Inflammatory memory restrains intestinal stem cell regeneration

Intestinal stem cells (ISC) encounter inflammatory insults in immune mediated gastro-intestinal (GI) diseases. It remains unknown whether, and how, they adapt, and if the adaptation leaves scars on the ISCs that affects their subsequent regeneration capacity. We investigated the consequences of inflammation on Lgr5+ISCs in well-defined clinically relevant models of gastro-intestinal acute graft-versus-host disease (GI GVHD). Utilizing single cell transcriptomics, organoid, metabolic, epigenomic and in vivo models we found that Lgr5+ISCs undergo metabolic changes that lead to accumulation of succinate, which reprograms its epigenome. These changes reduced the ability of ISCs to differentiate and regenerate ex vivo in serial organoid cultures demonstrating the persistence of the maladaptive impact of an in vivo inflammatory encounter by the ISCs. Thus, inflammation from GI GVHD leaves a memory of its effects on ISCs that persist and are likely to affect their sensitivity to adapt to future stress or challenges.


Introduction
The memory paradigm of non-immune cells, particularly in adult tissue speci c stem cells that remember their in ammatory encounters and its impact on their tness and functions has only been recently appreciated (1)(2)(3)(4). However, the molecular mechanisms underpinning this phenomenon, and whether such memories are germane to intestinal stem cells (ISCs) remain unknown. ISCs renew the different cell types of the intestinal epithelial barrier (5). ISCs undergo a multitude of choregraphed changes, demonstrate plasticity and integrate environmental cues to maintain normal intestinal homeostasis (6, 7). Non-infectious causes of intestinal in ammation and injury, speci cally gastrointestinal graft-versushost disease (GI GVHD) target ISCs resulting in their loss and thus cause severe morbidity and mortality from allogeneic hematopoietic stem cell transplantation (Allo-HSCT) (8, 9). GI GVHD is the major cause of mortality and morbidity from Allo-HSCT, a potentially curative therapy for many hematological diseases. The biology of GVHD is complex. GVHD is caused by in ammation that is induced by cytopathic alloreactive T cells that target host epithelial cells in GI tract, skin and liver. This in ammation is ampli ed or molli ed by complex microenvironment characterized by several other host cells and its microbiome (8, 10,11). The host ISCs are the chief targets of the in ammation that causes GI GVHD (12)(13)(14)(15). GI GVHD is characterized by reduction in ISCs. However, it remains unknown whether the ISCs that survive or tolerate in ammation from GVHD are fully functional, or can return to their full functionality after the resolution of GVHD, with fundamental implications for host resilience and repair (16). Cellular metabolism plays a central role in regulation of ISC function and regeneration (17)(18)(19)(20)(21)(22)(23). Emerging data have demonstrated a role for in ammation in causing metabolic adaptations in various immune and non-immune cell targets (24)(25)(26)(27)(28)(29)(30)(31)(32). However, whether ISCs show altered metabolism in the context of in ammation induced stress remains unclear. Metabolism has been shown to regulate epigenetic reprogramming (33,34). Speci cally, metabolites derived from oxidative phosphorylation (OXPHOS) such as alpha-ketoglutarate (alpha-KG) and succinate have been shown to regulate the epigenome (35)(36)(37). ISCs undergo epigenetic reprogramming that regulate regeneration and function (Saxena M, Ann Rev Physio 2021). However, whether in ammation regulates OXPHOS and its metabolites in ISCs and if this metabolic adaptation mediates epigenetic reprogramming and function of the ISCs remain unexplored. The concept of epigenetic memory of in ammatory exposure by non-immune cells has been shown for tumor cells (38), and in non-hematopoietic stem cells such as the skin and nasal epithelial stem cells (1)(2)(3)(4). It is unknown whether ISCs can retain an epigenetic memory of an in ammatory encounter and whether that modi es their function and sensitivity to future in ammatory encounters.
We investigated whether in ammation induces qualitative metabolic changes in the ISCs and the consequence of such an impact in vivo. Utilizing single cell transcriptomics, metabolic analyses, organoid cultures, epigenomics and in vivo model systems of GI GVHD we found that in ammation induced metabolic change, reduction in OXPHOS with an increase in succinate leads to epigenetic reprograming of the ISCs, which is retained by the ISCs even in absence of active in ammation. These data thus demonstrate evidence for memory of an in ammatory encounter on by ISCs, provide a molecular mechanism and illustrate a functional consequence on the potential for ISC regeneration.

Results
Single cell transcriptomic analysis of Lgr5 + ISCs following intestinal tract of GVHD In ammation-mediated ISC loss is well-documented, however there is little data on functional changes of ISCs that survive the in ammatory insult from GI GVHD. Therefore, to characterize the impact of in ammation on ISCs, we examined the transcriptomes of intestinal crypt epithelial cells using single cell RNA sequencing (scRNA-seq) in the well-established lethally irradiated major histocompatibility complex (MHC)-disparate BALB/c into C57BL/6 (B6) model of GVHD ( g. S1A). The single cells were isolated from the intestines of allogeneic (Allo) (BALB/c→ C57BL/6) and syngeneic (Syn) (C57BL/6→C57BL/6) transplant 7 days after bone marrow transplantation (BMT), and then barcoded with the 10x Chromium technique for scRNA-seq ( g. S1A).
scRNA-seq data was analyzed in R using the Seurat package and integrated using the Seurat canonical correlation work ow (39,40). All cells in the integrated object were plotted with the Uniform Manifold Approximation and Projection (UMAP) (Fig. 1, A and B). Eight clusters in Allo-and Syn-recipients were identi ed and annotated based on published gene expression data sets (41,42), that include the populations of crypt-base columnar cells (CBCs), transit amplifying cells (TA), Paneth cells (Paneth), goblet cells (Goblet), enteroendocrine cells (EE), tuft cells (Tuft), absorb enterocytes (Enterocytes), and T lymphocytes (T) ( Fig. 1A and g. S1B).
To analyze the transcriptomes' pro le of speci cally ISCs ( g. S1, B to I), we focused on Lgr5 + ISCs ( Fig. 1C). We used a cut-off of 0.5 and con rmed the Lgr5 + ISCs population to be within CBC-ISCs by expression of lgr5 ( Fig. 1D and g. S1J). A total of 1128 Lgr5 + ISCs across the conditions (760 syngeneic condition and 368 allogeneic condition) were identi ed ( Fig. 1D and g. S1J), and the top differential genes between the Syn-and Allo-recipients were clustered (Fig. 1E). Amongst the highest differentially expressed genes, were the increase in the expression of genes such as ido1, iigp1, cxcl9, and gbp2 (Fig. 1E, and G to H). By contrast, the most differentially expressed genes that were reduced, were the genes such as fth1, chchd2, slc25a5, gpx4, and kif5b ( Fig. 1, E, I, and J), that are involved in mitochondrial function and cellular metabolic pathways (43)(44)(45)(46)(47). 1071 downregulated genes were involved in cellular bioenergetics, including ATP metabolic process, generation of precursor metabolites and energy, oxidative phosphorylation (OXPHOS), cytoplasmic translation, and mitochondrial organization ( Fig. 1, F, and K to M). Thus Lgr5 + ISCs that survived in Allo-GVHD demonstrate differential genes expression in scRNA-seq analysis.
Next, to further con rm these results, we utilized Lgr5-eGFP-IRES-CreERT2 B6 (B6 GFP-Lgr5) animals as BMT recipients and sorted GFP-Lgr5 + ISCs (Lgr5 + ISCs) from Allo-or Syn-recipients and validated the differentially expressed genes by qPCR ( g. S1, K to Q). Thus, single cell transcriptomic analysis indicates that Lgr5 + ISCs that survived in Allo-BMT change their metabolic process, speci cally in OXPOS pathway as a response to the in ammation from GVHD.

In ammation causes metabolic adaption in ISCs
To assess the functional impact of the changes observed in the scRNA analyses, we next utilized the ex vivo intestinal organoids culture technique to recapitulate the in vivo functional properties of ISCs (48).
Because scRNA analyses suggested metabolic alterations in the Lgr5 + ISCs, to address whether the transcriptomic changes indeed altered the metabolic function of ISCs in context of GI GVHD, we analyzed their bioenergetic pro les by Seahorse XF in organoid cultures (27,49) (Fig. 2A). Seven days after Syn (C57BL/6→C57BL/6) and Allo (BALB/c→C57BL/6)-BMT, intestine crypts were isolated from allogeneic or syngeneic recipients and cultured in Seahorse 96-well plate with 25-50 crypts per well for ve days.
Organoid growths were monitored daily and analyzed on day 4 for growth, size and budding. Consistent with previous reports, the numbers of organoids and their buds were signi cantly decreased in Allorecipients when compared to that of Syn-recipients ( g. S, A and B), indicating the downregulation of ISCs regeneration capability-mediated by in ammation (15,50). On day 5 after the initial culture, organoids were sequentially exposed to oligomycin, Carbonyl cyanide-4 (tri uoromethoxy) phenylhydrazone (FCCP) and rotenone, and antimycin A for analyzing oxygen consumption rates (OCRs) via Seahorse platform. Signi cantly lower levels of OCR and OCR/extracellular acidi cation rate (ECAR) ratio were observed in organoids derived from Allo-recipients when compared with Syn-recipients ( Fig. 2C and g. S2C).
Next, to con rm whether this observation was generalizable and not limited to a single mouse strain or a model, we generated intestinal organoids from irradiated minor antigen mismatched C3H.SW→B6 model of GVHD and observed once again a signi cant decrease in OCR and OCR/ECAR ratio in GVHD group when compared with the syngeneic controls (Fig. 2D).
It is formally possible that the reduction in organoids and subsequent differences in OCR might be a result of reduced numbers of Lgr5 + ISCs in isolated crypts from Allo-recipients. Therefore, to de nitively determine that the reduction in organoids and OCR is because of a qualitive change in Lgr5 + ISCs (and not a re ection of their quantitative loss in the crypts), we directly tested organoid reconstitution and its bioenergetics pro les from sorted Lgr5 + ISCs in the BALB/c→B6-GFP-Lgr5 model of GVHD. Brie y, seven days after BMT, single cells from small intestine crypts were isolated from Allo-or Syn-recipients and GFP-Lgr5 + ISCs were sorted by FACS and cultured in Seahorse 96-well plate with 1000 cells per well for 9-10 days and then analyzed by Seahorse. The ISCs-organoid reconstitution was signi cantly decreased with Lgr5 + ISCs from Allo-recipients compared with Syn-recipients ( Fig. 2E and g. S2D). Moreover, the level of OCR and OCR/ECAR ratio was also signi cantly reduced in the organoids reconstituted with Lgr5 + ISCs from Allo-recipients than that in Syn-recipients (Fig. 2F). Taken together, these ndings demonstrate that Lgr5 + ISCs isolated from allogeneic environment are qualitatively distinct and demonstrate poor regeneration and OXPHOS ex vivo even after they have been removed from in ammatory milieu.

OXPHOS in ISCs regulates severity of GI GVHD
We next explored whether the reduced OCR in Lgr5 + ISCs has functional impact in vivo GI GVHD severity. Because recent report demonstrated de ciency in SDHA dependent OXPHOS in context of T cell mediated colitis (27) and because Lgr5 + ISCs show reduced OXPHOS, we rst explored the effect of SDHA de ciency on ISCs in vitro and in vivo. The SDHA protein in Lgr5 + ISCs was measured from Allo-and Syn-B6 GFP-Lgr5 recipients after BMT by ow cytometry. SDHA expression in Lgr5 + ISCs was signi cantly decreased in Allo-recipients than Syn-recipients, and the magnitude of SDHA loss in Lgr5 + ISCs was greater than that in Lgr5 − intestinal epithelial cells (IECs) (Fig. 3, A and B) from the Allo-recipients, suggesting that the SDHA protein in ISCs is reduced to a greater extent than in other IECs in the in ammatory milieu of GVHD.
To determine the role of LGR5 + ISC speci c expression of SDHA, we bred Sdha oxp/ oxp mice (27) with GFP-Lgr5 mice to generate mice that lack SDHA only in Lgr5 + ISCs (SdhaΔ/ISC) upon tamoxifen administration ( g. S3A and Fig. 3C). E cient reduction of SDHA exclusively in the ISCs was con rmed in these SdhaΔ/ISC mice (Fig. 3, C and D). We next utilized tamoxifen-treated ISC-speci c knockout (KO) and diluent-treated (WT) SdhaΔ/ISC littermate animals as BMT recipients. Both the KO-and WTsyngeneic recipients survived. By contrast, allogeneic ISC-speci c KO-SdhaΔ/ISC-recipients demonstrated signi cantly greater mortality when compared to the Allo-WT littermate recipients (Fig. 3E). The increase in GVHD mortality was associated with the signi cant decrease in Lgr5 + ISCs in the crypts of Allo ISCspeci c KO SdhaΔ/ISC-recipients when compared with Allo-WT-recipients (Fig. 3F). However, no signi cant differences were noted in donor T cell expansion, Tregs or cytokine production ( g. S3, B to E).
These data demonstrated that the alteration in SDHA dependent OXPHOS in Lgr5 + ISCs regulates GVHD severity in vivo.
Next, we compared the organoid formation by the harvested intestinal crypts derived from the ISC-speci c KO SdhaΔ/ISC recipients and WT recipients (Fig. 3, G to H). Interestingly, the genetic deletion of SDHA in ISCs signi cantly reduced the numbers of organoids and buds in both ISC-speci c KO SdhaΔ/ISC Alloand Syn-recipients when compared with WT-recipients (Fig. 3, G to H). But the reduction was much greater in the KOs than the WT Allo-recipients. Seahorse assay further showed a signi cant decrease of OCRs in Allo-KO-recipients than Allo-WT-recipients (Fig. 3I), directly demonstrating that Lgr5 + ISC intrinsic alteration in SDHA dependent OXPHOS enhanced intestinal tract sensitivity to in ammation induced damage during GVHD.
Succinate regulates ISC function and DNA methylation OXPHOS is a critical process that generates ATP and is indispensable for ISC fate decisions (17,19,23,(51)(52)(53). To understand the relationship of SDHA dependent metabolism and Lgr5 + ISCs' regeneration capability in context of GVHD, we next quanti ed succinate levels, a metabolic byproduct of tricarboxylic acid (TCA) cycle that accumulates in context of SDHA de ciency (27). We observed signi cantly increased levels of succinate in organoids from Allo-recipients than from Syn-recipients (Fig. 4A), consistent with the reduction in SDHA and OCR (27).
Succinate induces DNA methylation, therefore, we next explored whether succinate accumulation directly regulated Lgr5 + ISCs regeneration and its DNA methylation (54,55). To this end, we treated organoids with cell-permeable dimethyl succinate or diluent control to directly evaluate the impact of succinate on organoid formation and epigenome. Seventy-two hours after the succinate treatment, we observed organoids formation from sorted Lgr5 + ISCs was inhibited (Fig. 4B-4C). The amount of global DNA methylation (5-mC) in organoids measured by assessment of 5-mC levels was signi cantly increased after treatment with succinate when compared with control organoids (Fig. 4D). To investigate whether succinate treatment directly regulated gene expression as observed from the transcriptomic analysis from above, we analyzed the expression of slc25a5, chchd2, gpx4, and Lgr5 with qPCR (Fig. 4, E to H). Their expression was signi cantly reduced in succinate treated organoids when compared with control treated organoids. Thus, the genes that were downregulated from the Lgr5 + ISCs transcriptomes analysis in context of GVHD were also downregulated by succinate treatment suggesting that alteration of SDHA dependent metabolism during GVHD regulates ISC transcriptome.
Lgr5 + ISC epigenome is reprogramed in context of GI GVHD Because succinate, the byproduct of the metabolic de cit in ISCs after Allo-BMT, modi ed DNA methylation and regulated organoid growth, we next determined whether ISC epigenome was reprogramed after Allo-GVHD. To this end, we performed ATAC-seq in sorted Lgr5 + ISCs harvested from the BALB/c→B6-GFP-Lgr5 model of GVHD. Brie y, fty thousand Lgr5 + ISCs were sorted by FACS from Allo-and Syn-recipients seven days following BMT as above. To generate the ATAC-Seq library, we used the Omni-ATAC-Seq protocol (56). The libraries were pooled, quantitated, and sequenced. ATAC-seq data were analyzed using the Diffbind package in R, and showed a change in landscape, including increased chromatin accessibility in 6252 peaks of and a reduction in chromatin accessibility in 3167 peaks in Allorecipients, compared with Syn-recipients (Fig. 5, A and B and g. S4A). To further explore the position of genome chromatin accessibility, we annotated the peaks using the ChipSeeker R package and split the reads based on differential accessibility in all conditions. We observed that there was a difference in distance of peaks to closest transcription start sites (TSS) and different binding sites (Fig. 5, C and D) between Allo-and Syn-recipients.
To understand which signal pathway was involved in the chromatin sites that were differential accessible, we performed Gene Ontology (GO) enrichment analysis using Protein ANalysis THrough Evolutionary Relationships (PANTHER). Mapping the hyper-accessible regions of the corresponding genes (3kb neighborhood of TSS) resulted in 1493 gene targets, most of which are known to be involved in immune responses and antigen processing and presentation (Fig. 5E). In contrast, the corresponding genes to the hypo-accessible regions are mostly related to the Rho GTPase pathway and actin cytoskeleton pathway, which participate in a wide variety of cell processes including proliferation and adhesion (Fig. 5F). Importantly, the epigenetic changes from the ATAC-seq were consistent with expression of genes observed in transcriptomic analysis such as ido1, gbp2, gpx4, and kif5b ( g. S4, B to E) after Allo-BMT.

Epigenetic reprograming impacts ISC memory
We next analyzed whether the Lgr5 + ISCs retain the memory of the functional de cits induced by epigenome reprograming. We hypothesized that the reduction in organoid formation and OCR will be retained in serial organoid passages. To determine the functional de cit, Lgr5 + ISC crypts harvested from Syn-and Allo-recipients were harvested and placed in organoid cultures as above. Five days later, these primary organoids cultures were split and passaged for the next generation to determine whether the de cits in Allo-organoids persisted in subsequent generations of organoids following secondary and also tertiary passage. We monitored organoid growth daily and evaluated their organoids formation and bioenergetics from Allo-and Syn-recipients. The numbers of organoids and their buds were signi cantly decreased in Allo-recipients compared to Syn-recipients in primary organoids and also in the secondary and tertiary passaged organoids ( g. S2, A and B and Fig. 6, A and B). Seahorse assay showed the reduction of OCRs observed in primary organoids from Allo-recipients was sustained at a signi cantly lower level in the secondary and tertiary organoids than the Syn-recipients (Fig. 6, C and D). Furthermore, the lower expression level of fth1, chchd2, and gpx4-expression levels in ISCs after GVHD were also reduced in primary, secondary, and tertiary organoids in Allo-recipients compared to Syn-recipients in serial organoids cultures (Fig. 6, E to G). Thus, ISCs from Allo-animals demonstrated persistent de cits in their organoid regeneration, OCR and gene expression indicating that epigenetic rewiring of the ISCs, induced by their encounter with in ammation from GVHD in vivo, is remembered by the ISCs for multiple generations despite the absence of active in ammation in the ex vivo organoid milieu. We next analyzed whether the persistent cellular impact induced from the in ammatory encounter by the Lgr5 + ISCs is linked to epigenetic reprograming. Because 5-ethynyl-2'-deoxyuridine (EdU)-labeling of Lgr5 + ISCs in crypts were increased in vivo in the GVHD recipients compared to no-GVHD recipients (57), and because organoids showed reduction in growth ex vivo, we next analyzed whether the high fold changes (FC) of chromatin accessibility peaks were noted in genes that regulate cell cycle and cell biosynthetic process ( g. S5A). We observed upregulation of gene signatures for G2-M check point and in ammation response pathways in Lgr5 + ISCs by scRNA-seq (Fig. S5, B and C). Therefore, to determine whether the epigenome changes had functional impact on ISCs that encountered in ammation and survived in vivo, we analyzed Lgr5 + ISCs proliferation in serial organoids culture from Allo-and Synrecipients of GFP-Lgr5 mice. We pulsed EdU into organoids culture for 1 hour and then analyzed EdU + Lgr5 + ISCs by FACS (see Methods and g. S6A). We observed signi cantly higher percentage of EdU + Lgr5 + ISCs in Allo-recipients in primary and secondary organoids compared with Syn-recipients ( Fig. 6H). Furthermore, higher levels of Annexin V + Lgr5 + ISCs was observed in Allo-recipients in both primary and secondary organoids compared with Syn-recipients ( Fig. 6I and g. S6B). Thus in ammation-mediated epigenetic reprograming altered Lgr5 + ISCs propensity for proliferation and apoptosis that persisted in serial organoid cultures.
Next, to directly link metabolic de cit to epigenome induced cellular function changes we utilized malonate, a SDHA inhibitor, in organoids culture as described in the methods, and assessed for changes in organoids formation and Lgr5 + ISCs proliferation. The organoids formation was reduced ( g. S7A) and the percentage of EdU + Lgr5 + ISCs and Annexin V + Lgr5 + ISCs were increased in malonate-treated organoids compared to control ( g. S6, A and C, and S7, B and C). To rule out impact of off-target effect of malonate, and to further con rm that the metabolic de cit in SDHA regulated Lgr5 + ISCs proliferation in vivo, we once again utilized ISC-speci c KO SdhaΔ/ISC mice as above. Fifteen days after treatment of tamoxifen, we pulsed EdU for 2 hours and then harvested small intestine to analyze EdU + Lgr5 + ISCs by FACS ( g. S6, A to C). We observed that the percentage of EdU + Lgr5 + ISCs were substantially higher in the Sdha-deleted mice compared with Sdha-intact SdhaΔ/ISC mice that were treated with diluent ( g. S7D).
Furthermore, the percentage of Ki67 + Lgr5 + ISCs and Annexin V + Lgr5 + ISCs were signi cantly increased in Sdha-deleted mice than Sdha-intact mice (Fig. S7, E and F). Taken together, these results demonstrate in ammation mediated-metabolic alteration in SDHA, and the subsequent-succinate accumulation induced epigenetic rewiring result in Lgr5 + ISCs dysfunction that are retained and regulate the severity of GI GVHD.

Discussion
Adult Lgr5 + ISCs reside at the bottom of the crypt and are critical for the regeneration and repair of intestinal epithelium at homeostasis and following injury (6, 58). Lgr5 + ISCs display context-dependent functionality and adapt to different requirements over their lifetime, as dictated by their microenvironment cues (7,22). Cellular response to in ammation-induced stress in ISCs and its impact on their long-term ability is not well understood. Because emerging data demonstrate that adult stem cells harbor in ammatory memory, we examined whether Lgr5 + ISCs might also possess an in ammatory memory and if so, its mechanisms and their cellular functional impact (1)(2)(3)(4). Utilizing well-characterized in vivo models of Allo-BMT wherein ISCs are known to be bona de targets of GI GVHD in ammation, we found that in ammation induced OXPHOS de ciency in Lgr5 + ISCs leading to accumulation of succinate, that reprogrammed the epigenome and regulated their ability to regenerate in multiple passages of organoid development.
GI GVHD is known to be caused by the depletion of ISCs caused by alloreactive T cell induced in ammation in the context of intestinal microbiome and innate immune activation (8, 10). Several recent studies have explored the cause and impact of loss of ISCs in vivo, and ex vivo in organoid cultures, by their inability to grow, bud and differentiate (12)(13)(14)(15). However, whether GI GVHD in ammation induces any qualitative changes in the ISCs besides the quantitative change caused by their apoptosis remained unclear. Our data demonstrate that GI GVHD induced in ammation causes qualitative changes in the ISCs. Speci cally, single cell transcriptomics show signi cant changes in the transcriptomics of the harvested ISCs the allogeneic recipients. We found signi cant gene downregulation in the pathways of ATP metabolic process, generation of precursor metabolites and energy, and OXPHOS. The qualitative differences in the Lgr5 + ISCs from the GVHD and control recipients ISCs were observed in the organoid reconstitution and crypts organoids formation from the sorted post-BMT Lgr5 + ISCs in the absence of ongoing ex vivo in ammation. Furthermore, consistent with reduction in pathways related to OXPHOS and ATP production from the single cell transcriptomic analysis, the Lgr5 + ISCs from GVHD also demonstrated a reduction in their OCRs indicating a functional metabolic defect.
Metabolic alterations induced from and by in ammation in immune cells and tumors is increasingly appreciated (59,60). Whilst adult stem cells may have a metabolic signature that is distinct from more differentiated cells (61, 62), our demonstrate a similar metabolic defect in SDHA dependent OXPHOS metabolism in Lgr5 + ISCs as was reported in IECs (27), albeit to an even greater degree of reduction in SDHA in the ISCs. Interestingly, Lgr5 + ISCs show a higher OXPHOS activity compared to other differentiated epithelial cells (51,63,64) thus, ISCs may be more dependent on SDHA. However, whether greater levels of SDHA in ISCs than the IECs makes them more sensitive to alloreactive T cell mediated in ammation will need to be analyzed in future studies. Nonetheless our data unequivocally demonstrate that SDHA in Lgr5 + ISCs modulates organoid formation and, also regulates GI GVHD severity in vivo, thus demonstrating it to be a critical in ammation induced metabolic checkpoint in ISCs and IECs. ISCs include several subtypes which may play a context dependent role (42,(65)(66)(67)(68). It is important to note that our study focused on, and therefore is limited only to the Lgr5 + ISCs, which are the actively cyclingtype stem cells and are sensitive to irradiation and in ammation (65, 69, 70). Furthermore, progenitor intestinal cells and mature cells subtypes such as Paneth cells demonstrate metabolic heterogeneity (6, 71,72). Therefore, the impact of in ammation on SDHA induced metabolic checkpoint on various intestinal epithelial cell subsets may be distinct and will also need to be determined in future studies.
Our data demonstrates that reduction in SDHA led to accumulation of succinate as a byproduct of in ammation-induced metabolic adaptation in Lgr5 + ISCs. Succinate is a known oncometabolite that affects DNA methylation in tumors, affects monocyte/macrophage responses and promotes posttranslational modi cation of histones and other proteins (54,73,74). Our data now link succinate to ISC epigenetic remodeling and cellular functions. Succinate accumulation impaired organoids formation and increased DNA methylation in organoids. ATAC-seq of ISCs showed changes driven by the alteration in DNA methylation after exposure to in ammation. The epigenomic changes are demonstrated by changes in the chromatin accessibility peak changes in their position and binding sizes of genes that could regulate ISC growth and differentiation. Integration with transcriptomics showed that the changes in epigenome most signi cantly regulated genes that affect cell cycling. These changes in gene expression were con rmed in SDHA deleted Lgr5 + ISCs, thus linking succinate with epigenetic regulation of cell cycling genes. The differential impact of succinate accumulation on DNA methylation and the effects on post-translational modi cation of proteins in ISCs or any other cell types remain to be addressed. Nonetheless, utilizing serial organoids culture, we demonstrate that the changes in the expression of these genes, and the functional impact on cell cycling persist, suggesting that the rewiring the epigenome of ISCs is retained following their exposure to in ammation in vivo.
To our knowledge, this is a rst demonstration of in ammation induced metabolic change that remodels the epigenome and leaves its memory in ISCs. Whether similar or distinct reprograming occurs in other tissues and other types of in ammation will need to be analyzed. Importantly, metabolic adaptations in OXPHOS play a role in generation of immune memory (75)(76)(77)(78). When considered in light of our study it suggests that OXPHOS is one common mechanism of epigenetic reprogramming that regulates memory in both immune and non-immune cells such as ISCs. However, whether the impact on cellular memory by OXPHOS alterations is speci cally only from an SDHA dependent process, or whether the mechanisms are variable, or distinct depending on the type of stimuli will need to be addressed in future studies. Regardless, because our data suggests the epigenetic rewiring persists even after in ammation subsides, the residual effects on ISCs may affect their tness and subsequent tolerance of future stressors thus having critical implications for intestinal tissue resilience and repair. Cellular adaptation is a fundamental cellular property that maximizes tness to environmental pressures. ISCs are exposed to a number of environmental cues that regulate their tness, function and fate at homeostasis. Therefore, whether the environment of ISCs contributes to its epigenetic imprinting induced by in ammatory metabolic changes and if that determines the maladaptive or a protective nature of the in ammatory memory remains to be investigated. A recent seminal study demonstrated that exposure to an infection during pregnancy triggered in ammatory memory within in fetal intestinal epithelial cells and protected the offspring from later infection demonstrating protective intergenerational memory in the intestinal epithelial cells (3). Our data now extend the notion of in ammatory memory to ISCs and suggest that a memory may be detrimental or maladaptive because of the persistent defect in the ability of organoid formations by these ISCs in serial organoid cultures. Our data suggest that the ISCs retain the scars of in ammation, and these impair their ability to regenerate and repair. Thus, following an in ammatory encounter, the GI tract may become more sensitive to the next stressor or even to homeostatic environmental cues. Our data therefore suggest that next generation of therapies targeting metabolic adaptation of epigenetic rewiring could allow for better tissue tolerance and regeneration in the context of immune mediated damage. Figure 1 Single cell transcriptomics of Lgr5 + ISC metabolic on GI GVHD.