Development of antifibrotic therapy for stricturing Crohn’s disease: lessons from randomized trials in other fibrotic diseases

Abstract

Intestinal fibrosis is considered an inevitable complication of Crohn’s disease (CD) that results in symptoms of obstruction and stricture formation. Endoscopic or surgical treatment is required to treat the majority of patients. Progress in the management of stricturing CD is hampered by the lack of effective antifibrotic therapy; however, this situation is likely to change because of recent advances in other fibrotic diseases of the lung, liver, and skin. In this review, we summarize data from randomized controlled trials (RCTs) of antifibrotic therapies in these conditions. Multiple compounds have been tested for antifibrotic effects in other organs. According to their mechanisms, they were categorized into growth factor modulators, inflammation modulators, 5-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors, intracellular enzymes and kinases, renin-angiotensin system (RAS) modulators, and others. From our review of the results from the clinical trials and discussion of their implications in the gastrointestinal tract, we have identified several molecular candidates that could serve as potential therapies for intestinal fibrosis in CD.

CLINICAL HIGHLIGHTS

1. Fibrosis associated with inflammatory bowel disease can induce pathological fibrosis in the digestive tract if the intestinal wall sustains chronic injury, which disrupts the production and balance of extracellular matrix.

2. Crohn’s disease persists as an unpredictable fibrotic disease of the digestive tract and therefore poses continuing difficulties in the development of effective antifibrotic therapies.

3. Already developed antifibrotic therapies in organs such as the lung, liver, kidney, bone, heart, and skin can shed light on intestinal fibrosis and ultimately contribute to the progress toward effective antifibrotic therapies for Crohn’s disease.

4. Potential antifibrotic therapies for Crohn’s disease can also be developed through improved understanding of the cellular mechanisms that underlie intestinal fibrosis.

1. INTRODUCTION

Crohn’s disease (CD) is a relapsing and remitting disease that mainly affects the digestive tract (1). Despite the fact that most patients have a purely inflammatory disease course at diagnosis, disease progression to intestinal strictures or fistulas is common over time (2). Approximately 40% of CD patients with ileal-predominant disease experience obstructive symptoms, and 70–80% require intestinal surgery within 20 yr of diagnosis (3). Although in the past two decades there has been increased use of immunosuppressants and biologics in CD, rates of stricturing complications remain high and bowel resection continues to be a common treatment for the disease (4). Recent studies have shown improved mechanistic insights into intestinal fibrosis (2, 5, 6); medical therapeutic options for intestinal strictures remain limited and rely exclusively on anti-inflammatory agents, primarily tumor necrosis factor (TNF) antagonists, which are of unproven efficacy for the treatment of fibrosis. Consequently, patients may require multiple surgeries and/or balloon dilations to manage their stricturing disease, with associated morbidity and costs to the health care system (7, 8). Thus, an important unmet medical need is understanding of the pathophysiology of intestinal fibrosis and development of specific antifibrotic drug therapies. In this regard, clinical observations indicate that fibrogenesis in the intestine is not a one-way chain of pathophysiological events inevitably culminating in obstruction but a reversible process that may be amendable to specific preventive or therapeutic approaches (9). Although understanding of intestinal fibrosis lags far behind that in other organs such as lung, liver, kidney, and skin (5), similar histological features and mechanisms, including activated mesenchymal cells and excessive deposition of extracellular matrix (ECM) (5, 10), are observed in these diseases. Hence new concepts acquired from research in other areas could be valuable in understanding intestinal fibrosis and developing antifibrotic drugs in CD.

This narrative review summarizes the pathogenesis of fibrosis in CD and previous as well as ongoing clinical trials of antifibrotic therapy in organs other than the gut and speculates on how this new knowledge can be applied to the intestine. We therefore conducted a comprehensive search for the relevant clinical trials that were registered on ClinicalTrials.gov from 1990 to September 2019. Randomized, controlled clinical trials (RCTs) that focus on antifibrotic interventions were included in this review. A screen was performed with the following exclusion criteria: 1) noninterventional or observational trials, 2) nonrandomized controlled studies, 3) intervention not related to antifibrotic agents, 4) no targeting of preselected organs (liver, kidney, skin, heart, bone, and lung), 5) lack of clear primary end points. References of clinical trials that had published results were subsequently retrieved from PubMed. The data from these articles, their references, as well as our own experiences form the basis of this review article.

2. MOLECULAR TARGETS AND ORGAN DISTRIBUTION OF ANTIFIBROTIC TRIALS

The screening and selection processes are summarized in Supplemental Figure S1 (all Supplemental Materials are available at https://doi.org/10.6084/m9.figshare.14061767). Two hundred seventy-eight RCTs that focus on antifibrotic intervention were identified. Information on these RCTs was updated until June 1, 2021. The molecular targets of the antifibrotic drugs in the selected clinical trials comprised growth factor modulators (21.2%), inflammation modulators (15.8%), intracellular enzymes and kinases (24.1%), ECM modulators (5.8%), renin-angiotensin system (RAS) modulators (6.8%), 5-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors (3.6%), and others (22.7%). The development phase of these clinical trials is shown in FIGURE 1A.

FIGURE 1.

Overview of all clinical studies included. A: clinical trials grouped by status and mechanisms. B: chart for proportion of clinical trials with antifibrotic therapies based on organ of interest. C: number of clinical trials over time formatted based on the type of investigated drugs. ECM, extracellular matrix; HMG-CoA, 5-hydroxy-3-methylglutaryl-coenzyme A; RAS, renin-angiotensin system.

Liver was the most commonly reported organ (36.0%), followed by lung (34.9%), skin (19.4%), kidney (4.0%), heart (5.0%), and bone (0.7%) (FIGURE 1B). No clinical trials of antifibrotic targets in the gut were reported. As shown in FIGURE 1C, the number of clinical trials evaluating antifibrotic drugs has increased substantially since 1995; however, a plateau was reached in the last 10 yr. Completed clinical trials with potential antifibrotics grouped by mechanism of action, phase of development, and mode of administration are summarized in FIGURE 2.

FIGURE 2.

Completed clinical trials with potential antifibrotics grouped by mechanism of action, phase of development, and mode of administration. ECM, extracellular matrix; HMG-CoA, 5-hydroxy-3-methylglutaryl-coenzyme A; RAS, renin-angiotensin system.

3. TISSUE REPAIR IN THE GASTROINTESTINAL TRACT AND ITS RELATION TO FIBROSIS

The gastrointestinal (GI) tract is a highly vulnerable site for tissue damage because of its close contact with environmental factors such as the microbiome, exogenous antigens from the diet, and a large immune system. The human intestinal mucosa functions beyond purely selective nutrient absorption. Only one layer of simple columnar epithelium separates the intestinal immune system from the environment, which forms a critical barrier against the invasion of bacteria, pathogens, or other antigens (11). This important physiological interface is maintained by various factors including the secretion of mucus, intercellular junctions (e.g., tight junction, subjacent adherens junction, and desmosomes), renewal of epithelial cells, tolerance to commensal bacteria, and immune response against harmful organisms (12). Defects of the intestinal barrier can lead to the development of different diseases, such as inflammatory bowel disease (IBD), celiac disease, food allergy, and irritable bowel syndrome (13).

Tissue repair is a critical process ensuring survival of higher living organisms. To repair an injured intestinal barrier, a series of tissue repair programs are initiated to restore tissue integrity (14). This can in fact be observed in response to a variety of common and short-lived insults, such as nonsteroidal anti-inflammatory drug (NSAID)-induced injury or infectious enteritis, that are followed by complete healing in most instances. One of the major mechanisms of tissue repair is the activation of an inflammatory response, induced by multiple cellular participants, including immune cells (e.g., monocytes, macrophages, and neutrophils) and nonimmune cells (e.g., endothelial cells, fibroblasts). During injury of the intestinal mucosa, polymorphonuclear neutrophils are the first responders and undergo activation and transepithelial migration toward the intestinal lumen, accompanied by the release of inflammatory cytokines and chemokines (12). Immune and nonimmune cells can produce inflammatory mediators (e.g., H₂O₂) and multiple cytokines [e.g., interleukin (IL)-4, IL-13, and transforming growth factor (TGF)-β1] that induce repair (14). Other released cytokines like IL-17A (15) and IL-22 (16) have the capacity of mucosal barrier restoration and protection. Besides the direct effects of the immune response against exogenous pathogens, postmigrated neutrophils that are closely aligned with the apical epithelium can secrete adenosine 5′-monophosphate that is subsequently converted to adenosine, which is an inducer of chloride secretion. This process causes diarrhea and facilitates excretion of potential pathogens (17).

Although a certain degree of inflammation is necessary for mucosal healing, a delicate balance exists between the inflammatory response, healing, and progressive tissue damage. Prolonged healing due to an insufficient inflammatory response can drive tissue damage and fibrogenesis as much as excessive chronic inflammation (18). For example, recruitment of additional neutrophils after injury could support clearance of dead immune cells and microbes through macrophages, which attenuates inflammation and promotes repair of the epithelial barrier (19). Thus, tight control of inflammation during tissue repair is critical. Under physiological conditions, inflammation resolves after the healing is completed through mechanisms such as immune cell apoptosis (20), reverse migration of neutrophils from the inflamed site (21), depletion of chemoattractants (22), and the transition from proinflammatory mediators such as leukotrienes to anti-inflammatory mediators such as IL-10, TGF-β1, or resolvins (23, 24). In addition to inflammation, nonimmune cells are other important participants in tissue repair. For example, epithelial stem cells, residing in the base of the intestinal crypts, undergo enhanced proliferation, migration, and differentiation to replace those dead and damaged cells during injury (25). Activation of myofibroblasts is observed in normal wound healing, which leads to ECM production and deposition (26). Similarly to immune cells, the participation of nonimmune cells needs to be fine-tuned and terminated when tissue repair is complete, again through mechanisms such as cell clearance (27) or deactivation (28). This ensures restitution, in fact in most instances of short-term intestinal injury without scars. However, once this process is dysregulated, as is observed in CD, it can promote a persistent chronic inflammation and eventually drive tissue fibrosis.

4. MECHANISMS OF INTESTINAL FIBROSIS IN CROHN’S DISEASE

Fibrosis is not a disease in its own right but a pathological condition that results from recurrent injury and dysregulated tissue repair, especially in chronic inflammatory disorders (10). In health, the components and structure of ECM are precisely regulated through balanced degradation and production of ECM proteins by matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs). However, this balance is disrupted by repetitive injury of the intestinal wall that ultimately triggers fibrosis (29) (FIGURE 3).

FIGURE 3.

Imbalance of extracellular matrix (ECM) degradation and deposition in fibrosis. MMP, matrix metalloproteinase; TIMP, tissue inhibitor of metalloproteinase.

In line with this concept, CD-associated fibrosis is a complex process characterized by excessive accumulation of collagen-rich ECM produced by different types of mesenchymal cells in response to specific mediators (5). Its pathogenesis is dynamic and multifactorial (30).

First, a dysregulated immune response is one of the key mechanisms in IBD-related fibrosis. Cytokines secreted by T helper (Th)1 cells [interferon (IFN)-γ and IL-12] have inhibitory effects on fibrosis, whereas Th2 cytokines (IL-4, IL-5, and IL-13) have the opposite activities. This shifting balance toward Th2 signifies a switch toward excessive repair and fibrosis (31). In strictured CD, intestinal tissue expression of IL-13 is elevated, leading to a inhibition of fibroblast MMP synthesis and a reduced degradation of ECM (32). This shift in immunophenotype may also apply to the Th17 pathway. IL-17A, which is a key factor secreted by Th17 cells, is elevated in strictured CD tissues compared with nonstrictured segments. It significantly upregulated ECM production by myofibroblasts in CD (33). Distinct responses within the same cell type can also be observed with macrophages. M1 macrophages reduced fibrosis by inducing myofibroblast apoptosis and digesting ECM, whereas M2 macrophages activated fibroblast activation and proliferation via TGF-β1 and platelet-derived growth factor (PDGF) signaling (31). Persistent activation or sustained recruitment of activated macrophages led to a proregenerative and profibrotic phenotype (34). In experimental intestinal fibrosis the number of CD16+ macrophages increased in strictured tissue and correlated with the expression of fibrogenic markers, corroborating the link between macrophages and fibrosis in IBD (35).

Second, a key participant in CD-related fibrosis are nonimmune cells, especially the mesenchymal cells. Although mesenchymal cells include multiple subtypes, like fibroblasts, myofibroblasts, and smooth muscle cells, myofibroblasts are considered the chief effector cell in fibrogenesis (30). Myofibroblasts are derived from different sources, including through proliferation of resident mesenchymal cells (fibroblasts and smooth muscle cells), differentiation from other mesenchymal cell types (pericytes, stellate cells), epithelial and endothelial cells [epithelial- or endothelial-mesenchymal transition (EMT or EndoMT)], bone marrow-derived cells (bone marrow stem cells and myeloid cells), or circulating precursors such as fibrocytes (5, 36) (FIGURE 4). Abnormal persistent activation of myofibroblasts is a hallmark of fibrosis. The subsequent activation of myofibroblasts is driven by multiple agents including TGF-β1 (37), connective tissue growth factor (CTGF) (38), fibroblast growth factors (FGFs) (39, 40), inflammatory cytokines, such as TNF (41), ILs (42), but also endothelins (43), peroxisome proliferator activator receptor (PPAR) (44), and angiotensin (ANG) II (43).

FIGURE 4.

Major cellular sources of myofibroblast in intestinal fibrosis: mesenchymal cells, epithelial cells, endothelial cells, and bone marrow-derived cells.

Third, the intestinal ECM in strictures itself also has a profibrotic effect on mesenchymal cells through a positive feedback mechanism. This is believed to be mediated by tissue mechanoproperties (stiffness). Increased ECM stiffness, even in the absence of any profibrotic mediators, drives mesenchymal cell activation and ECM production (45), and an increased wall stiffness can be found in stricturing CD (46). Another mechanism is the composition of the ECM itself or the presence of ECM-bound profibrotic enzymes (47). Given the fact that CD is a transmural disease process, mesenchymal cells located outside the mucosa and submucosa are also important players in the process of fibrosis. For instance, activated intestinal muscle cells in stricturing CD secreted more fibronectin, which induced the migration of preadipocytes, resulting in the formation of creeping fat (48). As yet another positive feedback mechanism, creeping fat-derived mediators could potentially promote muscular hypertrophy and fibrosis (49).

It is worth noting that environmental factors could also be a trigger for fibrosis. Evidence derived from chronic obstructive pulmonary disease (COPD) patients indicates an activation of mesenchymal cells in human bronchial epithelial cells from smokers compared with nonsmokers, indicating that cigarette smoke-induced EMT participates in airway remolding (50). Cigarette smoke-induced airway remodeling was selectively inhibited by the blockade of the TGF-β1/Smad3 pathway (51, 52). Cigarette smoke extract could also directly activate human lung myofibroblasts by the induction of endoplasmic reticulum stress (53). Although few data on the direct effect of smoking on the promotion of intestinal fibrosis in CD are available, smoking is associated with CD complications, such as fistulizing and stricturing behaviors in the first 8 yr of CD (54). Smoking is a risk factor for the recurrence of strictures after endoscopic balloon dilation (EBD) (55), and in postsurgical CD patients smoking significantly increased the risk of clinical, endoscopic, and surgical recurrence (56). All of these findings support the rationale that smoking could enhance intestinal fibrosis in CD.

Medications may be another factor influencing intestinal fibrosis. Systemic glucocorticoids are frequently used in CD as an effective induction therapy of active CD (57). Interestingly, their effects on intestinal fibrosis are poorly explored. In fibroblasts derived from the other organs or tissues, like skin, bone marrow, amnion, and mouse granulomas, collagen synthesis was suppressed after exposure to glucocorticoids (58–63). Glucocorticoids demonstrated an antiproliferative effect on fibroblasts via glucocorticoid receptor and nuclear factor of activated T cells 5(64). In contrast, evidence from fetal lung fibroblasts indicates that glucocorticoids could synergize with TGF-β in the enhancement of collagen gel contraction (65), a mechanism with potential relevance to CD-related strictures. This effect might be related to the inhibition of endogenous prostaglandin E₂ production by glucocorticoids (66). Although highly effective in the short-term relief of symptoms of intestinal obstruction (67) mediated likely by the anti-inflammatory effect of glucocorticoids, a better understanding of the direct effect of glucocorticoids on intestinal fibrogenesis is desired.

The mechanisms of fibrosis in the digestive tract do not appear to be unique to CD but can also be observed in other fibrotic conditions. For example, eosinophilic esophagitis is an immune-mediated disorder characterized by esophageal infiltration with eosinophils, leading to chronic inflammation and fibrotic strictures (68). Similar to CD-related fibrosis, the fibrosis in eosinophilic esophagitis is initiated by chronic inflammation and driven by overactivated mesenchymal cells (e.g., fibroblasts) (69). In this case, however, the role of eosinophils appears to be more prominent (70) compared with other immune cells like T cells and macrophages in CD-related fibrosis (30). Lines of evidence support the fact that eosinophilic esophagitis is closely linked to allergic susceptibility and exhibits enhanced Th2 cell activity with the overexpression of IL-5 and IL-13, at least partly explaining its fibrotic complications (71). Another example for an intestinal fibrotic condition is collagenous colitis. As one type of microscopic colitis, this disease is characterized by the local accumulation of a layer of collagen underneath the intestinal epithelial layer. Infiltration of mononuclear inflammatory cells is seen in the lamina propria (72). Although the etiology of collagenous colitis is inadequately studied, it is believed to be mediated by immune cells (73), and at least an overlap of gut microbiotal changes with IBD has been noted (74).

Despite the previously reported high number of clinical trials for antifibrotic therapies in other organs as mentioned above, no clinical trials for stricturing CD using a drug with known antifibrotic efficacy have been performed. Multiple lines of evidence suggest that fundamental mechanisms of fibrogenesis are shared across organs (75), and a “central” or “core” fibrosis pathway has been proposed previously (FIGURE 5) (10). Thus, in the following sections we review ongoing and completed trials with antifibrotics in multiple diseases and discuss how their mechanisms of action could be applied to the clinical problem of stricturing CD. FIGURE 6 provides a potential mechanistic overview with selected compounds being tested in fibrotic diseases other than those of the gut and the pathway they modulate.

FIGURE 5.

Core pathological mechanisms of wound healing that are shared across organs. IL, interleukin; TGF, transforming growth factor; Th, T helper.

FIGURE 6.

Potential mechanisms of action of antifibrotic therapies from organs other than the gut. Available therapies modulate growth factor and cytokine signaling, intracellular kinases, and extracellular matrix (ECM) remodeling. AKT, protein kinase B; ASK, apoptosis signal-regulating kinase; CASP, caspase; CTGF, connective tissue growth factor; EMT, epithelial-to-mesenchymal transition; ET, endothelin; ETR, endothelin receptor; FGF, fibroblast growth factor; FGFR, FGF receptor; HIMF, human intestinal myofibroblast; IL12, interleukin-12; IL12R, IL12 receptor; JAK, Janus kinase; JNK, c-Jun NH₂-terminal kinase; MKK, mitogen-activated protein kinase kinase; mTOR, mammalian target of rapamycin; PDGF, platelet-derived growth factor; PDGFR, PDGF receptor; PI3K, phosphoinositide 3-kinase; PPAR, peroxisome proliferator-activated receptor; Rho, Ras homolog family member; RIP, receptor-interacting serine/threonine protein; RXR, retinoid X receptor; STAT, signal transducer and activator of transcription protein; TAK, transforming growth factor-β activated kinase; TGF, transforming growth factor; TGFR, TGF receptor; Th, T helper cells; TNF, tumor necrosis factor; TNFR, TNF receptor; TRADD, TNFR1-associated signal transducer; TRAP, TNFR-associated factor; VEGF, vascular endothelial growth factor; VEGFR, VEGF receptor.

5. ANTIFIBROTIC AGENTS FROM DISEASES OTHER THAN IN THE INTESTINE AND THEIR POTENTIAL APPLICATION IN CROHN’S DISEASE-RELATED INTESTINAL FIBROSIS

5.1. Growth Factor Modulators

5.1.1. Transforming growth factor-β inhibitors.

TGF-β, which belongs to a superfamily comprised of three separate isoforms (TGF-β1, TGF-β2, and TGF-β3) (76), is synthesized as a precursor form and secreted as a latent complex. After binding to ECM, this complex can be activated by specific molecules including MMPs, plasmin, and integrin α_vβ₆, resulting in release of active TGF-β (77). Activated TGF-β binds to TGF-β receptors (TGF-β-R1 and R2) and subsequently induces Smad signaling pathways. Smad2 and 3 are directly phosphorylated and translocate into the nucleus, where fibrosis-related genes are regulated (78). FIGURE 7 shows the classical canonical and noncanonical pathways in TGF-β signaling. In addition to this mechanism, activated TGF-β interacts with mitogen-activated protein kinase (MAPK), Wnt/β-catenin, epidermal growth factor receptor, and mammalian target of rapamycin (mTOR) pathways (37). TGF-β1 plays a central role in multiple fibrogenic diseases through induction of fibroblast proliferation and differentiation, enhancement of ECM production, and inhibition of ECM degradation (79). Several observations are consistent with the notion that TGF-β1 plays an important role in stricture formation. In CD the expression of TGF-β1 and its receptors is increased (80). TGF-β1 signaling pathways enhance the deposition of ECM components (81, 82). The concentrations of profibrotic phosphorylated Smad2 and 3 and TIMP1 are higher in the intestinal mucosa, whereas antifibrotic Smad7 and MMPs are found in reduced concentrations (83). Thus, blockade of TGF-β1 is an attractive strategy for prevention or treatment of CD-related fibrosis.

FIGURE 7.

Overview of canonical and noncanonical pathways of transforming growth factor (TGF)-β signaling in fibrosis. AKT, protein kinase B; ECM, extracellular matrix; LIMK, LMK kinase; PI3K, phosphoinositide 3-kinase; Rho, Ras homolog family member; SMA, smooth muscle actin; TGFR, TGF receptor.

5.1.1.1. pirfenidone.

Pirfenidone is an oral small molecule targeting the activity of TGF-β, TNF (84), and nuclear factor (NF)-κB (85). The inhibitory effects of pirfenidone on fibroblast proliferation and collagen deposition have been shown in several diseases (86–90). The earliest clinical proof for the benefits of pirfenidone in idiopathic pulmonary fibrosis (IPF) came from a phase II placebo-controlled, double-blind study of 107 Japanese patients (91), which showed a statistically significant increase from baseline in a 6-min steady-state exercise/walk test at 6 and 9 mo relative to placebo. These encouraging results led to the phase III clinical trial that evaluated 275 Japanese patients. In this study both the mean change in vital capacity at 52 wk and progression-free survival time were improved in patients assigned to pirfenidone (92) compared with those who received placebo. These findings are supported by other findings from two concurrent trials (CAPACITY studies, NCT00287716 and NCT00287729) conducted in the United States, Europe, and Australia.

In study 004 (NCT00287729), pirfenidone at 2,403 mg/day reduced the decline in percentage of forced vital capacity (FVC) compared with the placebo group at week 72 (−8.0% vs. −12.4%, P = 0.001). Although the difference in the decline of FVC could not reach significance at week 72 between pirfenidone and placebo groups, in study 006 (NCT00287729) (−9.0% vs. −9.6%, P = 0.501) an evident treatment effect was observed from week 12 until week 48 (P = 0.005). Furthermore, the pooled analysis of these two studies showed a benefit of pirfenidone treatment at 2,403 mg/day, with predicted FVC mean change of −8.5% in the pirfenidone group and −11.0% in the placebo group (P = 0.005). A total of 779 patients were enrolled, and the results demonstrated a significant reduced mean decline in percent predicted FVC compared with placebo at week 72 (93). Similarly positive results were identified by additional phase III trials (ASCEND study, NCT1366209) (94), leading to approval of pirfenidone as the first treatment of IPF in Europe and the United States and in fact the first approval of an antifibrotic drug across organs. Further evidence for the antifibrotic effects of pirfenidone has been provided by positive studies in improving pulmonary fibrosis in patients with Hermansky–Pudlak syndrome (NCT00001596) (95) and hepatic fibrosis in hepatitis C patients (NCT02161952) (96). Other RCTs are evaluating the potential benefits of treatment in skin fibrosis [keloid (NCT02823236) and systemic sclerosis (NCT03068234)] and cardiac failure (PIROUETTE Trial, NCT02932566) (97). Although mechanistic confirmation of the effects of the proposed inhibitory actions on fibroblast proliferation and differentiation in humans is currently lacking, the data from translational substudies of the described RCTs are anticipated in the near future.

5.1.1.2. prm-151.

Pentraxins (PTXs), a protein family that includes C-reactive protein, PTX2, and PTX3, contribute to both immune defense against bacterial pathogens and tissue repair (98). A member of the pentraxin family, PTX2, also known as serum amyloid P (SAP), has demonstrated antifibrotic effects including inhibition of fibrocyte differentiation and regulation of macrophages and peripheral blood mononuclear function based upon in vitro studies using cells from pulmonary fibrosis and myelofibrosis patients (99). PTX2 has been shown to reduce fibrosis in TGF-β-driven experimental lung fibrosis (100). The first clinical study with intravenous PRM-151 (a recombinant human PTX-2), which was performed in a small number of healthy volunteers and IPF patients, demonstrated that PRM-151 was associated with a reduction of fibrocyte number by 30–50% in IPF patients, suggesting an antifibrotic effect (101). Improvement of predicted FVC and 6-min walking distance relative to placebo was subsequently evaluated in both short-term and long-term studies. PRM-151 was assessed in two randomized double-blind phase II trials (NCT01254409 and NCT02550873) (102–104). In NCT01254409, improved pulmonary function, including FVC and forced expiratory volume in 1 s (FEV₁) on day 57 was noted, whereas the control group exhibited worsening of pulmonary function (102). The effects of PRM-151 were further confirmed in NCT02550873, showing a reduced change of FVC percentage of predicted value from baseline in the treatment group compared with the placebo group {difference, +2.3 [90% confidence interval (CI), 1.1–3.5], P = 0.001} at week 28 (103). The effect of long-term treatment persisted at week 52 (104). PRM-151 was well tolerated; the most common adverse events were cough, fatigue, and nasopharyngitis (102, 103). A phase II study on primary myelofibrosis trial is currently underway (NCT01981850).

5.1.1.3. avotermin.

TGF-β3 has an importance similar to that of TGF-β1 in wound healing processes. However, despite a related molecular structure, TGF-β3 has antifibrotic potential, in contrast to TGF-β1, which is profibrotic (105). In contrast to TGF-β1, which has been linked to keloid formation, TGF-β3 reduces the formation of skin scar after injury (106). Evidence shows that TGF-β3 may reduce ECM deposition by multiple pathways, such as bone morphogenic protein signaling and the Wnt pathway (107, 108)^,. These observations provide the rationale for evaluation of TGF-β3 as an antifibrotic agent, and multiple clinical studies have proved this in humans. The human recombinant form of TGF-β3 (avotermin) was first shown to reduce skin fibrosis by decreasing excessive ECM, reducing proinflammatory cytokines, and changing cellular behaviors in both in vitro and in vivo models (109, 110). Subsequently, in three double-blind, placebo-controlled clinical phase I trials (NCT00629811, NCT00847795, and NCT00847925), avotermin was intradermally administered to full-thickness skin incision margin in healthy volunteers at different concentrations (111). At the end of months 6, 8, and 12, patients assigned to avotermin injection achieved significantly lower cutaneous scarring scores on a visual analog score relative to placebo. Abnormal orientation of collagen fibers was also reduced in subjects who received the active drug (111). These findings showed the potential of intradermal avotermin to provide an accelerated and permanent improvement in scars after surgery, which was later evaluated in three phase II studies, which assessed total scar score difference between avotermin and placebo in NCT00430326 (at dose of 500 ng/100 μl per linear cm, 16.49 mm, P = 0.036), NCT00432211 (at dose of 200 ng/100 μl per linear cm, 21.93 mm, P = 0.04), and NCT00594581 (at dose of 200 ng/100 μl per linear cm, 7.5 mm, P = 0.003) (112–114). For details on other TGF-β inhibitors, see Supplemental Table S1.

5.1.2. Connective tissue growth factor inhibitors.

Connective tissue growth factor (CTGF) is a member of the CNN (cysteine-rich angiogenic inducer 61, CTGF, and nephroblastoma overexpressed protein) family, which serve as regulatory proteins for ECM production (115). CTGF binds to ECM components and mediates cellular adhesion and motility, a process that modulates ECM turnover (116). Also, CTGF can induce myofibroblast differentiation and activation (117), which is consistent with the well-established role it plays in tissue remodeling and fibrosis in various organs (115). For example, administration of CTGF led to increased fibrosis in a “fibrosis-resistant” mouse model of bleomycin-induced lung injury (118). Furthermore, CTGF has been shown to act as a cofactor for TGF-β1-induced renal fibrogenesis (116). On the basis of these observations, CTGF can be considered a potential target for antifibrotic agents.

5.1.2.1. pamrevlumab.

Pamrevlumab (FG-3019) is a human monoclonal antibody directed against CTGF. In a multicenter phase II study (PRAISE Trial, NCT01890265), patients with IPF were randomly assigned pamrevlumab or placebo. Pamrevlumab was shown to significantly reduce the decline in percentage of predicted FVC (−2.9% with pamrevlumab vs. −7.2% with placebo, P = 0.033) and to prevent disease progression at week 48 compared with placebo (10.0% with pamrevlumab vs. 31.4% with placebo, P = 0.013) (119). Therapy was well tolerated, and no substantive adverse events were identified in this short-term study. A phase III study in IPF (NCT03955146) is currently underway. The only other clinical study reported with this agent was a phase II study conducted in patients with chronic hepatitis B-related liver fibrosis (NCT01217632) that was terminated prematurely. No details regarding the results of this study could be found. For other connective tissue growth factor inhibitors, see Supplemental Table S1.

5.1.3. Fibroblast growth factor 19 analogs.

FGFs are a family of polypeptides with multiple biological functions, including wound healing and tissue repair, which bind to FGF receptor (FGFR) (120). Among FGFs, FGF19, an endocrine GI hormone, participates in bile acid metabolism and is considered a potential drug for liver fibrosis (120). NGM282 is an analog of FGF19 that has been shown to improve histological features and periductal fibrosis in an animal model of primary sclerosing cholangitis (PSC) (121). A RCT has evaluated this molecule in human PSC. This study, which evaluated patients for 12 wk of treatment with NGM282, did not show decreased serum alkaline phosphatase (ALP) concentrations relative to placebo. However, drug therapy was associated with significant decreases in fibrosis biomarkers including enhanced liver fibrosis score and serum fibrosis biomarkers, neoepitope-specific NH₂-terminal propeptide of type III collagen (Pro-C3) (122). Similarly, recent trials demonstrated that NGM282 reduced nonalcoholic fatty liver disease [nonalcoholic steatohepatitis (NASH)] activity score and Pro-C3 and enhanced liver fibrosis score in comparison to placebo (123, 124). Recombinant PEGylated human FGF21 (pegbelfermin) has been investigated for the treatment of NASH (FALCON 1, NCT03486899; FALCON 2, NCT03486912), showing reduced hepatic fat and improved metabolic factors and biomarkers of hepatic injury and fibrosis (125). Additional details on other growth factors in clinical development as antifibrotic targets can be found in Supplemental Table S1.

5.1.4. Potential applications of growth factor modulators in Crohn’s disease-related intestinal fibrosis.

Multiple growth factors are involved in the development of fibrosis, among which TGF-β1 is considered a central player. TGF-β1 and its receptors are elevated in intestinal tissues from CD patients (80). TGF-β transcripts, phosphorylated Smad2/3 (p-Smad2/3), and TIMP-1 protein were reported to be higher in both mucosa and myofibroblasts from strictured CD (83). TGF-β1 exerts profibrotic effects on intestinal myofibroblasts, including enhancement of cell activation, proliferation, migration, and ECM production (126). It induces fibrosis by promoting EMT and EndoMT (37). In vivo, vaccination targeting TGF-β1 ameliorated intestinal fibrosis in a chronic colitis model (127). Transferring the TGF-β1 gene to the mouse colon resulted in obstructive intestinal fibrosis (128). In dextran sulfate sodium (DSS)-induced murine colitis, reduced collagen deposition could be found with oral pirfenidone administration (129). Pirfenidone also inhibited intestinal fibroblast proliferation, differentiation, and collagen production (129–131).

Smad7, which was reported to be decreased in stricturing CD (83), works as a competitive inhibitor of the TGF-β/Smad signaling pathway. Antisense oligonucleotide targeting Smad7 (mongersen) induced remission and clinical response in active CD (132) but was found not to be efficacious in later-stage development programs. Since Smad7 is a negative regulator of p-Smad3, the risk of developing intestinal fibrosis by blocking Smad7 led to concerns among clinicians. A phase I open-label trial showed that short-term mongersen treatment was not associated with the development of small bowel stricture (133). The same was shown in a murine chronic colitis model. It is interesting to note that oral administration of Smad7-specific antisense oligonucleotide did not increase the risk of intestinal fibrosis but reduced collagen deposition and diminished fibrosis in chronic 2,4,6-trinitrobenzenesulfonic acid solution (TNBS)-induced colitis (134). This might be explained by the fact that p-Smad3 levels were not elevated despite the treatment with Smad7-specific antisense oligonucleotide, likely because of diminished production of TGF-β1 in the colitis model (134).

A major safety concern for the inhibition of TGF-β is the potential to induce carcinogenesis, which might be related to its antiproliferative activity on most epithelial cell types. This was observed in clinical trials not only in pan-TGF-β blockage (like fresolimumab) but also in TGF-β1-selective blockade (like metelimumab), leading to the termination of the clinical studies (10). Fortunately, this adverse effect was not documented with the use of pirfenidone (94). Careful thought also needs to be given to the risk for exacerbation of intestinal inflammation upon TGF-β1 blockade, since TGF-β1 has immunoregulatory properties (135). This effect may, however, be dose and context dependent. The antifibrotic efficacy and safety of TGF-β1 inhibition need further investigation in clinical trials. Also, strategies to de-risk potential side effects should be considered by carefully crafted clinical development programs and through optimizing the dose and duration of TGF-β1 inhibition, using drugs in combination with other anti-inflammatory therapies and use of local delivery systems with specific administration sites, hence reducing systemic exposure.

CTGF, another critical growth factor in intestinal fibrosis, is elevated on the mRNA level in severely fibrotic CD tissue. This correlated with an increased expression of TGF-β1, collagen I, and fibronectin (136). In CD, CTGF localized to fibroblasts within the submucosal layer (137). Moreover, CTGF in myofibroblasts is induced by TGF-β1 and TNF in CD (138), indicating the potential interactions between these molecules in profibrotic pathways. Increased expression of CTGF was found in the acute DSS-induced colitis model (139); however, data in chronic fibrosis models are lacking. align="center"

FGFs serve as modulators for intestinal fibroblast proliferation in both health and IBD (140, 141). FGF-10 induced collagen expression in intestinal myofibroblasts but failed to show effects on experimental intestinal fibrosis (142). In CD patients, serum levels of basic FGF correlated with bowel wall thickness (143). Neutrophils producing basic FGF were observed in ulcerative colitis (UC) stenotic tissue (144). Tocotrienols, a diet constituent with vitamin E activity, inhibited fibroblast proliferation induced by basic FGF in IBD but not in healthy control subjects (145).

In summary, specific growth factors such as TGF-β1 and CTGF induce fibrosis in different organs, including the gut. On the basis of these observations, antifibrotic agents targeting growth factors have entered the clinic in several fibrotic diseases, among which pirfenidone and pentraxin have the highest potential as an antifibrotic therapy for CD intestinal fibrosis.

5.2. Inflammation Modulators

5.2.1. Tumor necrosis factor inhibitors.

TNF is a proinflammatory cytokine produced by multiple cell types that stimulates fibroblast proliferation, modulates cellular chemotaxis, and increases fibrogenesis (146). Evidence from animal models of lung fibrosis and obliterative bronchiolitis showed that treatment with TNF antagonists significantly reduces pulmonary fibrosis (147, 148). Similar results have been demonstrated in experimental models of fibrosis in other organs (149–151).

A limited number of human studies have evaluated TNF antagonists as antifibrotics. In a phase III trial (NCT00385086), infliximab and placebo were randomly administered to 35 patients (18 infliximab and 17 placebo) with sciatica from postoperative peridural lumbar fibrosis. No benefit of infliximab therapy was demonstrated. The proportion of patients with 50% reduction in sciatica pain was not significantly different between the infliximab and placebo groups (17.6% vs. 27.8%, respectively, P = 0.69) (146).

Pentoxifylline, a small molecule that has been used in vascular diseases for many years, was reported to decrease serum TNF concentrations (152), and potential antifibrotic effects have been shown in both in vivo and in vitro studies (153–157). However, a phase III clinical trial (NCT00119119) of pentoxifylline in liver fibrosis patients was terminated. Therefore the interpretation of this result and extrapolation to any potential benefits of TNF antagonists in stricturing CD are difficult. Possible explanations for the negative findings include lack of specific relevance to the specific pathophysiology of hepatic fibrosis, inadequate drug exposure, and an unresponsive patient population. The only relevant clinical experience in CD was a small open-label study of pentoxifylline in corticosteroid-dependent disease that failed to show a symptomatic benefit (158). We are unaware of any active development of pentoxifylline for stricturing CD.

5.2.2. Interleukin modulators.

ILs are a group of immunomodulatory cytokines that mediate several different biological responses in cells and tissues, especially cell growth, differentiation, and activation during inflammation (159). Different families of ILs exert distinct effects, including in fibrosis, which makes them potential targets for antifibrotic therapy.

5.2.2.1. interleukin-12.

IL-12 is a proinflammatory cytokine produced by multiple immunoregulating cells (160). IL-12 can reduce pulmonary fibrosis induced by bleomycin (161). Interestingly, in a phase II trial (NCT01389973), ustekinumab, a human monoclonal antibody targeting IL-12 and IL-23, could modestly decrease enhanced liver fibrosis scores in primary biliary cirrhosis (PBC), but only 20 individuals were registered in this trial (162).

5.2.2.2. interleukin-10.

IL-10 is a cytokine that suppresses excessive immune responses and defense against tissue damage (163). IL-10 also showed an antifibrotic effect both in vitro and in vivo (164, 165). A phase I (NCT01115868) and a phase II (NCT00984646) clinical trial using recombinant IL-10 on skin scars have been registered.

5.2.2.3. interleukin-4 and interleukin-13.

IL-4 and IL-13 are mainly secreted by activated Th2 cells and elicit comparable biological responses by sharing a common receptor chain (166, 167). Previous studies demonstrated that IL-4 and IL-13 could promote fibrosis through IL-4Rα/STAT6 and upregulating the TGF-β/Smad3 pathway (167). Inhibition of the IL-4/IL-13 pathway has been attempted as an antifibrotic strategy but without showing efficacy to date. For example, a phase II trial showed that SAR156597, a bispecific IgG4 antibody against IL-4 and IL-13, showed no benefit to prevent decline of the FVC in patients with IPF (168). Similar results were found with a different IL-13 inhibitor, tralokinumab, in IPF, which led to early termination of the study (169).

5.2.2.4. interleukin-36.

IL-36 is a newly characterized member of the IL-1 family that has demonstrated regulatory effects on immune responses in multiple chronic diseases, including psoriasis, arthritis, obesity, and IBD (170). IL-36 receptor stimulation promoted, whereas its inhibition suppressed, tissue fibrosis in lung, kidney, heart, and pancreas. IL-36 inhibition as a potential antifibrotic therapy has hence drawn the attention of researchers, including for the indication of stricturing CD (170, 171). Blockade of IL-36 ameliorated experimental intestinal fibrosis (172). Spesolimab, a monoclonal antibody against IL-36 receptor, was beneficial in the IL-36-mediated disorder generalized pustular psoriasis in a phase I study. However, no fibrotic biomarkers were assessed (173). The efficacy and safety profiles of spesolimab in IBD are still under investigation.

5.2.3. Alkylating agents.

Alkylating agents are conventional anticancer drugs that cause cell death through DNA alkylation (174). Cyclophosphamide, a synthetic alkylating agent, has potent anti-inflammatory effects through interference with DNA, RNA, and protein synthesis in rapidly dividing immune cells. Cyclophosphamide is effective in a wide range of immune diseases (175–179). Evidence from preliminary studies that suggested cyclophosphamide might have benefit in pulmonary fibrosis (180–184) were confirmed in a multicenter RCT (NCT00004563) in scleroderma lung disease (185). Enrolling 158 patients, this study showed that the adjusted mean absolute difference in FVC was 2.53% between the cyclophosphamide group and the placebo group after 1-yr treatment, suggesting a significant but modest benefit of cyclophosphamide (185). Recently, a systematic review that comprised four trials with 495 patients indicated a small benefit of cyclophosphamide treatment compared with placebo, but the evidence showing that cyclophosphamide was superior to other drugs (mycophenolate mofetil) was not clear (186). Two additional phase III clinical trials of cyclophosphamide for IPF (EXAFIP Trial, NCT02460588) (187) and scleroderma interstitial lung disease (NCT01570764) are continuing. In CD, cyclophosphamide has only been evaluated in small uncontrolled studies that reported positive effects on symptom-based measures in patients with refractory CD (188, 189), and its effect on intestinal fibrosis likewise remains unproven.

5.2.4. Chemokines and their receptors.

Chemokines are small cytokines that drive leukocyte migration and immune responses by binding to CC-chemokine receptors (CCRs) (190). Chemokines and CCRs participate in both inflammatory pathways and fibrosis. For example, CCR1, CCR2, and CCR5 were initially described as important factors in liver fibrosis both in vivo and in vitro (191–193) before their role in monocyte/macrophage chemotaxis was identified. Accordingly, animal experiments were performed to explore the antifibrotic effects of cenicriviroc (CCR2/CCR5 antagonist) in liver and kidney fibrosis (194), which showed that cenicriviroc reduced collagen synthesis and deposition in three fibrosis models (thioacetamide-induced liver fibrosis, diet-induced NASH, and unilateral ureter obstruction-induced renal fibrosis). On the basis of these findings, a phase II clinical trial (NCT02217475) was completed in patients with NASH. At the end of a 2-yr observation, patients who switched to cenicriviroc tended to have a higher chance to achieve fibrosis improvement (>1 stage) than those who remained on placebo (24% vs. 17%, P = 0.37). The majority of responders (60%) at the first year maintained the benefit at the second year (195).

Human CC-chemokine ligand (CCL)2 is considered a chemoattractant for fibroblasts participating in fibrogenesis. Carlumab is a human monoclonal antibody that specifically binds and neutralizes profibrotic activities of CCL2 (196). However, results of a phase II clinical trial (NCT00786201) could not support the benefit of carlumab to IPF patients (197). A total of 126 patients were evaluated in this study, but no therapeutic effect of percent change in FVC was observed with carlumab treatment at week 52 (placebo −0.582%, 1 mg/kg carlumab −0.533%, 5 mg/kg carlumab −0.799%, and 15 mg/kg carlumab − 0.470%, P = 0.261) (197). Thus, no clinical data have been generated that support the efficacy of CCL2 antibody therapy in fibrotic diseases.

5.2.5. Other inflammation modulators.

CD20 is a surface marker on the B cell membrane, serving as an ion channel (198). Rituximab is a human mouse chimeric monoclonal antibody targeting CD20. Since approval for non-Hodgkin lymphoma and chronic lymphocytic leukemia, intravenously administered rituximab has become the treatment of various B cell-related malignancies (199). Previous observational studies showed a potential efficacy of rituximab in the treatment of interstitial lung disease, implying an antifibrotic potential (200–203). On the basis of these findings, several RCTs [NCT01862926 (204), NCT01969409 and NCT02990286 (205)] on interstitial lung disease are recruiting. The efficacy of rituximab for UC therapy has been explored in a phase II clinical trial (NCT00261118). Although rituximab was well tolerated in patients, it was not effective in remission induction (206). In addition, cases of rituximab-induced CD have been reported, which suggested that anti-CD20 therapy might not be feasible for CD (207, 208).

The properties of thalidomide as an angiogenesis inhibitor and inflammation modulator make it a possible candidate drug for fibrosis. In a randomized phase III study, thalidomide could improve cough and respiratory quality of life in IPF patients (209). Since no fibrosis-related markers were compared in this study between treatment and placebo groups, the antifibrotic potential of thalidomide needs further verification in future studies. Thalidomide was shown to be an effective alternative therapy in CD patients, but a potential antifibrotic effect has not been determined (210). Details on therapies directed toward other IL modulators are found in Supplemental Table S2.

5.2.6. Applications of inflammation modulators in Crohn’s disease-related intestinal fibrosis.

TNF inhibitors are among the most widely used biologics in CD treatment. The applications of anti-TNF are induction of remission and maintenance therapy (211). The rationale for evaluating TNF antagonists in CD intestinal fibrosis is based upon both their well-characterized anti-inflammatory effects and emerging evidence that TNF can increase collagen deposition and intestinal myofibroblast proliferation in the intestine (212). TNF stimulation increases the level of TGF-β1 secreted by colonic epithelial cells (213). In a peptidoglycan-polysaccharide-induced colitis rat model, treatment with TNF antagonists prevented the development of bowel wall fibrosis (214). There is no RCT studying the antifibrotic effects of anti-TNF in CD. An early study reported no progression of small bowel stenosis and no appearance of new strictures with infliximab treatment (215). In another prospective trial, during a mean follow-up duration of 23 mo, <10% of CD patients developed sub/obstructive symptoms under anti-TNF treatment, with no significant sonographic changes of existing intestinal lesions (216). Furthermore, adalimumab was evaluated in symptomatic obstructed CD patients due to small bowel strictures who participated in the CREOLE study (NCT01183403). Although this study had several limitations—most importantly, it used a single-arm nonrandomized design, and no attempt was made to quantify the degree of fibrosis in the stricture by cross-sectional imaging—it showed that adalimumab may be effective in this patient population. The observation that 50.7 ± 5.3% of patients did not need bowel surgery during a follow-up of 4 yr is consistent with a potential antifibrotic effect (217). Moreover, it is noteworthy that adalimumab might improve the long-term outcomes of stricturing disease by simply reducing inflammation rather than fibrosis because of the persistence of inflammation in stricturing areas in the gut. Supporting this notion, a different study demonstrated that intestinal fibrosis was associated with the lack of response to infliximab therapy (218). The degree of remaining inflammation may be a factor for the progression of fibrosis. In a subanalysis of the CALM study, a multicenter RCT evaluating the effect of tight control management of anti-TNF therapy in CD compared with conventional management (219), patients who did not achieve endoscopic remission or deep remission at 1 yr had a higher risk of disease progression, including stricture development and surgery (220). This suggests that patients with persistent but low inflammation may still have an elevated risk to develop intestinal strictures. A well-designed RCT is needed to further evaluate this question.

ILs also play important roles in the pathology of CD. IL-12 is secreted by activated phagocytes and dendritic cells in response to pathogens during intestinal inflammation and infection (221). Elevated expression of IL-12 was found in lamina propria mononuclear cells (LPMCs) in CD. IFN-γ production in LPMCs was blocked by inhibiting IL-12 (222). IL-23, which also belongs to the IL-12 family, is increased in innate lymphoid cells and upregulates IL-17A and IFN-γ production in CD (223, 224). Targeting IL-12/IL-23 via a p40 peptide-based vaccine ameliorated collagen deposition in the murine chronic TNBS-induced colitis model (225). This implies a potential antifibrotic effect of IL-12/IL-23 inhibition on intestinal fibrosis. In clinical trials ustekinumab has been successfully utilized as induction and maintenance therapy in CD patients (226), but its antifibrotic effects need to be further explored. A reduced production of IL-10 was found in patients with severe CD (227). In addition, IL-10 knockout mice spontaneously develop chronic colitis, suggesting that IL-10 can be a protective factor in IBD (228). Preclinical administration of IL-10 showed efficacy in the DSS-induced mouse colitis model (229). The IL-10 knockout mouse develops intestinal fibrosis, which is rescued with IL-10 treatment (230). However, no clinical trials have been registered to explore IL-10 therapy in fibrostenosing CD patients.

An additional antifibrotic target in intestinal fibrosis could be IL-13, which plays a pivotal role in the Th2 immune response that is closely related to fibrosis (231). Although inhibition of IL-13 did not show a solid benefit on fibrotic diseases in past clinical trials, its antifibrotic effects on stenotic CD should be explored (168, 169). Higher levels of IL-13 transcripts were identified in the muscularis propria layer of CD compared to patients with healthy mucosa. The production of matrix remodeling-related genes in fibroblasts derived from the intestinal muscle layer, which expressed IL-13 receptor α1, was altered after stimulation with IL-13, via phosphorylation of STAT6 (32). The profibrotic effects of IL-13 were validated in the chronic TNBS-induced colitis model, potentially mediated via IL-13 receptor α2 and TGF-β1 (232). Unfortunately, anti-IL-13 therapy using tralokinumab failed to improve clinical response in moderate to severe UC patients at week 8 compared with placebo (38% vs. 33%, P = 0.406). Only a higher clinical remission rate could be found in the tralokinumab group compared with placebo (18% vs. 6%, P = 0.033), suggesting that tralokinumab might benefit some patients with UC (233). In addition, another IL-13 monoclonal antibody, anrukinzumab, did not show therapeutic effects in active UC (234). Given that no target engagement data were provided and the fact that fibrosis and inflammation can take an independent course and likely need to be targeted separately, anti-IL-13 therapies are still a viable option in stricturing CD.

Within the IL-1 family, several ILs have proven to contribute to intestinal fibrosis. Both IL-1α and IL-1β were reported to be profibrotic in other organs, such as lung (235–238), liver (239), and kidney (240). IL-1α enhanced the effects of intestinal epithelial cell lysate-induced activation of human intestinal fibroblasts (241). IL-1α-deficient mice experienced milder disease in the DSS-induced colitis model, and neutralization of IL-1α in wild-type mice showed a similar effect (242). With direct relevance to intestinal fibrosis, IL-1α induced the production of TGF-β1 by colonic epithelial cells (213). Considering a direct effect of IL-1 on fibroblasts, IL-1β stimulated the production of collagen I and IV, IL-8, monocyte chemoattractant protein, and MMP-1 by colonic myofibroblasts (243). However, this was challenged by other studies showing the inhibition of collagen synthesis and induction of collagenase production in intestinal smooth muscle cells after IL-1β stimulation (244, 245). In line with this, IL-1β-deficient mice showed exacerbation of colitis during DSS induction (242). Together, these results suggest potentially distinct roles of IL-1α and IL-1β in intestinal fibrosis. This may be context dependent rather than driven by molecular function, as the biologic activity of both molecules is believed to be identical. Further studies are needed to shed additional light on these initial observations.

In addition, IL-33, which also belongs to the IL-1 family, is associated with fibrotic diseases in multiple organs by driving proinflammatory and Th2-associated immune responses through its receptor ST2 (246, 247). IL-33 is synthesized by multiple cell types, such as epithelial cells, endothelial cells (248), and adipocytes (249). Elevated expression of IL-33 and ST2 can be found in IBD patients (250, 251). In pediatric patients with stricturing CD, epithelial IL-33 was increased and strongly associated with fibrostenotic disease (252). In experimental adherent-invasive Escherichia coli fibrosis, IL-33 promoted intestinal fibrosis and collagen deposition in fibroblasts via epithelial IL-33-ST2 signaling. Hence, blocking this pathway could attenuate intestinal fibrosis (253).

IL-36 has been identified to have both proinflammatory as well as profibrotic properties in the intestine (254). IL-36α and IL-36γ are increased in the intestinal mucosa of UC patients (255, 256). Significantly higher levels of IL36A were also noted in the intestinal wall from patients with fibrostenotic CD (172). IL-36 receptor activation induced expression of fibrosis-related genes in both human and mouse fibroblasts (172). Of note, IL-36 receptor deficiency leads to higher susceptibility to acute DSS colitis and impaired mucosal healing, implying an important role in intestinal tissue repair (257), but, conversely, IL-36 blockade ameliorates intestinal fibrosis. Although multiple lines of evidence indicate a potential antifibrotic therapeutic effect of targeting IL-1 family members in intestinal fibrosis, no agent has been designed for this mechanism to date.

IL-17A is mainly secreted by the Th17 subset of CD4+ T cells and closely related to the Th2 immune response (258). IL-17A has been identified as profibrotic in different organs, such as liver, skin, lung, and heart (126, 259). However, its role in intestinal fibrosis is still controversial (260). On one hand, IL-17A is elevated in CD strictured tissue compared with the nonstrictured portion and upregulates matrix remodeling enzymes and collagen production by myofibroblasts derived from strictured CD segments (33). IL-17A also stimulates the expression of collagen I in subepithelial myofibroblasts, via heat shock protein 47 (261). In addition, IL-17A also induced EMT in intestinal epithelial cells (262). In the TNBS-induced colitis model, intestinal fibrosis was alleviated by treatment with anti-IL-17A neutralizing antibody (262). On the contrary, secukinumab, a human monoclonal antibody against IL-17A, showed an unfavorable response in moderate to severe CD, which was accompanied by significant side effects (263). This could be potentially explained by the fact that IL-17A is a critical molecule for intestinal homeostasis and barrier integrity.

Chemokines and their receptors are another group of factors that contribute to intestinal inflammation (264). Some important receptors and their ligands were detected in the intestinal tissue and increased in response to inflammation. For example, CCR9-CCL25 is constitutively expressed in the small bowel and upregulated during inflammation, whereas the expressions of CCR5-CCL3/4/5/8 and CCR2-CCL2/7/8 were more prominent in the colon (264). Among these, CCR9-CCL25 is the most studied. Administration of CCX282-B, which is an orally bioavailable CCR9 antagonist, resulted in attenuation of intestinal inflammation in TNFΔARE mice (265).The encouraging preclinical results led to an RCT in CD showing higher response rates of 61% at week 12 with 500 mg CCX282-B twice a day compared with 47% in the placebo group (P = 0.039). The percentage of patients who remained in remission was higher in the CCX282-B group compared with the placebo group (47% vs. 31%, P = 0.012) at the end of week 52 (266). A more recent phase II randomized, double-blind trial investigated the therapeutic effects of CCR9-targeted leukapheresis in UC. Evaluating 23 patients, it showed a reduction of proinflammatory HLA-DRhi cells in the treatment group (from 13.5% at baseline to 10.1% after treatment, P = 0.0391), whereas no significant difference was found in the placebo group (from 11.2% at baseline to 11.5% after treatment, P = 0.46). Also, Mayo score dropped from 8.8 at baseline to 5.7 with CCL25 treatment (P = 0.0156 ) (267). The antifibrotic effects of this target need to be studied in the future.

As discussed above, inflammation is a necessary process in tissue repair (14). Thus, the administration of inflammation modulators, at least in theory, might lead to impaired tissue repair and delayed mucosal healing. This hypothesis, however, to date has not been shown to be the case in IBD and in particular in CD. A different potential consequence of using inflammation modulators is the risk of opportunistic infections (268), due to the immunosuppressive effects of these agents, making a risk/benefit evaluation necessary (269).

In summary, targeting specific inflammatory mediators has shown some promise in fibrotic diseases in other organs. However, definitive evidence is lacking and the potential value of optimizing anti-inflammatory therapy as a strategy to halt disease progression in fibrostenotic CD has not been shown in controlled studies. Potential targets that have been tested preclinically in experimental fibrosis are listed in TABLE 1. IL-36 blockade, IL-13 blockade, and IL-10 administration might be the most promising approaches in CD.

Table 1.

Preclinically tested approaches for intestinal fibrosis in animal models based on antifibrotic clinical trials in other organs

	Model System	Mechanism	Prevention or Reversal of Fibrosis	Effects	Year	Reference
TGF-β1 vaccination	Murine chronic TNBS-induced colitis	TGF-β1 antagonism	Both preventive and reversal (from week 0 or 2 of TNBS colitis)	Reduced collagen deposition and decreased levels of TGF-β1, IL-17, and IL-23	2010	(127)
Smad7-specific antisense oligonucleotide	Murine chronic TNBS-induced colitis	TGF-β1 signaling modulator	Reversal (from week 5 or 9 of TNBS colitis)	Reduced signs of colitis, reduced collagen deposition, and diminished fibrosis	2018	(134)
Pirfenidone	Murine chronic DSS-induced colitis	TGF-β, TNF-α, and NF-κB modulator	Both preventive and reversal (from day 0 or 14 of DSS colitis)	Reduced collagen deposition, suppressed the mRNA expression of COL1A2, COL3A1, and TGF-β	2016	(129)
IL-10	Spontaneous colitis of IL-10 knockout mice	IL-10 administration	Preventive (from week 12 of age)	Decreased tissue fibrosis, decreased inflammatory cytokines and TGF-β1	2013	(230)
IL-13 receptor α2-specific small interfering RNA	Murine chronic TNBS-induced colitis	IL-13 inhibitor	Therapeutic (from day 35 of TNBS colitis)	Alleviated fibrosis and TGF-β1 production	2007	(232)
IL-36 receptor antibody	Murine chronic DSS- and TNBS-induced colitis	IL-36 inhibitor	Preventive (from day 1 of DSS or TNBS colitis)	Reduced fibrosis score, submucosal thickness, and α-SMA-positive cells	2019	(172)
AMA0825	Murine chronic TNBS-induced colitis and adoptive T-cell transfer colitis	Local ROCK inhibitor	Both preventive and reversal (from week 2 of T-cell transfer or week 7 of DSS colitis)	Inhibited myofibroblast accumulation, expression of profibrotic factors, and accumulation of fibrotic tissue; reversed established fibrosis	2017	(365)
GED-0507-34 Levo	Murine chronic DSS-induced colitis	5-ASA analog (mimicked by PPAR-γ)	Preventive (from day 12 of DSS colitis)	Improved macroscopic and microscopic intestinal lesions; reduced the profibrotic gene expression of ACTA2, COL1A1, and FN1; reduced protein levels of α-SMA and collagen I/II, TGF-β/Smad pathway components, IL-13, and CTGF	2016	(369)
GED-0507-34 Levo	Murine chronic DSS-induced colitis	5-ASA analog (mimicked by PPAR-γ)	Preventive (from day 12 of DSS colitis)	Reduced the expression of the main fibrosis markers (α-SMA, collagen I–III, and fibronectin) as well as the pivotal profibrotic molecules IL-13, TGF-β and Smad3; increased the antifibrotic PPAR-γ	2017	(368)
LOX inhibitor	Murine chronic DSS-induced colitis	LOX inhibitor	Preventive (from day 0 of DSS colitis)	Reversed the ECM contraction and MMP3 activity in stenotic myofibroblasts grown in fibrotic environment	2018	(47)
Captopril	Murine chronic TNBS-induced colitis	ACEI	Preventive (from day 0 of TNBS colitis)	Reduced the score of macroscopic and histological lesions, colonic tissue levels of collagen α1, hydroxyproline, ANG II, and TGF-β1 proteins, and TGF-β1 mRNA	2004	(410)
Losartan	Murine chronic TNBS-induced colitis	ARB	Preventive (from day 0 of TNBS colitis)	Improved macro- and microscopic scores of intestinal fibrosis and reduced TGF-β1 concentration	2012	(411)
Simvastatin	Murine chronic TNBS-induced colitis	HMG-CoA reductase inhibitor	Preventive (from day 2 of TNBS colitis)	Attenuated intestinal fibrosis by lowering CTGF and inducing apoptosis of fibroblasts and myofibroblasts	2012	(442)

ACEI, angiotensin-converting enzyme inhibitor; ANG, angiotensin; ARB, angiotensin II receptor blocker; 5-ASA, 5-aminosalicylic acid; CTGF, connective tissue growth factor; DSS, dextran sulfate sodium; ECM, extracellular matrix; HMG-CoA, 5-hydroxy-3-methylglutaryl-coenzyme A; IL, interleukin; LOX, lysyl oxidase; MMP, matrix metalloproteinase; NF, nuclear factor; PPAR, peroxisome proliferator activator receptor; ROCK, rho-associated coiled-coil forming protein kinase; α-SMA, α-smooth muscle actin; TGF, tumor growth factor; TNBS, 2,4,6-trinitrobenzenesulfonic acid solution; TNF, tumor necrosis factor.

5.3. Intracellular Enzymes and Kinases

5.3.1. Endothelin receptor antagonist.

Endothelins (ETs) include three 21-amino acid peptides with similar molecular structures. By binding to endothelin receptors, ETs serve as potent vasoconstrictors regulating vascular tone (270). In addition, ET-1 plays an important role in fibrogenesis by inducing myofibroblast differentiation and ECM component deposition (271, 272). Endothelin receptor antagonist treatment could suppress the expression of ECM components and also prevent EndoMT (273, 274). Furthermore, administration of endothelin receptor antagonist could attenuate fibrosis in vivo (275–278).

Bosentan is a small-molecule ET-1 antagonist under development as an antifibrotic agent in IPF. An initial phase II clinical trial, BUILD-1 (NCT00071461), showed no significant difference but a trend for delayed time to death or disease progression in IPF patients with the administration of bosentan (n = 74) in comparison to placebo (n = 84) [hazard ratio (HR) 0.613; 95% confidence interval (CI) 0.328–1.144, P = 0.119]. However no benefit was demonstrated in 6-min walk distance (bosentan treatment effect of −18 ± 20 m, P = 0.226) or decline in FVC (bosentan treatment effect, hazard ratio 2.34; 95% CI 0.22–25.25, P = 0.5952) compared with placebo at 1 yr (279). On the basis of these results the phase III trial, BUILD-3 (NCT00391443), was designed to further explore the potential benefit of bosentan in IPF. This trial, which evaluated 616 patients over a 1-yr period, failed to demonstrate a benefit of bosentan therapy (hazard ratio 0.85, 95% CI 0.66–1.10, P = 0.2110) (280). Subsequently, data from a phase IV study (NCT00637065) also could not identify improvement in objective measures of pulmonary hypertension in IPF following treatment with bosentan (281). Furthermore, no improvement was found in a single study conducted in patients with systemic sclerosis-related interstitial lung disease (NCT00070590) (282). Consistent with these results, macitentan (NCT00903331; ET-A and ET-B inhibitor) did not benefit the mean change of FVC [macitentan −0.20 L (−0.29 to −0.16 L) vs. placebo −0.20 L (−0.28 to −0.13 L), P = 0.9631] (283). Another endothelin receptor antagonist, ambrisentan (ET-A inhibitor; NCT00768300), could not demonstrate benefit in IPF and in fact increased the risk of disease progression and side effects. Patients with administration of ambrisentan had a higher percentage of respiratory hospitalization than control subjects (13.4% vs. 5.5%, P = 0.007) (284).

5.3.2. Kinase inhibitors.

Kinases, a group of ATP-dependent phosphotransferases that transfer a single phosphate group from the γ position of ATP to protein substrates, participate in diverse cellular activities including fibrogenesis through regulation of cellular signal pathways (285).

5.3.2.1. mitogen-activated protein kinase inhibitors.

MAPKs are a group of ubiquitous proline-directed protein-serine/threonine kinases that mediate cellular proliferation, differentiation, and responses through signal transduction. Several pathways are involved in MAPK activation, including the extracellular signal-regulated kinase (ERK)1 and ERK2, c-Jun NH₂-terminal kinase/stress-activated protein kinase (JNK/SAPK), and p38 pathways (286). Previous data demonstrated that MAPKs are crucial in CD pathogenesis. Among the MAPKs, apoptosis signal-regulating kinase (ASK)1 has recently received increasing attention in fibrogenesis research. In vivo evidence that ASK1 participates in tissue fibrosis came from knockout mouse models in angiotensin II-induced heart hypertrophy and remodeling (287). Supportive data for a relevant role in fibrosis have also come from models of liver, kidney, lung, and skin disease (288–292). A recent phase II clinical trial (NCT02466516, conducted in patients with nonalcoholic steatohepatitis) showed that administration of selonsertib, an ASK1 inhibitor, improved fibrosis as measured by matrix stiffness and collagen content (293). On the basis of these encouraging results, two recent phase II studies with large sample sizes (STELLAR-3, NCT03053050; STELLAR-4, NCT03053063) were performed. However, no antifibrotic effects of selonsertib monotherapy were demonstrated in NASH patients, who were required to have bridging fibrosis or compensated cirrhosis for eligibility (294). These findings are contrary to results obtained in animal experiments that found that ASK knockout mice or mice given ASK1 inhibitor (MSC 2032964) had increased lipid storage and were more prone to develop NASH fibrosis (295). These data suggest a protective role of ASK1 in fibrogenesis. One potential explanation is that the difference of molecular structure of the inhibitors MSC 2032964 and selonsertib resulted in different observations. More studies are needed to explore the role of ASK1 in fibrogenesis.

Another MAPK, c-Jun NH₂-terminal kinase (JNK), also has potential effects on fibrogenesis. Data to support this pathway come from the observations that the profibrotic action of TGF-β1 on myofibroblasts can be blocked by JNK pathway inhibition (296) and JNK1 deficiency protects bleomycin-treated mice from pulmonary fibrosis (297). Inhibition of the JNK pathway by CC-930 showed potential activity in reducing pulmonary fibrosis in a phase II trial (NCT01203943) in IPF. However, the benefit/risk profile was 46% of 28 mild to moderate IPF patients experiencing adverse events leading to drug withdrawal. The trial was prematurely terminated because of an unacceptable tolerability and safety profile (298).

5.3.2.2. receptor tyrosine kinase inhibitors.

Receptor tyrosine kinase (RTK) is one of two subtypes of tyrosine kinases, which includes multiple cell surface receptors such as platelet-derived growth factor receptors (PDGFRs), vascular endothelial growth factor receptors (VEGFRs), and fibroblast growth factor receptor (FGFR) (299). Inhibition of RTK has been demonstrated to attenuate tissue fibrosis in different organs (300–302). The most tested RTK inhibitor, nintedanib (BIBF 1120), blocks PDGFR, VEGFR, and FGFR, which have all been implicated in in lung fibrosis (303). The randomized double-blind clinical trial in humans with IPF (TOMORROW Trial, NCT00514683) demonstrated that nintedanib could significantly slow down the decline of lung function in patients with IPF (304). A total of 432 patients were enrolled in this study and randomly assigned to receive different concentrations of nintedanib or placebo. The results showed that annual decline of FVC was significantly lower in the nintedanib group with the dose of 150 mg twice a day compared with the placebo group (P = 0.01) (304). On the basis of these encouraging results, two phase III clinical trials (NCT01335464 and NCT01335477, INPULSIS-1 and INPULSIS-2) were carried out. A total of 1,006 patients were randomly allocated into the nintedanib group or the placebo group in these two studies. In INPULSIS-1, adjusted annual rate of FVC change was −0.11 L with nintedanib treatment (150 mg twice daily) versus −0.24 L with placebo (P < 0.001). A similar result was observed in INPULSIS-2, with −0.11 L in the nintedanib group versus −0.21 L in the placebo group (P < 0.001). Both studies demonstrated and showed that nintedanib could protect IPF patients from disease progression (305). Nintedanib was subsequently approved by the US Food and Drug Administration (FDA) in 2014 for the treatment of IPF. In addition, as the only two approved drugs for IPF, the combination of nintedanib and pirfenidone was shown to be safe and well tolerated in two clinical trials (NCT01136174 and NCT02579603) (306, 307). Disappointingly though, the addition of sildenafil to nintedanib was not shown to have greater efficacy than nintedanib monotherapy. A second study is currently further evaluating the potential benefits of combination therapy in a targeted patient population with right-sided heart failure (INSTAGE Trial, NCT02802345) (308, 309). Also, nintedanib has been studied in patients with systemic sclerosis-associated interstitial lung disease in a phase III clinical trial (NCT02597933), showing less decline in FVC in the nintedanib group but no benefit being achieved in manifestations (310).

5.3.2.3. pi3k/akt/mtor signaling pathway inhibitors.

The phosphoinositide 3-kinase (PI3K)/protein kinase B (AKT)/mTOR pathway is a kinase-related pathway of relevance to fibrogenesis through regulation of cell proliferation, autophagy, EMT, and other cellular interactions (311). Similar findings have been shown in the lung, kidney, and liver, suggesting that mTOR might be an attractive target for antifibrotic agents (312–315).

Everolimus is an mTOR kinase inhibitor that has been studied in malignant tumors including breast cancer, gastric cancer, liver cancer, neuroendocrine tumors, and fibrosis (316). However, a recent phase IV clinical trial (NCT02096107) demonstrated that graft fibrosis scores in kidney transplant recipients managed with tacrolimus/everolimus-based regimens were not different from those treated with tacrolimus/mycophenolate treatment-based protocols (317). Of the two clinical trials designed to evaluate the antifibrotic effects of everolimus in liver transplant recipients, one (NCT00582738) was terminated prematurely; the second trial (NCT01707849) was completed, but no results are yet available. Another novel inhibitor of mTOR, omipalisib (GSK2126458), was shown to attenuate fibrosis in vitro by modulating proliferation and inhibiting the TGF-β1-induced profibrotic responses in fibroblasts derived from IPF (318). On the basis of preclinical results, another phase I clinical trial (NCT01725139) showed pharmacodynamic activity and acceptable tolerability of omipalisib in patients with IPF (319).

5.3.2.4. other kinase inhibitors.

Rho-associated coiled-coil forming protein kinases (ROCKs) are serine/threonine kinases that interact with actin cytoskeleton and regulate the shape of cells as well as sensing their cellular environment through integrin-based interactions (320). ROCK kinase could act as a profibrotic modulator and a potential therapeutic target in pulmonary fibrosis (321). Interestingly, the selective ROCK2 inhibitor KD025 ameliorated lung collagen deposition and normalized pathogenic pulmonary function in a murine multiorgan system chronic graft-versus-host disease (cGVHD) model with bronchiolitis obliterans syndrome (BOS) when administered in a therapeutic experimental setting (322). In addition, KD025 reversed clinical signs of sclerodermatous cGVHD, and both phenomena were related to STAT3 signaling(322). These results imply the potential antifibrotic property of ROCK inhibitors in fibrogenic disorders. A clinical trial (NCT02688647) is currently ongoing to explore the therapeutic effects of a ROCK inhibitor, belumosudil. However, concerns exist regarding potential cardiovascular side effects of this approach.

Activation of BCR-ABL tyrosine kinase has a core role in the pathology of Philadelphia chromosome-positive chronic myeloid leukemia (323). Imatinib is the first tyrosine kinase inhibitor (TKI) developed to treat chronic myeloid leukemia (324). Consequently, extensive clinical experience has evolved with this molecule. The antifibrotic effects of this traditional TKI were explored in IPF patients in a phase II trial (NCT00131274); however, imatinib could not reduce the risk of disease progression, defined by 10% decline in percent predicted FVC from base or elongated time of death compared to placebo (log rank, P = 0.89) (325). No data on the effects of TKI in CD are available.

5.3.3. Caspase inhibitors.

Caspases are a group of intracellular proteases involved in cell apoptosis and proteolytic activation of cytokines that regulate the immune response in autoinflammatory disease (326). Apoptosis activated by the caspase cascade fulfills critical roles in fibrogenesis (327). In preclinical studies, lung fibroblasts from IPF were more resistant to apoptosis related to FasL-induced pathway (via caspase) (328). In vitro, administration of caspase inhibitors could attenuate pulmonary and hepatic fibrosis (329–332). Two phase II clinical trials focusing on the antifibrotic effects of emricasan, a pan-caspase inhibitor, were completed. One of them (NCT02230670) showed that emricasan improved liver function in patients with hepatitis C and NASH-associated cirrhosis, suggesting that emircasan could be a potential antifibrotic agent in liver fibrosis (333). However, another phase II randomized trial showed that emricasan could not improve the prognosis of patients with cirrhosis (334).

An increase of apoptosis in intestinal epithelium of CD was found in several studies, whereas a decreased apoptosis of ECM-producing mesenchymal cells may exist in fibrotic CD intestine (335, 336). Since apoptosis regulates the pathology of CD bidirectionally, the potential therapeutic benefits of caspase inhibitors should be carefully assessed in well-designed clinical trials in the future.

5.3.4. Peroxisome proliferator-activated receptor agonists.

PPAR belongs to a family of nuclear receptors regulating lipid metabolism. They are involved in pathological processes including tissue fibrosis (337, 338). Several PPAR agonists improve fibrosis in preclinical experiments, whereas their antagonists could inhibit the antifibrotic effects (339–342). In a phase III clinical trial (NCT01654731, BEZURSO), bezafibrate, a pan-PPAR agonist, could achieve a higher rate of complete biochemical response with less liver stiffness and lower enhanced liver fibrosis score in PBC patients who had inadequate response to ursodeoxycholic acid (UDCA) (343). Selective antagonists of PPAR also show antifibrotic potential in several clinical trials. First, a higher rate of NASH patients without fibrosis worsening was found with the use of elafibranor, an agonist of both PPAR-α and -δ in a phase II RCT (NCT01694849) (344). Similarly, seladelpar targeting PPAR-δ was shown to benefit PBC patients with limited response to ursodeoxycholic acid by normalizing ALP levels (NCT02609048) (345). However, the significant increases of aminotransferases led to an early termination of this trial (345). For PPAR-γ agonists, in a phase II trial (NCT00492700) rosiglitazone had antisteatogenic effect in NASH patients, but no improvement of hepatic fibrosis could be found compared with placebo (346). Another clinical trial on farglitazar (a PPAR-γ agonist) also demonstrated no evidence of antifibrotic actions in patients with chronic hepatitis C infection (347).

5.3.5. Lysophosphatidic acid receptor 1 antagonist.

Lysophosphatidic acids (LPAs) are a group of glycerophospholipids modulating various cellular events including cell proliferation, motility, and chemotaxis (348). One of their corresponding receptors, lysophosphatidic acid receptor 1 (LPA1), takes part in fibrogenesis of both lung and kidney (349, 350). The antifibrotic effects of LPA1 antagonist were found in animal fibrosis models (351–353). A recent clinical trial showed that the antagonist BMS-986020 could slow down the decline of FVC in IPF patients with 26-wk treatment (354). For now, still no studies have demonstrated the role of LPA1 in CD. Further investigation should be carried out to check the potential antifibrotic role of its antagonist in CD patients.

5.3.6. Other intracellular enzymes and kinases.

Sapropterin (a phenylalanine hydroxylase stimulant) (355) and riociguat (a guanylate cyclase stimulant) (NCT02138825) (356) could not reduce the hepatic venous pressure gradient (HVPG) in liver cirrhosis and 6-min walking distance in IPF, respectively. Details on intracellular enzymes and kinases are found in Supplemental Table S3.

5.3.7. Applications of intracellular enzymes and kinases in Crohn’s disease-related intestinal fibrosis.

High ET-1 immunoreactivity was identified in both CD and UC compared with control subjects (357). One study explored the effects of endothelin receptor antagonism in the iodoacetamide-induced colitis rat model. Although the inhibition of endothelin receptors attenuated inflammation in the intestine and the severity of peritonitis, the study results did not evaluate any antifibrotic effects (358) and endothelin receptor antagonism remains an antifibrotic strategy to be tested in CD.

The MAPKs and their related signaling pathways contribute to the pathogenesis of IBD (359), whereas their role in intestinal fibrosis is inadequately investigated. Although the antifibrotic effect of an ASK1 inhibitor has been reviewed in multiple conditions as described above, very few studies evaluated this pathway in IBD. TI-1-162, a hydroxyindenone derivative, attenuates inflammation in a TNBS-induced colitis mouse model through inhibition of the receptor-interacting serine/threonine protein (RIP)/ASK1/MAPK signaling pathway (360). No data have demonstrated the functional role of ASK1 in intestinal fibrosis in humans. RTKs and their signaling were also identified as potentially important mechanisms in intestial fibrosis. For example, PDGFR is highly expressed in both active inflammation and fibrotic areas of CD (361). Also, higher than normal serum concentrations of vascular endothelial growth factor (VEGF) and VEGFR have been demonstrated in CD patients (362). Although nintedanib has not been investigated in CD, its relevant receptors have been shown to be involved in CD pathogenesis. Given the established efficacy and safety profile, nintedanib should be evaluated in fibrostenosing CD. Activation of the PI3K/AKT/mTOR signaling pathway has been demonstrated in stricturing CD (363). A multicenter, randomized, double-blind trial evaluating 96 patients with active CD showed safety and tolerability of everolimus (6 mg/day) comparable to azathioprine but no benefit in achieving steroid-free remission at month 7 compared with placebo (P = 0.610) (364). However, no clinical data are available in fibrostenosing disease. Accordingly, the role of mTOR inhibition in fibrotic CD still needs to be further explored. Compared with the kinase inhibitors mentioned above, topical application of Rho/ROCK blockade could be a very promising method to treat CD intestinal fibrosis. Local ROCK inhibition by AMA0825 prevented and reversed experimental intestinal fibrosis by decreasing TGF-β1-induced myocardin-related transcription factor, p38 MAPK activation, and increasing autophagy in fibroblasts that are critical to the pathogenesis of stricturing CD (365) all while minimizing systemic drug exposure (365).

PPAR-γ is the most widely studied PPAR isoform in fibrotic diseases, and the therapeutic value of targeting PPAR-γ has been studied. PPAR-γ has been demonstrated to be able to reduce the production of the profibrotic molecules PDGF, IL-1, and TGF-β1 (366). The therapeutic effect of a PPAR-γ agonist has been shown in DSS-induced colitis (367). In addition, PPAR-γ can directly antagonize Smad3 of the TGF-β1 pathway and decrease CTGF expression, which play an important role in fibrogenesis in CD (368, 369). PPAR-γ exerts anti-inflammatory action by mimicking 5-aminosalicylic acid (5-ASA) (370). Later studies further demonstrated that GED-0507-34 Levo, a 5-ASA analog, could ameliorate inflammation-driven intestinal fibrosis in DSS-induced colitis (368, 369). PPAR-γ agonists can become a candidate for antifibrotic therapy in CD but need to be explored in future clinical trials.

Additional novel agents targeting intracellular enzymes emerge as potential therapies for intestinal fibrosis. Hydroxylases are a group of oxygen-sensing enzymes that confer hypoxic sensitivity in the tissue via the hypoxia-inducible factor (HIF) pathway (371). A high degree of hypoxia in the IBD intestine was observed during chronic inflammation, and hydroxylase inhibitors ameliorated colitis and promoted enhanced intestinal epithelial barrier function (372). In addition, inhibition of hydroxylases ameliorated experimental inflammation-induced intestinal fibrosis, potentially through the suppression of ERK-mediated TGF-β1 signaling (373). Although this suggests a novel pathway of intestinal fibrogenesis, future human trials need to confirm its efficacy.

In summary, among intracellular enzymes and kinases, topical ROCK inhibition as well as PPAR-γ agonists have the most potential to serve as novel antifibrotic therapies in intestinal fibrosis. The antifibrotic activity of nintedanib and other mechanisms such as ASK1 inhibition and mTOR inhibitors are also worthy of further exploration.

5.4. Extracellular Matrix Modulators

5.4.1. Lysyl oxidase modulators.

The lysyl oxidase (LOX) family comprises five members including LOX and four lysyl oxidase-like proteins (LOXL1–4). LOX mediates cross linking of the main components of the ECM including collagens and elastin (374). LOX and LOXL-mediated collagen cross-links contribute to matrix mechanical properties including stiffness (375), which in turn regulates mesenchymal cell activation (376). Among members of the LOX family, LOXL2 catalyzes the cross-linkage of collagen and elastin but initiates the cross-link of collagen IV (375, 377). Multiple preclinical studies have demonstrated that LOX and LOXLs were involved in the development of fibrosis by regulating cellular differentiation, promoting collagen stabilization, and increasing matrix stiffness (378).

Simtuzumab, a humanized monoclonal antibody directed against LOXL2, was studied in a randomized, double-blind, controlled phase II trial (NCT01769196) conducted on 544 patients with IPF. However, simtuzumab failed to achieve a therapeutic effect in progression-free survival relative to placebo [hazard ratio (HR) 1.13, 95% CI 0.88–1.45, P = 0.329], and drug-related adverse effects occurred with sufficient frequency that the trial was terminated (379). Two additional clinical trials (NCT01672866 and NCT01672879) that evaluated this antibody in nonalcoholic steatohepatitis were terminated prematurely without identifying a benefit of simtuzumab treatment in decrease of hepatic collagen content (−0.2%, 95% CI −1.3 to 1.0, P = 0.77 in 75 mg simtuzumab vs. placebo; −0.4%, 95% CI −1.5 to 0.8, P = 0.52 in 125 mg simtuzumab vs. placebo) (380). Furthermore, a clinical trial (NCT01672853) performed in PSC also did not show a benefit of using simtuzumab (381). Another trial in IPF (NCT01362231) completed recruitment; however, no result was reported. In these studies no target engagement markers were evaluated, and on this basis it remains unclear whether LOX inhibition is still a viable antifibrotic target.

5.4.2. Integrin α_vβ₆ modulators.

Integrins are heterodimeric cell surface receptors that facilitate adhesion of cells to the ECM (382). Among these molecules, integrin α_vβ₆ is believed to be closely involved with fibrosis. Overexpression of α_vβ₆ was induced in epithelial tissues of patients with fibrotic kidney, lung, and liver diseases relative to normal control subjects (383). TGF-β1 can also be upregulated through α_vβ₆-mediated mechanisms (384). In animal experiments, fibrosis could be inhibited through administration of integrin α_vβ₆ antagonists (385, 386). GSK3008348 is a selective, high-affinity inhibitor of integrin α_vβ₆ that was well tolerated in a phase Ib clinical trial (NCT02612051) conducted in IPF patients (387). Although the clinical efficacy of this agent has not been evaluated in human disease, RCTs are planned in IPF in future studies. Clinical trials evaluating other integrin α_vβ₆ inhibitors including BG00011 (NCT03573505), IDL-2965 (also inhibits α_vβ₁ and α_vβ₃; NCT03949530), and PLN-74809 (also inhibits α_vβ₁; NCT04072315) are underway.

Details on other ECM modulators are found in Supplemental Table S4.

5.4.3. Applications of extracellular matrix modulators in Crohn’s disease-related intestinal fibrosis.

Increased LOX gene expression was observed in TNBS-induced colitis rat models (388). In vivo, expression of LOX was identified in myofibroblasts derived from stenotic sites of CD. In addition, LOX inhibition could reverse enhanced ECM contraction and MMP3 activity of myofibroblasts grown in an increased stiffness environment (47). Given these findings, the antifibrotic effects and safety of LOX modulators should be explored in future clinical trials.

Integrins mediate various bioactivities and may be closely related to disease progression and fibrosis in IBD (389). After administration of TNBS, wild-type mice developed colonic fibrosis with elevated integrin α_vβ₆ levels, whereas Smad3 knockout mice did not have fibrosis and the expression level of integrin α_vβ₆ in intestinal tissue was very low (390). These observations indicate a plausible link between the TGF-β1/Smad3 pathway and integrin α_vβ₆ in colonic fibrosis. However, the observation that lower expression of integrin α_vβ₆ was found in the mucosa from active IBD colonic biopsy samples compared with those from control subjects is inconsistent with this hypothesis (391). Whether integrin α_vβ₆ inhibition could be a potential antifibrotic approach in CD requires further investigation. Importantly, vedolizumab, which was designed for targeting the selectively expressed gut integrin α₄β₇, was approved by the FDA in 2014 for both moderate to severe UC and CD (392). However, little is known about its antifibrotic activity in intestinal fibrosis. One recent case reported successful treatment of refractory stricturing CD using the combination of vedolizumab and ustekinumab (393). This suggests that this combined regimen could be a viable option for antifibrotic therapy in intestinal fibrosis. This is particularly true as vedolizumab has a very favorable safety profile (394).

Among extracellular matrix modulators, both LOX modulation and integrin blockade could be candidates in antifibrotic therapy of CD, but well-designed clinical trials are needed in the future.

5.5. Renin-Angiotensin System Modulators

5.5.1. Clinical trials of renin-angiotensin system modulators in fibrotic diseases.

A growing body of evidence supports that the renin-angiotensin-aldosterone system (RAS) takes part in fibrogenesis. Angiotensin (ANG) II increases ECM deposition by decreasing MMPs and increasing TGF-β1 production in heart and kidney (395, 396). ANG II can also induce contraction and proliferation of hepatic stellate cells by binding to angiotensin II type (AT)1 receptor (397). Then as the major components of RAS, ANG II and aldosterone can regulate the process of EMT and contribute to organ fibrosis (398).

Enalapril is a kind of angiotensin-converting enzyme inhibitor (ACEI) that inhibits the conversion of ANG I to ANG II. In a phase III clinical trial (NCT02432885), enalapril could significantly slow down the progression of myocardial fibrosis in Duchenne and Becker muscular dystrophy (399). This suggested an antifibrotic action of ACEI. One clinical trial (NCT02682459) on a different ACEI, lisinopril, in kidney fibrosis is still recruiting. Angiotensin II receptor blocker (ARB), as another kind of drug targeting RAS, was proven to have antifibrotic potential in clinical trials. For example, candesartan improved histological fibrosis in alcoholic liver disease (400). However, another clinical trial on losartan could not recruit sufficiently, so the antifibrotic effects in the liver could not be determined (401). In addition, the aldosterone antagonist spironolactone failed to show antifibrotic effects on heart fibrosis in a phase IV clinical trial (NCT00879060) (402).

5.5.2. Applications of renin-angiotensin system modulators in Crohn’s disease-related intestinal fibrosis.

All main components of the RAS are located in the human colonic mucosa. AT1 receptor is expressed in myofibroblasts and epithelium, whereas AT2 receptors are found in mesenchymal cells (403). Renin is highly expressed in colonic biopsies from UC and CD patients. Increased levels of colonic renin and angiotensin II were found in the TNBS-induced colitis model (404). In addition, ANG I and II expression are also elevated in the mucosa of human CD intestine (405). Administration of the ACEI enalapril reduced histological severity in acute DSS-induced colitis (406). These results were confirmed in other animal models (407–409). Renin transgenic mice, overexpressing renin, experienced enhanced colonic inflammation during TNBS colitis induction, which is potentially mediated via the ANG II-JAK2/STAT1/3 pathway (404). The enhanced RAS in colitis promoted mucosal Th17 activation and induced profibrotic cytokine secretion (TGF-β1 and IL-6), which implies a profibrotic role of RAS in intestinal fibrosis (404). More direct evidence is derived from chronic murine colitis models. The ACEI captopril significantly reduced the score of macroscopic and histological lesions in chronic TNBS-induced colitis, as well as inhibition of collagen I, hydroxyproline, ANG II, and TGF-β1 expression (410). The ANG II receptor antagonist losartan showed a similar antifibrotic effect in chronic TNBS-induced colitis (411). Although no confirming evidence from a well-designed clinical trial is available, recent data indicate that collagen deposition in IBD patients correlated with ANG II and inversely correlated with ACE2 activity. Furthermore, patients who required hospitalization or surgery over a time frame of 2 yr took ACEIs and ARBs less often compared with those who did not require hospitalization or surgery (412). Together, these findings support the rationale of using RAS modulators as antifibrotic therapy in intestinal fibrosis.

RAS modulators have been long used as effective therapy for various kinds of disorders, such as hypertension, chronic kidney disease, and heart failure (413–415). They are well tolerated and safe. Concerns for long-term therapy are induction of hypotension, kidney injury, and hyperkalemia (416). However, these adverse events happen significantly more often in patients who have impaired kidney function. In addition, ARBs might be better options for disease therapy than ACEIs, because of the lower incidence of withdrawal-related adverse effects (417). With the recent spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)-induced coronavirus disease 2019 (COVID-19) (418), ACE2, which is widely expressed in type II alveolar cells, macrophages, and epithelial cells in the lung, was found to be a functional receptor for the entry of SARS-CoV-2 (419, 420). Although the administration of ACEIs and ARBs might theoretically enhance the expression of ACE2 and hence risk for SARS-CoV-2 infection or severe COVID-19, evidence to that effect is missing to date. To the contrary, pretreatment with ACEIs or ARBs prevented SARS-CoV-induced acute lung injury and could be used as a potential therapy of COVID-19 (421). Specific to intestinal fibrosis in CD, no current evidence is available demonstrating an increased risk for COVID-19 after RAS modulator treatment.

Details on RAS modulators can be found in Supplemental Table S5.

5.6. 5-Hydroxy-3-Methylglutaryl-Coenzyme A Reductase Inhibitor

5.6.1. Clinical trials of 5-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor in fibrotic diseases.

HMG-CoA reductase is the key enzyme in the cholesterol synthetic pathway (422). HMG-CoA reductase inhibitors were primarily designed for cholesterol-lowering therapy, especially in coronary heart disease (423). In addition to lowering cholesterol, antifibrotic effects of HMG-CoA reductase inhibitors in vitro were revealed in recent studies. Initial evidence came from rat models with chronic cyclosporine-induced nephropathy, in which pravastatin could attenuate interstitial fibrosis (424). The antifibrotic effects include inhibition of collagen expression, α-smooth muscle actin expression, and CTGF synthesis in fibroblasts (425, 426). In vivo, HMG-CoA reductase inhibitors suppress fibrosis in different organs (427–432).

A phase II RCT showed that simvastatin could significantly decrease HVPG and improve liver blood perfusion in patients with cirrhosis (433). Although additional use of simvastatin could not reduce the risk of variceal rebleeding, it improved survival in patients with decompensated cirrhosis (NCT01095185) (434). Simvastatin at a dose of 20 mg/day in combination with rifaximin in treating decompensated cirrhosis had fewer adverse effects compared with a higher dose of 40 mg/day (LIVERHOPE-SAFETY Trial, NCT03150459) (435). In addition, administration of statin in heart failure patients could significantly reduce propeptide of procollagen type III, a surrogate marker of fibrosis, suggesting an antifibrotic effect of statin (NCT00795912) (436). Another phase IV clinical trial (BATTLE-AMI Trial, NCT02428374) on myocardial fibrosis has been completed, but no result was found (437).

5.6.2. Applications of 5-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor in Crohn’s disease-related intestinal fibrosis.

The role of HMG-CoA reductase inhibitors in IBD has been investigated in previous studies. Statin treatment could decrease inflammatory biomarkers and suppress the inflammation in DSS-induced colitis animal models (438, 439). As the first statin being tested in CD patients, atorvastatin was considered to be an effective anti-inflammatory therapy (440, 441). Although the evidence for antifibrotic potency of statins is limited, evidence is accumulating. First, simvastatin showed not only decrease in the degree of inflammation in TNBS-induced colitis models but also improved fibrosis by lowering CTGF and inducing apoptosis of fibroblasts and myofibroblasts (442). Second, pravastatin also attenuated intestinal fibrosis induced by radiation in animal models, implying the efficacy of statins in fibrotic intestines (443). Third, adipose tissue contributed to the pathology of CD, and the fat wrapping around the involved colon (creeping fat) was closely associated with the stricture formation in CD (49). This suggests that disordered metabolism of fat tissue might be a key component to fibrotic CD and statin might have the potential to reverse it. Since HMG-CoA reductase inhibitors have been long used to lower blood cholesterol for the prevention of heart attacks and strokes, their safety properties have been well demonstrated in clinical practice (444). Myopathy, new-onset diabetes mellitus, and probably hemorrhagic stroke are identified as rare serious adverse events during long-term treatment with statins (444). Close care and monitoring of myopathy-related symptoms (e.g., muscle pain or weakness), stroke-related symptoms (e.g., dizziness or fainting), and blood glucose should be performed for patients who take statins.

In conclusion, HMG-CoA reductase inhibitors have shown antifibrotic effects in different conditions. They might become possible antifibrotic agents for stricturing CD, but future clinical trials are expected. Details on HMG-CoA reductase inhibitors can be found in Supplemental Table S6.

5.7. Other Antifibrotic Agents

5.7.1. Granulocyte colony-stimulating factor.

Granulocyte colony-stimulating factor (G-CSF) is a glycosylated polypeptide that enhances the proliferation and maturation of white blood cells (445). Animal experiments showed that administration of G-CSF could attenuate fibrosis in different organs (446–448). Two RCTs (NCT02642003 and NCT02451033) demonstrated that G-CSF could not only improve the survival in patients with decompensated cirrhosis but reduce liver fibrosis as well (449, 450). However, another recent clinical trial (REALISTIC) made a contrary conclusion, suggesting that G-CSF might not improve liver dysfunction but increase adverse events (451). The mechanism of therapeutic effects of G-CSF in CD has been investigated, including restoring impaired neutrophil migration (452) and protecting colonic epithelial cells from apoptosis (453). Recombinant human G-CSF was effective and safe in treating both recurrent and active CD (454, 455). However, whether G-CSF could attenuate fibrosis in CD needs to be further explored.

5.7.2. N-acetylcysteine.

N-acetylcysteine is an acetylated variant of amino acid l-cysteine that acts as an antioxidant and is widely used in different diseases (456). N-acetylcysteine exhibited antifibrotic effects in multiple animal models such as pulmonary, renal, and cardiac fibrosis (457–459). Most of the evidence came from studies in patients with IPF. An early multicenter RCT, the IFGENIA trial, tested administration with acetylcysteine at a high dose (600 mg 3 times a day), which helped preserve vital capacity and single-breath carbon monoxide diffusing capacity in IPF patients when added to azathioprine and prednisone (460). However, the benefit of N-acetylcysteine in IPF was challenged in another trial (NCT00650091). N-acetylcysteine in combination with prednisone and azathioprine did not have physiological or clinical benefit but showed more adverse events compared with placebo in treating IPF patients (461, 462). Furthermore, a meta-analysis including five studies was not significant for benefit for N-acetylcysteine-treated IPF (463). Contrary to the results in IPF, one recent RCT (NCT01537926) suggested small benefit of N-acetylcysteine in cardiac fibrosis (464). Although the antifibrotic effect of N-acetylcysteine in CD is not reported, this molecule still has multiple functions in the intestine including anti-inflammation, antioxidative capacity, and improvement of intestinal mucosa energy status (465). Therefore, N-acetylcysteine could become a potential antifibrotic therapy in CD.

5.7.3. Interferon-γ_1b.

IFN-γ_1b is a biological response modifier with a single-chain polypeptide containing 140 amino acids (466). It was considered a potential antifibrotic agent by reducing collagen synthesis and deposition in vivo and in vitro experiments (467–469). Although the expression of fibrosis biomarkers (like PDGF and type I procollagen) in bronchoalveolar lavage fluid changed in IPF with IFN-γ_1b treatment (470), two well-designed placebo-controlled trials (INSPIRE and NCT00075998) could not find benefits of IFN-γ_1b in IPF (471, 472). In the intestine, IFN-γ induces the expression and activation of protein tyrosine phosphatase nonreceptor type 22, a protein related to reduced risk of CD (473). IFN-γ_1b might play a protective role in CD, but its antifibrotic effects need to be explored in the future.

5.7.4. Ursodeoxycholic acid.

Ursodeoxycholic acid (UDCA) is a natural hydrophilic bile acid with multiple biological functions including antioxidation, apoptosis inhibition, and stimulation of bile flow (474). UDCA is applied to patients with PBC, restoring defective natural killer cell activity (475). In a phase II RCT, UDCA administration significantly reduced FibroTest serum fibrosis marker in NASH compared with placebo, but no histological end points were available (476). There is no evidence for the use of UDCA in CD.

5.7.5. Obeticholic acid.

Obeticholic acid is a selective agonist with high affinity for farnesoid X receptor that has protective effects for liver injury and inflammation (477). Evidence from animal experiments supported the antifibrotic effects of obeticholic acid in liver fibrosis (478, 479). Clinical trials (NCT00550862, NCT00570765, and NCT01473524) demonstrated that obeticholic acid could significantly reduce the level of liver enzymes including ALP, γ-glutamyl transpeptidase, and alanine aminotransferase in PBC, suggesting antifibrotic effects on hepatic fibrosis, but risks for adverse events were also increased (480–482). Obeticholic acid could also improve the histological features in NASH, but long-term benefits and safety still need to be investigated (NCT01265498) (483).

5.7.6. Copper-chelating drug.

Tetrathiomolybdate, as a potent copper-chelating drug, was first developed for the treatment of Wilson’s disease (484). Tetrathiomolybdate was demonstrated to reduce the plasma levels of TNF and TGF-β1 and attenuated bile duct ligation-induced cholestatic liver fibrosis (485). In a phase III clinical trial, tetrathiomolybdate was shown to be a safe agent to improve serum biomarkers and biopsy in PBC (486), but a future clinical trial with a longer observing period is needed. Details on other antifibrotic agents can be found in Supplemental Table S7.

6. INSIGHTS AND CHALLENGES IN DEVELOPING ANTIFIBROTIC THERAPIES FOR CROHN’S DISEASE PATIENTS

According to our review of antifibrotic agents in other organs and those in preclinical studies in intestinal fibrosis, several candidates with different mechanisms of action stand out as potential therapies for CD intestinal fibrosis, as summarized in Table 1. Although they all have potential antifibrotic effects in intestinal fibrosis, targets can be prioritized on the basis of several factors. For example, drugs that directly influence the immune response and inflammation, such as TNF inhibitors and IL modulators, may act preferentially at the early phases of intestinal fibrosis for preventive treatments, since inflammation is likely a major initiator of intestinal fibrosis (5, 487). Early intensive medical therapy in CD may lower the risk for developing small bowel stenosis in CD patients (215, 216). One counterargument to this point is the already significant tissue damage at the time of diagnosis due to a long preclinically asymptomatic phase. In the opinion of the authors the largest area of opportunity for antifibrotic therapies is targeting inflammation-independent mechanisms, given that once established fibrosis can progress irrespective of the suppression of inflammation. This includes agents targeting effector mesenchymal cells and ECM deposition, such as TGF-β modulators, FGF modulators, ROCK inhibitors, RAS modulators, and HMG-CoA reductase inhibitors. Those drugs are likely suitable for both prevention and reversal of intestinal fibrosis. In vivo examples for this approach include inhibition of intestinal fibrosis in chronic TNBS-induced colitis after TGF-β1 vaccination before and after colitis was established (127). Pirfenidone also attenuated DSS-induced fibrosis in the early and late stages of fibrosis (129).

Careful consideration during drug development should be given to the safety of the therapeutic strategy. Beside the adverse events specific to each individual mechanism discussed above, the safety of the drugs may also be influenced by their effect on wound healing. For example, inflammation modulators or growth factor inhibitors may affect important pathways in inflammation and healing, may drive delayed, impaired tissue repair, and at least theoretically may inhibit restitution (79, 268, 488). Given the transmural nature of CD, concern has been raised about inducing internal penetrating disease in this setting. One may argue that other mechanisms, such as RAS modulators and HMG-CoA reductase inhibitors, would show more favorable safety properties and tolerability as therapies. In addition, most of the targeted pathways are not selective to the gut, and may then affect healing broadly and result in systemic adverse events. Therefore, an ideal potential antifibrotic agent would selectively target mechanisms unique to intestinal fibrosis with either organ-selective delivery or therapeutic activity only in the areas affected by fibrosis. This may be the case for the ROCK inhibitor AMA0825, which is topically delivered, exhibits a long retention time in the gut, and is degraded quickly by esterases when absorbed into the circulation (365), avoiding systemic side effects in animal models. Another potential approach is to target gut-selective integrins through, e.g., vedolizumab or etrolizumab (489). Although their therapeutic effects are well established in IBD inflammation, little is known about their antifibrotic effects in the gut. Most current drugs are given by oral or intravenous infusion, and most enter the systematic circulation. Thus, local administration and local release formulas are an option, and these could include enemas (490) or oral drug formulations that are released in select areas of the GI tract, such as the terminal ileum, the most common site for stricturing disease (491).

The development of antifibrotic drugs for the management of stricturing CD requires innovative solutions to multiple challenges, some of which are unique. First, the GI tract is one of the organs most exposed to exogenous agents derived from food and the microbiome (492). In IBD, the composition and function of the microbiome is altered and microbiome-derived metabolites act as key initiators and perpetuators of pathological responses (493). Although conclusive evidence is lacking that shows that alteration of microbiota can contribute to gut fibrogenesis, some important clues have emerged from both preclinical and clinical studies. For example, it is intriguing that TNF-mediated intestinal fibrosis is dependent on the close relation between the microbiome and several organisms such as Mucispirillum schaedleri and Ruminococcus (494), an observation that may be exploited by future antifibrotic therapies. Furthermore, the bacterial component flagellin has been demonstrated to enhance ECM production by intestinal fibroblasts in a MyD88-dependent manner (495). Given that immune responses against flagellin are common in CD patients, this observation suggests that commonality exists for activation of the pathological inflammatory and fibrotic pathways. Although the complexity of the microbiome may present a challenge for the development of antifibrotic drugs in the gut, it also offers novel opportunities for drug targets. As a further example, either whole body deletion or hepatocyte-specific deletion of bacterial cell wall sensor nucleotide-binding oligomerization domain-containing 2 (NOD2) exacerbated liver fibrosis in diet-induced nonalcoholic steatohepatitis in mice (496). Second, a unique consideration comes from the presence of creeping fat in CD, which is a pathognomonic process in which mesenteric fat wraps around the intestine in stricturing CD. Given the specificity of this hallmark anatomical finding, it is likely that it is integral to the pathophysiology of intestinal fibrosis and stricture formation because it is not characteristic of other chronic inflammatory intestinal diseases that do not cause pathological fibrosis. Recent studies showed that mediators derived from creeping fat could enhance the proliferation of human intestinal fibroblasts and human intestinal muscle cells, indicating that mesenteric fat may be a key metabolic engine that drives fibrogenesis in CD (49). Finally, although chronic inflammation is considered one of the most important activators of intestinal fibrosis (5), abrogation of inflammation is likely necessary but insufficient to prevent the progression of fibrosis. Hence it is noteworthy that complications and surgery rates remain significant despite the emergence of anti-inflammatory therapy (4). Therefore, investigating inflammation-independent mechanisms of intestinal fibrosis will provide important clues for developing antifibrotic therapy.

For clinical trials, several big challenges exist. First, the translatability from the bench (cell culture and animal models) to the bedside (approval of antifibrotic drugs) has been poor. Here efforts need to be undertaken to better characterize the existing animal and cell culture models as to their similarities with human disease on a molecular level. Second, there is a lack of validated measurement tools and end points to accurately measure disease activity and the degree of intestinal fibrosis, differentiated from inflammation (FIGURE 8). Selection of patients at high risk for development or progression of fibrosis is critical to the design of efficient trials. Although multiple clinical factors, as well as serologic and genetic markers, have been identified as having prognostic value, none can discriminate future stricturing from penetrating behavior or need for surgery with sufficiently high accuracy to be able to stratify CD patients at a high risk of strictures for recruitment into clinical trials (497). Therefore, development of a robust clinical prediction rule for stricturing disease is a research priority. In planning a RCT of an antifibrotic, it is paramount to accurately measure the total burden of fibrosis or change in fibrosis burden over time (497). However, validated modalities to quantify fibrosis are currently lacking, which accounts for the current variability in definitions of stricturing CD (498). This issue makes it extremely difficult to assess the response to antifibrotic therapy in clinical trials. Cross-sectional imaging techniques, including intestinal ultrasound, CT enterography (CTE), and MR enterography (MRE), are accurate in detecting intestinal strictures or assessing the degree of inflammation. However, their relative capabilities for quantification of fibrosis are unknown. Specifically, responsiveness to change following treatment has not been evaluated for any of the cross-sectional imaging modalities (499). Several IBD interest groups have engaged in discussions to address these needs (498, 500). The Stenosis Therapy and Anti-Fibrotic Research (STAR) Consortium was created to establish accepted clinical trial end points for fibrostenosing CD. Definitions for diagnosis of fibrostenosis, improvement following treatment, and treatment targets were established with RAND/UCLA methodologies (498). The need for intervention (EBD or surgery) within 24–48 wk has been proposed as an appropriate end point to assess antifibrotic agents in pharmacological trials of patients with stricturing CD (498). A program for development of valid trial end points, including a patient-reported outcome tool, radiology indexes, and histopathology index, is now well underway, and it will facilitate the performance of future trials of antifibrotic therapies. The ultimate success will be achieved by combining the pipelines for preclinical targets and the development of validated clinical trial end points with partnerships with industry to properly test novel compounds (FIGURE 9).

FIGURE 8.

Challenges and solutions for development of antifibrotic therapy for intestinal fibrosis.

FIGURE 9.

Ultimate success for antifibrotic drug development will be achieved by combining the pipelines for preclinical targets and the development of validated clinical trial end points with partnerships with industry to properly test novel compounds. IBD, inflammatory bowel disease.

7. CONCLUSIONS

Despite advances in the treatment of CD, the incidence of structural bowel damage, including strictures, fistulae, and abscesses requiring surgical resection, remains high. No specific antifibrotic therapy is available in CD; however, multiple mechanisms and compounds have been tested in organs other than the gut. Based upon these novel mechanisms, several molecular candidates could be used for the treatment of intestinal fibrosis. The STAR Consortium currently addresses the lack of clinical trial end points to allow for testing of novel antifibrotic drugs in the near future.

SUPPLEMENTAL DATA

Supplemental Figure S1 and Supplemental Tables S1–S7: https://doi.org/10.6084/m9.figshare.14061767.

GRANTS

This work was supported by the Helmsley Charitable Trust through the Stenosis Therapy and Anti-Fibrotic Research (STAR) Consortium (No. 3081 to F.R.), the Crohn’s and Colitis Foundation (No. 569125 to F.R.), the National Institutes of Health (NIDDK K08DK110415 and R01DK123233 to F.R.), the Cleveland Clinic through the LabCo program to F.R., and the National Science Foundation of China (No. 81970483 to R.M.).

DISCLOSURES

D.B. is on the advisory board of or is a consultant for AbbVie, Amgen, Arena, BNG Service, Celltrion, Dr. Falk Foundation, Ferring, Galapagos Janssen-Cilag, Medical Tribune, MSD, Pfizer, Pharmacosmos, Roche, Takeda, Thieme, Tillotts Pharma, and Vifor. V.J. has received consulting fees from AbbVie, Eli Lilly, GlaxoSmithKline, Arena Pharmaceuticals, Genetech, Pendopharm, Sandoz, Merck, Takeda, Janssen, Alimentiv Inc., Topivert, and Celltrion and speaker’s fees from Takeda, Janssen, Shire, Ferring, Abbvie, and Pfizer. B.G.F. has received grant/research support from AbbVie Inc., Amgen Inc., AstraZeneca/MedImmune Ltd., Atlantic Pharmaceuticals Ltd., Boehringer-Ingelheim, Celgene Corporation, Celltech, Genentech Inc/Hoffmann-La Roche Ltd., Gilead Sciences Inc., GlaxoSmithKline (GSK), Janssen Research & Development LLC., Pfizer Inc., Receptos Inc./Celgene International, Sanofi, Santarus Inc., Takeda Development Center Americas Inc., Tillotts Pharma AG, and UCB, consulting fees from Abbott/AbbVie, Akebia Therapeutics, Allergan, Amgen, Applied Molecular Transport Inc., Aptevo Therapeutics, Astra Zeneca, Atlantic Pharma, Avir Pharma, Biogen Idec, BioMx Israel, Boehringer-Ingelheim, Bristol-Myers Squibb, Calypso Biotech, Celgene, Elan/Biogen, EnGene, Ferring Pharma, Roche/Genentech, Galapagos, GiCare Pharma, Gilead, Gossamer Pharma, GSK, Inception IBD Inc, JnJ/Janssen, Kyowa Kakko Kirin Co Ltd., Lexicon, Lilly, Lycera BioTech, Merck, Mesoblast Pharma, Millennium, Nestle, Nextbiotix, Novonordisk, Pfizer, Prometheus Therapeutics and Diagnostics, Progenity, Protagonist, Receptos, Salix Pharma, Shire, Sienna Biologics, Sigmoid Pharma, Sterna Biologicals, Synergy Pharma Inc., Takeda, Teva Pharma, TiGenix, Tillotts, UCB Pharma, Vertex Pharma, Vivelix Pharma, VHsquared Ltd., and Zyngenia, and speaker’s bureau fees from Abbott/AbbVie, JnJ/Janssen, Lilly, Takeda, Tillotts, and UCB Pharma; is a scientific advisory board member for Abbott/AbbVie, Allergan, Amgen, Astra Zeneca, Atlantic Pharma, Avaxia Biologics Inc., Boehringer-Ingelheim, Bristol-Myers Squibb, Celgene, Centocor Inc., Elan/Biogen, Galapagos, Genentech/Roche, JnJ/Janssen, Merck, Nestle, Novartis, Novonordisk, Pfizer, Prometheus Laboratories, Protagonist, Salix Pharma, Sterna Biologicals, Takeda, Teva, TiGenix, Tillotts Pharma AG, and UCB Pharma; and is the Senior Scientific Officer of Alimentiv Inc. F.R. is a consultant to Agomab, Allergan, AbbVie, Boehringer-Ingelheim, Celgene, Cowen, Genentech, Gilead, Gossamer, Guidepoint, Helmsley, Index Pharma, Jannsen, Koutif, Metacrine, Morphic, Pfizer, Pliant, Prometheus Biosciences, Receptos, RedX, Roche, Samsung, Takeda, Techlab, Thetis, and UCB and receives funding from the Crohn’s and Colitis Foundation of America, the Helmsley Charitable Trust, Kenneth Rainin Foundation, and the National Institutes of Health. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

S.-N.L., the STAR Consortium, and F.R. conceived and designed research; S.-N.L., the STAR Consortium, and F.R. analyzed data; S.-N.L., the STAR Consortium, and F.R. interpreted results of experiments; S.-N.L., the STAR Consortium, and F.R. prepared figures; S.-N.L., R.M., C.Q., D.B., J.W., J.L., D.H.B., V.J., B.G.F., M.-H.C., the STAR Consortium, and F.R. drafted manuscript; S.-N.L., R.M., C.Q., D.B., J.W., J.L., D.H.B., V.J., B.G.F., M.-H.C., the STAR Consortium, and F.R. edited and revised manuscript; S.-N.L., R.M., C.Q., D.B., J.W., J.L., D.H.B., V.J., B.G.F., M.-H.C., the STAR Consortium, and F.R. approved final version of manuscript.