Stromal Niche Signals That Orchestrate Intestinal Regeneration

Abstract

Stromal cell populations have a central role in providing signals that support the maintenance, differentiation, and function of the intestinal epithelium. The behavior and fate of epithelial cells is directed by the spatial organization of stromal cells that either sustain stem and progenitor cell identity or drive differentiation. A combination of single-cell analyses, mouse models, and organoid coculture assays have provided insight into the diversity of signals delivered by stromal cells. Signaling gradients are established and fine-tuned by the expression of signaling agonists and antagonists along the crypt-villus axis. On epithelial injury, there are disruptions to the abundance and organization of stromal populations. There are also distinct changes in the signals originating from these cells that impact remodeling of the epithelium. How these signals coordinate to mediate epithelial repair or sustain tissue injury in inflammatory bowel diseases is beginning to emerge. Understanding of these processes may lead to opportunities to target stromal cell populations as a strategy to modify disease states.

Summary.

This review encapsulates the role of stromal cells in providing key niche signals that impact the intestinal epithelium and mediate repair following injury. The spatial organization of signals mediates the balance between proliferation and differentiation and restoration of the epithelium.

The intestinal tract is lined by a highly dynamic epithelial monolayer with undifferentiated stem cells, proliferative progenitor cells, and specialized differentiated cells organized into specific domains. Epithelial turnover is fueled by a population of stem cells that reside at the base of crypts and relies on signals from the surrounding niche environment to ensure the balance between cell production and differentiation is maintained.¹ Different signals supplied by the niche determine the equilibrium between expansion and survival of stem/progenitor cells and commitment toward either secretory or absorptive cellular states. Original studies pioneered in the laboratory of Hans Clevers identified Paneth cells, in direct contact with stem cells, as a source of many regulatory molecules in the mouse small intestine.² Some studies using single-cell transcriptional profiling in human tissue have not detected these molecules so it remains unclear whether human Paneth cells express substantial niche factors.^3,4 Several studies have now revealed a large variety of signals are supplied by the supportive stroma underlying the epithelium.5, 6, 7, 8, 9, 10 The core signaling pathways that coordinate epithelial cell decisions are the WNT, NOTCH, bone morphogenic protein (BMP), and epidermal growth factor (EGF)/neuregulin-1 (NRG1) signaling pathways. Because the epithelial lining of the intestine is in direct contact with luminal contents and is highly vulnerable to damage, elucidating how stromal cells and the signals they secrete adapt to restore the tissue following injury is key to the understanding of intestinal disease. Identifying specific stromal niche populations and the factors they produce has been challenging. This is because stromal markers are expressed in more than 1 cell population, cellular phenotypes are influenced by signals within their microenvironment, and injury and disease states impact stromal cell populations. A clearer picture of the cellular heterogeneity within the mesenchyme is emerging using a combination of high-resolution tissue imaging, single cell analyses, and functional studies.

Signals From the Cells Within the Mesenchyme Regulate Epithelial Cell Fate

Different niche signals coordinate to regulate epithelial cell fate. Canonical WNT signaling is essential for maintaining stem and proliferative cells within crypts. This has been demonstrated in the murine small intestine by knockout of the key intracellular mediators Ctnnb1¹¹ and Tcf4¹² in the epithelium and blocking of WNT secretion.^9,13 WNT3 is secreted by Paneth cells in the mouse gut, but deletion of WNT3 in intestinal epithelial cells of transgenic mice is dispensable for the maintenance of intestinal stem cells. Accordingly, the predominant source of WNTs is supplied by subepithelial cells in the mesenchyme.14, 15, 16 WNT ligands interact with FZD receptors located on the cellular surface of receiving stem and progenitor cells.¹⁷ There is considerable evidence that specific thresholds of WNT signaling determines different cell fates. Secreted R-spondin (RSPO) signals interact with LGR5 receptors specifically located within stem cells to potentiate WNT signals and generate the high threshold of WNT signaling required to maintain stem cells in an undifferentiated state.^1,18, 19, 20 A combination of secreted WNT inhibitors and activating WNT ligands and potentiators produce a gradient of WNT signaling that is highest at the crypt base and lowest within the differentiated cell compartment.^8,21 This results in stem and proliferating progenitors to be restricted to crypts, whereas noncycling, mature cells reside closer to the luminal surface.

In contrast to WNT, BMP signaling attenuates cell proliferation. Analysis of BMP pathway components along the crypt-villus axis has revealed an opposing gradient of activators and inhibitors to that of WNT, with the highest signals maintained in the zone of differentiated epithelial cells (Figure 1). Specifically, stromal niche-derived BMP inhibitory molecules, such as Noggin and Gremlin 1 (GREM1), are required to curtail BMP signaling at the crypt base and allow proliferation of stem and progenitor cells, whereas BMP activators BMP2, BMP4, BMP5, and BMP7, produced by stromal cells located in the villus compartment, support the differentiation of enterocyte subtypes in a location-dependent manner.²²

Figure 1.

Distribution of cell populations in the mesenchymal niche. Stromal cell populations are distributed along the crypt-villus axis. This includes subepithelial PDGRFA^hi telocytes that express BMP, PDGFRA^lo stroma, and deep crypt PDGFRA^lo, CD81⁺ trophocytes that express WNT2B and RSPO. Other cell types, such as immune cells, vascular, and lymphatic endothelial, also provide key signals.

Figure created with BioRender.com.

Another essential pathway that governs cell decisions is the NOTCH signaling pathway, which operates via lateral inhibition within epithelial cells to control cell fate.23, 24, 25 This pathway seems not to be controlled by a gradient of subepithelial niche signals.

Finally, proliferative signals supplied by the EGF family of ligands are essential to maintain the epithelium.^2,26, 27, 28 Although it was initially demonstrated that EGF is secreted by Paneth cells within the murine epithelium,² there is emerging evidence that other family members, such as stromal-derived NRG1, are important signaling molecules.^26,28, 29, 30 Indeed, inducible loss of Nrg1 in transgenic mice led to a significant decrease in the number of intestinal stem and proliferative cells.²⁹

Cellular Organization and Sources of Signals in Supporting Stromal Cells

The epithelium is supported by numerous cell types, including stromal cells, that produce growth factors and extracellular matrix (Figure 1). The diversity of stromal cell populations has been characterized by high-resolution microscopy, knockin mouse models that mark cells with fluorescent tags, and single-cell RNA sequencing (scRNA-seq). The combination of these techniques is leading to a spatial and molecular understanding of the organization of the niche, although each of these approaches have their limitations. Immunostaining relies on antibody specificity and isolation methods for single-cell analyses can damage and/or change the expression profile of cells. This is particularly significant under conditions of inflammation where cells are exposed to cytotoxic signals that can lead to cell loss when cell dissociation techniques are used.³¹ The functional significance of niche cells and signals has been studied using Cre-mediated knockout mouse models and organoid cocultures. These phenotypes can also be confounded by the limitations in cell specificity of Cre-drivers, the resistance of some stromal cells to tamoxifen.³²

Despite these challenges, different mesenchymal stromal cell populations defined by their morphology, spatial location, and expression profile can be described (Figure 1). An important subepithelial population known as telocytes, characterized by an extensive cytoplasm with long processes,^9,33 have been visualized from the bottom of the crypts to the top of the villus using a Foxl1-cre-driven GFP system.⁹ Importantly, ablation of Foxl1-positive cells, using a diphtheria toxin receptor-mediated cell knockout transgenic mouse model, led to dramatic changes to the intestinal architecture, with a strong decrease in crypt and villus length and almost complete loss of stem and proliferative cells.⁹ A similar phenotype was observed when WNT secretion was blocked specifically in the Foxl1-positive cells.^6,33 Importantly, the expression of Foxl1 was likely overlapping with several distinct stromal cell populations, as revealed by the concomitant expression of both WNT activators and BMP inhibitors by bulk RNA sequencing. Therefore, it was suggested that the molecular signature of telocytes was defined by their position along the crypt-villus axis.

scRNA-seq analysis of mouse intestinal tissues allowed for this cellular heterogeneity to be identified and characterized based on the differential expression of a stromal marker (platelet-derived growth factor receptor alpha [PDGFRA]) (Figure 2A). Indeed, telocytes, defined by high expression of PDGFRA aggregate at the top of crypts and villi and express several BMPs, LGR5, noncanonical WNT5A but little WNT2B³⁴ (Figure 2C).

Figure 2.

Expression of niche signals during homeostasis and regeneration. Data were integrated from mouse PDGFRA sorted cells and whole mesenchyme from GSE130681,⁶ mouse small intestine GSE116514,²⁵ mouse healthy, and DSS-treated colon GSE114374.⁵ (A) UMAP visualization of distinct cell clusters based on PDGFRA and CD81 levels. Data were processed and integrated using the Seurat package version 4.9.9 in R.³⁵ Potential doublets were removed using DoubletFinder³⁶ and cells were annotated using CellMarkerDB.³⁷ (B) Nrg1 expression in the integrated data set. Separation of the data set based on PDGFRA and CD81 levels. (C) Relative expression of key signals in control compared with DSS-treated mice.⁵

In contrast, 1 population of PDGFRA^lo cells reside in the lamina propria, whereas deep below the crypts, directly above the muscularis externa, PDGFRA^lo trophocytes marked by CD81 can be found.^8,21 Trophocytes express WNT2B, GREM1, and RSPO3 (Figure 2C) and sustain mouse small intestinal organoids when cocultured ex vivo. PDGFRA^lo stromal cells can also support organoids in cocultures but this is inhibited when exposed to BMPs, which reduce the expression of RSPO3 and GREM1 in these cells.

Similar stromal cell clusters have been identified in scRNA-seq analysis of human intestinal mesenchymal populations.^7,28,38,39 The abundance of these cell populations is not fixed along the length of the human intestine because regional differences in cellular composition are observed in distinct tissue segments.⁴⁰ Another key signal expressed within the PDGFRA^lo CD81^lo and PDGFRA^high populations is the EGF family member NRG1 (Figure 2B). NRG1 has been shown to impact the growth and differentiation of mouse and human intestinal organoid cultures.28, 29, 30

Cellular Changes and Key Signals Instigated Following Injury

The apical surface of the epithelium is exposed to a complex environment of food products, bacteria, bacterial by-products, toxins, and mechanical stress. This can cause injury to the epithelium, which is rapidly repaired by activity of stem cells. Injury can disrupt the epithelial barrier, resulting in inflammation and influx of microbes and toxins, causing further damage to the intestinal epithelium.41, 42, 43 Following damage, there is a distinct program of epithelial cell regeneration where resident stem cells are replaced by dedifferentiation of progenitor cells. This recovery mechanism involves considerable cellular plasticity and is dependent on Yes-associated protein signaling and reactivation of fetal reprogramming signatures.^29,35, 36, 44, 45, 46, 47, 48, 49, 50

The role of supporting niche cells in mediating this repair process is emerging and changes in the composition of niche cells during the injury and regeneration time course and the signals impacting the epithelium are apparent.

scRNA-seq analysis of the murine colon at different timepoints during dextran sulfate sodium (DSS)-mediated acute damage and the repair phase suggests complex remodeling of the colonic microenvironment, including a gradual decrease in the number of epithelial cells, stromal cells, myofibroblasts, and glial cells over the course of inflammation, damage, and repair.³⁷ Strikingly, stromal cells are major providers of signaling molecules and interact with a variety of cell populations during DSS-mediated injury and repair, which suggests that these specialized cells are a potential cell communication hub (Figure 2C).³⁷ Similarly, in human tissue from patients with ulcerative colitis, cellular remodeling of the mesenchyme is also observed, including expansion or appearance of a proinflammatory stromal subset.^7,51 Stromal cell populations expressing WNT are altered in ulcerative colitis, which likely impacts the maintenance of epithelial cells.⁷ The inflammatory stromal subset highly expresses interleukin (IL)33, IL6, and TNFSF14 (LIGHT), which may have impacts on epithelial stem cell populations.⁷ These cells also express Lox and Loxl1, which can promote oxidative stress and tissue damage. Another study also found that inflammation-associated fibroblasts express IL11, IL24, and IL13RA2.⁵¹

A wide range of essential niche-derived signals are produced in response to injury, including WNT ligands, WNT potentiators (RSPOs), BMP inhibitors, and NRG1 (Figure 2). Several approaches, including the use of transgenic mouse models, have been used to demonstrate the requirements of these signals in tissue repair. For example, in the absence of stromal WNTs, intestinal epithelial regeneration following radiation-induced injury is compromised in mice.¹⁵ In DSS-induced colitis, a higher expression of Rspo3 is detected in the colon⁵ and loss of Rspo3 in PDGFRA-positive cells impairs tissue regeneration.⁶ Similarly, higher levels of stromal Grem1 are detected in the mouse colon following DSS treatment and genetic manipulation of BMP signaling by Grem1 overexpression supports efficient tissue repair.⁵² Finally, in both 5-FU and irradiation injury mouse models, the expression of Nrg1 is induced in PDGRFA-positive stromal cells and loss of Nrg1 decreases the number of proliferative and stem cells in regenerating crypts.²⁹

Implications for Developing New Approaches to Treat Inflammatory Bowel Disease

Remodeling of the stromal niche microenvironment and rewiring of its molecular networks are fundamental features of the inflamed colonic tissue. Characterizing these processes is a critical step toward elucidating mechanisms of drug resistance or the development of novel therapies. Recent advances in the field have demonstrated that inflammatory-associated stromal cells, which are unique to patients with ulcerative colitis, are enriched in patients who are resistant to anti–tumor necrosis factor treatment.⁵¹ In addition, the gene signature of anti–tumor necrosis factor resistance is positively correlated with the signature of these unique stromal cells.⁵¹ Similarly, the molecular features of therapy nonresponse across multiple inflammatory bowel disease treatments are enriched in inflammatory stromal cells,⁵³ suggesting that the appearance of this specialized cell population drives therapy resistance in patients. Inflammatory-associated stromal cells are suggested to exacerbate colitis through local redox disturbance and/or neutrophil recruitment.^7,53 Interestingly, mice with colitis that were treated with β-aminopropionitrile, an irreversible inhibitor of the LOX enzymes, showed a reduction in oxidative damage and an improvement in disease. This supports the notion that targeting stromal cells may be a viable pathway for the development of new therapies for intestinal disease.⁷

Future Directions

The deep understanding of cellular remodeling and cell-cell communication during colonic inflammation and repair reveals a potential therapeutic opportunity to dampen inflammation mediated by specific subpopulations of stromal cells while at the same time promoting tissue repair by activating stemness and epithelial proliferation mediated by other fibroblast populations. There is evolving evidence of the role of many other cell types that reside in close proximity to stromal and epithelial cells providing signals that impact regeneration. This includes immune,54, 55, 56 endothelial,⁵⁷ lymphatic,58, 59, 60 and neural cells.⁶¹ Analysis of crosstalk between all the different cell types is likely to be facilitated by the recent improvements in spatial technologies^56,62 that can define molecular changes at single-cell resolution.