Understanding fibrosis in eosinophilic esophagitis: Are we there yet?

Abstract

Eosinophilic esophagitis (EoE) is an immune/antigen-mediated, progressive fibrostenotic disease characterized by symptoms of esophageal dysfunction and abnormal eosinophilic infiltration in the esophagus. Despite current treatment modalities of dietary antigen elimination or topical corticosteroids, a subset of patients do not have clinical or histologic response. Even with resolution of superficial epithelial eosinophilia, patients may still have progressive subepithelial fibrosis, which may lead to esophageal strictures over time. Histologic identification of subepithelial fibrosis requires deep esophageal biopsies, which are not routinely obtained. Herein, we review the challenges in diagnosing and treating fibrosis in EoE. We propose the novel concept of vitamin D supplementation to treat fibrosis in EoE through downregulation of profibrotic mediator, transforming growth factor-beta.

Transforming growth factor-beta, a central mediator in esophageal remodeling in eosinophilic esophagitis, may be downregulated by vitamin D.

1 INTRODUCTION

Eosinophils are multifunctional cells of the innate immune system linked to allergic inflammation. The role of eosinophils in the gastrointestinal (GI) tract is poorly defined, although the presence of eosinophils can be normal in the stomach (< 10), duodenum (< 20), and colon (10–30). However, eosinophils are not present in a normal esophagus,1 and their migration to the esophagus has been linked to a variety of GI disorders (Table 1), including eosinophilic esophagitis (EoE).2

Eosinophilic gastroenteropathies are a broad spectrum of diseases in which there is abnormal infiltration of eosinophils to the GI tract without other known causes of eosinophilia.1 They are immune-mediated diseases, which appear to arise from environmental influences, such as food allergy in genetically predisposed individuals. Common treatment strategies include the removal of food antigens, preventing eosinophil infiltration, and stimulating eosinophil apoptosis.1–5 Prognosis of these disorders varies from self-limited, such as infant-onset eosinophilic proctocolitis from cow’s milk protein allergy, to progressive and relapsing such as EoE (Table 2).1

Of these eosinophilic gastroenteropathies, EoE is described as a progressive fibrostenotic condition.6 Despite conventional therapies of corticosteroids or dietary elimination, a subset of patients are nonresponsive7–11 and progress to stenosis requiring recurrent esophageal dilations (Fig. 1),10,12–15 that can have a significant impact on quality of life.16 As understanding of the pathophysiology of EoE begins to unfold, TGF-β1 has been identified as the central mediator of fibrosis in EoE.11,17–20 Herein, we will focus on advances in the identification, treatment, and pathophysiology of fibrosis in EoE and discuss the potential role of TGF-β1 as a target for treatment.

2 BACKGROUND AND GENETIC PREDISPOSITION

EoE is a chronic immune-mediated clinicopathologic condition of the esophagus characterized by increased eosinophils (> 15 per high power field) isolated to the esophagus despite high-dose acid suppression.2,5 First described in the 1970s,21 EoE was identified as a distinct disease from gastroesophageal reflux disease (GERD) in the 1990s,5 and now considered to be a progressive fibrostenotic disease.6 With an estimated disease prevalence of 1/200022 at a 3:1 male to female ratio5,22 and annual healthcare cost of $1.4 billion,23 EoE places a significant burden on society. Defined as a T_H2-associated chronic allergic inflammatory disease, EoE has a strong heritability pattern with relative risk ratios being high among family members.5 Twin studies suggest that common environmental factors24 such as preterm labor, maternal fever, caesarian delivery, or infant antibiotic or acid suppression use, may increase the risk of EoE.25 Additionally, the expression of several candidate genes (TSLP, CAPN14, EMSY, LRRC32, STAT6, and ANKRD27) together with allergic triggers, is associated with the development of EoE in susceptible individuals.26 Gene–environment interactions can be advantageous to reduce the risk of disease if understood, as seen in the reduced risk of EoE development in breastfed patients with a mutation in CAPN14.27 Genetic markers for disease severity or fibrosis are yet to be identified.

3 ASSESSMENT AND DIAGNOSIS OF FIBROSIS IN EoE

Since subepithelial esophageal tissue is not routinely sampled during esophagogastroduodenoscopy, the exact prevalence of fibrosis in EoE is unknown. Analysis of biopsy specimens that include subepithelial tissue in EoE suggest that the prevalence of subepithelial fibrosis ranges from ∼57–90%.8,28,29 Symptomatically, patients with fibrosis are more likely to have signs of esophageal dysfunction including dysphagia and food impaction, although they can also present with nonspecific upper GI symptoms such as pain and vomiting (Table 3). Confirmatory diagnosis requires histologic analysis with deep esophageal biopsies, although there is a fibrostenotic phenotype diagnosed by visual changes seen on endoscopy.6,30 There may be benefit to using other diagnostic modalities such as barium esophagram, endoscopic ultrasound, or high resolution manometric testing.

3.1 Symptom presentation of fibrosis in EoE

Symptom presentation varies with age, as younger children present with symptoms of GERD such as abdominal pain and vomiting, while older children and adults present with signs of esophageal dysfunction such as food impaction and dysphagia,2,5 which is thought to mirror the progressive fibrostenotic nature of the disease (Table 3). Symptoms of dysphagia and food impaction are associated with fibrosis in EoE,31 although pediatric patients can have subepithelial fibrosis even in the absence of esophageal dysfunction symptoms.11 The density of subepithelial fibrosis is associated with age in adults9 but not in pediatric patients.31 Fibrosis can lead to esophageal dysmotility,32 and delayed diagnosis is associated with stricture formation that is time dependent,33,34 suggesting that patients with esophageal fibrosis may develop stricture with time.

3.2 Endoscopic findings of fibrosis in EoE

Endoscopic findings in EoE include loss of vascular pattern, longitudinal furrowing, rings, white plaques, fragile mucosa, and reduced esophageal caliber.5 Esophageal furrowing or ridging can be associated with subepithelial fibrosis in pediatric patients, although endoscopic findings can be visually normal as well.11,29 The validated EoE EREFS (exudate, rings, edema, furrows, and strictures) system characterizes visual features endoscopic scoring6 reflective of the fibrostenotic phenotype in adults (Fig. 2).30

3.3 Biopsy recommendations in EoE

There is substantial variability in biopsy practices among gastroenterologists evaluating eosinophilic diseases.35 Current consensus guidelines are to obtain two to four biopsies from multiple sites of the esophagus focusing on areas of visual abnormalities.2,5,36 Obtaining biopsies of the esophagus is technically more challenging compared to other parts of the GI tract because of the tubular nature of the esophagus.37 Since the squamous epithelium easily detaches from underlying soft tissue, biopsies are frequently small and only contain superficial squamous epithelium.37 The average depth of esophageal biopsy is 0.5–1 mm compared to the average full thickness of adult esophageal wall of 10 mm.37 While, there are no recommendations for specimen depth to diagnose EoE, the lamina propria should be present to diagnose fibrosis in EoE.31

In a direct comparison, large capacity biopsy forceps were shown to take larger tissue samples than standard, but not deeper.38 However, needled biopsy forceps obtain deeper biopsies than those without a needle.38 Bussann et al. systematically evaluated four types of biopsy forceps to assess the safety and adequacy of deep esophageal biopsies in 30 adult patients with EoE.39 Despite attempting to obtain deep esophageal biopsies by positioning the biopsy forceps tangentially, opening the forceps, and suctioning the mucosa towards the tip, only 55% of biopsies using large capacity radial alligator jaw forceps included subepithelial tissue. This percentage was higher for two static jaw type and alligator jaw type forceps (93, 97, and 80%, respectively).39 Additional information is required to delineate the safety of obtaining deep esophageal biopsies in EoE, particularly in pediatric patients.

3.4 Histologic findings of fibrosis in EoE

Eosinophils may be found in every layer of the esophagus in EoE. Surgical resection specimens indicate that allergic inflammation in EoE is not restricted to the squamous epithelium but also extends to the lamina propria of the mucosa, submucosa, and even the muscularis and adventitia.40,41 Esophageal fibrosis is diagnosed based on collagen deposition in the lamina propria of esophageal biopsies31,42 (Fig. 3), and may be evidence of early fibrostenotic remodeling before endoscopic changes are visible. While a correlation between eosinophil count in the lamina propria and fibrosis has not been established,31 intraepithelial eosinophil counts have been significantly associated with subepithelial fibrosis in pediatric patients.31,43 However, a causative relationship has not been established in adults, as a reduction in intraepithelial eosinophil counts does not correlate to improvement in fibrosis.9 Epithelial eosinophilia in EoE is known to have a patchy distribution,36 but the fibrosis in EoE is hyalinized.29

3.5 Differentiating fibrosis in EoE from other eosinophilic GI diseases

Inflammatory changes in GERD are most prominent in the distal esophagus, and improve proximally compared to EoE where the inflammation is patchy.37,44 In addition, fibrosis is more likely to be distal when compared EoE.29 Fibrosis occurs in direct relation to lymphoid cell aggregates within submucosal space in GERD, while fibrosis in EoE is more homogenous.29

Subsets of patients with esophageal eosinophilia respond to high dose proton pump inhibitor (PPI) and are thus termed proton pump inhibitor responsive esophageal eosinophilia (PPI-REE). Current guidelines recommend differentiating PPI-REE from EoE.2,5 However, evolving evidence from clinical and histologic findings as well as transcriptome analyses showing significant overlap, suggests that PPI-REE may be a subphenotype of EoE.45,46 While there may be collagen deposition in the lamina propria in PPI-REE patients that have some similarity to that in EoE patients (Fig. 3), studies investigating the pathogenesis of fibrosis in PPI-REE are lacking.

3.6 Imaging for esophageal stricture diagnosis

Presently, recommendations for the use of imaging to identify esophageal stricture are not available. Sensitivity of radiologic detection of stricture in EoE patients compared to endoscopic findings ranged from 29% in pediatric patients and 71–77% in adults.32,47,48 In a retrospective review of EoE patients with confirmed esophageal stricture who underwent esophagram and esophagogastroduodenoscopy, Menard-Katcher et al. noted that 55% of strictures identified on esophagram had not been identified on endoscopy alone.49 However, in a retrospective study of pediatric EoE patients who had undergone barium studies, of five who had previously undergone endoscopy for food impaction, only one demonstrated stricture.32 These data suggest that barium esophagram may be beneficial to identify strictures in persistently symptomatic patients. However, impaired esophageal motility may also be the cause of food impaction rather than esophageal stricture.

3.7 Endoscopic ultrasound for esophageal stricture

Information on the potential use for endoscopic ultrasound (EUS) in diagnosis and management of EoE is limited. Although EUS is not routinely performed to monitor EoE, Yamabe et al., demonstrated its use in the identification of diffuse esophageal wall thickening in areas of esophageal stricture in an 85-year-old male with EoE.50 Muscular wall thickening can persist, despite resolution of both epithelial eosinophilia and symptoms one month after topical corticosteroid use.50 However, since subepithelial fibrosis was not considered in this case report, it is unclear if EUS may also be a useful tool to identify fibrosis in EoE. Pediatric EoE patients have increased total wall thickness, combined mucosa and submucosa, and muscularis propria compared to healthy controls.51 One potential mechanism of increased wall thickness could be from fibrosis, but additional information is required to determine utility of EUS in this setting.

3.8 Esophageal dysmotility in the setting of fibrosis with EoE

Patients with EoE may have esophageal dysmotility, and one theory for the pathogenesis of symptomatology is that patients with fibrosis undergo changes to the underlying esophageal musculature leading to dysmotility (Fig. 4).52

Early studies using conventional manometry showed esophageal dysmotility of circular esophageal muscle ranged from 10–33% of adult EoE patients.53–55 However, using EUS, Korsapati et al., concluded that it may be longitudinal rather than circular muscles that may lead to dysphagia.56 Using prolonged esophageal manometry, Nurko et al., determined that children with EoE are more likely to have ineffective peristalsis during swallowing as well as resting states57 that could be a marker of underlying fibrosis. The use of high resolution manometry was used to demonstrate that patients with fibrostenotic features were more likely to have increased intrabolus pressures without affecting the distal contractile integral or integrated relaxation pressure values when compared to EoE patients with the inflammatory phenotype.58

More recently, functional lumen imaging probe (FLIP) technology has been used in pediatric patients with increased epithelial eosinophil density, lamina propria fibrosis, and fibrostenotic phenotypic to show decreased esophageal distensibility and altered function.59 In an adult cohort, FLIP evaluation showed that ring severity rather than eosinophil count, was associated with esophageal distensibility.60

4 TREATMENT OF FIBROSIS IN EoE

EoE was initially treated with systemic steroids similar to other allergic diseases.61 However, topical steroids were found to be equally effective in achieving initial histologic and clinical improvement62 although a comparison of topical and systemic steroids on the impact of fibrosis is unavailable. While there are ongoing trials of novel therapies for EoE, current guidelines recommend that patients are treated with dietary restrictions or topical corticosteroids.2,5 Both elimination diet and topical steroid therapy can reduce esophageal eosinophilia9,28,62–77 and distensibility.78

While topical steroid therapy has shown sustained reductions in epithelial eosinophils in pediatric population,11 the effect of topical steroids in adults diminishes over time.7,11 The lasting effect on esophageal distensibility using FLIP is yet to be determined. The optimal timing of therapy is not established, but given its progressive fibrostenotic nature6 and recurrent relapse off therapy,62 it seems these patients may require treatment for the duration of their lives. The development of novel therapeutics that target both eosinophil migration (CRTH2 antagonist) and the inflammatory pathway (anti-IL4RA, anti-IL-13, anti-IL-5RA)26 in EoE is ongoing as the understanding of EoE pathophysiology advances. While there are no Food and Drug Administration-approved treatments for EoE to date in the United States, Jorveza, an orodispersible topical budesonide was recently (November 9, 2017) approved by The European Medicines Agency Committee for Medicinal Products for Human Use79 for EoE treatment. However, its effectiveness in reducing subepithelial fibrosis is yet to be determined.

4.1 Dietary therapy of fibrosis in EoE

There is limited evidence that dietary elimination therapy may reverse esophageal remodeling in pediatric patients. While the elemental diet was shown to reverse esophageal fibrosis in a single pediatric patient by Abu-Sultaneh and colleagues,28 Lieberman et al., showed that of 17 EoE patients with subepithelial tissue on both pre- and posttherapy biopsies undergoing dietary therapy alone, only three (17.6%) had fibrosis resolution.8

Studies in adult patients have also shown that dietary therapy alone can have favorable clinical, endoscopic, and histologic responses in EoE,80,81 but suggest that fibrostenotic phenotype may be less likely to respond. In a prospective cohort of adults evaluating the efficacy of six food elimination diet, Gonsalves et al. reported that patients with the fibrostenotic phenotype with strictures and rings were less likely to respond favorably to the elimination diet than those with the inflammatory phenotype.80 Others have shown that despite a reduction of esophageal eosinophils in response to the elimination diet, the majority of responders show evidence of fibrostenotic phenotype (i.e. rings), although the presence of fibrosis based on histologic diagnosis was not reported.81

4.2 Topical corticosteroids of fibrosis in EoE

Treatment of esophageal fibrosis with both topical fluticasone and budeosnide has been shown to significantly decrease subepithelial fibrosis in pediatric patients with EoE28,43 with reduction in epithelial eosinophils shown as a predictor of resolution of remodeling.43 This effect on tissue remodeling in pediatric patients may not be sustained over time. Rajan et al. conducted a longitudinal study of pediatric EoE patients over 10 years. These authors showed that patients who initially responded to topical corticosteroids had sustained reduction in epithelial eosinophils, but starting around 4.1 years into the study, those patients had increased TGF-β1 positive cells and lamina propria fibrosis.11 In adults, despite a reduction in epithelial eosinophils with topical corticosteroid therapy, there was no significant decrease in subepithelial remodeling9 or esophageal rings.63 Similarly, there was no significant change in pre- and posttherapy subepithelial fibrosis scores in adult patients prospectively studied with topical budesonide treatment.71

4.3 Esophageal dilation as treatment of esophageal strictures

Esophageal dilation is reserved for patients with treatment-resistant disease, dysphagia, and/or food impactions. Although early case series reported high rates of perforations in adults with EoE,82,83 recent meta-analysis indicates that esophageal dilation (both balloon and bougie) is safe for both pediatric and adult patients.15,84 The most frequent complications reported are mucosal tear and chest pain.10,84 Despite dilation and concurrent treatment with topical corticosteroids, a high percentage of patients require repeat dilations suggestive of possible ongoing remodeling.10,15,84

5 PATHOGENESIS OF TISSUE REMODELING AND FIBROSIS IN EoE

Histologically, esophageal remodeling is defined as basal zone hyperplasia, dilated intercellular spaces, rete peg elongation and desquamation, and increased vascularization and fibrosis in the subepithelial lamina propria26 (Fig. 3). Esophageal lamina propria fibrosis is related to collagen density.31 Esophageal eosinophilic inflammation and remodeling progress to fibrosis in some patients with EoE despite treatment. Understanding the pathophysiology of fibrosis in EoE is vital to develop more effective treatment regimens.

While histologic measure of disease activity focuses on eosinophil density,2,5 there may be a cooperative role between eosinophils and mast cells in the pathogenesis of EoE wherein both cells increase TGFB1 expression.26 TGF-β1 signaling via the phospho-SMAD2/3 pathway, may be a central molecular mediator of EoE,17,42 as it has been implicated in the five main areas–smooth muscle dysfunction, epithelial remodeling, barrier dysfunction, collagen deposition, and angiogenesis–of disease pathogenesis (Fig. 5).85,86

Antigen presentation of food allergens in EoE may be related to direct contact. In vitro studies have proven that esophageal epithelial cells have the ability to engulf, process, and present antigen to T_H lymphocytes via the MHC class II system after exposure to interferon-γ.87 Esophageal biopsies from EoE patients also have elevated interferon-γ87 suggesting that antigen presentation in EoE may be, in part, through direct contact.

Dilated intercellular spaces is a hallmark of a dysregulated epithelial barrier and reflects increased paracellular permeability.88 Secreted interleukin (IL)-13 during esophageal inflammation triggers expression of CAPN14 that can lead to a reduction in desmoglein-1 thereby leading to a loss of barrier integrity.26 Furthermore, TGF-β contributes to epithelial barrier dysfunction by reducing the epithelial tight junction protein, claudin 7.86 Increased angiogenesis contributes to increased edema while secretion of proangiogenic factors such as vascular endothelial growth factor (VEGF) and TGF-β1 may promote vascular remodeling in the esophagus.85,89

Applying concepts from the field of allergic asthma,90–92 and taking into consideration the importance of the innervation to normal esophageal function, we present the novel concept for the role of vasoactive intestinal peptide (VIP) as a regulator of eosinophil recruitment and degranulation in EoE. Eosinophils produce VIP and express VIP receptors93 thereby being susceptible to both autocrine and paracrine regulation. Our own ongoing studies initiated in 2014 show that peripheral and esophageal eosinophils from pediatric EoE patients produce VIP and have a distinct expression pattern for VIP receptors when compared to GERD or PPI-REE patients (unpublished). Since TGF-β induces VIP expression,94 we are exploring the hypothesis that TGF-β promotes a positive feedback loop for persistent recruitment and activation of eosinophils through VIP in EoE.

Basal zone hyperplasia is more severe in EoE patients when compared to reflux esophagitis and may be a marker of underlying cell proliferation. TGF-β1-regulated epithelial cell expression of plasminogen activator inhibitor-1 in EoE biopsies correlates with basal zone hyperplasia as well as other markers of esophageal fibrosis.19

Eosinophil involvement in the GI tract is associated with smooth muscle hypertrophy.41 Phospholamban activation in the presence of TGF-β1 causes smooth muscle contraction.18 In addition, release of granule mediators such as major basic protein-1 by activated eosinophils alters smooth muscle contractility through inhibition of M2 muscarinic receptors.95 As a profibrogenic mediator, TGF-β1 promotes fibroblast-mediated collagen-matrix contraction96 and the transition of fibroblasts to myofibroblasts thereby promoting type 1 collagen deposition.97

5.1 TGF-β1 as a potential target of treatment

While TGF-β is important in homeostatic functions and is crucial in regulation of the healthy immune system, dysregulation has been implicated in many disease states.98–101 TGF-β signaling pathway is considered the “master regulator of fibrosis”.100 Numerous pharmaceutical agents have been developed to target the TGF-β signaling pathway, primarily for cancer treatment, but also for vascular symptoms of Marfan syndrome and fibrosis associated with systemic sclerosis.102 TGF-β inhibitors, galunisertib, and fresolimumab, are in Phase I and II clinical trials for use in treatment of various cancers.103 Galunisertib, a small molecule inhibitor of TGF-β1 receptor, can inhibit SMAD2 phosphorylation, thereby blocking effects of TGF-β1 in ex vivo in liver fibrosis.104 Fresolimumab, a neutralizing monoclonal antibody that targets all three TGF-β isoforms, can rapidly inhibit TGF-β-regulated gene expression and reduce skin fibrosis in systemic sclerosis.105

Treatment with vitamin D may reduce TGF-β through vitamin D receptor (VDR) regulation.106–111 Although VDR is best known for its role in calcium homeostasis, recently vitamin D has been described as an immunomodulator in various diseases.106–111 VDR signaling is a known inhibitor of TGF-β/Smad3 signal transduction on fibroblasts in fibrotic states of many organs including the bowel in Crohn’s disease, liver in hepatic fibrosis, skin in systemic sclerosis, and ovaries in polycystic ovarian syndrome.107,109–111 Recently, studies of VDR in skin fibrosis in systemic fibrosis showed that VDR is a negative regulator of the TGF-β/Smad pathway.111

EoE patients are known to be vitamin D insufficient,112 however, the role of VDR in immunomodulation of fibrosis in EoE is yet to be explored. Preliminary work in our laboratory suggest that peripheral blood eosinophils from pediatric EoE patients have downregulated expression of the VDR and upregulated expression of TGFB indicating a profibrotic state (unpublished), and we propose that the downregulation of VDR on eosinophils in EoE patients may contribute to fibrosis in EoE.

Treatment of vitamin D deficiency in various disease states has been shown to induce VDR expression and reduce TGF-β in a SMAD-dependent manner in mouse model and human studies.109–111 In vitro treatment of skin fibroblasts with vitamin D₃, induced VDR expression and decreased activity of the TGF-β/Smad pathway.111 In a randomized placebo-controlled clinical trial, vitamin D supplementation decreased circulating levels of TGF-β and VEGF.109 Vitamin D treatment reduced intestinal fibrosis by regulating the TGF-β/Smad pathway in a mouse model of fibrosis in Crohn’s disease.106 Similarly, treatment of vitamin D deficiency in EoE may lead to induction of VDR and reduction of TGF-β suggesting a novel function for vitamin D therapy as an inhibitor of fibrosis in EoE (Fig. 5).

6 CONCLUSION

EoE is a progressive fibrostenotic disease for which there is no cure. We suggest that subepithelial fibrosis may be an early finding of remodeling that progresses to esophageal stricture over time. Herein, we review the pathophysiology of and challenges in diagnosing fibrosis and strictures in EoE. Given the technical difficulties of obtaining deep esophageal biopsies for evaluation of subepithelial fibrosis, the use of advanced endoscopic techniques and barium esophograms may aid in diagnosis of strictures with persistence of symptoms despite therapy. TGF-β signaling is likely a central mediator in the pathophysiology of fibrosis in EoE. We suggest that delineating the role of VDR on fibrosis in EoE may lead to potential therapies and for the first time, propose that vitamin D may be a potential therapy to reverse fibrosis in EoE.

ACKNOWLEDGMENTS

J.A.-L. would like to acknowledge Drs. Dennis Black, Sandra Arnold, Alicia Diaz-Thomas, Mark Corkins, Jonathan McCullers, and the late Russell Chesney (all in the Department of Pediatrics, UTHSC) for their support throughout her fellowship. This work was partially supported through the Department of Pediatrics Clinical Fellowship Grant (J.A.-L.), and the Institute for Research, Innovation, Synergy and Health Equity (iRISE) Grant (A.E.S.).

DISCLOSURES

The authors declare no conflicts of interest.

Abbreviations

 
• EoE

  eosinophilic esophagitis

 
• EUS

  endoscopic ultrasound

 
• FLIP

  functional lumen imaging probe

 
• GERD

  gastrointestinal reflux disease

 
• GI

  gastrointestinal

 
• PPI

  proton pump inhibitor

 
• PPI-REE

  proton pump inhibitor responsive esophageal eosinophilia

 
• VEGF

  vascular endothelial growth factor

 
• VIP

  vasoactive intestinal peptide

REFERENCES