Epithelial wound healing in inflammatory bowel diseases: the next therapeutic frontier

Abstract

Patients with one of the many chronic inflammatory disorders broadly classified as inflammatory bowel disease (IBD) now have a diverse set of immunomodulatory therapies at their disposal. Despite these recent medical advances, complete sustained remission of disease remains elusive for most patients. The full healing of the damaged intestinal mucosa is the primary goal of all therapies. Achieving this requires not just a reduction of the aberrant immunological response, but also wound healing of the epithelium. No currently approved therapy directly targets the epithelium. Epithelial repair is compromised in IBD and normally facilitates re-establishment of the homeostatic barrier between the host and the microbiome. In this review, we summarize the evidence that epithelial wound healing represents an important yet underdeveloped therapeutic modality for IBD. We highlight three general approaches that are promising for developing a new class of epithelium-targeted therapies: epithelial stem cells, cytokines, and microbiome engineering. We also provide a frank discussion of some of the challenges that must be overcome for epithelial repair to be therapeutically leveraged. A concerted approach by the field to develop new therapies targeting epithelial wound healing will offer patients a game-changing, complementary class of medications and could dramatically improve outcomes.

Introduction

If the therapies developed in the past thirty years for inflammatory bowel disease (IBD) represent the fruits of intense research into intestinal mucosal immunology, then the next thirty years may well mark the advent and profusion of therapies stemming from basic research in wound healing. The discoveries supporting this translational medicine could not be timelier. Despite access to an arsenal of medications that suppress the immune system, many IBD patients continue to experience reduced quality of life and poor outcomes that may require surgical intervention. The goal of any medical therapy for IBD, and the universally recognized gold standard that must be achieved to induce long-term remission of disease, is mucosal healing [1–3]. Central to mucosal healing is the restoration of the barrier function of the epithelium through wound healing processes. Experimental models of intestinal inflammation have highlighted important actors, including epithelial stem cells, stromal niche factors such as cytokines, and the microbiome, in the multi-scene play that restores the damaged intestinal mucosa to health. Discoveries of molecular crosstalk between these systems bring hope for a new generation of therapies that directly target epithelial wound repair. These new therapies could complement the current immune-targeting medications. Optimal outcomes in IBD patients will be achieved only after basic research and translational investments into the epithelial repair processes, and the stromal and host-microbe interactions controlling them, have yielded a new class of therapies.

With nearly 7 million people diagnosed with IBD globally [4], developing innovative approaches and interventions is an important public health matter. IBD represents a collection of many diseases that arise from the convergence of multiple factors, which by themselves are usually insufficient to cause disease. They present as two predominant phenotypes, ulcerative colitis (UC) and Crohn’s disease (CD), which have as their hallmark chronic immune activation, mucosal inflammation, and destruction. Current therapies are almost exclusively focused on reducing mucosal inflammation by acting on the immune system, although there is growing interest in modifying the gut microbiome which is typically skewed in patients with active disease. However, the importance of promoting healing of the gut epithelium and other mucosal subsystems in an injurious microenvironment has largely been neglected or understudied. Unsuccessful or inadequately treated chronic disease is often associated with a lack of mucosal healing; impaired healing can give rise to anomalous or compensatory responses. These can have serious sequelae that contributes to the chronicity of disease, treatment failure, and higher relative risk for gastrointestinal adenocarcinoma. Intestinal fibrosis can result in stricturing and fistula formation that are no longer medically manageable. In addition, the microbes comprising the intestinal microbiome must adapt to the inflammatory environment. In doing so, they change their metabolic outputs, and different taxa emerge [5, 6]. The result is a microbial dysbiosis that may sustain mucosal inflammation and further impair wound healing.

And so, the term “mucosal healing,” which refers to the restoration of normal intestinal architecture and homeostasis, has a definition that can be simultaneously narrow and broad and ambitious yet obvious. To be clear, it has not always been the endpoint of clinical treatment for IBD. For many years, it was common practice to assess a patient’s response by clinical indices based on symptomatology. However, there were often disconnects between symptom-based scoring and actual status of disease. Thus, direct endoscopic and histological criteria were developed to assess mucosal healing; these criteria are aggregated into scoring systems with defined cutoffs under which the mucosa are deemed healed (e.g., Mayo endoscopic subscore ≤ 1 [7, 8]). Endoscopic scoring systems, such as the Crohn’s Disease Endoscopic Index of Severity (CDEIS) [9] and Simple Endoscopic Score for Crohn’s Disease (SES-CD) [10], use refined criteria to qualify the depth of the lesions and approximate percentage of surface-area involvement. At the histological level, the Geboes score [11, 12], Robarts Histopathology Index [13], or Nancy Histological Index [14] are used to grade the status of mucosal healing. These systems are similar in that they consider both the status of immune cell infiltration into the mucosa and the morphology of the epithelium. To be considered healed, both the epithelial abnormalities and the immune infiltration into the mucosa must be resolved. The typical histological characteristics of inflamed mucosa and epithelial healing are shown in Figure 1. The highest grades of disease are characterized by crypt abscesses and marked attenuation of epithelium. Lower grades of disease are typified by mucosal infiltration of different types of immune cells, such as neutrophils, plasma cells, or eosinophils, into the lamina propria, and the presence of bifurcating or multifocal crypts. These scoring systems acknowledge that inflammation and epithelial damage go hand-in-hand. One notable assumption is that an epithelium exhibiting normal histological morphology exemplifies restored mucus production and tight junction assembly. Molecular mediators of wound healing have demonstrated key roles in restoring barrier function [15]. However, these aspects are not easily captured by standard hematoxylin-and-eosin staining, and no epithelium can realistically be considered fully healed without proper regulation of cell-cell junctions and the protective mucus layer.

Fig 1.

Histological features of colonic epithelium during murine colitis and mucosal healing. Displayed are transverse colonic sections stained with hematoxylin and eosin. These tissues were obtained from HSP25−/− mice, which exhibit delayed wound healing responses and thereby facilitate visualization of different aspects of injury and healing within the same tissue section. The left panel shows a transverse section of uninjured mouse colon. Note that crypts are uniform and undistorted, and few immune cells are present in the lamina propria. In contrast, the right panel shows the colon 1 4 d after the induction of colitis via a 5-d treatment with 3% dextran sulfate sodium (DSS). DSS treatment caused thickening of the colon and severe epithelial damage. This colon is in various states of wound healing. Restitution, the rapid resealing of eroded mucosa by epithelial cells, is accompanied by a massive immune infiltration into the lamina propria at multiple sites. Lymphoid follicles are enlarged. Regenerative epithelial changes include the formation of 3-dimensional wound channels, the morphological distortion of crypts, including their adoption of bifurcating/branching structures, and crypt hyperplasia, which may be associated with expansion of the progenitor cell zone. Photomicrograph credit: Yun Tao, PhD.

Given the attention already paid to immunomodulation as first-line therapy, it seems that targeting the epithelium during the repair process could lead to an alternate and complementary avenue of treatments. We therefore focus this review on the epithelium-targeted mechanisms and opportunities. However, one should note that targeting other mucosal systems, for example through mesenchymal stem cells, could also indirectly promote epithelial wound healing and therefore broadly restore homeostatic function to the mucosa.

Epithelial repair is critical for breaking the vicious cycle of events underlying IBD pathology. During an active flare, a storm of cytokines and immune cells invades the intestinal mucosa. Although the exact etiology is unknown and could have idiosyncratic origins, this immune response is believed to primarily target gut luminal contents including the commensals comprising the normal microbiome. The epithelium is destroyed in concert with the immune reaction. The breakdown of the epithelial barrier results in the loss of a critical mucus layer (e.g, containing trefoil factors [16]) and ablates homeostatic regenerative functions that normally help to promote wound healing. As a result, the host immune system is further exposed to luminal contents [17], propagating the cycle of inflammation and wounding. It follows that to break this cycle, the antigenic stimulation, the immune overreaction, or the wound healing response need to be modulated. A measure of success has been achieved with immunomodulatory strategies. These include older agents such as mesalamine, corticosteroids, and antimetabolites (e.g., 6-mercaptopurine), as well as newer-generation therapies targeting TNF (e.g., infliximab), integrin subunits (e.g., vedolizumab), IL-12/23 (ustekinumab), and JAK/STAT (tofacitinib). An important limitation of these approaches is that they induce remission in only a minority of patients [18–22]. Thus, there is ample room for therapeutic innovation.

The case for wound healing

Do IBD patients really exhibit defective epithelial wound healing, and can wound healing really be therapeutically leveraged? The evidence that the intestines of IBD patients may have underlying defects associated with epithelial repair comes from a few sources.

• Genetics: Genome-wide association studies [23–25] have indicated risk alleles for both CD and UC in genes involved in intercellular junctions needed for barrier maintenance (reviewed in [26]) and in intestinal cell restitution, the initial migratory step necessary for wound closure. Risk loci encoding genes with plausible roles in wound healing include: 1) PTGER4, the EP4 prostaglandin receptor that is an essential mediator of the epithelial cell-fate change required for restitution [27], 2) ERRFI1, a negative regulator of epidermal growth factor (EGF) receptor signaling [28], and 3) HNF4A, a broad transcription factor with demonstrated roles in intestinal epithelial repair and differentiation [29]. First-degree relatives of CD patients are also more likely to exhibit permeability defects after challenge than spouses of patients or the general population [30–32].

• Histopathology of IBD samples: IBD represents a unique challenge to the intestinal mucosa, in which a loss of tolerance to luminal contents leads to mucosal infiltration of both lymphoid and myeloid cells. This challenge requires an epithelial repair response that can contend with high levels of cytokines and an injurious microenvironment. The inability to resolve persistent ulcers and the appearance of chronic structural changes in crypts suggests an insufficient repair response.

• Preclinical studies in mice: Disruption of cellular pathways regulating epithelial cell migration, proliferation, survival, barrier formation, and differentiation, key functional components of the wound healing process, exacerbates outcomes in experimental colitis models (see below). Cytokines directly modulate epithelial barrier integrity [33–36], and targeted disruption of barrier integrity and accelerates and exacerbates experimental colitis [37, 38].

• Clinical outcomes: Large-scale longitudinal studies have shown that mucosal healing is a clinically significant predictor of long-term response to medical therapy. IBD patients who achieve mucosal healing after treatment with an immunomodulator have longer periods of steroid-free remission and a reduced risk of surgery [39–42]. Even among patients who are considered “healed,” with Mayo scores ≤1, additional benefits in longitudinal outcomes are observed in patients who are fully healed (score = 0) compared to those who are mostly healed (score = 1) [43]. Thus, complete healing is the ultimate goal of medical management of IBD.

Preliminary studies and small-scale clinical trials have suggested that targeting epithelial wound healing would be efficacious. Enemas containing the epidermal growth factor (EGF) peptide, when used alongside mesalamine, induced remission in >80% of UC patients [44]. The EGF receptor pathway is critical for intestinal epithelial wound healing, enhancing the migration, proliferation, and survival of epithelial cells [45–48]. Similarly, irrigation of the distal colon with butyrate, a microbial metabolite that promotes epithelial barrier function [49–52], improved the stool frequency and endoscopic scores of UC patients [53]. As promising as these clinical results are, these therapies target multiple mucosal systems related to IBD pathogenesis, and their effects are not restricted to epithelial wound healing. Thus, their efficacy cannot be taken yet as direct proof that epithelial targeting in isolation would be beneficial in IBD. Other studies of wound healing candidates have shown promise but are ultimately inconclusive due to the small size of the patient cohort [54, 55]. A better understanding of wound healing processes could strengthen the foundation for translational investment into this approach.

Below we summarize additional avenues that could lead to therapies for IBD based on activation of epithelial wound healing. The full development of these therapies and the evaluation of their potential efficacy in IBD patients will provide answers as to whether and how epithelial wound healing can be directly targeted. We have categorized these approaches as related to: 1) epithelial stem cell responses to injury and inflammation, 2) role of cytokines and immune signaling in epithelial wound healing, and 3) microbial signals to generate a favorable environment for host wound repair. A summary schematic of how these systems can work together to mediate wound healing is shown in Figure 2. Furthermore, key therapeutic approaches leveraging wound healing through these systems are listed in the Table. This review is not meant as a full treatment of the scientific principles behind each of these topics; rather, we aim to provide adequate background to contextualize some of the exciting avenues and outstanding issues.

Fig 2.

Schematic of key mucosal tissue systems involved in colonic epithelial wound healing. Interactions between these systems and the epithelium could be therapeutically leveraged to restore normal mucosal architecture and barrier function. The inflammatory process of 1BD induces ulcers and epithelial erosions. At the site of the open wound, crypts adopt a “wound channel” structure, through which intestinal epithelial cells quickly migrate to re-fonn a rudimentary barrier. This process is known as restitution and does not require cell proliferation. Rapid closure of ulcers is critical to prevent gut microbes from further entry into the host. Crypts undergo morphological changes associated with stem cell proliferation to sustain wound closure and to restore the normal pattern of epithelial structure, cellular differentiation, and barrier function. Signals associated with immune cells, wound-associated mesenchyme, and the microbiome promote epithelial wound healing. For example, growth factors, cytokines, microbial metabolites and short-chain fatty acids, and microbial-associated molecular patterns modulate wound healing responses. Further understanding of these cellular and molecular interactions, and distinguishing their pathological versus beneficial effects, will advance potential therapies for mucosal healing in 1BD. Created using Biorender (biorender.com).

Table.

Candidates for epithelial wound healing-targeted therapy in IBD

	Agent	Proposed mechanism of action+	Delivery / Formulation	Preclinical or Clinical Evidence	Example References
Growth factors, restitutive signals, and stem cells	EGF	Epithelial migration, proliferation, and survival through EGFR signaling	Enema	Induces clinical remission in 10/12 UC patients	[44, 48, 77]
	FGF10/KGF-2	Epithelial restitution through FGFR2	Systemic injection	Promotes healing of indomethacin-induced small-intestinal ulcerations in rats	[83]
	TFF3	Essential for epithelial restitution, but details of molecular mechanism unknown	Enema	Mild improvement in symptoms in a minority of 8 enrolled UC patients, comparable to mesalamine	[54]
	IGF-1	Epithelial cell proliferation and goblet cell regeneration through extracellular signal-related kinase (ERK) signaling	Systemic injection	Improved recovery from DSS-induced colitis in rodents	[80, 81]
	HGF	Epithelial proliferation	Systemic injection	Accelerated healing of inflammatory ulcers in rats	[78, 79]
	Mongersen	Migration of epithelial cells through inhibition of SMAD7 (i.e., potentiation of TGFbeta signaling) via antisense targeting	Oral	Clinical remission in ~50% of 83 enrolled CD patients receiving higher doses	[104]
	GB004	Epithelial integrin expression and migration through inhibition of prolyl hydroxylase, resulting in stabilization of HIF-1alpha	Oral	Accelerated wound closure in TNBS-induced colitis in rodents	[88]
	Organoids	Direct engraftment of new epithelium onto injured area	Enema or endoscopic	Improved outcomes in mice with DSS-induced colitis	[107]
Cytokine-inspired	Interleukin 10	Epithelial proliferation through Wnt-inducible signaling protein 1 (WISP1); design of IL-10 variants with specific receptor affinities	Systemic injection	Accelerates closure of biopsy-induced wounds in mice; specific IL-10 variants decouple anti- and proinflammatory signaling	[139, 140]
	Interleukin 2	Activation of Tregs and epithelial restitution	Low-dose systemic injection	Ameliorates DNBS-induced and DSS-induced colitis in mice	[142, 143]
	Interleukin 22	Epithelial wound closure through activation of STAT3	Local gene delivery; treatments with variants to specifically activate STAT3	Attenuates TCRalpha−/− colitis in mice; variants provide selective tissue healing without inflammation	[133, 134, 140]
	STAR2	Selective targeting of TNF receptor 2 to promote colonic epithelial proliferation in injury	Systemic injection	Ameliorates graftvs-host disease in mice	[145, 148]
	IL-36R ligands	Epithelial proliferation and secretion of antimicrobial proteins	Local submucosal injection	Accelerates ulcer closure in biopsy-induced wounds in mice	[132]
	Rivenprost (ONO-4819CD)	Selective activation of EP4 prostaglandin receptor involved in adaptive differentiation of wound-associated epithelia	Intravenous	Improved histological score in 4 UC patients	[27, 55]
Microbe-derived	Butyrate	Promotes epithelial tight junction integrity	Enema	Reduces stool frequency and endoscopic score in trial with 10 UC patients	[50, 51, 53]
	p40 protein from Lactobacillus rhamnosus GG	Activates host EGFR signaling for wound healing	Oral delivery of p40 on hydrogel bead system	Improved preclinical outcomes and epithelial cell survival in DSS-induced and oxazolone-induced murine colitis	[173]
	Faecalibacterium prausnitzii	Preserves epithelial stem cell pool, proliferation, and barrier function	Intragastric delivery of F. prausnitzii strain A2–165	Protects murine colons from radiation-induced damage	[176]
	Microbiome purine reconstitution	Epithelial cell metabolism, proliferation, and mucin secretion	E. coli K12	Monocolonized mice are resistant to DSS-induced colitis compared to germ-free	[174]
	VSL #3	Treatment of dysbiosis	Lactobacilli (4 strains), Bifidobacteria (3 strains), Streptococcus thermophilus	Remission in ~40% of 77 enrolled UC patients	[177, 178, 185]

⁺Agents can have multiple mechanisms of actions complementing their wound healing activity; only the wound healing-relevant mechanism is listed here.

Intestinal epithelial stem cells and wound healing

Mechanisms

Much of what is known about the sequence of events mediating regeneration of the intestinal epithelium comes from mouse models of biopsy punch injury or chemically-induced colitis. Damage through either of these mechanisms induces a temporary loss of epithelial barrier function, reminiscent of human IBD patients. The first stage of epithelial repair is characterized by structural rearrangements of actin filaments within differentiated cells to facilitate rapid cellular migration into the wound. This migratory response, known as restitution, occurs without requiring proliferative changes in the stem cells that normally reside at the base of the crypt. A sheet of cells, each with flattened morphology representative of what has been proposed to be a “wound-associated epithelial” (WAE) phenotype and marked by expression of claudin-4 (Cldn4), emerges from the field of surrounding crypts [56]. Over time the three dimensional shape of surviving crypts extends toward the wound bed and resembles a series of “wound channels” that are derived from horizontal elongations of wound-adjacent crypts [57]. The goal of the restitutive process is to rapidly restore a rudimentary barrier over the ulcer. Unlike wound healing in skin, intestinal epithelial restitution is not believed to involve formation of a scab.

The ”mass balance” of intestinal injury means that the epithelial cell population must eventually be renewed by proliferative activity. In biopsy injury models, upregulation of mitosis is restricted to the epithelial cell population at the base of wound channels and neighboring crypts [57]. The proliferation of epithelial cells occurs with the reshaping of crypts and wound channels: furrows near the base of these structures initiate repetitive fission events that ultimately restore the regular crypt patterning of the mucosa. The position of these furrows is, in part, specified by the location of wound-specific mesenchymal cells expressing Wnt5a [57], which in turn activates pro-repair TGFbeta signaling. Thus, neighboring mesenchymal cells supply cues (e.g., [58]) that promote epithelial repair behaviors and crypt morphogenesis after injury.

Much attention has been given in recent years to addressing whether there is a specialized epithelial stem cell population that is activated during injury. Although the homeostatic turnover of intestinal epithelial cells is sustained by the proliferation of an Lgr5+ stem cell population located at the base of the crypt [59], some studies have suggested a “reserve” or “revival” stem cell population with distinct molecular identity. These cells may be located at the +4 position (i.e., immediately above the base of the crypt) or represent “label-retaining” cells that share properties of both stem cells and Paneth cells [60–65]. In contrast, a competing hypothesis is that the broad plasticity of intestinal epithelial cell fate confers the ability of differentiated cells to revert to a stem-like state during times of physiological challenge [66–72]. This is associated with the adoption of a fetal-like state in the epithelium [73–75]. Unlike the profound epigenetic changes that accompany mitosis and differentiation in fetal development, the differentiation status of an adult intestinal epithelial cell does not appear to be associated with a specific epigenetic configuration; that is, the lack of an epigenetic signature in differentiated epithelial cell types vs. epithelial stem cells essentially confers a fluidity to cell fate specification in the intestinal epithelium [76]. One implication of these findings is that the effective size of the targetable stem cell pool for wound healing could be larger than previously anticipated, as it may include partially differentiated cells that are competent for reversion (de-differentiation).

Therapeutic opportunities

Based on the framework described above, one would predict that signals promoting the “fab five” of epithelial repair - cell survival, migration, proliferation, de-/differentiation, and barrier integrity - would have some positive impact on mucosal healing. One simple approach to enhancing wound healing therapeutically would involve directly treating IBD patients with growth factors or small-molecule regulators shown to enhance these characteristics in mouse models. A variety of bioactive agents and pathways, including EGF [48, 77], HGF [78, 79], insulin growth factor [80, 81], fibroblast growth factors [82, 83], transforming growth factor beta (TGFbeta) [84–86], HIF-1alpha [87, 88], and focal adhesion kinase (a key mediator of cell survival, migration, and barrier function) [89–92] have demonstrated key roles in epithelial wound healing. The efficacy of EGF in a small clinical trial with UC patients [44] lends substantial promise that this approach could be used to improve outcomes in IBD through the enhancement of mucosal healing.

However, the progress with this direct treatment approach has admittedly been slower than anticipated. There are three main reasons for this:

1. Difficulty restricting the effect on the bioactive agent to the epithelium – Receptors and intracellular targets leveraged for epithelial wound healing are found in many other mucosal cell types, especially immune cells. Signals that promote epithelial wound healing behaviors may also promote inflammatory function of immune cells, which may hinder the therapeutic benefit. For example, p38 kinase is essential for epithelial cell migration [93, 94], but it also represents a potent signal involved in the inflammatory pathophysiology of experimental colitis [95–97]. Likewise, EGFR signaling in macrophages may partially drive colitis [98], suggesting that the overall efficacy of EGF-based therapies could be improved if their activity could be skewed away from immune cells. Thus, at least conceptually, the ideal target will have expression restricted to the epithelium, or have complementary roles in immunosuppression and wound repair.

2. Concerns about oncogenesis – Many signaling pathways such as Wnt (APC), Ras, and EGFR that have beneficial roles in mucosal healing are implicated in the pathogenesis of colorectal cancer. However, recent preclinical studies have shown that suboptimally treated inflammation poses a higher risk for cancer than the use of mitogenic agents to aid inflammatory resolution [48, 77]. Expanded preclinical and longitudinal studies will need to be performed for medications targeting repair.

3. Uncertain intellectual property landscape – Growth factors were initially identified in the 1950s and are naturally occurring proteins, limiting their opportunities for intellectual property protection. However, some of these issues could be alleviated by developing novel scalable ways of production, such as using agricultural methods to produce peptides [99, 100], or devising new encapsulation strategies to target these agents to the intestinal mucosa [101, 102]. Moreover, recent approaches have turned towards using novel and patentable chemical species to “lock” enzymes within an activated state or to inhibit the activities of inhibitory proteins within the target pathway. For example, although it failed a phase 3 clinical trial for IBD, a synthetic antisense oligonucleotide to block inhibitory SMAD7 signaling, thereby potentiating reparative TGFbeta signals [103, 104], demonstrates how some creativity can be utilized to generate patentable candidates for clinical studies. Another example undergoing clinical trials is the new compound GB004, which acts as a stabilizer of the hypoxia-inducible HIF-1alpha transcription factor critical for epithelial restitution [87, 88].

The molecular identification of the intestinal epithelial stem cell population, characterization of their niche, and subsequent expansion in vitro as organoids has highlighted a new approach [105–108] to mucosal healing. Its concepts are rooted in tissue engineering. Here, patient-specific organoids are grown from a biopsy of healthy colonic tissue, then endoscopically transplanted to the ulcerated region to directly heal it. A proof of principle was demonstrated in colonic organoids grown from single Lgr5+ stem cells in mice; these fluorescently labeled donor organoids could be successfully engrafted into the colon of a recipient mice afflicted with DSS-induced colitis. The engraftment was associated with accelerated recovery from the acute colitis and provided a long-lasting, self-renewing transplant [107]. Organoids can be grown in culture indefinitely and do not appear to acquire oncogenic mutations, and new techniques have optimized their growth to reduce the number of required exogenous factors and to improve crypt patterning [109–114]. Clinical trials have been initiated using IBD patient-autologous transplants, which would minimize the risk of immunologic rejection.

A complementary source of intestinal organoids is patient-derived induced pluripotent stem cells (iPSCs). iPSCs can be isolated from non-GI tissues and subsequently differentiated to intestinal lineages through a defined and step-wise differentiation protocol that recapitulates regional cues during fetal development [115–117]. The use of iPSCs also enables the cogeneration of blood vessels and enteric neurons [118, 119], important support structures that could facilitate the engraftment and function of the organoid transplant. In organoids grown from either adult biopsied GI tissue or iPSCs, gene editing could be performed to correct genetic defects that may have contributed to the development of IBD. Whether such defects can be identified in most patients and whether the transplanted epithelium will resist future IBD-like injury remain open questions. Accumulating evidence suggests that while both iPSC-derived and adult GI-derived organoids exhibit significant plasticity enabling engraftment, the engrafted tissue may retain epigenetic hallmarks of its original tissue source [108]. In the case of iPSC-derived organoids, their transcriptional profile in culture resembles that of fetal epithelium, but their engraftment is associated with the acquisition of adult epithelial gene expression [120]. The potential long-term side effects of functional mismatches between donor organoids and target engrafted epithelium need to be studied. Moreover, in some patients the pre-existing damage to the epithelium may be too severe to establish robust organoid cultures; these patients would require a different therapeutic approach.

Cytokines and intestinal regeneration

Mechanisms

Although a hyper-inflammatory response is associated with IBD, basic studies have demonstrated the essential role of immune responses in the promotion of wound healing. Many cytokines thought to be central to the pathogenesis of IBD have, in fact, been shown to support epithelial repair in cell culture systems and mouse models. The result is a more-complex set of connections between the various cell types that secrete cytokines and the multitude of effects these cytokines can have on target tissues, including intestinal epithelium, which precludes a simple assignment of whether a particular cytokine is “friend” or “foe.”

Nearly every IBD-associated cytokine has some context in which it can boost epithelial wound healing behaviors. This has been demonstrated in both recent and classic studies of interferons [121], IL-1 [122], IL-2 [122, 123], IL-6 [124], TGFbeta [84, 86, 122], TNF [125–127], IL-15 [128], IL-17 [82, 129, 130], IL-33 [131], IL-36 [132], IL-22 [133, 134], and others, all of which act at some level by promoting epithelial cell migration, proliferation, survival, or differentiation. Common signaling intermediaries that regulate the wound healing response include members of the TGFbeta pathway [84, 86], STAT3/5 [133, 135, 136], and downstream targets of NF-kappaB [137]. Given what is known now about the importance of cytokine signals to intestinal regeneration, it never ceases to amaze that some of the modern therapies which inhibit these same pathways work at all! Indeed, the benefit of an immunomodulating therapy must be considered and balanced against its potential deleterious effects on mucosal healing. For example, inhibition of the IL-17 pathway seemed so promising from the immunologic standpoint but failed clinical trials [138], in part due to this cytokine’s pro-healing effects on the epithelium. These cautionary examples demonstrate the need for more-precise targeting of both the immunologic and the epithelial aspects of the IBD pathophysiological process.

Therapeutic opportunities

Due to the moderate clinical success achieved by anti-TNF therapies and JAK/STAT inhibitors, it seems unlikely that direct treatment with large doses of IBD-associated cytokines will become a primary treatment paradigm for patients who present with severe acute colitis, even if there are some positive effects of these cytokines on epithelial wound healing. In these patients, epithelial repair is not the immediate priority - one does not put out a forest fire by planting new trees. One exception may be administration of interleukin 10, which mediates immune tolerance and also has recently been shown to promote epithelial wound healing in cell lines and mouse models [139]. A recent study has demonstrated how the structure of interleukin 10 can be modified to improve its anti-inflammatory properties [140]. Similar perturbations to the cytokine structure-function relationship have also been recently engineered for interleukin 22 and allow specific activation of downstream STAT isoforms involved in tissue repair [141]. Some gains may also be made by administering a low dose of classically pro-inflammatory cytokines, such as interleukin 2 [142, 143]. Even so, there are additional intricacies in how overlapping immune and wound healing pathways could be activated for mucosal healing. These strategies can be roughly categorized as targeting receptor-specific signals, cell-specific signals, and time/physiology-specific signals.

Cytokine signaling can be distributed downstream across several cellular receptors. These receptors may be linked to different cellular functions which could enable discrimination of pro-inflammatory processes from epithelial wound healing. For example, TNF signaling is executed through two receptors, TNFR1 (Tnfrsf1a) and TNFR2 (Tnfrsf1b). Whereas TNFR1 can have mixed pro- and anti-inflammatory effects (e.g., [144]), selective activation of TNFR2 signaling exerts strong anti-inflammatory effects and induces epithelial repair responses in a variety of autoimmune conditions [145–148]. As another example, prostaglandin signaling through the EP4 receptor acts as a “gatekeeper” in the conversion of intestinal epithelial cells into the migratory WAE phenotype involved in restitution [27], and improves preclinical outcomes [149, 150]. Promising outcomes of UC have been obtained in a small-scale clinical trial [55] with the EP4-selective agonist rivenprost (ONO-4819CD). This strategy of selective receptor targeting could help to reduce activation of classically pro-inflammatory prostaglandin signaling [151].

Recent interrogation of mucosal cell composition using single-cell “omics” techniques has revealed a rich diversity of cytokine-secreting immune and mesenchymal cells that may each have specialized functions, including, possibly, the promotion of epithelial wound healing. As immunosuppressive strategies can have cytotoxic effects on a broad range of cells (e.g., antibody-dependent cellular cytotoxicity) [152], in regards to mucosal healing the current complement of medications may be removing some of the “good” cell types with the “bad.” The varied repertoire of stromal cells is reminiscent of the recent elaboration of different kinds of macrophages, including M1- and M2-polarized subsets, that mediate pro-inflammatory and wound healing-type responses, respectively. Recent single-cell profiling of the IBD-afflicted colon [153] has demonstrated the emergence of a specialized subpopulation of inflammation-associated mesenchymal cells. Intriguingly, this subpopulation expresses IL-33, a cytokine that promotes epithelial proliferation during wound healing via the activation of microRNAs (e.g., miR-320) [131]. Thus, further elaboration of the molecular pathways defining the different stromal cell types involved in IBD-associated inflammation may highlight new approaches to target immune or mesenchymal cells to promote wound healing.

An IBD patient’s clinical history will change over time and is dependent on the effectiveness of the medical treatments they are offered. All of which is to say that the timing of a treatment matters. In this sense, there may be an opportunity to leverage cytokine-based treatments for mucosal healing in one disease context versus another. The colonic mucosa of a patient admitted with a flare of severe acute colitis will be in a fundamentally different biological state than one that has been treated with a powerful immunosuppressive regimen, which will in turn be different from an otherwise asymptomatic individual who is beginning to show early signs of coming out of remission or one who has chronic but mild under-treated inflammation. Animal models of intestinal inflammation have been useful for breaking down the differential roles of cytokines at different timepoints in the natural history of intestinal injury. For example, in acute chemical models such as DSS-induced or TNBS-induced colitis, there is an early phase of injury onset, followed usually by a spontaneous recovery that follows over the span of several weeks. As has been shown in regard to colitis-associated tumorigenesis [77], depending on when growth factor- or cytokine signals are administered, they may have different outcomes. In the future, one can envision that patients may be eligible for low-dose cytokine treatment after certain histological or clinical criteria have been met. This timing-based strategy respects the biological complexity of inflammation and wound healing, and takes advantage of specific windows of time in which certain immune signals could provide a big benefit towards mucosal healing.

Healing through microbial signals

Mechanisms

The intestinal epithelium resides in proximity to trillions of luminal and crypt-associated microbes that comprise the human microbiome. The dynamical nature, adaptability, and critical functions of the microbiome relative to the host mean that any meaningful mucosal healing vis-a-vis epithelial wound repair also needs be accompanied by the restoration of microbiome homeostasis. IBD is almost always associated with a microbiome state known as dysbiosis, in which the overall diversity, composition, stability, and metabolic activities of the microbiome have been perturbed. It is not known whether dysbiosis causes the initial onset of IBD, but it may contribute to delayed healing. Dysbiotic states are associated with the loss of commensals producing important homeostatic short-chain fatty acids (SCFAs) such as butyrate, propionate, and acetate. These SCFAs have shown the ability to promote intestinal epithelial restitution, proliferation, differentiation, and barrier function [49, 154, 155]. As multiple taxa can produce SCFAs, it may be more appropriate to think of microbial metabolic signatures, rather than specific taxa, that are conducive to wound healing.

Preclinical studies in mice have demonstrated that microbial signals are important for intestinal epithelial repair. Germ-free mice exhibit severe exacerbation of DSS-induced colitis [156]. Toll-like receptor signaling, which is activated upon binding by microbe-associated molecular patterns, such as endotoxin/lipopolysaccharide (TLR4), flagellin (TLR5), and unmethylated DNA (TLR9), improves outcomes in experimental colitis through the promotion of wound healing [157–162]. The microbiome can also act to promote wound healing in a localized manner. Specific microbes in proximity to an ulcer activate host epithelial proliferative signaling through a formyl peptide receptor pathway [163, 164]. The spatial topography and organization of the crypt and surrounding mucus also means that the epithelial cells are exposed to different commensal microenvironments, with implications for both host and microbial signaling [165, 166]. Differentiated cells near the top of the crypt metabolize much of the microbially derived SCFAs; as a result, the stem cells at the base of the crypt are relatively untouched by this microbe-derived signal [167]. Likewise, the presence of Paneth and deep crypt secretory cells, which secrete antimicrobial enzymes, at the crypt base changes the nature of the reciprocal signals that characterize the host-microbe relationship [168, 169]. Through symbiosis, the crypt can therefore simultaneously provide an environment facilitating disparate epithelial behaviors along its vertical axis, with proliferative stem cells at the base and differentiated cells capable of restitution at the top, matching the diversity of cell behaviors needed for wound healing.

Therapeutic opportunities

The attractiveness of the microbiome as a therapeutic target for wound healing is rivaled only by the sheer theoretical diversity of the ways it could be targeted. By now, key microbes associated with the IBD-afflicted microbiome have been identified, fueling speculation that adding back so-called “symbionts” could counteract the dysbiosis represented by the presence of “pathobionts” (e.g., [170, 171]). A simple approach could be the administration of a prebiotic or a probiotic compound. There are a few examples of this. Butyrate enemas have been shown to be effective in treating UC [53]. Even single microbial proteins can have profound effects on intestinal epithelial signaling and stromal responses. p40, a protein produced by Lactobacillus rhamnosus GG, activates host epithelial EGFR signaling and mediates wound healing [172, 173]. Restoration of microbe-sourced purines by colonization with purine-competent strains of E. coli protects the colonic epithelium against apoptosis and promotes proliferation and mucosal healing [174]. A microbe commonly depleted in IBD, Faecalibacterium prausnitzi [175], may protect epithelial stem cells during challenge [176] and may thus represent a target for restoration. Beyond single microbial species or metabolites, groups of microbes may be targeted for supplementation with probiotic mixtures. The probiotic mixture known as VSL #3, containing 4 strains of Lactobacilli, 3 strains of Bifidobacteria, and 1 strain of Streptococcus has been shown effective in preventing pouchitis and in treating flareups of UC [177–179], and may do so by partially upregulating expression of host regeneration-associated growth factors [180]. We note that while antibiotics are not classically associated with an epithelial repair response in IBD, in principle the elimination of certain sets of microbes resulting in broad shifts in the community phenotype (e.g., change in IgA status [181] or eliminating oral taxa [5, 182]) could make a more-conducive environment for wound healing.

As with any new therapeutic modality, targeting the microbiome for wound healing has some challenges. First, the details matter. Preclinical studies of the efficacy of certain microbes may apply only to certain strains. Moreover, differences in the structures of human versus mouse microbiomes may challenge the clinical translation of discoveries made primarily in mice. Second, it is not necessarily “easy” to colonize the adult colon with an exogenous microbe, as the microbial community has become adapted to the inflammatory milieu. Successful colonization likely requires pre-treatment with antibiotics to partially clear the microbial community, which may exacerbate dysbiosis. Third, and perhaps a more philosophical question, can one trust the long-term effects of an exogenously introduced microbe? Unlike a protein factor or prebiotic, a living microbe can adapt, mutate, and potentially cause unwanted side effects long after its benefits to mucosal healing have been realized. Ideally we would have some measure of control over the microbe after its introduction. One can envision that this justifies the engineering of microbes with designer molecular circuits that encode complex behaviors [183] to optimize therapeutic delivery and control.

With advances in metabolomic, lipidomic, and proteomic technologies, it should be possible to identify and develop small molecule effectors that promote mucosal healing. The advantage of this approach is that these compounds are no longer dependent on directed colonization or functional properties of probiotics or fecal microbiota transplant, all of which can be unpredictable and difficult to dose. Small molecules, on the other hand, can be administered at optimal dose-responsive levels and targeted to regions in need of mucosal healing. More study will be needed to overcome these potential hurdles and to unlock these new approaches to wound healing.

Concluding remarks

IBD is likely a collection of diseases that are more stratified than simply UC vs. CD. For example, there is growing recognition that colonic CD tends to respond to a different set of therapies than ileal-dominant CD [184]. Combined with the individuality of patient responses and the sheer number of environmental, microbiome, and genetic factors that contribute to risk of disease, it is becoming clear that personalized and precision therapies will be the future. In addition to an approved therapy to enhance wound healing, it will be important to find precise ways to assess and predict healing responses early within the treatment regimen, allowing wound healing therapies to be deployed earlier. The current practice of waiting 4–12 weeks to assess clinical response to therapy is quite hard on the patient; after all, these are real weeks, with real suffering. But with recent advances in our understanding of wound healing and a promising therapeutic pipeline, help is on the way. To be sure, the task at hand is very challenging. The dynamic and precise nature of the wound healing process means that there are many potential failure-points for newly proposed therapies. However, the reward, a generational class of therapeutics that complements emerging immunomodulatory strategies to improve patients’ lives, is well-worth the investment of scientific careers and resources to achieve it.