The functional role of the extracellular matrix in inflammatory bowel disease associated gut fibrosis

Abstract

Intestinal fibrosis is characterized by the excessive accumulation of extracellular matrix (ECM) in the bowel wall. Complications, such as strictures that require surgical intervention in a large proportion of patients, are considered an inevitable consequence of chronic inflammation in inflammatory bowel disease (IBD) and leads to severe complications. The study of intestinal fibrosis in IBD has been traditionally focused on the associated immune process, and the role of the ECM itself has been largely overlooked. More recent studies have now clearly demonstrated that ECM is not simply a passive bystander of inflammation-driven fibrosis but is instead an active participant in the initiation and progression of the fibrogenic process. In this narrative review, we first describe the composition and function of the ECM components under physiological and pathological conditions of the gut. Then, we review the alterations of the intestinal ECM in IBD-associated fibrosis and the impact of fibrotic ECM on intestinal biology and function. We next critically evaluate the existing experimental systems to study the intestinal ECM, both in vitro and in vivo. We conclude by discussing the unique challenges that still exist to better understand the role of the ECM in intestinal fibrosis, and its potential diagnostic and therapeutic implications.

Introduction

Inflammatory bowel diseases (IBD) are a group of conditions typified by chronic relapsing and remitting inflammation of the gastrointestinal tract [1]. The two main forms of IBD are Crohn’s disease (CD) and ulcerative colitis (UC), both of which are highly complex at the pathogenic, diagnostic, clinical, evolutionary and therapeutic level, and are frequently accompanied by multiple complications, among which fibrosis is one of the most common. In fact, narrowing (stricturing or stenosis) of the intestinal lumen often develops, and frequently brings about bowel obstruction [2]. More than half of the patients with ileal CD eventually manifest obstructive symptoms unresponsive to conventional anti-inflammatory treatment, and most of them will need an operation for effective relief [3]. Intestinal fibrosis with or without obstruction is not unique to the transmural inflammatory process typical of CD, but also occurs in UC, where inflammation is restricted to the mucosa and submucosa [4]. Although in the past decades numerous new biologics or small molecules have been developed that effectively control the inflammatory activity in IBD [5], the proportion of patients who progress to stenosis and the rate of bowel surgery remain high [6]. Thus, understanding the specific mechanisms underlying the inflammation-driven fibrotic process that leads to a stricturing phenotype is a requisite to develop specific antifibrotic therapies and improve patient prognosis.

The morphological substrate of stricturing CD is an excessive deposition of a disorganized extracellular matrix (ECM) predominantly in the submucosa but also in the muscularis propria [7] (Fig. 1). This abnormal accumulation of ECM has been traditionally considered a consequence of long-term stimulation of local mesenchymal cells by a chronic immune response, but emerging evidence suggests that the ECM itself functionally contributes to the progression of fibrosis [8]. First, clinically, a disconnect exists between suppression of inflammation and progression of fibrosis and, second, an improved understanding of ECM composition and its mechanoproperties indicate that the interaction of specific ECM components with mesenchymal cells can directly promote fibrogenesis. While in other fibrotic diseases, such as idiopathic pulmonary fibrosis [9], hepatic fibrosis [10], renal fibrosis [11] and cardiac fibrosis [12] the role of the ECM is better comprehended, this is not so in intestinal fibrosis. Thus, this review aims at critically assessing the role of ECM in IBD and its relations with intestinal mesenchymal cells to provide the basis for future studies in the field of inflammation-driven intestinal fibrosis.

Fig. 1.

Bowel wall changes in stricturing Crohn’s disease (CD). Left panel: Normal intestinal wall with its layers of mucosa, muscularis mucosa, submucosa, muscularis propria and mesentery, covered by serosa. Right panel: In stricturing CD a thickening of the intestinal wall occurs. This is mediated by thickening of all layers, particularly the muscularis propria, which happens in conjunction with an expansion of the mesenteric fat. Inflammation is transmural and the extracellular matrix mainly accumulates in the submucosa and in the muscularis propria.

Composition and physiological function of the intestinal ECM

The so called ‘core’ matrisome comprises all major components of the ECM, and in mammals is constituted by about 300 proteins, with collagens, proteoglycans and glycoproteins being its main components [13]. The human core matrisome components are listed in Table 1 [14]. With the development of high-throughput (“omics”) technologies, more proteins are identified that take part in the remodeling of the ECM [15]. These proteins are further categorized by their structures and functions into ECM-affiliated proteins (e.g. mucins), ECM regulators (e.g. matrix metalloproteinases [MMPs]) and secreted factors bound to the ECM (e.g. transglutaminases, transforming growth factor [TGF]-β1, etc.) [15]. Although qualitatively the components of ECM are largely similar in different organs and tissues, there are major quantitative variations from organ to organ [16]. Unlike other organs with solid and fairly homogeneous structures like the liver [17] and the lung [18], the intestine has a more complex composition comprising multiple different layers which confer heterogeneity to the ECM depending on location, as exemplified by the distinct ECM composition in the basement membrane and the interstitial matrix [19]. While other non-parenchymatous organs such as the aorta and bronchi also exhibit fibrotic remodeling and share certain histological features (e.g., layered structure, presence of smooth muscle and connective tissue), the intestine is uniquely characterized by its rapid epithelial turnover, mucosal immune regulation, and the constant interplay between epithelial, stromal, immune, and microbial elements. This is of particular relevance because the basement membrane underlying epithelial cells is a key barrier that maintains the integrity of the mucosa, protecting the lamina propria and the resident immune and non-immune cells from an excessive and potentially injurious exposure to luminal contents [20]. A basement membrane is also found underneath the endothelial cells of blood vessels [19]. Collagen IV is the main ECM component of the basement membrane, but other collagens like type VI, VII, XV, XVII and XVIII can also be present [21]. The lamina propria and submucosa, which are located directly underneath the epithelial basement membrane, contain a variety of immune and non-immune cells types, as well as other ECM-rich interstitial structures [22] and are regarded as major sites of fibrogenesis [2]. They contain a mixture of interstitial ECM components, mainly collagens I, III, V and elastin [19]. A recent proteomic analysis using decellularized small intestinal submucosa revealed that the main ECM components in the submucosa include collagen I, II, IV, V, VI, VII, XV and fibronectin [23]. Fibroblasts and myofibroblasts are considered the major producers of ECM. An investigation demonstrated that the most abundant ECM deposited by primary human intestinal myofibroblasts derived from intestinal submucosa is fibronectin, followed by tenascin C, elastin microfibril interfacer 1, and TGF-β induced protein [24]. Via matrisome analysis of decellularized human intestinal tissue, our group has identified several main ECM components in the gut, including fibronectin, collagen I, collagen VI, fibrillin 1, decorin and laminin among others [25]. In regard to the intestinal muscle layers, a study showed that their main secreted ECM components are collagen I, III and V [26]. Another recent study based on the analysis of the IBD and control muscle cell secretome revealed that collagen I, decorin and fibronectin are the most abundant ECM components [27]. The major ECM components in the different intestinal layers are summarized in Fig. 2.

Table 1.

Core matrisome of the human colon from MatrisomeDB database [14].

ECM categories	Gene symbols
Collagens	COL1A1, COL1A2, COL2A1, COL3A1, COL4A1, COL4A2, COL4A3, COL4A4, COL4A5, COL4A6, COL5A1, COL5A2, COL5A3, COL6A1, COL6A2, COL6A3, COL6A5, COL7A1, COL8A1, COL9A1, COL9A2, COL9A3, COL10A1, COL11A1, COL11A2, COL12A1, COL13A1, COL14A1, COL15A1, COL16A1, COL17A1, COL18A1, COL19A1, COL20A1, COL21A1, COL22A1, COL23A1, COL25A1, COL26A1, COL27A1, COL28A1,
Glycoproteins	ABI3BP, AEBP1, CILP, CILP2, COLQ, DPT, ECM1, EFEMP1, ELN, EMID1, EMILIN1, EMILIN2, EMILIN3, FBLN1, FBLN2, FBLN5, FBN1, FBN2, FGA, FGB, FGG, FGL2, FN1, HMCN1, HMCN2, IGFBP7, LAMA1, LAMA2, LAMA3, LAMA4, LAMA5, LAMB1, LAMB2, LAMB3, LAMB4, LAMC1, LAMC2, LAMC3, LGI4, LTBP1, LTBP4, MATN2, MFAP2, MFAP4, MFAP5, MXRA5, NID1, NID2, NPNT, POSTN, SRPX, SRPX2, TGFBI, THSD4, TINAGL1, TNC, TNXB, VTN, VWA1, VWF
Proteoglycans	ASPN, BGN, DCN, FMOD, HSPG2, LUM, OGN, PRELP, PRG2, PRG3, VCAN

ECM, extracellular matrix

Fig. 2.

Major extracellular matrix (ECM) components in different layers of the intestine. Delineated are epithelium and basement membrane (green), lamina propria and submucosa (blue) and muscularis propria (red). Each respective box contains the major ECM molecules in the respective layers. Abbreviations: COL, collagen; DCN, decorin; FN, fibronectin; LAM, laminin; TNC, tenascin-C.

Because the ECM was conventionally considered as a blend of secreted proteins with a purely physical support function [16], the active contribution of ECM to intestinal homeostasis has been long overlooked [28]. We know now that ECM remodeling is a vital process that maintains tissue homeostasis by replacing old and damaged ECM molecules with new ones [29], a physiological process that requires a closely orchestrated interaction between ECM proteins, cytokines and various cells. In the intestine, the major source of ECM are the mesenchymal cells, i.e., fibroblasts, myofibroblasts and smooth muscle cells [2]. Under physiologic condition the ECM forms a well-oriented meshwork that mediates critical functions, such as cell attachment, adhesion and enabling normal cellular function. During homeostasis, ECM deposition is physiologically regulated by an ongoing balance between its biosynthesis and breakdown [16]. MMPs, a group of zinc-dependent endopeptidases, are mainly responsible for ECM degradation [30] and, under physiological conditions, these enzymes are kept at low levels and in latent forms [31]. Moreover, the function of MMPs is also closely regulated by another family of proteins called tissue inhibitor of metalloproteinases (TIMPs) [30]. In the gut, both MMPs and TIMPs are secreted by mesenchymal cells [32]. Finally, other proteins such as adamlysins [33] and meprins [34] also cleave the ECM and participate in ECM turnover.

The ECM mediates multiple protective functions in the gut. The intestinal tract can frequently suffer tissue damage due its constant exposure to multiple exogenous substances. When this occurs, the ECM shows remarkable responsiveness allowing for a rapid and complete tissue repair [35]. As evidence, fibronectin and collagen VI, two major components of the ECM, promote the restitution of the intestinal epithelium after injury in vitro [36]. Wounded epithelial cells exhibit an altered distribution of fibronectin and collagen VI, and blocking their functional domains leads to inhibition of epithelial cell migration, confirming the reparative role of these two ECM molecules in epithelial injury [36]. Additionally, the ECM can mediate cell-cell or cell-ECM interactions and regulate cell signal transduction pathways via integrins or other adhesion molecules. Integrins are heterodimeric cell surface receptors responsible for mediating cell interactions with the ECM and other cells [37]. Depending on their binding regions, they are further categorized into arginine-glycine-aspartic acid (RGD) receptors (binding ECM components that contain a RGD domain, like fibronectin), collagen receptors, laminin receptors and leukocyte-specific receptors [37,38] (Fig. 3; adapted from [38]). As evidence for the importance of integrins in intestinal inflammation and repair, laminin is essential for the focal adhesion of epithelial cells to the basement membrane via integrins [39], and the gut-selective integrin, integrin α4β7, has been successfully developed as a therapeutic target for IBD because its blockade limits migration of leucocytes to the gut [40].

Fig. 3.

Classical integrins and their most-established binding partners. The figure delineates the recognition sequence, integrins and binding partners. ECM components are bolded. Abbreviations: GFOGER, integrin-specific glycine-phenylalanine-hydroxyproline-glycine-glutamate-arginine; ICAM-1, intercellular adhesion molecule 1; LDV, Leu-Asp-Val sequence; MAdCAM-1, mucosal addressin cell adhesion molecule 1; MFGE8, Milk fat globule-EGF factor 8 protein; RGD, Arg-Gly-Asp sequence; VTN, vitronectin, VCAM-1, vascular cell adhesion molecule 1.

Alteration of the intestinal ECM in IBD-associated fibrosis

Pathological remodeling of the ECM in IBD-associated fibrosis

In IBD the production and function of the ECM is qualitatively and quantitatively altered due to a multitude of factors involved in the local inflammatory process. Prominent among these factors is the cytokine TGF-β whose level is likely increased as a counteracting factor to inflammation, because of its immunosuppressive and anti-inflammatory activity. To fulfill this task, TGF-β acts also as a potent pro-fibrotic factor that activates intestinal mesenchymal cells and differentiates them into myofibroblasts [41]. Mesenchymal cell-derived myofibroblasts are not the only source of excessive ECM production. Activated fibroblasts can exert the same function and are derived from multiple sources, including epithelial cells via epithelial to mesenchymal transition (EMT) [42], endothelial cells via endothelial to mesenchymal transition (EndoMT) [43], circulating fibrocytes [44], interstitial stellate cells [45] and stem cells from the bone marrow [46]. Even some immune cells contribute to alterations of the ECM [47]. For instance, a recent study revealed that CC chemokine receptor 2 (CCR2)-positive monocytes can inhibit collagen degradation via TIMP-1 [48], with a negative feedback response that leads to increased ECM breakdown. The transcriptional level of MMP-1, -3, -7, -9, and -10 increases in cells from inflamed IBD mucosa with a consequently enhanced proteolytic activity [49]. These multiple concurrent processes are yet to be fully elucidated, but it is clear that the result is the unwarranted deposition of a disorganized ECM culminating in intestinal fibrosis [41].

Alterations of ECM components

Collagens

Collagens are the most abundant ECM proteins of connective tissues, where they may interact with other ECM proteins to create supramolecular aggregates [50]. Sirius red and Masson Trichrome staining are commonly used to assess the content of collagen in the intestinal tissues, a surrogate technique that assists in evaluating the degree of fibrosis in the intestine of humans and animals [48,51-56]. Collagen I and III are considered the key ECM components of intestinal fibrosis [57]. In an early study, collagens extracted from intestinal tissue revealed that collagen I was the major subtype identified in healthy intestine (68%), while collagen III and V comprised 20% and 12% of the collagen content, respectively [58]. Tissue from strictured CD had a higher collagen content, especially collagen V, compared to controls [58]. These findings were further validated by a more recent study, showing increased transcript levels of collagen I, III, IV, V and VI in the terminal ileum of fibrotic CD tissue [59]. Serum levels of peripheral N-terminal propeptide of collagen III are higher in patient with stenotic CD than controls, suggesting not only a higher amount of collagen III, but also a higher ECM turnover. After surgical resection of the stricture serum levels of peripheral N-terminal propeptide of collagen III drop, suggesting that surgical removal of the stricture removed the source of the increased collagen turnover [60]. It is relevant to note that an upregulation of collagens I and III is also found in the intestinal wall of long-lasting UC, defined by a disease duration of longer than 10 years [61] . ECM deposition in UC follows a gradient from proximal to distal [62]. This supports the notion that both UC and CD are progressive diseases [63]. Our own recent study demonstrated that the expression of collagen VI was elevated in CD and UC patients compared to controls. The elevated collagen VI serves as a powerful contributor for the adhesion of T cells, whose retention could contribute to intestinal inflammation [24]. Among all collagens, only collagen I, III, IV, V, VI have been reported in the context of intestinal fibrosis. Other collagens which have been reported to be expressed in other fibrotic diseases, such as collagen X, XVI, XVIII, remain to be explored [50]. Besides, recent prospective study identifies serum collagen components as promising noninvasive biomarkers reflecting endoscopic severity and therapeutic outcomes in UC [64]. As mentioned above, collagens are the integral part of intestinal fibrosis, contributing to tissue stiffness and impaired function. Also, specific fragments of collagen can potentially serve as noninvasive indicators for disease progression in the future.

Glycoproteins

Fibronectin is a high molecular weight glycoprotein exerting multiple regulatory effects on cell migration, adhesion and differentiation [65], and, similar to collagens I and III, fibronectin production increases in intestinal fibrosis [66]. High plasma levels of fibronectin are associated with stricturing disease and those levels drop significantly after surgical resection of the stricture, suggesting a correlation between fibronectin and intestinal fibrosis [67]. As for fibronectin tissue levels, this correlation has been explored in both CD and UC. Fibrotic CD contains a higher transcript level of fibronectin in the terminal ileum compared to non-fibrotic CD [59], and immunostaining for fibronectin indicates increased expression in UC and CD [68].

Laminins are a group of glycoproteins mainly present in the basement membrane [39]. In addition, mast cells, which increase in number in the muscularis propria of strictured CD, colocalize with laminin, implying that the expression of laminin may be related to mast cell activity [69]. The level of serum laminin is elevated in CD and UC patients compared to that of healthy controls [70]. A study investigating levels of laminin subtypes found a deficiency of laminin α2 and an up-regulation of laminin α3 and laminin α5 in the crypts of CD patients, implicating a dysregulation of laminin expression in gut inflammation [71]. Nevertheless, so far, no direct correlation of laminin and a fibrotic disease phenotype has been established.

Elastin is a glycoprotein that provides elasticity and resilience to tissue [72], and it is also found at higher levels in fibrotic than non-fibrotic CD tissue [59]. Another glycoprotein called secreted protein, acidic, cysteine-rich (SPARC), which is expressed in tissues that actively undergo remodeling, was found to be inversely related to experimental murine intestinal fibrosis but not inflammation [73]. Recent study also reported that milk fat globule-EGF factor 8 (MFGE8), which is involved in anti-inflammation and anti-fibrosis, is also elevated in fibrotic CD intestinal tissue [74].

Proteoglycans

Proteoglycans are another important class of ECM molecules that play roles in tissue homeostasis and remodeling [75]. They consist of a core protein covalently linked to glycosaminoglycan (GAG) chains, which interact with ECM components, growth factors, cytokines and cell surface receptors [75]. Biglycan, decorin and fibromodulin, have received special attention in fibrosis associated with chronic intestinal inflammation. In the case of CD, enhanced gene expression of these proteoglycans have been recorded in fibrotic ileal tissue versus its non-fibrotic counterparts, indicating a potential involvement within the fibrotic response [59]. These findings suggest that specific proteoglycans may contribute to the fibrotic remodeling process in chronic intestinal inflammation.

Proteoglycans play multifaceted pathophysiological roles in fibrosis. Decorin has antifibrotic activity by preventing the activation of TGF-β. This inhibition results in decreased canonical Smad2/3 signaling, which is required for inducing expression of profibrotic genes like COL1A1, fibronectin, and α-smooth muscle actin (α-SMA) [76]. Moreover, decorin was demonstrated to cross talk with other receptor tyrosine kinases such as epidermal growth factor receptor (EGFR), and Met, exhibiting a wider control on cell proliferation, apoptosis, and inflammation contributing to the fibrotic response [75]. Biglycan, is another proteoglycan that is closely related to organ fibrosis. For example, biglycan can promote hepatic fibrosis through activating heat shock protein 47 [77]. In addition, biglycan serves as a proinflammatory factor via Toll like receptor (TLR)2 and TLR4, which potentially contribute to fibroblast activation and matrix deposition [78]. Fibromodulin is associated with collagen fibril organization that might contribute to the ECM deposition and the mechanical properties of fibrotic tissue [79].

However, data on the altered expression of proteoglycans in fibrotic intestine tissue is limited. Most of the evidence that proteoglycans participate in fibrogenesis comes from other organs. Further investigations need to be carried out for further illustration about their roles in the gut.

Changes of the core matrisome in the intestinal fibrosis versus normal tissue are summarized in Table 2.

Table 2.

Alteration of the Core Matrisome in inflammatory bowel diseases-related intestinal fibrosis.

ECM components	Descriptions	References
Collagens
Collagen I	↑in the tissue of fibrotic CD patients	[55,59,148,194]
	↑in the tissue of long-lasting UC patients	[68]
	↑in human intestinal fibroblasts isolated from fibrotic CD	[93]
	↑in the tissue of fibrotic colitis mouse model	[48,52-54,66]
Collagen III	↑in the tissue of chronic DSS mouse model	[53,54]
	↑in the fibroblasts isolated from stricturing CD tissue	[195]
	↑in the tissue of fibrotic CD patients	[59,194]
	↑in the plasma in patients who later developed stricturing disease	[196]
	↑in the tissue of long-lasting UC patients	[68]
Collagen IV	↑in the tissue of fibrotic CD patients	[59,194]
Collagen V	↑in human intestinal fibroblasts isolated from fibrotic CD	[148]
	↑in the tissue of fibrotic CD patients	[58,59,194]
Collagen VI	↑in the tissue of fibrotic CD patients	[55,59,194]
	↑in the tissue and in human intestinal myofibroblasts isolated from IBD patients	[197]
Collagen VII	↑in the tissue of fibrotic CD patients	[194]
Collagen VIII	↑in the tissue of fibrotic CD patients	[194]
Collagen XI	↑in human intestinal fibroblasts isolated from fibrotic CD	[148]
Collagen XII	↑in the tissue of fibrotic CD patients	[194]
Collagen XIII	↑in the tissue of fibrotic CD patients	[194]
Collagen XV	↑in the tissue of fibrotic CD patients	[194]
Collagen XXVII	↑in the tissue of fibrotic CD patients	[194]
Glycoproteins
Fibronectin	↑in the tissue of chronic DSS mouse model	[66]
	↑in the tissue of fibrotic CD patients	[59]
	↑in the tissue of long-lasting UC patients	[68]
Laminin	Mast cells localize in the patches of laminin	[69]
Elastin	↑in the tissue of fibrotic CD patients	[59]
SPARC	inversely related to fibrosis in TNBS model	[73]
MFGE8	↑ in the tissue of fibrotic CD patients ↑in human intestinal myofibroblasts isolated from fibrotic CD	[74]
Proteoglycans
Biglycan	↑in the tissue of fibrotic CD patients	[59]
Decorin	↑in the tissue of fibrotic CD patients	[59]
Fibromodulin	↑in the tissue of fibrotic CD patients	[59]

CD, Crohn’s disease; DSS, dextran sodium sulfate; ECM, extracellular matrix; MFGE8, milk fat globule-EGF factor 8 protein; SPARC, secreted protein, acidic, cysteine-rich; TNBS, trinitrobenzene sulfonic acid; UC, ulcerative colitis;

Alterations in ECM mechanical properties

With repetitive cycles of inflammation and repair during the evolution of concomitant inflammatory and fibrotic processes, the progressive increase in ECM accumulation is accompanied by changes of its mechanical properties. Because of its important clinical implications, ECM stiffness is one of the most common properties that has attracted the attention of investigators. Multiple studies show that excessive deposition of ECM increases the stiffness of the tissue in different organs, including the liver [80-82], lung [83-86], kidney [87] and heart [88, 89]. The stiffness of the intestine in colon cancer has also been examined. Kawano et al. used a tactile sensor to measure the stiffness of normal and colon cancer tissue and found that cancer tissue had a higher stiffness than normal tissue, and this difference increased with tumor progression [90]. The stiffness of the intestine in IBD has been assessed in a recent study that utilized a multi-scale indentation system, revealing that the stiffness of inflamed tissue was higher than that of non-inflamed tissue in both the ileum and colon [91]. In a combined study, in a rat model of trinitrobenzene sulfonic acid (TNBS)-induced intestinal fibrosis, stiffness was also increased compared to controls, and stiffness of fibrotic areas from CD intestine stiffness was higher than those from adjacent grossly normal areas [92]. In another human study, a microelastometer was used to measure the stiffness of intestinal tissue from strictured CD patients and normal controls, and found that the stiffness of stricturing CD tissue was nearly 6-fold greater than that of CD uninvolved and control tissue [93].

The recent advances in technological imaging have also contributed to detect and confirm increased stiffness in fibrotic tissues noninvasively. Ultrasound elasticity imaging is a noninvasive way to assess the tissue mechanical properties and has been applied to measure the stiffness of different kinds of tissue [94]. This technique may differentiate fibrotic and normal intestine in CD patients [92]. Another novel technique, real-time shear wave ultrasound elastography (SWE), has also been employed to measure the stiffness of fibrotic bowel wall in CD, with SWE value being the highest in severe fibrosis, followed by moderate and then mild fibrosis [95]. In conclusion, all these combined reports using diverse methodologies confirmed that an increased ECM stiffness is a typical feature of fibrotic IBD intestine.

Impact of fibrotic ECM on intestinal biology and function

Cell morphology

Cell morphology plays a key role in multiple biological functions [96]. Quiescent human colonic fibroblasts (CCD-18co cells) present a rounded shape on a soft ECM, as found in the normal bowel, but, when cultured on a stiff matrix, similar to conditions found in a fibrotic intestine, they acquire a typical stellate morphology [93]. This morphological change reflects an alteration of the biological cytoskeletal architecture, including increased actin filament and focal adhesions. A similar morphological alteration is also exhibited by primary human intestinal fibroblasts [93,97]. Inhibition of Rho/ROCK signaling can suppress the effect of stiffness on cell morphology, and concomitantly on actin stress fiber formation and mature focal adhesions [98]. Although the evidence above come from intestine derived cells, these stiffness-induced morphological changes are not unique to intestinal fibroblasts, and are also observed in other cell types, including hepatocytes [80], dermal fibroblasts [99] and alveolar epithelial cells [100].

Cell migration

Cell migration is another process essential for wound healing and is implicated in tissue fibrosis. Migration is mediated primarily by integrins and pro-fibrotic factors like TGF-β [101]. Mesenchymal cells isolated from fibrotic tissue exhibit an enhanced migration capacity, suggesting that a fibrotic environment drives a migratory phenotype [102]. In the gut, fibronectin acts as a major promoter of colonic fibroblasts migration [103], a response mediated by binding to the cell- and heparin-binding domains of fibronectin [104]. This fibronectin binding site and cell with ECM interaction is not restricted to fibroblasts but is also essential for the migration of colonic epithelial cells [105]. Laminins are a different group of ECM proteins that can also enhance cell migration. As an example, the expression of laminin 5 in the basement membrane is associated with migration and differentiation of enterocytes [106], while the stimulatory effect on enterocyte migration by the chemokine C-X-C motif chemokine ligand 12 (CXCL12) is mediated by the laminin-specific integrins α3β1, α6β1, and α6β4 [107]. Various other ECM components, such as insulin-like growth factor binding protein 1(IGFBP1) [108], MFGE8 [109,110], reelin [111] and tenascin [112] display an enhancing effect on cell migration. This evidence suggest that ECM components participate in cell migration in the intestine.

In addition to the effects of ECM components, other studies suggest that the direction of cell migration can be guided by the gradient of rigidity of the ECM, moving from a softer to a stiffer one, a process termed durotaxis that requires the polymerization of dynamic actin at focal adhesions [113-116]. As mentioned above, reorganization of the cytoskeleton and increased expression of mature focal adhesions induced by ECM stiffness could be the reason for durotaxis in the intestinal fibrosis [93,98]. Although there is still no direct evidence on the effects of ECM stiffness on cell migration in intestinal fibrosis, some studies show that high stiffness inhibits cell migration in other organs [80,117]. Decreased directedness, distance and trajectory velocity in cell migration are observed with increasing stiffness of the ECM, a response regulated by the enhanced expression of integrin β1 and translocation of β-catenin to the nucleus [80]. These combined observations imply that mesenchymal cells migrate from a normal soft to a rigid fibrotic area where they settle down to produce more ECM, and thus creating positive feedback in an ongoing fibrogenesis response.

Epithelial to mesenchymal transition

EMT is a phenomenon in which an epithelial cell is transformed into a mesenchymal cell type under the pressure of long-acting inflammatory, neoplastic, or other biological forces [118]. Chronic inflammation is typical of IBD and, in addition to activating local mesenchymal cells, is also responsible for inducing EMT, therefore creating an additional source of ECM components. The role of EMT in fibrosis has been investigated across multiple organs, including lung [119], liver [120], and kidney [121], indicating that EMT contributes to fibrogenesis by producing fibroblasts of epithelial origin, which subsequently drive ECM deposition and tissue scarring. Whereas growing evidence suggests that EMT plays a key role in the development of intestinal fibrosis associated with IBD [42,122,123], till now, there is few formal in vivo proof that epithelial or endothelial cells fully transition into bona fide mesenchymal cells in intestinal fibrosis. Instead, what has been more convincingly shown is the acquisition of an EMT-like cellular state, often characterized by intermediate phenotypes and gene expression, without terminal mesenchymal differentiation [124,125]. Most data on ECM driven EMT is derived from cancer studies. Examples include cartilage oligomeric matrix protein (COMP), an ECM component, which is uniquely overexpressed in early-onset colon cancer patients, and correlates with the expression of key EMT genes, such as CDH2, FN1, VIM, TWIST1, TWIST2, SNAI1, SNAI2, ZEB1, ZEB2, POSTN, MMP2, MMP9, and COL1A1 [126]. Another ECM component, fibrinogen-like protein 2 (FGL2), plays an oncogenic role in colon cancer aggressiveness by inducing EMT [127]. R-spondin 2 (RSPO2) belongs to the roof plate-specific spondin protein family and activates the canonical Wnt/β-catenin signaling pathway involved in EMT through a positive feedback loop [128]. A recent report implicated RSPO2 in the enrichment of LGR5+ spheroid colon cancer stem cells and promotes metastasis via EMT. In the established model of colitis-associated cancer induced by administering azoxymethane/dextran sodium sulfate (DSS) to adult mice, the deficiency of decorin promotes EMT and colon cancer metastasis [129]. Beside the role of ECM in driving EMT in carcinogenesis, additional evidence indicates comparable effects of ECM in other diseases. For example, connective tissue growth factor (CTGF), which belongs to ECM glycoproteins, mediates high-glucose induced EMT in diabetic nephropathy [130]. CTGF also induces EMT in mouse primary alveolar type II epithelial cells [131]. In carditis, osteoglycin silencing leads to the inhibition of EMT and EndoMT via the Wnt signaling pathway [132]. Lumican, a type of proteoglycan, regulates EMT in ventilation-induced lung injury [133] and also in lens epithelial cells responding to injury [134]. Other matrisome components, like vimentin, argin, decorin or osteopontin, can also regulate the process of EMT [129,135-138]. Some of the ECM genes inducing EMT are summarized in Table 3. While data on ECM driven EMT in IBD is still not available, the presence and upregulation of multiple ECM genes with the capability to lead to EMT makes its involvement in intestinal fibrosis likely. Although no data are available for the effect of ECM stiffness on EMT in intestinal fibrosis, stiffness may also be a driver based on some clues from studies in other organs. For example, mouse primary epidermal keratinocytes seeded into polyacrylamide hydrogels with different stiffness show enhanced TGF-β1 induced EMT [139]. Enhanced EMT induced by ECM stiffness can also be found in alveolar epithelial cells [100] and lung adenocarcinoma [140].

Table 3.

Key epithelial-to-mesenchymal transition related extracellular matrix (ECM) genes.

ECM Name (Gene Symbol)	Function in EMT	References
AEBP1	Expression was positively correlated with EMT metascores in colon adenocarcinoma	[198]
CRIM1	EMT was inhibited in gastric cancer by targeting CRIM1	[199]
COMP	Co-expression of COMP with the EMT markers CDH2, FN1, VIM, TWIST1, TWIST2, SNAI1, SNAI2, ZEB1, ZEB2, POSTN, MMP2, MMP9, and COL1A1 in colon cancer	[126]
CTGF	Mediates high-glucose induced EMT through activation of β-catenin in podocytes	[130]
	Induce in alveolar type II epithelial cells	[131]
CTHRC1	enhance EMT in colorectal cancer.	[18]
	Knockdown CTHRC1 inhibits EMT in glioblastoma cells.	[200]
EDIL3	Promotes EMT through its interaction with integrin alpha(V)beta(3) in cancer cells	[201]
	Regulator of EMT controlling early recurrence of hepatocellular carcinoma.	[202]
EFEMP1	Inhibits the EMT via Wnt/beta-catenin signaling in endometrial carcinoma.	[203]
EFEMP2	Suppresses EMT via Wnt/beta-catenin signaling pathway in human bladder cancer.	[204]
	Inhibit EMT in lung cancer cells	[205]
ELN	Prognostic biomarker correlated with EMT in gastric cancer	[206]
FGL2	The depletion of FGL2 expression inhibited tumor progression and EMT in vitro and in vivo	[127]
FN1	EMT biomarker	[126]
	Regulates TGF-beta1-induced EMT	[207]
GAS6	Prevents EMT in alveolar epithelial cells	[208]
	Attenuates lung fibrosis via inhibition of the EMT	[209]
KCP	Overexpression of the KCP gene blocked EMT in renal tubular cell	[210]
LTBP1	Promotes esophageal squamous cell carcinoma progression through EMT	[211]
	Associated with EMT in gastric cancer.	[212]
MFAP2	Promotes EMT in gastric cancer cells	[213]
	Induces EMT of osteosarcoma cells	[214]
MFAP5	Promotes EMT in Gastric Cancer Cells.	[215]
MFGE8	Regulates TGF-beta-induced EMT in endometrial epithelial cells	[216]
	Induces EMT in gastric cancer	[217]
MXRA5	Silencing or knockout MXRA5 suppressed EMT in pancreatic cancer cells	[218]
NTN1	Blockade of NTN1 inhibits EMT features in endometrial cancer	[219]
	Pharmacological targeting of netrin-1 inhibits EMT in cancer	[220]
POSTN	EMT biomarker	[126]
RSPO2	RSPO2 enriches LGR5(+) spheroid colon cancer stem cells and promotes its metastasis by epithelial-mesenchymal transition	[221]
RELN	Induces EMT via p38/GSK3beta/Snail signaling in non-small cell lung cancer	[222]
SMOC2	Promotes EMT in epithelial cells of renal cell carcinoma origin	[223]
SPARC	Participate on parathyroid hormone-related peptide modulated EMT in colorectal cancer	[224]
SPON2	Increases EMT in lung adenocarcinoma	[225]
SPP1	Promotes EMT activation in prostate cancer	[226]
SRPX2	Enhances the EMT in glioblastoma cells.	[227]
THBS1	Enhances EMT in colorectal cancer	[228]
DCN	Decorin deficiency promotes epithelial-mesenchymal transition and colon cancer metastasis	[129]
TNC	Elevates expression of EMT-related markers by activating NF-κB signaling	[229]
	Enhances EMT in pancreatic cancer-associated fibroblasts	[230]
WISP1	WISP1 silencing protects against EMT of renal tubular epithelial cells in uremia.	[231]
COL10A1	Promotes invasion and metastasis in gastric cancer via EMT	[232]
	Downregulation of COL10A1 inhibit renal fibrosis after AKI by regulating EMT	[233]
COL11A1	Facilitates the occurrence of EMT in pancreatic cancer cells	[234]
COL15A1	Regulates the EMT markers	[235]
COL5A1	Promotes EMT in cholesterol-resistant ovarian cancer cells	[236]
COL5A2	Promotes malignant phenotypes in gastric cancer cell via inducing EMT	[237]
COL6A2	Regulates EMT in in glioblastoma	[238]
COL8A1	Promotes the EMT in glioblastoma	[239]
COL8A2	Facilitates the malignant development of glioblastoma cells via inducing EMT	[240]
ASPN	Promotes pancreatic cancer cell invasion and migration by regulating the EMT	[241]
DCN	DCN deficiency promotes EMT in colon cancer	[129]
ESM1	ESM1 facilitates the EGFR/HER3-triggered EMT in gastric cancer	[242]
LUM	Regulates ventilation-induced EMT	[133]
	Regulates EMT in lens epithelial cells responded to injury	[134]
	Regulates EMT in breast cancer	[243]
OGN	Reverses EMT in colorectal cancer	[244]
	OGN silencing exerts inhibitory effects on EMT in a mouse model of myocarditis	[132]
PODNL1	Regulates EMT in in bladder cancer.	[245]
PRELP	Inhibits colorectal cancer progression by suppressing EMT	[246]
SPOCK1	Induces EMT in drug-induced gingival overgrowth.	[247]
	associated with EMT in lung adenocarcinoma	[248]
	Promotes metastasis in pancreatic cancer via EMT	[249]
	Promotes the metastasis of gastric cancer through EMT	[250]
SRGN	Involves in TGF-beta induced EMT in breast cancer tissue.	[251]

AEBP1, AE binding protein 1; ASPN, asporin; COMP, cartilage oligomeric matrix protein; CRIM1, cysteine rich transmembrane BMP regulator 1 (chordin-like); CTGF, connective tissue growth factor; CTHRC1, collagen triple helix repeat containing 1; CYR61, cysteine-rich, angiogenic inducer, 61; DCN, decorin; ECM, extracellular matrix; EDIL3, EGF-like repeats and discoidin I-like domains 3; EFEMP1, EGF-containing fibulin-like extracellular matrix protein 1; EFEMP2, EGF-containing fibulin-like extracellular matrix protein 2; ELN, elastin; EMT, epithelial to mesenchymal transition; ESM1, endothelial cell-specific molecule 1; FGL2, fibrinogen-like 2; FN1, fibronectin; GAS6, growth arrest-specific 6; KCP, kielin/chordin-like protein; LTBP1, latent transforming growth factor beta binding protein 1; LUM, lumican; MFAP2, microfibrillar-associated protein 2; MFAP5, microfibrillar-associated protein 5; MFGE8, milk fat globule-EGF factor 8 protein; MXRA5, matrix-remodelling associated 5; NTN1, netrin 1; OGN, osteoglycin; PODNL1, podocan-like 1; POSTN, periostin; PRELP, proline/arginine-rich end leucine-rich repeat protein; RELN, reelin; RSPO2, R-spondin 2; SMOC2, SPARC related modular calcium binding 2; SPARC, secreted protein, acidic, cysteine-rich (osteonectin); SPOCK1, sparc/osteonectin, cwcv and kazal-like domains proteoglycan (testican) 1; SPON2, spondin 2; SPP1, secreted phosphoprotein 1; SRGN, serglycin; SRPX2, sushi-repeat-containing protein, X-linked 2; THBS1, thrombospondin 1;; TNC, Tenascin-C; WISP1, WNT1 inducible signaling pathway protein 1

Fibroblast activation, proliferation and differentiation

Fibroblast activation and proliferation are essential for fibrogenesis [2]. Selective ECM components can induce these effects in specific cell types as shown, for example, by the enhanced proliferation of epithelial cells after exposure to fibronectin via an integrin α5β1-mediated pathway [141]. There is no direct evidence showing ECM can regulate the proliferation and activation of fibroblasts, but some studies indicate that one of the potential anti-fibrotic mechanisms for pirfenidone in a radiation-induced rat intestinal fibrosis model is through the inhibition of intestinal fibroblast proliferation through the TGF-β1/Smad/ CTGF signaling pathway [142]. The proliferation of human colonic CCD-18co fibroblasts as well as primary human intestinal fibroblasts increases upon exposure to stiff as opposed to soft ECM [93]. Additionally, increased ECM stiffness can induce the differentiation of colonic fibroblasts into myofibroblasts with an enhanced expression of αSMA, a response accompanied by an increased expression of MKL1 and myocardin-related transcription factor A (MRTF-A) nuclear localization [93]. Beside the evidence from the intestine above, similar examples of fibroblast proliferation and activation can also be observed in other fibrotic condition, and several studies indicate that the YAP/TAZ pathway is a key coordinator of stiffness-induced fibroblast proliferation and activation [85,143-146].

ECM production and deposition

ECM production and deposition are at the center of the intestinal fibrogenic process, which is primarily regulated by modulatory enzymes, including MMPs and TIMPs [30]. However, some other ECM proteins also participate in ECM production and deposition. For instance, fibrinogen, can enhance ECM deposition in the intestine, as shown in fibrinogen knock-out mice which have less collagen deposition in the radiation induced-intestinal fibrosis model [147]. Another matrisome component, the ECM bound growth factor, CTGF, also serves as a direct regulator of collagen expression in colonic fibroblasts [32]. Recent research has found that MFGE8, a member of glycoproteins, can inhibit intestinal fibrosis by regulating the focal adhesion kinase (FAK) pathway via integrin αvβ5, which suppresses fibroblast production of ECM [74].

ECM production and deposition in the intestine can also be regulated by ECM mechanical properties. Both gene expression of MMP-1 and MMP-3 in human colonic fibroblasts decrease when ECM stiffness increases, and the enzymatic activity of MMP-3 is considerably suppressed by a stiff compared to a soft ECM [93]. These findings suggest that an increased ECM stiffness promotes ECM deposition in a fibrotic intestine by the inhibition of proteolytic events. A recent study demonstrated that human intestinal myofibroblasts isolated from normal colon express more MMP-3 with higher stiffness, while those from a fibrotic colon behave in a paradoxical manner, as less MMP-3 expression is detected on exposure to a stiffer ECM, a response that limits ECM degradation [148]. Other than the studies in intestine, some studies demonstrate the ECM production could be mediated by ECM mechanical properties in other organs. It is worthy to point out that inhibition of lipoxygenase (LOX) can restore the activity of MMP-3 in fibrotic myofibroblasts, implying that LOX signaling could be a potential mechanism for stiffness-induced ECM deposition. ECM production by fibroblasts has been explored in human lung fibroblasts placed on ECM with different degrees of stiffness [84]. On a very stiff ECM, fibroblasts enhance the expression of collagen I and III, which can be blocked by the inhibition of Rho/ROCK signaling [84]. Another potential mechanism contributing to the balance of ECM production and deposition could be the YAP/TAZ pathway [85,146].

Modulation of inflammation

ECM remodeling (degradation and de novo synthesis) may have an important effect on the inflammatory process of IBD by altering cellular function through sequestration of cytokines, chemokines, and growth factors [8]. Intestinal epithelial cells are situated on a specialized basement membrane ECM scaffold (mainly type IV collagen and laminins), with abundant collagens, such as type I and type III collagen, situated below the basement membrane as the interstitial matrix [149]. During tissue homeostasis, the ECM is constitutively remodeled by the breakdown of ECM proteins mediated by MMPs and deposition of newly synthesized collagen [19]. In IBD, the ECM remodeling is altered and associated with transmural inflammation [19]. Generally speaking, the ECM molecules, acting as the basement membrane proteins, provide the structure and stability of the epithelium, controlling its behavior (adhesion, polarization, proliferation, and differentiation) by maintaining barrier integrity and normal growth and by interacting with epithelial integrin receptors, and thus promoting the overall maintenance of intestinal homeostasis in response to injury [35,39,150-152]. Meanwhile, interstitial ECM proteins also provide stability and maintain the tissue structural integrity, and regulate mesenchymal and leukocyte homeostasis by serving as a reservoir of cytokines that are stored and released during tissue remodeling [19,35,150]. Syndecan, a transmembrane proteoglycan, mediates cell adhesion and participate in cell proliferation, cell migration and cell to ECM interaction to also promote wound healing and modulating the immune responses [153]. In fact, loss of syndecan-1 increases mucosal permeability by disrupting epithelial tight junctions and causing an increased protein leakage and bacterial influx [154,155,12,13]. Abnormal ECM expressions in response to pro-inflammatory molecules can also change the recruitment of circulating leukocytes by endothelial cells to areas in which inflammatory cells are needed [8].

ECM also regulates inflammation via its effects on intestinal endothelial cells [8]. In IBD, dysregulation of the perivascular ECM promotes adhesion and extravasation of leukocytes and platelets, and release of pro-inflammatory cytokines [156]. ECM components, such as fibronectin, collagens and laminin, interact with cell adhesion molecules expressed on the surface of endothelial cells and modulate their permeability by regulating recruitment of circulating leukocytes and intracellular signaling pathways [8]. Microvascular endothelial cells isolated from IBD patients express high levels of platelet endothelial cell adhesion molecule 1 (PECAM-1, also known as CD31) and the binding of PECAM-1 to fibronectin or collagen increases endothelial permeability and leukocyte transmigration [157-159]. Furthermore, endothelial ECM components, such as hyaluronan, play a direct role in the recruitment of leukocytes in response to inflammatory stimuli [160]. In addition, fragmentation of ECM molecules contributes to inflammation by impairing endothelial barrier integrity and transmigration of immune cells into the submucosal space [161], but further study needs to be carried out for direct evidence in the gut.

In the submucosa, proline–glycine–proline (PGP) derived from collagen by the combined action of MMP8 and/or MMP9, is increased in the intestinal tissue of IBD patients and displays chemotactic effects on neutrophils promoting neutrophilic infiltration in the intestine [162]. Finally, tumor necrosis factor (TNF) binds fibronectin to generate a complex which stimulates monocyte MMP9 expression and chemotaxis and limits the cytokine’s availability and bioactivity to target areas of inflammation [163]. Altogether, all of the above experimental evidence convincingly demonstrates that the ECM remodeling process is profoundly altered in active IBD with important pathophysiological consequences.

To sum up, ECM molecules play a pivotal role in the development and progression of intestinal fibrosis. ECM components modulate cellular process such cell morphology, migration, EMT, fibroblast activation, proliferation and differentiation. Moreover, ECM components regulate their own production and deposition, as well as modulate inflammatory responses. Beyond the biochemical properties of ECM components, their mechanical properties, such as stiffness and fiber orientation could contribute to intestinal fibrosis. These effects share similarities with the fibrosis in other organs.

Extracellular matrix experimental systems

Unlike most traditional experimental systems utilized to study the composition and function of cells, including mesenchymal cells, the ECM, by virtue of being an acellular system, requires special experimental conditions to investigate its components and their respective biological functions. This is particularly true to gain realistic pathophysiological insights into the mechanisms of fibrogenesis as they occur in IBD. Various in vitro models have been created to experimentally study the ECM, and two of them will be discussed below because of their usefulness in mimicking in vivo fibrogenic events.

Three-dimensional cell culture platforms

Traditionally, two-dimensional (2D) culture systems have been applied to in vitro ECM investigations [164]. However, such systems have important limitations, including an artificial and incorrect dimensionality of the ECM, a property that is critical to physiologically regulate surrounding cell behavior [164]. In fact, there is strong experimental evidence showing that human intestinal fibroblasts behave differently in 2D compared to three-dimensional (3D) cultures, including their responses to pro-fibrotic stimuli [165], demonstrating that a 2D system does not adequately mimic the tissue microenvironment where fibrosis develops. Multiple attempts have been made to establish experimental platforms that better reproduce the 3D properties of the ECM and one of the most widely used is that of polyacrylamide gels [166]. In this platform, cells are “sandwiched” in 3D between a layer of fibrillar ECM and an ECM-coupled polyacrylamide gel with defined matrix stiffness [166]. This sandwich gel and its modified versions have been successfully used to culture the human myofibroblast cell line (CCD-18co cells) [93,97,98] as well as primary human intestinal myofibroblast isolated from CD tissue [98,148].

Decellularized intestinal scaffolds

Although 3D matrigel systems can provide some dimensionality and mechanical property of the ECM, they still have limitations when studying intestinal fibrogenesis. First, the native ECM is a complex of various components with specific organization, which artificial 2D or 3D matrigels do not reproduce. Second, the intestine is a unique microenvironment composed by different layers with multiple ECM components of variable stiffness, while matrigel is a uniform complex with single predefined stiffness. To overcome these limitations and better understand the intestinal fibrosis process, novel cell culture platforms are clearly needed. With the rapid development of organ transplantation and bio-engineering in the past few decades, an interest emerged for the use of biomaterial scaffold in gastrointestinal disorders [167]. One of the most popular and successful scaffolds relies on the decellularization of native intestinal tissue. There is no universally accepted definition of decellularization, but this term usually refers to the removal of all cellular and nuclear components while preserving the ultrastructure and components of the ECM [168]. To accomplish this, multiple methods have been employed, including chemicals (sodium deoxycholate [SDC] 165,169,170], sodium dodecyl sulfate [SDS] [171], Triton X-100 [171, 172], et al.), biological enzymes (DNase [169,170], RNase [173], et al.) and physical destruction (freeze and thaw cycle [174], mincing [173]). Based on the system, the process can be further categorized by two main methods, perfusion and immersion [168].

Specific to the gastrointestinal tract, a decellularized scaffold was initially developed from porcine small intestine with sodium azide and SDC [175]. Since then, decellularized intestinal scaffolds have been progressively utilized across many different species, such as mice, rat, cow and sheep, as summarized in Table 4. It is worthy to note that decellularized systems developed from small intestine submucosa have been successfully adopted to study damaged tissue repair in other organs, like the cornea [176], the skin [177], and the urethra [178,179]. Decellularization has been used in small and large intestinal animal tissue, but only a few reports exist on human tissue, with the first preparation of decellularized human colon mucosa reported in 2014 by L. Genovese et al. [172]. Briefly, freshly harvested healthy and cancerous colon tissue was dissected into small fragments and treated with different decellularizing solutions under sterile conditions. Multiple assays including immunohistochemistry, histochemistry and western blot were applied to the treated tissues to validate the efficiency of the method, showing the complete removal of cellular components and the preservation of the ECM structure. Then, the biological function of the decellularized ECM was demonstrated by seeding of cancer cell lines [172]. SDS- or SDC-based decellularized methods were also applied to human intestinal samples [165,169-171,180]. Despite these advances, only one report used decellularization to study intestinal fibrosis [165]. In this report, primary human intestinal myofibroblasts were reseeded into decellularized healthy duodena created with the SDC-based protocol. Results showed that expression of mRNAs related to ECM turnover was distinct when cells were seeded on 3D scaffolds as opposed to conventional 2D cell culture, demonstrating that a well-designed 3D cell culture method can closely mimic the native tissue of intestinal fibrosis [165].

Table 4.

Decellularized scaffold in gastrointestinal tract.

Origin	Tissue	Year	Method	Detergent/ Process	Reference
Mouse	Colon	2019	Perfusion	ddH2O, SDC, DNase	[252]
Rat	Small intestine	2012	Immersion	ddH2O, SDC,	[253]
		2013	Perfusion	DNase, NaCl ddH2O, SDC, DNase	[254]
		2016	Perfusion	ddH2O, Triton X-100, ammonium hydroxide and peracetic acide	[255]
		2017	Perfusion	ddH2O, SDS, Triton X-100, hydrogen peroxide and peracetic acid	[182]
		2018	Immersion	SDS, TritonX- 100	[178]
	Colon	2015	Immersion	Liquid nitrogen, ddH2O, SDS	[174]
		2020	Perfusion	SDS	[256]
Porcine	Small intestine	2005	Immersion	Sodium azide,	[175]
		2011	Immersion	SDC, DNase Sodium azide, SDC	[257]
		2014	Immersion	SDC, sodium azide	[258]
		2016	Perfusion	SDC, ddH2O, DNase, freeze dried and make gelation	[259]
		2019	Immersion	SDC, ddH2O, DNase, freeze dried and make gelation	[260]
		2020	Immersion	SDC, DNase, gamma radiation	[261]
		2022	Immersion	SDC, Peracetic acid, DNase	[262]
	SIS	2011	Immersion	Methanol, chloroform, ddH2O, SDS, peroxyacetic acid, ethanol	[263]
		2011	Immersion	Peracetic acid, ddH2O	[179]
		2018	Immersion	Triton X-100, Ampuwa water, gamma radiation	[264]
		2019	Immersion	Methanol, chloroform, ddH2O, SDS, peroxyacetic acid, ethanol, liquid nitrogen	[265]
		2020	Immersion	Peracetic acid, ddH2O	[173]
		2020	Immersion	SDS	[176]
Sheep	SIS	2018	Immersion	DSS	[266]
		2020	Immersion	SDS, ddH2O	[267]
Bovine	SIS	2017	Immersion	NaOH, peracetic acid, H2O2, Ethanol	[177]
Human	Healthy duodenum	2019	Immersion	ddH2O, SDC	[165]
	Colon	2019	Immersion	SDS, Triton-X	[171]
	Small intestine and colon	2020	Immersion	ddH2O, SDC, DNase, NaCl	[169]
	Normal colon and colorectal cancer tissue	2014	Immersion	EDTA, TrytonX-100, NaCl,Sodium cholate hydrate, HCl	[172]
	Normal colon and colorectal cancer tissue	2018	Immersion	SDC, DNase, ddH2O	[170]
	Intestinal tissue from patients with stricturing CD, UC, and normal controls	2024	Immersion	SDC, DNase, RNase, ddH2O	[74]

CD, Crohn’s disease; ddH2O, double-distilled water; EDTA, ethylenediaminetetraacetic acid; SDC, sodium dodecyl sulfate; SDS, sodium deoxycholate; SIS, small intestine submucosa; UC, ulcerative colitis

Due to the limited research in this field, knowledge on the most suitable decellularized ECM systems for intestinal fibrosis is incomplete due to various challenges. First, intestinal fibrotic tissue contains an ECM with a much denser structure which is difficult for detergents to penetrate. Second, the delicate balance between elimination of all cellular components while maintaining the integrity of ECM is easily lost by an insufficient or an excessive use of detergents. To overcome this technical difficulty, our team, drawing on the experiences of previous researchers, optimized the intestinal decellularization protocol by comparing different decellularization methods and applied this optimized approach to construct a decellularized model for human IBD fibrotic intestinal tissue [74]. However, reliable methods for recellularization of the depleted ECM still need to be optimized. Recellularization is considered the best method to repopulate the decellularized bioscaffolds with different kinds of cells and demonstrate the preserved biological function of the ECM [181]. Most studies reseed cells by a simple drop-on method which is easy to perform, but most cells remain on the surface of the ECM instead of penetrating inside its 3D structure even after prolonged incubation [165,169]. Since fibrotic tissue is thicker and much denser than normal tissue, it might be even harder for seeded cells to penetrate a fibrotic bioscaffold. Microinjection of cells into decellularized ECM scaffold might be a solution, as shown by a report that reseeded endothelial cells, epithelial cells and colonic fibroblasts back into different layers of decellularized tissue cubes [171]. An alternate method might be attempting repopulate cells via the vasculature [182].

Animal models

A huge variety of animal models exist to study and replicate events occurring in healthy and diseased humans [183]. This is also true for intestinal fibrosis and several animal models have been adopted to gain a deeper understanding of its pathophysiology and explore anti-fibrotic therapies [184]. These experimental models can be categorized into spontaneous, gene-manipulated, chemical-, immune-, bacteria-, and radiation-induced and postoperative fibrosis [184]. Among them, the chronic DSS-[55,56,185,186] and the TNBS-induced [187,188] models are by far the most commonly used. Although there are limitations to these two models, including the variability of induction in different species or even substrains of mice and unclear mechanism of chemical-induced inflammation, they are the easiest and most reproducible way to induce intestinal fibrosis [184]. Beside the chemical-induced chronic intestinal fibrosis models, a different mouse strain, the SAMP1/YitFc mouse strain, represents a model of spontaneous CD-like ileitis with both inflammation and fibrosis in the intestine [189]. However, the use of this model is largely challenged by its significant breeding limitations. Another model, the TNF^ΔARE mouse model, which develops spontaneous ileal inflammation, is proved to be an practical model for both mechanistic and therapeutic studies for intestinal fibrosis [190]. As the deletion of AU-rich elements (ARE) in the 3’ untranslated region strengthens the stability of TNF mRNA, leading to over-activation of inflammation, this model well mimics the inflammation-induced intestinal fibrosis in CD. The limitations of this model include the necessity to age the mice to a specific time point for the development of intestinal fibrosis, and its failure to replicate the intestinal fistulas and obstructions in real patients [190]. SAMP1/YitFc and TNF^ΔARE model provide critical complementary value to chemically induced colitis models, which induce spontaneous inflammation-related fibrosis that are better recapitulating the disease progression of CD. Without utilization of chemicals, these models might offer a more disease relevant platform for immune-matrix crosstalk in intestinal fibrosis. A weakness of these two models may be considered the challenging interpretation of mechanisms of inflammation-independent fibrosis.

Furthermore, with the increasing recognition of the role of ECM components in fibrosis, ECM gene-manipulated models have been widely applied to fibrosis research. One example comes from MFGE8 knockout mice, which was carried out to investigate the anti-fibrotic property of MFGE8 in intestinal fibrosis [74]. These models have been more widely used in other organs, including liver, heart, lung and kidney (Table 5). For example, the deletion of collagen VI in mice alleviates cardiac fibrosis after myocardial infarction, suggesting a pro-fibrotic role of collagen VI [191]. With the adoption of these ECM knockout or overexpressing mice, the effects of specific ECM components to particular disease processes, including intestinal fibrosis, can be studied in vivo. The most important strength of these gene-manipulated models is their ability to direct investigation of the function of specific matrix molecules. With the technical maturity of murine gene manipulation, the establishment of the models would be straightforward and cost-effective. Nevertheless, limitations for these gene-manipulated models include: 1. Given the critical role of the ECM molecules in organismal development, some specific gene manipulation might cause systemic impact, potentially on the developmental abnormalities and impaired reproductive capacity in animals. Some would lead to lethality. For these genes, conditional knockout models need to be carried out. Some gene knockout, while not lethal, can lead to varying degrees of congenital defects, affecting the normal physiological processes, which may influence the results in fibrosis research; 2. Some ECM molecules are prominently expressed in specific tissues or organs, introducing tissue or organ specificity into the fibrosis research; 3. Knockout of certain genes may lead to comprehensive changes in the model, such as alterations in metabolism and inflammation, making it challenging to exclude the interference of these systemic responses in fibrosis research. The key pros and cons of these models are summarized in Table 5. To sum up, the animal models for intestinal fibrosis are far from perfect. To take advantage of the different current models, the combination of gene manipulated mice and chemical-induced colitis might be a valid experimental approach for future exploration.

Table 5.

Key extracellular matrix (ECM) gene manipulated mice applied in studies on fibrosis.

Category of ECM	ECM Name (Gene Symbol)	Knockout/ Overexpressing	Organ	Pros vs Cons	Year	Reference
Glycoproteins	Adiponectin (ADIPOQ)	Knockout	Liver	Pros: Direct investigation; Cons: Systemic metabolic disturbances [268]; acute loss of ADIPOQ can be fatal [269]	2003	[270]
					2011	[271]
					2009	[272]
			Heart		2008	[273]
					2011	[274]
					2014	[275]
					2017	[276]
			Kidney		2007	[277]
					2013	[278]
		Overexpressing	Kidney		2018	[279]
	AE binding protein 1 (AEBP1)	Knockout	Liver	Pros: Direct investigation; Cons: Severe developmental defects in general knockout [280] and requires conditional knockout	2018	[281]
			Lung		2009	[282]
	Collagen triple helix repeat containing 1 (CTHRC1)	Knockout	Liver	Pros: Direct investigation; Cons: Directly related to metabolism [283] and osteoblastic bone formation [284]	2019	[285]
			Lung		2017	[286]
	Elastin (ELN)	Knockout	Aorta	Pros: Direct investigation; Cons: Severe developmental defects in general KO, organ-specific limitations (aorta) [287]	2017	[288]
	Fibulin 1 (FBLN1)	Knockout	Lung	Pros: Direct investigation; Cons: Severe developmental defects in general knockout, influencing neural crest cell -dependent development [289]	2019	[290]
	Fibulin 2 (FBLN2)	Knockout	Heart	Pros: Direct investigation; viable and fertile; Cons: Organ-specific limitations (heart, aortic arch vessels and cartilage) [291]	2014	[292]
					2016	[293]
	Growth arrest-specific 6 (GAS6)	Knockout	Liver	Pros: Direct investigation; evidence from human lung fibrosis [294]	2011	[296]
			Heart	Cons: Dual roles in inflammation and fibrosis [295]	2016	[297]
		Overexpressing	Heart		2016	[297]
	Insulin-like growth factor binding protein 3 (IGFBP3)	Knockout	Liver	Pros: Direct investigation; Cons: Organ-specific limitations (kidney, liver, lung, heart, spleen and muscle) [298]; Monoaminergic and synaptic dysfunction [299]	2020	[300]
	Leucine-rich alpha-2-glycoprotein 1 (LRG1)	Knockout	Heart	Pros: Direct investigation; Cons: Critical for vascular development [301]	2016	[302]
	Matrilin 2 (MATN2)	Knockout	Liver	Pros: Direct investigation; Cons: Develop mild skeletal abnormalities [303]	2015	[304]
	Milk fat globule-EGF factor 8 protein (MFGE8)	Knockout	Intestine	Pros: Direct investigation; Cons: Neurogenesis Impairment [305]	2024	[74]
			Heart		2017	[306]
			Lung		2009	[307]
		Overexpressing	Heart		2017	[306]
	Netrin 1 (NTN1)	Knockout	Lung	Pros: Direct investigation; Cons: Critical component of embryonic development with functions in axon guidance, cell migration, morphogenesis and angiogenesis [308]	2016	[309]
		Overexpressing	Kidney		2013	[310]
	Periostin, osteoblast specific factor (POSTN)	Knockout	Liver	Pros: Direct investigation; Cons: Develop abnormalities in fetal mice [311]	2015	[312]
					2016	[313]
			Lung		2017	[314]
			Kidney		2017	[315]
					2019	[316]
			Adipose tissue		2018	[317]
		Overexpressing	Kidney		2018	[318]
	Slit homolog 3 (SLIT3)	Knockout	Heart	Pros: Direct investigation; Cons: Congenital diaphragmatic hernia, kidney agenesis and cardiac defects [319]	2020	[320]
	SPARC related modular calcium binding 2 (SMOC2)	Knockout	Lung	Pros: Direct investigation; Cons: Potentially affect skeletal development[321]	2018	[322]
	Secreted protein, acidic, cysteine-rich (osteonectin) (SPARC)	Knockout	Lung	Pros: Direct investigation; Cons: Serve as a key player in metabolism [323]	1999	[324]
	Secreted phosphoprotein 1 (SPP1)	Knockout	Muscle	Pros: Direct investigation; Cons: Two-sided mediator of intestinal inflammation [325]	2009	[326]
					2019	[327]
	Thrombospondin 1 (THBS1)	Knockout	Liver	Pros: Direct investigation; Cons: Lead to prolonged bleeding, defective thrombosis [328]	2019	[329]
			Lung		2011	[330]
			Muscle		2013	[331]
	Thrombospondin 4 (THBS4)	Knockout	Heart	Pros: Direct investigation; Cons: Spontaneous dystrophic changes [332]	2012	[333]
					2018	[334]
	Tenascin C (TNC)	Knockout	Liver	Pros: Direct investigation; Cons: Impaired fetal lung branching [335]	2007	[336]
			Heart		2010	[337]
					2015	[338]
					2018	[339]
			Lung		1998	[340]
Collagen	Collagen, type VI, alpha 1 (COL6A1)	Knockout	Heart	Pros: Direct investigation; Cons: Develop early-onset myopathy [341]	2012	[191]
Proteoglycans	Biglycan (BGN)	Knockout	Heart	Pros: Direct investigation; Cons: Spontaneous aortic dissection and rupture [342]	2016	[343]
	Decorin (DCN)	Knockout	Kidney	Pros: Direct investigation; Cons: Develop spontaneous intestinal tumors [344]	2009	[345]
	Fibromodulin (FMOD)	Knockout	Liver	Pros: Direct investigation; Cons: Tendon and joint abnormalities [346]	2012	[347]
			Heart		2018	[348]
	Lumican (LUM)	Knockout	Heart	Pros: Direct investigation; Cons: Tendon and joint abnormalities [346]	2017	[349]
					2020	[350]

Challenges and future perspectives

As highlighted in the preceding sections, the deposition of a disproportionate amount of disorganized and dysfunctional ECM is at the core of the intestinal fibrogenic process. In the last few decades, the understanding of its pathogenesis has significantly improved as a result of increased knowledge of the physiological role of the ECM in the healthy gut, its quantitative and qualitative changes in IBD-associated fibrosis, the development of 3D ECM and decellularized tissue systems, and the use of animal models of intestinal fibrosis. Notwithstanding these advances, substantial challenges remain before the role of the ECM in the normal and diseased intestine is fully elucidated, challenges that are primarily due to peculiarities inherent to the gastrointestinal tract.

First, unlike solid organs with a relatively homogeneous parenchymal mass, the gastrointestinal tract is tubular and has multiple anatomical and functionally distinct layers with discrete ECM composition. This anatomical and structural complexity makes it difficult to separate and identify the specialized functions of the ECM in each layer undergoing a fibrogenic response. Second, while hepatic or kidney biopsy-performed sampling provide a representative and generally reliable picture of the overall fibrosis burden, a routine intestinal endoscopic biopsy is only a superficial mucosal sampling. Thus, it is not deep enough to retrieve a full-thickness tissue sample, which is only obtainable from surgical resections commonly performed in advanced disease stages when fibrosis is already evident and quantifiable. Third, although imaging techniques have substantially improved recently [192,193], there is still no reliable non-invasive methodology which can precisely assess the degree and status of ECM deposition in the various segments of the intestine.

Finally, the existing platforms to study the ECM are far from perfect. Because of the gut layered structure, the current decellularization and recellularization protocols are more problematic than those of other organs. In fact, it is challenging to make sure that all cellular components have been removed completely, that specific ECM components have not been damaged, fragmented or modified so that subsequent recellularization studies still reflect real life events. The development of novel technologies, such as 3D bio-printer [135] might be an option to build better systems soon.

Despite the above challenges, a full understanding of the ECM is still a key to intestinal fibrosis and must be aggressively pursued. Specific ECM components may turn out to be valid therapeutic targets, making prevention or therapy of intestinal fibrosis a feasible reality. For example, targeting pro-fibrotic ECM components, such as Collagen VI and CTGF, might alleviate intestinal fibrosis. Some other molecules that exhibit anti-fibrotic properties, like MFGE8 and decorin, might become therapy for intestinal fibrosis. Furthermore, modulating ECM stiffness through inhibition of enzymes like LOX would be a novel therapeutic target in the future. In addition, further investigation can lead to the identification of specific biomarkers detectable in the systematic circulation that may allow fibrotic degree assessment and monitoring. .

Abbreviations

2D

Two-dimensional

3D

Three-dimensional

α-SMA

α-smooth muscle actin

AEBP1

AE binding protein 1

ASPN

Asporin

CCR2

CC chemokine receptor 2

CD

Crohn’s disease

COMP

Cartilage oligomeric matrix protein

CRIM1

Cysteine rich transmembrane BMP regulator 1 (chordin-like)

CTGF

Connective tissue growth factor

CTHRC1

Collagen triple helix repeat containing 1

CXCL12

C-X-C motif chemokine ligand 12

CYR61

Cysteine-rich: angiogenic inducer: 61

DCN

Decorin

ddH2O

Double-distilled water

DSS

Dextran sodium sulfate

ECM

Extracellular matrix

EDIL3

EGF-like repeats and discoidin I-like domains 3

EDTA

Ethylenediaminetetraacetic acid

EFEMP1

EGF-containing fibulin-like extracellular matrix protein 1

EFEMP2

EGF-containing fibulin-like extracellular matrix protein 2

EGFR

Epidermal growth factor receptor

ELN

Elastin

EMT

Epithelial to mesenchymal transition

EndoMT

Endothelial to mesenchymal transition

ESM1

Endothelial cell-specific molecule 1

FAK

Focal adhesion kinase

FGL2

Fibrinogen-like protein 2

FN1

Fibronectin

GAS6

Growth arrest-specific 6

IBD

Inflammatory bowel disease

IGFBP1

Insulin-like growth factor binding protein 1

KCP

Kielin/chordin-like protein

LOX

Lipoxygenase

LTBP1

Latent transforming growth factor beta binding protein 1

LUM

Lumican

MFAP2

Microfibrillar-associated protein 2

MFAP5

Microfibrillar-associated protein 5

MFGE8

Milk fat globule-EGF factor 8

MMP

Matrix metalloproteinase

MRTF-A

Myocardin-related transcription factor A

MXRA5

Matrix-remodelling associated 5

NTN1

Netrin 1

OGN

Osteoglycin

PECAM-1

Platelet endothelial cell adhesion molecule 1PGP: Proline–glycine–proline

PODNL1

Podocan-like 1

POSTN

Periostin

PRELP

Proline/arginine-rich end leucine-rich repeat protein

RELN

ReelinRGD: Arginine-glycine-aspartic acid

RSPO2

R-spondin 2

SDC

Sodium dodecyl sulfate

SDS

Sodium deoxycholate

SMOC2

SPARC related modular calcium binding 2

SPARC

Secreted protein acidic cysteine-rich (osteonectin)

SPOCK1

Sparc/osteonectin: cwcv and kazal-like domains proteoglycan (testican) 1

SPON2

Spondin 2

SPP1

Secreted phosphoprotein 1

SRGN

Serglycin

SRPX2

Sushi-repeat-containing protein: X-linked 2

SWE

Shear wave ultrasound elastography

TGF

Transforming growth factor

THBS1

Thrombospondin 1

TIMP

Tissue inhibitor of metalloproteinases

TLR

Toll like receptor

TNBS

Trinitrobenzene sulfonic acid

TNC

Tenascin-C

TNF

Tumor necrosis factor

UC

Ulcerative colitis

WISP1

WNT1 inducible signaling pathway protein 1

Data availability

No data was used for the research described in the article.