tetra) condition. (E) Genome browser view of CUT&Tag tracks for H3K27ac-enriched regions in SI crypts at the p53 pathway-responsive genes Vdr, Fos, and Nupr1 under Fast (−Tetra) conditions. (F) ChIP-qPCR analysis of H3K27ac enrichment at Vdr, Fos, and Nupr1-specific peaks in Fed (−Tetra), Fed (+Tetra), Fast (−Tetra), and Fast (+Tetra) samples. (G) Motif enrichment analysis for enhancers unique to Fast (−Tetra) group using HOMER.

For gene annotation, we generated enhancer–promoter loop interaction data by integrating publicly available HiChIP datasets (36) with in-house H3K27ac CUT&Tag and ChIP-seq data (31) collected under both Fed and Fast conditions. Functional enrichment analysis of uniquely present H3K27ac and H3K9ac peaks in the Fast (−Tetra) group showed significant associations with pathways related to DNA repair and cell proliferation. Notably, genes involved in the p53 signaling pathway were highly represented (Fig. 5D and SI Appendix, Fig. S5C). These included well-characterized p53 target genes such as Vdr, Fos, and Nupr1, which showed increased H3K27ac and H3K9ac signals in the Fast (−Tetra) group compared to both Fed groups and the AKK depleted Fast (+Tetra) group (Fig. 5E and SI Appendix, Fig. S5D). H3K27ac ChIP–qPCR confirmed increased H3K27ac enrichment at each p53 target gene in the fasted group compared to all other groups, including the Fast (+Tetra) group, supporting a requirement for AKK in the establishment of this chromatin mark under fasted conditions (Fig. 5F).

To better understand the core regulatory programs that drive the radiation protection, we analyzed H3K27ac-marked enhancer regions unique to the Fast (−Tetra) group using motif enrichment analysis. This revealed a significant enrichment of binding sites for the KLF (Klf), HNF (Hnf4a, Hnf1b), and FOXA (Foxa3) TF families (Fig. 5G). Klf and Foxa3 are well-characterized pioneer factors capable of opening compacted chromatin and establishing enhancer activity during gut tube development and specification (37).

Accumulation of Clu+ Olfm4+ Primed Persister Cells in SI Crypts of Fasted Animals.

To investigate how fasting shapes the chromatin landscape and regenerative potential of the intestinal epithelium following injury, we performed single-cell ATAC-seq (scATAC-seq) on crypt epithelial cells isolated from fed or 24 h fasted mice exposed to TA-XRT (11.5 Gy). Samples were collected at baseline (0 h) and at 24, 48, or 72 h postirradiation (hpi) (Fig. 6A). UMAP projections of aggregated scATAC-seq data revealed distinct clustering by sample (Fig. 6B and SI Appendix, Fig. S6A) and by cell type (Fig. 6C and SI Appendix, Fig. S6B), with fasting associated with pronounced shifts in chromatin accessibility. We identified two clusters (C3 and C6) characterized by accessibility at the Clusterin (Clu) locus. Clu expression characterizes revival stem cells important for intestinal regeneration following irradiation (15). Marker gene accessibility analysis distinguished these two clusters. Cluster 3 (C3) displayed high Clu accessibility with no accessibility at Olfm4, whereas Cluster 6 (C6) showed accessibility at both the Clu and Olfm4 loci along with stronger accessibility at the Ascl2 and Hmgcs2 loci (Fig. 6D). Notably, in addition to Clu and Olfm4, Cluster 6 exhibited accessibility at multiple intestinal stem cell marker genes, including Lgr5, HopX, Lrig1, Sox9 (Fig. 6E) and IPA canonical pathway enrichment analysis identified Wnt/b-catenin signaling and stem cell pluripotency-associated pathways as significantly upregulated in C6 cells relative to C3 cells (SI Appendix, Fig. S6C). Analysis of genes associated with open chromatin regions in C6 identified Sox21, Nr5a2, Hnf1a, Fzd2, Axin2, and Ccnd1 revealing strong enrichment of Wnt, Notch, and other pathways critical for intestinal regeneration (SI Appendix, Fig. S6D). Analysis of genes associated with open chromatin regions in C3 identified Tcf4, Arrb2, Ago1, Hao1, and Spr, indicating enrichment of metabolic pathways in addition to modest enrichment of Wnt pathway (SI Appendix, Fig. S6D).

Fig. 6.

Fasting remodels the chromatin landscape of intestinal stem and revival-like cell populations. (A) Experimental design. Mice were fed ad libitum or fasted for 24 h, followed by total abdominal irradiation (11.5 Gy). Small intestinal crypt cells were collected at 0, 24, 48, or 72 hpi and subjected to single-cell ATAC sequencing (scATAC-seq). (B) UMAP projection of aggregated scATAC-seq data colored by sample, showing condition-specific shifts in chromatin accessibility. (C) UMAP colored by cluster identity, identifying major intestinal epithelial populations. (D) Bubble plot showing chromatin accessibility of stem and revival marker loci (Clu, Olfm4) across conditions. Cluster 6 corresponds to the Clu+Olfm4+ primed persister cell population. (E) Bubble plot showing chromatin accessibility (log2FC) of stem cell, revival, and injury-associated gene loci in Cluster 3 (C3) and Cluster 6 (C6). (F) C57BL/6J mice were fed or fasted for 24 h. Ileum tissues were collected at 0 h (no IR). Sections were stained for Lgr5 (white), Olfm4 (yellow), Clu (Red), and Chga (green) by RNAscope. Representative images are shown with individual channels and merged images. Boxed regions show higher magnification of the intestinal crypt. Nuclei are stained with DAPI (blue). (Scale bars, 50 µm). Magnification, 40×. Representative images from n = 2 mice per group. (G) Percentage of Lgr5+, Olfm4+, Clu+, and Clu+Lgr5+ double positive cells in SI crypts from fed and fasted mice at baseline (0 h) and at 24, 48, 72, and 96 hpi. Data are shown as mean ± SEM (n = 2 mice per group). *P < 0.05, **P < 0.01. (H) Bar graph showing the number of shared accessible regions between scATAC-seq and H3K27ac CUT&Tag datasets. “Common” refers to overlapping enhancer-like peaks. (I) Graphical representation of motif enrichment data (SI Appendix, Fig. S9B) for fed and 24 fasted mice. The Y axis represents binned −log10 P-values of TF motif activity. (J) Graphical representation of motif enrichment data (SI Appendix, Fig. S9B) for fed and fasted mice at 0, 24, 48, and 72 hpi. The Y axis represents binned -log10 p-values of TF motif activity.

Clusters 3 and 6 also differed in their temporal dynamics following irradiation (SI Appendix, Fig. S2 E and F). Cluster 6 was enriched in mice fasted for 24 h at baseline and more abundant following irradiation at 24 and 48 hpi before converging toward similar proportions by 72 hpi. In contrast, Cluster 3 represented a smaller and relatively stable population over time with only modest changes at the 24 and 48 hpi time points. By 72 h, cluster 3 declined to 1.2% in fasted mice but remained at 3.1% in fed mice. Collectively, these data demonstrate that Clusters 3 and 6 represent biologically distinct epithelial cell populations with divergent chromatin states and regenerative dynamics with preferentially expansion of Cluster 6 in response to fasting. ScRNA sequencing supported the conclusion that fasting enriches for stem cell and revival stem cell markers (SI Appendix, Fig. S6G).

To orthogonally validate these findings at the transcriptional level, we performed RNAscope analysis of Lgr5, Olfm4, and Clu expression (Fig. 6 F and G; baseline shown in the main figure, with postirradiation time points presented in SI Appendix, Fig. S7). We also included Chga as a marker of differentiated epithelial cells to serve as an internal control. In contrast to the marked depletion of Lgr5+ and Olfm4+ cells following irradiation, Chga+ cells were relatively preserved, consistent with the known radioresistance of differentiated epithelial lineages (38). These findings indicate that the loss of Lgr5+ and Olfm4+ expression reflects the selective radiation sensitivity of crypt stem and progenitor populations, rather than a general failure of epithelial RNA detection or compromised tissue quality.

At baseline (prior to irradiation), both Lgr5+ and Olfm4+ cells were present in SI crypts of both fed and fasted mice, while Clu+ cells were absent in the crypts of either group. Notably, Clu+ cells were detected at the villus tips in fasted but not fed mice. A previous study demonstrated that acute villus injury induces a population of Clu+ villus epithelial cells that contributes to resealing damaged villi and restoring epithelial barrier function (39).

By 24 hpi, Clu+ cells were no longer detectable at villus tips in fasted animals, and both Lgr5+ and Olfm4+ stem cells were markedly depleted from SI crypts in both fed and fasted mice; Clu+ cells were also absent from the crypts in both groups. By 48 hpi, Lgr5+ and Olfm4+ stem cells were further depleted in SI crypts of both groups, whereas Clu+ cells began to emerge in a subset of crypts in both fed and fasted mice, with a higher percentage observed in fasted animals (7% vs. 2% in fed mice). At 72 hpi, the number of Clu+ cells increased substantially in the crypts of both groups, again with a greater abundance in fasted mice. At this time point, only a small fraction of Clu+ cells coexpressed Lgr5+. By 96 hpi, Lgr5+ and Olfm4+ began to accumulate in regenerating crypts, and Clu+Lgr5+ double-positive cells were detected in the SI crypts of both groups, with significantly greater abundance in fasted animals (60% vs. 43.7% in fed mice). Collectively, these findings demonstrate that fasting alters chromatin accessibility at key stemness- and regeneration-associated loci in SI crypt epithelial cells without immediately inducing corresponding gene expression. Notably, the Clu promoter became accessible in fasted animals prior to radiation, yet Clu expression was only observed following irradiation-induced injury.

Next, to assess fasting and AKK associated transcriptional changes in SI crypts, we identified differentially expressed genes (DEGs; P ≤ 0.05 and log2 FC ≥2 or ≤–2) by bulk RNA-seq. Analysis was conducted on fed and 24 h fasted mice treated or not with tetracycline. DEGs exhibiting enrichment of H3K27ac or H3K9ac peaks within 5 kb upstream or 1 kb downstream of their transcription start sites were significantly upregulated under fasting conditions (SI Appendix, Fig. S8 A and B). Notably, Clu was selectively expressed in the crypts of fasted animals, and its expression was markedly reduced in tetracycline-treated fasted animals indicating that AKK contributes to Clu induction during fasting.

Trajectory (SI Appendix, Fig. S8C) and marker gene accessibility (Fig. 6 D and E) analyses distinguished cluster 3 from canonical revival stem cells. revSCs arise in the intestine only following irradiation and are characterized by expression of Clu and Ly6a (15, 40). In contrast, cluster 3 cells are present prior to irradiation and do not exhibit significant Ly6a chromatin accessibility. Cluster 3 may represent a transitional revSC-like state, rather than a fully defined revSC population. Fasting primarily expanded the Clu+Olfm4+ population (SI Appendix, Fig. S6E) which is consistent with an injury responsive cell pool that supports regeneration.

To gain insight into the regulatory programs that drive persistence of the Clu+Olfm4+ primed persister cell population (C6) upon radiation, we integrated scATAC-seq data with H3K27ac CUT&Tag profiles, which identified 4,042, 3,093, 5,157, and 2,149 common open active enhancer peaks in 0, 24, 48, and 72 hpi fasted samples, compared to 1,241, 2,100, 1,268, and 421 common peaks in the corresponding fed samples (Fig. 6H and SI Appendix, Fig. S9A).

To identify functional regulators, we filtered enriched TF motifs based on known roles in gut progenitor identity, injury response (IR) protection, and expression in bulk RNA-seq and highlighted candidate TFs potentially driving the fasting-induced state (SI Appendix, Fig. S9B). This analysis revealed a marked reprogramming of regulatory networks following 24 h of fasting (Fig. 6I). Fasting strongly increased chromatin accessibility surrounding motifs for TFs associated with epithelial cell differentiation and metabolic regulation, most notably Hnf4a, which shifted to the highest activity bin. Motifs for Gata factors (Gata4/6), Klf4, and the intestinal lineage regulators Cdx2 and Foxa3 were also enriched. In addition, motifs for the architectural proteins Ctcf and Boris increased, suggesting fasting-induced remodeling of chromatin organization. In contrast, accessibility of motifs associated with Atf3, Fos, and JunB declined to the lowest bin in fasted mice, likely due to nutrient deprivation.

Comparison of TF motif accessibility dynamics following irradiation revealed fundamental differences between fed and fasted states (Fig. 6J). Fasted animals displayed specific reprogramming of regulatory programs with persistent accessibility of Gata factors, Klf4, Cdx2. Foxa3, Hnf4a up to 48 hpi and sustained Ctcf/Boris accessibility through 72 hpi. Notably, motifs associated with Atf3, Fos, and JunB declined to the lowest accessibility bin after 24 h of fasting (SI Appendix, Fig. S9B) and robustly rebound when animals were returned to food immediately after irradiation (0 hpi) and resumed feeding (Fig. 6J). A pronounced increase in AP-1 family motif chromatin accessibility was observed in irradiated fasted animals (Fig. 6J) which likely reflects a combined effect of refeeding and radiation-induced signaling. In contrast, in fed animals, motif accessibility either persisted (Elf3, Ctcf) or transiently increased by 24 hpi, but most motifs lost accessibility by 48 hpi, and all motifs lost accessibility by 72 hpi, potentially due to excessive cell death.

Single-cell correlation data showed that Cluster 6 cells with high Clu and Olfm4 accessibility also exhibited elevated chromatin accessibility at regulatory regions associated with Gata4, Gata6, Sp1, Klf4, Foxa3, and Elf4 TFs (SI Appendix, Fig. S9C). These results indicate that the fasting-induced transcriptional reprogramming likely occurs in Clu+Olfm4+ primed persister cells, represented by Cluster 6.

Collectively, these data indicate that fasting reprograms transcriptional regulatory programs critical for intestinal stem cell persistence, metabolic adaptation, and regenerative competence following irradiation.

Discussion

Here we report that AKK becomes enriched in the microbiome of fasted mice leading to chromatin remodeling in small intestinal crypt cells. This occurs in a programmatic fashion leading to the reprogramming of key regulatory TF networks to establish a transcriptional output in primed persister cells that promote regeneration of the SI following high dose radiation.

Our functional studies demonstrated that AKK recolonization is sufficient to confer radioprotection in antibiotic-treated fasted mice. Notably, the radioprotection conferred by AKK was restricted to the fasting context, as reintroduction of AKK into antibiotic-treated fed mice did not confer protection. These results demonstrate that AKK is necessary but not sufficient to confer radioprotection. Beyond reshaping the microbiome, fasting induces β-OHB which contributes to the histone modifications and associated transcriptional reprogramming in small intestinal crypt cells (31). These changes, together with other fasting-induced adaptations, are absent in fed mice and likely explain their lack of radioprotection by the addition of AKK alone.

We also found that fasting enriched members of the Furmicutes, including Erysipelatoclostridium, Clostridium sensu stricto 1, Coprococcus, and Marvinbryantia, with all but Erysipelatoclostridium known to produce butyrate (but not propionate) (41, 42). Unlike AKK, these taxa are not consistently enriched across fasting studies (24). While it remains possible that these additional taxa contribute complementary metabolic or signaling functions, our data support AKK as the principal microbial mediator of the fasting phenotype. Future studies may clarify if these taxa interact with AKK or influence epithelial responses to injury. Interestingly, a previous study showed that Enterococcus and Lachnospiraceae can provide protection against whole-body radiation even in the absence of fasting, with tryptophan identified as a key metabolite mediating this radioprotective effect (41). In our mouse cohort, Enterococcus was not detected and although Lachnospiraceae was present at low levels, its abundance did not change in response to fasting.

The regenerative response of the SI to radiation injury has been defined by the acute loss of Lgr5+ stem cells, followed by activation of reserve, revival stem and progenitor cells that transiently sustain renewal. Prior studies have implicated p53, YAP/TAZ, ASCL2 activities and TGFb, Notch, and Hedgehog pathways in this revival function (15, 40, 43). These canonical pathways collectively restore crypt integrity, but they are only engaged after injury.

Our findings indicate that fasting does not replace established regenerative programs but instead primes the intestinal epithelium before injury through epigenetic reprogramming, thereby enhancing the regenerative response once DNA damage occurs. This fasting-induced epigenetic priming does not alter intestinal architecture on its own. In our prior studies villus height, crypt depth, and crypt numbers were shown to be indistinguishable in the SIs of fed and fasted mice (4, 5). Architectural differences emerged only after irradiation, with fasting leading to better preservation of crypt structure, consistent with radioprotection.

Fasting increased goblet cell numbers which supplies mucin, thereby enriching AKK in the gut microbiota. AKK produces propionate, while host metabolism contributes β-hydroxybutyrate (β-OHB). Together, these metabolites function as chromatin cofactors, inducing H3K27ac, H3K9ac, and β-hydroxybutyrylation (31), thereby remodeling enhancer–promoter interactions. This remodeling selectively increases accessibility at loci bound by pioneer TFs including Foxa, Gata, and Klf families (44, 45). Foxa and Gata are classical pioneer factors that bind condensed chromatin, displace nucleosomes, and establish enhancer competence in endodermal development (46). Their enrichment in SI crypts of fasted animals suggests that fasting re-engages developmental pioneer programs to establish regenerative competence prior to injury.

Clu+ revival cells are a rare, injury-induced population that emerge after intestinal injury, marked by clusterin (Clu) expression, activation of a fetal-like gene program including Ly6a, and activation of Yap/Hippo and IL6/STAT3 pathways. They serve as a transient regenerative reserve, replenishing lost Lgr5+ stem cells and contributing to epithelial repair (15, 40). The fasting-induced Clu+ cell populations identified in our study (Clusters 3 and 6) are distinct from canonical Clu+revSCs. Unlike revSCs, which emerge only after injury and express Ly6a, the Clu-high (Cluster 3) and Olfm4+Clu+ (Cluster 6) cells we describe are present in fasted animals prior to irradiation and exhibit low (Cluster 3) or absent (Cluster 6) Ly6a chromatin accessibility.

Based on marker gene accessibility-, and trajectory-analyses, Clu-high, (Cluster 3) cells may represent a transitional, revSC-like population. This population is enriched by fasting but does not expand during the injury responsive period. Clu+Olfm4+ (Cluster 6) cells are also enriched by fasting, but they expand during the early injury response period and exhibit enrichment of several SI stem cell markers. These properties are consistent with an injury responsive primed persister cell pool that supports regeneration.

These findings are supported by RNAscope analysis confirming that Clu is not expressed in crypts of fasted animals at baseline but is induced following irradiation. Clu+ cells emerge at 48 to 72 hpi and subsequently give rise to Clu+Lgr5+ double-positive cells during regeneration. This temporal pattern suggests that Cluster 6 cells do not represent preexisting Lgr5+ stem cells, but rather a regeneration-associated cell state that is primed by fasting and activated in response to injury.

Epigenetic profiling revealed a rewired core regulatory program in fasted mice that primes injury resilience, characterized by pioneer TFs (Foxa/Gata/Klf) that preemptively license chromatin accessibility; lineage-defining and metabolic regulators (Cdx2, Hnf4) that coordinate metabolic programs required for stem cell survival and regenerative capacity and architectural organizers (Ctcf, Boris) that stabilize higher-order chromatin structure.

Our data support a hierarchical model in which pioneer TFs first establish chromatin accessibility, enabling a subsequent binding of settler TFs (Cdx2 and Hnf4) at their cognate regulatory elements. Both the enhancer and promoter of Clu harbor conserved Hnf and Cdx binding motifs, directly linking fasting-induced chromatin remodeling to the regulation of Clu+ stem cell states (47). Architectural organizers (Ctcf, Boris) then stabilize these higher order, fasting-induced chromatin configuration, thereby priming Clu+ primed persister cells for a rapid response to irradiation-induced injury. In parallel, AP1-components c-Jun and Fos support proliferative programs during the regenerative phase.

Overall, we propose a model in which fasting-induced nutrient deprivation elicits a stress response in the SI, characterized by reduced epithelial cell proliferation and accumulation of host- and microbiome-derived metabolites. Intestinal crypt cells respond to this metabolic environment by remodeling chromatin accessibility to promote cellular plasticity. Under these conditions, C6 cells accumulate and become epigenetically “primed” for multiple differentiation trajectories. We hypothesize that this enhanced plasticity enables rapid and effective regeneration of the intestinal epithelium following irradiation-induced injury (Fig. 7).

Fig. 7.

Model summarizing findings.

Materials and Methods

Study Approval.

This study was carried out in accordance with the guidelines outlined in the Guide for the Care and Use of Laboratory Animals from the NIH. Ethical considerations regarding animal care and use were approved by the MD Anderson Institutional Animal Care and Use Committee under protocol number 00001101-RN04. Mice were humanely killed according to NIH and Association for Assessment and Accreditation of Laboratory Animal Care International guidelines by exposing them to carbon dioxide, followed by cervical dislocation.

Animals and Feeding.

C57BL/6J male mice purchased from Jackson Laboratories (JAX, 000664) were maintained at 72 °F ± 2 °F on a 12-h light/dark cycle. Male mice aged between 4 to 9 wk were randomly assigned to either fed or fasted groups. Fed mice had unrestricted access to both food (PicoLab 5053, #0007688) and water for 24 h, while fasted mice were deprived of food for 24 h but had unrestricted access to water. Mice were singly housed on aspen bedding for the duration of the experiments. The number of animals used is provided in figure legends.

Total Abdominal Radiation.

Mice were subjected to a single dose of 11.5 Gray (Gy) of total abdominal radiation therapy (TA-XRT). The radiation treatment was carried out using the X-RAD 225 Cx irradiator, a device designed for delivering controlled and targeted radiation therapy.

16S rRNA Sequencing.

Small intestinal ilea samples were collected from mice and weighed prior to DNA extraction. Synthetic spike-in was done as described (48). Detailed methods can be found in SI Appendix.

Culturing AKK.

AKK (MDAJAX AM001) was cultured in brain heart infusion agar containing 5 mg/mL of mucin under strict anaerobic conditions. Single colonies were then isolated and cultured in brain heart infusion broth for 48 to 72 h at 37 °C, under strict anaerobic conditions. AKK was diluted to a final concentration of 1 × 10⁸ CFU in 0.2 mL of anaerobic phosphate-buffered saline (PBS) for all mouse experiments.

PCR Amplification of AKK Specific Gene from Fecal Material.

Fecal material collected from small intestinal ilea was resuspended in 100 µL of nuclease-free water and incubated at 96 °C with shaking for 7 min to lyse bacterial cells and inactivate nucleases. Samples were then centrifuged at 12,000 rpm for 5 min at 4 °C and the supernatant isolated. Primers specific for genes encoding Amuc1483F (5′ GGCGGAGTCATGGTGTATATC 3′) and Amuc1483R (5′ CAGACCGGAGAGAAAGGAATAAA 3′) were used in a PCR reaction using the Q5^® High-Fidelity DNA Polymerase. An amplicon of 459 bp was visualized on a 1.5% agarose gel.

Broad-Spectrum Antibiotic Treatment.

C57BL/6J mice were administered antibiotics (Ampicillin 0.5 g/L and Enrofloxacin 0.25 g/L) in drinking water for 7 d followed by 3 d of antibiotic clearance. AKK (1 × 10⁸ CFU) or vehicle (PBS) were introduced by gavage. Gavaging was repeated 2 d later.

Tetracycline Treatment.

C57BL/6J mice were administered vehicle or antibiotics (Tetracycline 3 g/L and 10% sucrose) in drinking water for 21 d followed by 3 d of antibiotic clearance. AKK (1 × 10⁸ CFU) or vehicle (PBS) were introduced by oral gavage. Gavaging was repeated 2 d later.

Analysis of Short Chain Fatty Acids by Ultra-High-Resolution IC-MS.

Samples collected from small intestinal ilea or AKK conditioned media were obtained as described above. Approximately 50 mg of ilea or 2 mL conditioned media were snap frozen in liquid nitrogen, then homogenized with Precellys Tissue Homogenizer. Metabolites were extracted using ice-cold 0.1% Ammonium Hydroxide in methanol:water = 80:20 (v/v). Samples were centrifuged at 17,000×g for 5 min at 4 °C, and supernatants were transferred to clean tubes, followed by evaporation to dryness under nitrogen. Dried extracts were reconstituted in deionized water, and 10 μL were injected for analysis by ion chromatography (IC)-MS using the Thermo Scientific (Dionex ICS-6000+) system which includes a Thermo IonPac AS11 column (4 µm particle size, 250 × 2 mm) with a column compartment kept at 35 °C and an autosampler tray chilled to 4 °C. The IC mobile phase A (MPA; weak) was water, and the mobile phase B (MPB; strong) was water containing 100 mM KOH. A The mobile phase flow rate was 360 µL/min, and the gradient elution program was 0 to 2 min, 1% MPB; 2 to 25 min, 1 to 40% MPB; 25 to 39 min, 40 to 100% MPB; 39 to 50 min, 100% MPB; 50 to 50.5 min, 100 to 1% MPB. The total run time was 55 min. To assist with desolvation for better sensitivity, methanol was delivered by an external pump and combined with the eluent via a low dead volume mixing tee. Data were acquired using a Thermo Orbitrap IQ-X Tribrid Mass Spectrometer under ESI negative ionization mode. Raw data files were imported into Thermo Trace Finder 5.1 software for final analysis. The relative concentration of each compound was normalized to stool weight per sample.

Isolation of Small Intestinal Crypts.

Isolation of small intestinal crypts was conducted following established protocols as previously published (4) with minor adjustments. Detailed methods can be found in SI Appendix.

Generation of Stem Cell–Enriched Epithelial Spheroid Cultures.

Purified small intestinal crypts were mixed with 30 μL of matrigel and plated in a 24-well tissue culture dish following the procedure described (49). Detailed methods can be found in SI Appendix.

Immunohistochemistry.

Mice small intestinal samples were harvested as described previously (4) and ilea sections were used for all analyses. Detailed methods can be found in SI Appendix.

Histone Extraction and Analysis.

Histones were extracted and purified following methods outlined in ref. 50. Detailed methods can be found in SI Appendix.

RNA Isolation and Quantification from Stem Cell Enriched Spheroid Culture.

Spheroid cultures were lysed in RNA lysis buffer (PureLink™ RNA Mini Kit, ThermoFisher Scientific # 12183018A) and RNA isolated according to the manufacturer’s instructions. Total RNA was treated with DNase followed by column purification (RNA Clean & Concentrator-5, Zymo Research # R1013). Total RNA was quantified using a nanodrop spectrophotometer. The undiluted total RNA was then converted into cDNA using the SuperScript™ III First-Strand Synthesis SuperMix (ThermoFisher Scientific # 18080400). qPCR was performed on undiluted cDNA in duplicate for each primer and probe set using the TaqMan™ Fast Advanced Master Mix for qPCR (ThermoFisher Scientific # 4444557). qPCR data were normalized to an endogenous control, Gapdh. No-template controls were included for each probe set, and no amplification was observed for any samples. The qPCR analyses were conducted on the Applied Biosystems QuantStudio 6 Pro real-time PCR systems, and all assays were performed in biological triplicate to ensure the reliability of the results.

Bulk and Single-Cell RNA Sample Preparation and Analysis.

The detailed information for bulk and single cell RNA sample preparation and analysis can be found in SI Appendix.

CUT&Tag and Analysis.

CUT&Tag assays were conducted using the CUTANA CUT&Tag kit (EpiCypher) according to the manufacturer’s instructions. Detailed methods can be found in SI Appendix.

Single Cell ATAC Sample Preparation.

Nuclei were isolated from ~1 × 10⁶ cells using a modified 10× Genomics “Nuclei Isolation for ATAC” protocol (CG000169), lysed in chilled lysis buffer, washed, and resuspended in 1× nuclei buffer targeting 10,000 nuclei. Quality and concentration were assessed via trypan blue staining and microscopy. scATAC-seq libraries were prepared using the Chromium Next GEM Single Cell Multiome ATAC + Gene Expression Reagent Kits (CG000338) per the manufacturer’s instructions, including nuclei transposition, GEM generation/barcoding, post-GEM cleanup, and sample index PCR, followed by double-sided SPRIselect cleanup. Libraries were quality-checked (Agilent 4200 TapeStation HS D1000) and quantified (Qubit), then pooled and sequenced on a NovaSeq6000 S1 (Illumina) at ~25,000 read pairs/nuclei using 10x Genomics-recommended parameters (Read N1: 50 cycles; i7 Index: 8 cycles; i5 Index: 24 cycles; Read 2: 91 cycles).

Single Cell ATAC Sequencing Data Analysis.

Single cell multiome fastq files were aligned to the mouse (mm10) genome, cell barcodes were demultiplexed, and UMIs corresponding to genes were counted using the cellranger-arc (v2.0.2) count command using default parameters. Single-cell ATAC-seq data were processed using ArchR (v1.0.2) pipeline (51) in R (v4.2.2) with the mm10 genome. Fragment files from Cell Ranger ARC outputs were used to generate Arrow files (minTSS = 4, minFrags = 1,000). Predicted doublets were identified (addDoubletScores()) and removed (filterDoublets()). Dimensionality reduction was performed via iterative latent semantic indexing (varFeatures = 25,000, dimsToUse = 30, sampleCells = 5,000), followed by Louvain clustering (resolution = 0.2) and UMAP embedding with default parameter. Marker genes and peaks were identified using Wilcoxon rank-sum tests (FDR ≤ 0.01, log2FC ≥ 1.25) and visualized in heatmaps and UMAP projections. Peaks were called on pseudobulk replicates grouped by cluster using MACS2 (v2.2.9.1). We used the ArchR pipeline for most of our single-cell ATAC-seq datasets analysis. Integration of scATAC-seq and H3K27ac CUT&Tag peaks was performed using Bedtools (52) and homer for motif analysis.

Chromatin Immunoprecipitation (ChIP)-qPCR.

Small intestinal crypts were isolated from mice and cross-linked with 1% formaldehyde for 10 min at 37 °C, followed by quenching with 125 mM glycine for 5 min at 37 °C. Crypts were then collected, lysed, and processed as previously described (31). Immunoprecipitated DNA was purified, dissolved in water, and analyzed by qPCR using primers targeting the FAST (-Tetra) H3K27ac specific peaks.

Vdr (For) 5′-TGTGGAGAGTCTGCCAGGAT-3′, Vdr (Rev) 5′- TCTTGTTCTTCTGCCCACCC-3′,

Fos (For) 5′-CATGAACCTGTTCGTGCCAG-3′ Fos (Rev) 5′-GCTGTCTCGTTGAACTGTTGT-3′,

Nupr1 (For) 5′-ATGTTGTCCAGGTTGGCCTT-3′ Nupr1 (Rev) 5′-CCTGGCAGAGAGCAAGAGAG-3′,

GeneDesert (For) 5′-ACCAAGCACAGAAAAGGTTCAAAC-3′, GeneDesert (Rev) 5′-TCCAGATGCTGAGAGAAAAACAAC-3′

RNA In Situ Hybridization.

RNA in situ hybridization (ISH) was performed using RNAscope technology (Advanced Cell Diagnostics) to detect Lgr5, Olfm4, Clu, and Chga transcripts. Fluorescent detection was carried out using the RNAscope Multiplex Fluorescent V2 Assay. Signals were visualized with Opal Fluorophore reagent packs (Opal 520, 570, 690, and 780, Akoya Biosciences). Images were acquired using a Vectra Polaris platform (PerkinElmer). Positive cells were counted using QuPath (4) and statistically analyzed using GraphPad Prism 10.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All data have been deposited with NCBI GEO under Accession Nos. GEO: GSE306576 (53) (BulkRNA-seq), GEO: GSE306577 (54) (CUT&Tag), GEO: GSE306671 (55) (scRNA-seq), and GEO: GSE306673 (56) (scATAC-seq) are publicly available as of the date of publication. The code used in this research can be accessed via the following links: https://github.com/ajaykumarsaw/Fasting-primes-small-intestinal-regeneration-after-damage-via-a-microbiome-metabolite-chromatinaxis (57).