MTGR1 is widely expressed in the intestine.
As multiple epithelial cell types exist in the small intestine, cellular expression patterns can be used to infer genetic contributions to cell type-specific functions. To define MTGR1 expression patterns within the intestinal crypt, we first utilized in situ staining methods to spatially visualize Mtgr1 transcripts in the murine small intestine (Fig. 1A). As previously reported, Mtgr1 expression was dispersed throughout the intestinal crypt and did not appear to be specifically localized to distinct cell populations16. As in situ staining does not easily allow for overlay with multiple cell lineage markers, we next investigated Mtgr1 expression in specific cell types via single-cell RNA-sequencing (scRNA-seq) of the murine ileum29,36. These results confirm widespread Mtgr1 expression in various intestinal cells that was not restricted to specific cellular lineages (Fig. 1B and Supplemental Fig. 1). Similar results were observed in the human small intestine (Fig. 1C). Here, we queried publicly available scRNA-seq data from the Human Protein Atlas (GSE125970), and again, MTGR1 expression was observed in multiple differentiated and undifferentiated cell types37,38. Thus, Mtgr1 is widely expressed throughout the intestinal epithelium.
MTGR1 loss dysregulates ISC populations in vivo.
Next, we directly investigated the function of MTGR1 in ISC biology. Interestingly, higher levels of cellular proliferation have been reported in mice globally lacking MTGR1 (Mtgr1−/−), which may be due to loss of MTGR1-mediated downregulation of Wnt pathway genes via TCF4 repression15,19,39. Here, we first confirmed the expansion of proliferative cells in the crypts of Mtgr1−/− mice via quantifying Ki67 expression (Fig. 2A). We next hypothesized that this increase in proliferation may be associated with higher numbers of LGR5 + ISCs, due to their role in maintaining proliferation in the intestine at baseline15,19,39. Lgr5-expressing cells in the small intestine were identified by in situ hybridization. Here, we determined that Mtgr1−/− mice indeed had higher numbers of Lgr5 + cells per crypt as compared to WT mice (Fig. 2B). This was further confirmed by intercross of Mtgr1−/− mice with the Lgr5-Cre-EGFP reporter line (Fig. 2C), which expresses EGFP from the Lgr5 locus. Analysis of these mice again showed increased numbers of Lgr5-EGFP + cells in Mtgr1−/− versus WT Lgr5-Cre-EGFP mice. Higher levels of Lgr5, Ki67, and the Wnt target transcript Myc were also observed in Mtgr1−/− crypts by qPCR (Fig. 2D). Based on these results, it seems likely that the LGR5 + CBC population and overall ISC function may increase when MTGR1 is lost.
While Lgr5 is often regarded as the canonical identifier of CBCs and a robust ISC marker, we next assayed for expression of Ascl2 and Olfm4, which are highly expressed in CBCs along with Lgr540,41 (Fig. 2D). Interestingly, although Lgr5 transcript and Lgr5-EGFP + cells were consistently increased with MTGR1 loss, neither Ascl2 nor Olfm4 mirrored these changes. Here, we observed that Ascl2 transcript remained unchanged in Mtgr1−/− crypts, while Olfm4 was nearly undetectable in Mtgr1−/− crypts. Similarly, survey of markers associated with non-CBC ISC populations revealed further differences, such as greater numbers of Clusterin (Clu)-expressing cells (Fig. 2E)6.
As cell identity and function are complex, we next isolated WT and Mtgr1−/− crypts and expanded our transcriptomic analysis of MTGR1-dependent changes in intestinal cell types through bulk RNA sequencing. Differential expression data was then analyzed by gene set enrichment analysis (GSEA), and results from WT and Mtgr1−/− crypts were compared to gene sets established from prototypic Lgr5-expressing ISCs (Fig. 2F and Supplemental Fig. 2)9. Despite increased expression of Lgr5 itself in the setting of MTGR1 loss, total canonical Lgr5 + ISC gene signatures were significantly de-enriched in crypts collected from Mtgr1−/− mice. Instead, Mtgr1−/− samples were enriched for genes which indicated an expansion of TA populations, which is also consistent with increased proliferation noted in Mtgr1−/− crypts (Fig. 2A). Taken together, these results indicate that while more cells in Mtgr1−/− crypts express Lgr5, these cells may instead resemble more differentiated TA cells, rather than true Lgr5 + ISCs.
MTGR1 is required for enteroid viability.
In our in vivo analysis, loss of MTGR1 appeared to deregulate ISC-associated genes and promote TA-associated differentiation. Thus, we questioned whether Mtgr1−/− ISCs were fully functional as multipotent intestinal stem cells. We next specifically tested whether MTGR1 loss affected overall ISC function using the small intestinal organoid or “enteroid” system. Since enteroids rely on ISCs for their establishment and growth, enteroid formation efficiency can be used to assess general stem cell function and fitness3. Here, enteroids were established from duodenal crypts harvested from WT and Mtgr1−/− mice, and enteroid formation efficiency was assessed after 24 hours in culture (Figs. 3A and 3B). By dividing the number of enteroids formed by the number of crypts plated, we noted an approximately 2-fold enhancement of enteroid formation in the setting of MTGR1 loss. We also observed higher percentages of Mtgr1−/− enteroids with a cystic, spheroid morphology (Fig. 3C), a phenotype associated with increased Wnt tone20 ,compared to WT enteroids. Taken together, these results suggest that MTGR1-deficient ISCs are indeed initially functionally competent, despite de-enrichment of a canonical Lgr5 transcriptional profile.
While initial enteroid formation was augmented in Mtgr1−/− cultures, we observed striking viability defects in Mtgr1−/− enteroids within 48 hours of initial plating. Daily imaging (Fig. 3D) and viable enteroid counts (Fig. 3E) revealed that Mtgr1−/− cultures failed almost completely by day 5 post-plating. While WT enteroids formed crypt buds by day 3, Mtgr1−/− enteroids rarely developed crypt buds, even in the structures that survived until day 5 (Fig. 3F). These findings were also confirmed by live cell imaging, which showed no morphological changes in Mtgr1−/− enteroids throughout the 5-day period, until organoid death (Supplemental videos SV1 and SV2). Importantly, restoration of MTGR1 expression via lentiviral transduction rescued Mtgr1−/− enteroids and restored branching morphology, confirming the MTGR1 dependency of this phenotype (Fig. 3G and 3H). Thus, MTGR1 appears to be required for ex vivo enteroid survival and expansion.
Inhibition of programmed cell death does not rescue Mtgr1 −/− viability.
We next aimed to determine the mechanism driving the viability loss in Mtgr1−/− enteroids. As MTGR1 is a transcriptional co-repressor, we again utilized a bulk RNA-sequencing approach to broadly investigate MTGR1-dependent changes in gene expression. Briefly, crypts were isolated from age-matched WT and Mtgr1−/− mice, and mRNA was collected at the time of crypt isolation (day 0), or at day 1 and day 3 post-plating to yield matched RNA sets of crypts, day 1 enteroids, and day 3 enteroids (Fig. 4A). After RNA-sequencing, differential expression profiles were generated and analyzed using GSEA27,28.
Due to the rapid loss of established cultures, we hypothesized that MTGR1 loss may aberrantly activate programed cell death pathways to drive the observed decline in enteroid viability. Indeed, GSEA analysis from the Hallmark gene set collection identified a significant enrichment in apoptosis-associated genes in Mtgr1−/− enteroids at both day 1 and day 3 post-plating (Fig. 4B). Mtgr1−/− enteroids collected at day 1 post-plating also displayed higher numbers of apoptotic cells as compared to WT enteroids (Fig. 4C) as measured by fluorescent immunohistochemistry (IHC) against cleaved caspase-3. However, inhibiting apoptosis using the cell-permeable pan-caspase inhibitor, Z-VAD-FMK, failed to improve survival of Mtgr1−/− enteroids (Fig. 4D), even at concentrations which improved viability in WT cultures42. Likewise, inhibition of necroptosis, whose dysregulation has been noted in intestinal inflammatory diseases43–45, had no effect on Mtgr1−/− enteroid viability (Fig. 4E). Finally, we assessed the impact of p53 inhibition, as p53-related gene sets were also positively enriched in Mtgr1−/− samples by GSEA (Fig. 4F). As with Z-VAD-FMK, inhibition of p53-dependent apoptosis with pifithrin-α had no effect on Mtgr1−/− enteroid survival (Fig. 4G)46. Thus, despite increases in apoptosis, inhibition of known cell death mechanisms is insufficient to rescue Mtgr1−/− enteroid viability.
Proliferation and ISCs are lost in MTGR1-deficient enteroids.
Due to constant cell clearance, actively cycling stem cells and high levels of proliferation are necessary to maintain intestinal cell populations47. Thus, rather than aberrant apoptosis, we next hypothesized that the viability defect in Mtgr1−/− enteroids may instead be due to reduced proliferation and/or depletion of ISCs. To determine cell proliferation, sections from enteroids embedded at day 1 and day 3 post-plating were assessed by Ki67 IHC (Fig. 5A). Although we observed similar numbers of proliferating cells in day 1 enteroids, by day 3, the enteroid cultures established from Mtgr1−/− mice displayed a drastic, nearly 80% reduction in Ki67 + cells. Cell cycle- and proliferation-associated genes were also highly downregulated in Mtgr1−/− enteroids by day 3 (Fig. 5B), as well as ISC-associated genes and signaling pathways (Fig. 5C), as determined by GSEA. Interestingly, while numbers of Ki67 + cells were similar between WT and Mtgr1−/− enteroids at day 1 post-plating, proliferation-, ISC-, and Wnt-associated genes were still significantly downregulated at this early timepoint (Fig. 5D, 5E, and Supplemental Table S3). Mtgr1−/− enteroids, at either day 1 or day 3 post-plating, also demonstrated significant upregulation of the cell cycle inhibitors Cdkn1a, Cdkn1c, Cdkn2b. These results indicate that viability defects in Mtgr1−/− enteroids may arise from proliferation defects and the inability to maintain cycling ISC populations ex vivo.
MTGR1 loss promotes absorptive enterocyte differentiation.
After expansion in the TA zone and exit from the intestinal crypt, most ISC-derived cells rapidly undergo differentiation into non-proliferative cell lineages48. As stem and proliferative cell populations are not maintained in Mtgr1−/− enteroids, and their loss appears unlikely due to programed cell death, we hypothesized that the loss of enteroid viability may be due to augmented ISC differentiation into non-proliferative cells. Indeed, this would result in failed ISC amplification, inability to maintain enteroid cultures, and eventual death of terminally differentiated cells. Therefore, we next broadly surveyed differentiated intestinal cell types using GSEA. In agreement with the data presented in Fig. 5 and the results from intestinal crypts (Fig. 2F), cycling Lgr5 + cells again appeared to be greatly depleted, and were the most reduced cell population in Mtgr1−/− enteroids at both day 1 and day 3 post-plating (Fig. 6A and Supplemental Fig. 3). Conversely, more differentiated TA populations were greatly increased in Mtgr1−/− enteroids, as were fully differentiated populations of enterocytes.
To more clearly define whether MTGR1 loss indeed accelerates intestinal differentiation, we next determined the expression of genes associated with villi and enterocyte differentiation pathways. We first surveyed classical markers of the enterocyte lineage, such as intestinal alkaline phosphatase (Alpi), as well as genes associated with BMP and IHH pathways, as these work in opposition to Wnt-mediated signaling in order to promote differentiation3. In nearly all cases, these differentiation-associated genes were significantly enriched in Mtgr1−/− enteroids as compared to matched WT samples (Fig. 6B and Supplemental Table S4). GSEA analysis also determined significant enrichment of genes associated with features of absorptive enterocytes, such as the brush border, microvilli, and intestinal absorption (Fig. 6C). Finally, the cellular structure of day 1 enteroids was investigated using transmission electron microscopy (TEM). Comparison of WT and Mtgr1−/− enteroids illustrates an expansion of the apical cell surface as well as more pronounced and mature microvilli (Fig. 6D). Altogether, these results indicate that Mtgr1−/− ISCs are likely further differentiated at baseline than WT ISCs, and upon ex vivo culture, differentiation is so accelerated that proliferative cells are lost entirely to the absorptive enterocyte lineage, to the point that cultures cannot be maintained.
Secretory cells promote the survival of Mtgr1 −/− enteroids.
Previous research has indicated that MTGR1 is necessary for the differentiation of multiple secretory lineages in the small intestine, including Paneth, goblet, and enteroendocrine cells (EECs)15,18. Thus, we next confirmed the loss of these cell types in Mtgr1−/− crypts and enteroid cultures. A query of RNA-seq results indeed showed downregulation of genes associated with general secretory cell differentiation and those enriched in the Paneth, goblet, and EEC lineages (Fig. 7A, Supplemental Fig. 2, and Supplemental Table S5). Interestingly, loss of MTGR1 had the opposite effect on the tuft cell lineage in vivo, which was increased in the intestinal crypts; however, this same cell population appeared to be downregulated in Mtgr1−/− enteroids. Together, these results further define the variable effect of MTGR1 on intestinal cell differentiation.
Secretory lineages, specifically Paneth cells, are crucial regulators of CBCs and provide Wnt ligands that maintain ISC stemness, and crypts from mice lacking Paneth cells cannot form enteroid cultures without Wnt supplementation49–52. Thus, the loss of Mtgr1−/− CBCs, and by extension enteroid cultures, may be due to the lack of Paneth cells. While the majority of enteroids from Mtgr1−/− mice died by day 5, a small number of surviving enteroids occasionally could be maintained and passaged. We hypothesized that these few “passaged” Mtgr1−/− enteroids may provide clarity regarding mechanisms that could rescue MTGR1-dependent growth defects. To this end, we established and analyzed two independent Mtgr1−/− enteroid lines by RNA-sequencing to determine how these surviving cells differed from Mtgr1−/− cultures that subsequently died. Cell type analysis via GSEA found no deficiencies in Paneth or EEC populations in passaged Mtgr1−/− enteroids in comparison to WT enteroids (Fig. 7B), and Paneth cells could often be distinguished in the crypt base (Fig. 7C). Therefore, Paneth cells may drive the survival of passaged Mtgr1−/− cultures. However, despite the presence of Paneth cells, passaged Mtgr1−/− enteroids still displayed striking alterations in morphology, an inability to form enteroid buds (Fig. 7D), and little expansion in size over time (Fig. 7E). As observed in early enteroids at day 1 and day 3 post-plating, passaged Mtgr1−/− enteroids also showed significant reductions in ISC populations by GSEA (Fig. 7B) and proliferation by immunofluorescent IHC against the proliferative marker phospho-histone H3 (pH3, Fig. 7F). Thus, while Paneth cells likely aid the survival of passaged Mtgr1−/− enteroids, these results suggest that the presence of Paneth cells alone may not be sufficient to rescue growth, morphology, or ISC populations in the setting of MTGR1 loss.
Mtgr1 −/− enteroids can be rescued by high Wnt activity.
As our investigation of passaged Mtgr1−/− enteroid lines showed that Paneth cells may assist ISC survival in the context of MTGR1 loss, we next employed various combinations of small molecules to promote secretory cell differentiation and survival of Mtgr1−/− ISCs. First, we utilized DAPT, a γ-secretase inhibitor that has been shown to increase secretory cell numbers ex vivo at the expense of absorptive lineages. However, γ-secretase inhibitor treatment failed to rescue Mtgr1−/− enteroid growth ex vivo (Fig. 7G), even when treatment was begun in vivo prior to and continuing through enteroid establishment (data not shown). Next, we combined DAPT with the Wnt pathway agonist, CHIR 99021, as this combination should greatly promote Paneth cell differentiation53. These studies revealed only a modest induction of enteroid survival (Fig. 7H). Finally, we investigated whether Wnt pathway activation alone was sufficient to maintain Mtgr1−/− ISCs. Surprisingly, low-dose CHIR 99021 treatment (3µM), a concentration sufficient to promote WT ISCs and mimic the effects of recombinant Wnt3a, only had a modest effect on Mtgr1−/− enteroid survival (Fig. 7I)54. This could be overcome by increasing the concentration of CHIR 99021 (10 µM) to sustain survival and growth of Mtgr1−/− ISCs ex vivo. All together, these results indicate that while MTGR1 loss increases cell proliferation and Lgr5 expression, the overall functionality and differentiation state of Mtgr1−/− ISCs is severely compromised. While previously regarded as a transcription factor regulating secretory/absorptive lineage specification, these data indicate that MTGR1 is necessary for maintaining overall dedifferentiation and stem cell status in the small intestine.