Non-invasive biomarkers of eosinophilic esophagitis.

Eosinophilic esophagitis (EoE) is an emerging allergen-mediated disease characterized by symptoms of esophageal dysfunction and eosinophilic inflammation. EoE diagnosis requires 15 eosinophils per high power field (eos/HPF) in tissue biopsies endoscopically obtained. The need for several endoscopies to monitoring the disease and the absence of validated non-invasive biomarkers or tools are the main reasons for the significant burden on affected patients and the healthcare system. There is a critical need for non-invasive or minimally invasive biomarkers. In the last years, several efforts have been made to identify potential biomarkers for diagnosing and monitoring the disease that we summarized in this review. The future of EoE is exciting from both a diagnostic and therapeutic standpoint. Further research is required to confirm phenotypes and histological or serological biomarkers to provide a novel endotype classification based on different cytokine or genetic signatures relevant to precision medicine.

Introduction

Eosinophilic gastrointestinal disorders (EGIDs) are emerging inflammatory diseases which may involve any part of the gastrointestinal (GI) tract and lead to the eosinophilic mucosal infiltration in the absence of secondary causes of intestinal eosinophilia (1, 2). Based on the site of the eosinophil inflammations, EGIDs are classified into eosinophilic esophagi-tis (EoE) and nonesophageal EGIDs, distinct in eo-sinophilic gastritis (EoG), gastroenteritis (EoGE), and colitis (EoC) (1). While nonesophageal EGIDs still represent a clinical enigma for clinicians, EoE is considered the prototype of EGIDs with standardized guidelines (1, 3). EoE is a chronic/remittent, allergen-mediated disease characterized by esophageal dysfunction and eosinophilic infiltration, affecting both children and adults, with a male-female ratio of 3:1 (4). The prevalence of EoE is significantly increased in the last decade. It is currently considered one of the most common causes of upper gastrointestinal morbidity, detected in 12% — 23% of patients undergoing endoscopy for dysphagia and about 50% of subjects with food impaction (4, 5). EoE diagnosis requires 15 eosinophils per high power field (eos/HPF) in tissue biopsies endoscopically obtained, without concomitant eosinophilic infiltration in other GI tracts (3). The need for several endoscopies to monitoring the disease and the absence of validated non-invasive biomarkers or tools are the main reasons for a significant burden on affected patients and the healthcare system (6). In the last years, several efforts have been made to identifypotential non-invasive biomarkers for diagnosing and monitoring the disease. Biomarkers may provide new insight into the understanding of EoE pathogenesis and defining potential endotypes with relevant impact on precision medicine.

Biomarkers are measures of biological status. According to the Food and Drug Administration (FDA) — National Institutes of Health (NIH) definition, a bi-omarker is a “defined characteristic measured as an indicator of normal biological processes, pathogenic processes or responses to an exposure or intervention” (7). This definition is broad and encompasses therapeutic interventions and molecular, histologic, radiographic, or physiologic characteristics. According to their putative applications, several categories of biomarkers have been identified, and often, they may overlap each other (Table 1) (8). Notably, an ideal biomarker should present different features, such as reasonable costs and a significant impact on clinical management (Table 2). This review aimed to summarize current evidence on non-invasive biomarkers for EoE diagnosis and monitoring, highlighting promising tools and future potential candidates. We performed a non-systematic review of articles via the online database PubMed, combining the terms “eosinophilic esophagitis” AND “biomarkers.” The literature review was performed in May 2021. All studies that met the following criteria were included: 1) case series, cross-sectional and cohort studies, published in English in peer-reviewed journals in the last ten years, 2) participants were children and adult patients diagnosed with EoE, according to current guidelines (3). Articles were also required to assess non-invasive biomarkers. Potentially eligible publications were manually screened and reviewed, and non-relevant publications were excluded (figure 1).

Table 1.

Biomarker classification and definition.

Biomarker classification	Definition
Diagnostic Biomarker (DB)	A DB detects or confirms the presence of a disease or identifies an individual with a disease subtype.
Monitoring Biomarker (MB)	An MB assesses the status of a disease or detects the clinical (efficacy and safety) and pharmacodynamic effects of treatment (i.e., biological therapy).
Predictive Biomarkers (PreB)	A PreB assesses if the exposure to therapy or environmental agent induces favorable or unfavorable effects in a patient or group of individuals.
Prognostic Biomarkers (ProB)	A ProB can identify the likelihood of a clinical event, disease recurrence, or progression in affected patients.
Risk Biomarker (RB)	An RB indicates the potential for developing a disease in a healthy individual.

Table 2.

Features of an ideal biomarker for the diagnosis and monitoring of EoE.

Features of an ideal biomarker
Correlate with the EoE state
Connect with EoE severity
Non-invasive and easy to collect or perform
Standardized
Have high sensitivity
Carry high specificity
Cost-effective
Low biological variation

Figure 1.

Methods and search strategy.

Blood eosinophils, eosinophil granule, and cell-surface proteins

Considering the allergic pathogenesis, most studies have focused on the rationale that EoE patients may have elevated peripheral eosinophils compared to healthy controls or subjects with gastroesophageal reflux disease (GERD) (Table 3) (9-11). Many of these studies showed that peripheral eosinophil levels might increase during active disease, but whether this marker alone reflected mucosal inflammation is still unclear. Recently, Wechsler et al. have demonstrated that absolute eosinophil count (AEC), together with a panel of plasma biomarkers, such as galectin-10 (GAL-10), eosinophil cationic protein (ECP), eosinophil-derived neurotoxin (EDN), eotaxin-3 (EOT3), and major basic protein 1 (MBP-1) were useful to identify EoE subjects and predicted esophageal eosinophilia (10). Another study showed that AEC, ECP, EDN, and interleukin-(IL)-5 had statistically significant correlations with esophageal eosinophilia (11). Less recently, Rodriguez-Sanchez et al. assessed the potential usefulness of eosinophil activity markers (peripheral eo-sinophils, total serum IgE, ECP) as a predictor of diet response. Authors demonstrated that peripheral blood eosinophils decreased significantly in responders but not in non-responders patients (9).

Table 3.

Serum biomarkers of EoE.

Author, year	Population	Study	Biomarkers	Outcome
Rodriguez-Sanchez et al, 2013 (9)	30Adults	Cross-sectional	ECP, total IgE, peripheral blood eosinophils, and the maximum peak of eosinophils/hpf	Serum total IgE and ECP do not act as markers for EoE activity
Wechsler et al, 2021(10)	71Children and adolescents	Prospective case-control study	Blood AEC.Plasma EDN, ECP, MBP-1,GAL-10, EOT2, EOT3.Urine OPN and MMP-9	Plasma (GAL-10, ECP, EDN, Eotaxin-3, MBP-1), and urine (OPN) biomarkers were increased in EoE compared to control. Therefore, GAL-10 is a potential biomarker for EoE screening
Min et al, 2017 (11)	115Children and adults	Prospective case-control study	Serum analysis of AEC, EOT3 EDN, ECP, and IL-5	, AEC, ECP, and EDN were higher in EoE subjects compared to controls and correlated with the degree of esophageal eosinophilia
Nguyen et al, 2011 (12)	77Children and adolescents	Case-control study	CD66b, phospho-STAT1, and phospho-STAT6	Measurements of CD66b and phospho-STAT levels in peripheral eosinophils may be beneficial for identifying EoE
Morris et al, 2017 (13)	31Children and adolescents	Case-control study	Peripheral blood EoP.	EoP levels were increased in patients with active EoE and significantly correlated with esophageal eosinophilia
Johansson et al, 2020 (14)	25Adults	Prospective study	IIb-integrin (CD41)	CD41 associated with circulating eosinophils is a potential non-invasive biomarker for esophageal eosinophilic inflammation
Schwartz et al, 2019 (15)	31Children and adolescents	Retrospective study	Peripheral blood EoP	Blood EoP correlates with tissue pathology during activeEoE
Author, year	Population	Study	Biomarkers	Outcome
Henderson et al, 2020 (16)	34Children and adolescents	Prospective study	Circulating eosinophil progenitors	Blood EoP levels may be used as a biomarker to detect active EoE disease
Subbarao et al, 2011 (17)	80Children and adolescents	Case-control study	Serum IL-5 and EDN	Serum EDN levels were significantly higher in subjects with EoE than controls
Schlag et al, 2013 (18)	15Adults	Prospective observational study	ECP and TRP	ECP but not TRP could be a promising non-invasive biomarker to assess response to topical corticosteroid therapy
Domenech Witek et al,2017 (18)	19Adults	Retrospective study	Serum ECP	The serial determination of ECP was proper to monitor patients with EoE
Cengiz, 2019 (20)	29Adults	Case-control study	Serum ECP	Serum ECP level was significantly higher in patients with EoE than in controls. In addition, ECP is strongly correlated with EREFS and the symptom of food impaction
Wright et al, 2018 (21)	39Adults	Prospective case-control study	Serum EPX	EoE subjects had significantly lower median EPX levels
Lu et al, 2018 (23)	31Children and adolescents	Case-control study	Serum 15-HETE	15(S)-HETE may aid in the diagnosis of EoE
Dellon et al, 2016 (24) Dellon et al, 2015 (25)	61Adults	Case-control study	Serum periostin. Serum IL-4, IL-5, IL-6, IL-9, IL-13, TGF-a, TGF-P, TNF-a, EOT-1, -2, and -3,TSLP, MBP, and EDN	Serum periostin and cytokines levels were similar in cases and controls, and there were no changes post-treatment
Dellon et al, 2017 (27)	48Adults	Case-control study	Autoantibodies (IgG1 and IgG4) to DSG1, DSG3, and to collagen XVII (NC16A)	Anti-NC16A and anti-DSG3 IgG4 autoantibodies were strongly associated with EoE. Anti-NC16A levels decreased significantly in EoE cases with a histologic response after topical corticosteroid treatment

Other studies have evaluated blood eosinophil progenitors (EoP) and eosinophil-surface markers with promising results (12-14). Johansson et al. recently reported that platelet activation and platelet-eosinophil association pathways might be involved in EoE patho-genesis, showing that CD41 (aIIb-integrin subunit) expressed on eosinophils surface was a potential non-invasive biomarker for esophageal eosinophilic inflammation (14). Another study examined whether phe-notypic analysis of eosinophil surface markers could distinguish treated from untreated disease. In 2011, Nguyen et al. found elevated surface CD66 intracellu-lar phospho-STAT1 and phospho-STAT6, which differentiated children with active EoE from treated and healthy controls (12, 15, 16). Three studies recently assessed the levels of blood EoP as potential biomarkers of active EoE, esophageal inflammation, and response to treatments both in children both adults (13, 15, 16).

Eosinophil granule proteins have been investigated as other potential markers of disease, showing inconsistent and conflicting results (17-21). Subbarao et al. determined that EDN levels provided a sustained decrease following treatment in 66 children with EoE (17). More recently, a small prospective study of 15 adults showed that serum ECP, but not tryptase (TRP), significantly correlated with tissue eosinophils after swallowed steroid therapy (18). Moreover, ECP was high in adults with EoE, and its serial determination was also helpful in monitoring the disease (1920).

Recent evidence suggested a pathogenetic role for arachidonate 15-lipooxygenase (ALOX15) in EoE. ALOX15 is upregulated and overexpressed in mu-cosal biopsies of EoE patients (22). 15(S)-hydroxye-icosatetraenoic acid (15(S)-HETE), a metabolite of ALOX15, detectable in peripheral blood, was found elevated in the EoE compared to the non-EoE group, suggesting its potential role as a disease indicator (23).

Type 2 (T2) cytokines

With an advanced understanding of EoE patho-genesis, several studies sought to assess whether T2 cy-tokines, including interleukin (IL)-4, IL-5, IL-6, IL-9, IL-13, TGF-a, transforming growth factor (TGF)-P, tumor necrosis factor (TNF)-a, EOT-1, -2, -3, thymic stromal lymphopoietin (TSLP) and periostin were increased in the peripheral circulation of affected patients (24, 25). Therefore, peripheral cytokine measurements did not consistently characterize the esopha-geal inflammation or disease activity. In addition, the results of these studies are limited by the confounding influence of other concomitant allergic diseases.

Autoantibodies

EoE has been associated with a range of autoimmune conditions, such as inflammatory bowel diseases, coeliac disease, vasculitis, or type 1 diabetes mellitus (26). Moreover, esophageal epithelial barrier dysfunction is essential in EoE pathogenesis. Antibodies against epithelial adhesion molecules are founded in several autoimmune skin conditions. Therefore, EoE may even be associated with these specific au-toantibodies. Dellon et al. recently demonstrated that anti-collagen XVII (NC16A) and anti-desmoglein 3 (DSG3) IgG4 autoantibodies were strongly associated with EoE. Moreover, anti-NC16A levels decreased significantly in EoE patients after topical corticoster-oid treatment (27).

Immunohistsochemical markers

Diagnosis of EoE requires more than 15 eos/HPF in the esophageal tracts. Therefore, other diagnostic histological findings, including a thickened mucosa with basal layer hyperplasia and papillary lengthening, eosin-ophil surface layering, and eosinophilic microabscesses, have been proposed (28). Several studies assessing his-tological biomarkers have been reported. Extracellular deposition of eosinophil granule proteins, such as eo-sinophil peroxidase (EPX), is present in the esophagus of patients with EoE and positively correlates with the peak of tissue eosinophils (Table 4) (29, 30). Moreover, EPX levels decreased in treatment responders (29). On the contrary, Schroeder et al. demonstrated that the less invasive assessment of pharyngeal EPX did not correlate with the esophageal eosinophil count in children with EoE compared to healthy controls (31).

Table 4.

Immunohistochemical biomarkers.

Author, year	Population	Study	Biomarkers	Outcome
Wright et al,2021(29)	87Adults	Case-control study	EPX	EPX was strongly correlated with tissue eosinophils accurately identified subjects with EoE and decreases in treatment responders
Saffari et al, 2017 (30)	36Adults	Case-control study	EPX	EPX levels from esophageal mucosal samples correlated with eosinophilic inflammation
Schroeder etal, 2017 (31)	21Children and adolescents	Case-control study	Pharyngeal andnasal EPX	EPX levels from the throat swabs do not correlate with esophageal eosinophil counts
Peterson et al,2019 (32)	34Adults	Retrospective study	MBP1	MBP1 is increased in esophageal biopsy specimens from symptomatic patients with EoE and may be a marker of disease activity
Kim et al, 2019 (33)	72Adults	Retrospective study	TRP, EDN, and EOT3	TRP, EDN, and EOT3 could be promising biomarkers for disease activity, symptoms, and endoscopic response
Dellon et al,2020 (34)	110Adults	Retrospective study	MBP, EOT3, andTRP	Pretreatment MBP, EOT3, and TRP levels were not strongly associated with response to topical steroids. In contrast, elevated TRP levels may be associated with nonresponse compared with complete response
Dellon et al, 2014 (35)	196Adults	Case-control study	MBP, EOT3, andTRP	Esophageal tissues from patients with EoE have substantially higher MBP, EOT3, and tryptase than controls
Dellon et al,2012 (36)	105Children and adults	Case-control study	MBP and EOT3	Patients with EoE had substantially higher levels of MBP and EOT3 staining than GERD patients
Smadi et al,2018 (37)	94Children and adults	Prospectivecross-sectionalstudy	EDN	EDN in brushing samples is significantly higher in patients having active EoE compared to healthy controls, GERD, and EoE in remission
Hui et al,2017 (39)	21Children and adolescents	Retrospective case-control study	ALOX15	ALOX15 immunohistochemistry helped support the diagnosis of EoE in situations with strong clinical suspicion
Clayton et al, 2014 (40)	30Adults	Retrospective case-control study	IgG4	The level of IgG4-positive plasma cells was increased in the lamina propria and granular extracellular IgG4 deposits
Zukerberg etal, 2016 (41)	46Adults	Case-control study	IgG4 deposits	76% of EoE cases showed int extracellular IgG4 deposits, whereas all GERD cases were negative
Rosenberg etal, 2018 (42)	36Children and adolescents	Case-control study	IgG4	Tissue IgG4 levels correlated with esophageal eosinophil counts, histologic grade, stage scores, IL-4, IL-10, IL-13 expression, and had strong associations with a subset of the EoE transcriptome
ALOX, arachidonate lipoxygenase; EDN, eosinophil-derived neurotoxin; EPX, eosinophil peroxidase; GERD, gastroesophageal reflux disease; Ig, immunoglobulin; IL, interleukin; MBP-1, major basic protein-1; TRP, tryptase.

Other eosinophil granule proteins, such as MBP-1, TRP, EDN, and EOT-3, have been evaluated as potential histological biomarkers of EoE and response to therapy, with conflicting results. (32-36). Notably, EDN in brushing samples obtained with the nasogas-tric endoscopy was significantly higher in children and young adults with active EoE than patients in remission, healthy controls, and GERD. (37).

Other tissue markers

ALOX15 plays an essential role in the metabolism of fatty acids and the production of various cytokines and chemokines. ALOX15 is expressed in blood eosinophils and respiratory epithelium. ALOX15 is also upregulated in the esophageal epithelium from patients with active EoE in contrast to esophageal fragments from patients in remission, subjects with GERD, or healthy controls (38). Thus, ALOX15 immunohistochemistry may be helpful in the diagnosis of cases with clinical features of EoE but that do not meet the histological criteria (39).

IgG4

The role of immunoglobulin G4 (IgG4) in EoE pathogenesis has not been precisely defined, and available studies reported conflicting data. One of the first studies showed an increased level of IgG4-positive plasma cells (IgG4-PC) in the lamina pro-pria and granular extracellular IgG4 deposits (40). Zuckeberg et al. reported IgG4 deposits between the squamous cells in biopsies from patients with EoE.

Moreover, IgG4 in the esophagus showed a positive correlation with concurrent peak tissue eosino-philia, histological grade, and stage according to the EoE histology scoring system (EoEHSS) (42). However, the high amount of IgG4 in esophageal mucosa still represents a conundrum. Thus, current data do not conclusively determine if high tissue IgG4 titers could be good predictors of diet response in EoE patients.

Microribonucleic acids (miRNAs) and DNA methylation

MiRNAs are single-stranded RNA molecules of 19-25 nucleotides involved in the post-transcriptional gene silencing. Several studies reported that EoE patients had a marked change in tissue-specific gene expression (Table 5). Lu et al. investigated esophageal miRNA expression profile in patients with active disease and responsive to steroids, finding that the expression levels of the most upregulated miRNAs (miR-21 and miR-223) and the most downregulated miRNA (miR-375) strongly correlated with esophageal eosin-ophil levels (43). More recently, Bhardwaj et al. found that the expression of salivary miR-4668 is higher in EoE compared to non-EoE subjects, suggesting its potential role as a non-invasive biomarker (44).

Table 5.

Epigenetic biomarkers.

Author, year	Population	Study	Biomarkers	Outcome
Lu et al, 2012 (43)	29Children and adolescents	Case-control study	miRNAs	The expression levels of the most upregulated miRNAs (miR-21 and miR-223) and the most downregulated miRNA (miR-375) were strongly correlated with esophageal inflammation
Bhardwaj et al, 2020 (44)	44Adults	Case-control study	Salivary miR-4668-5p	The expression of miR-4668 is higher in EoE vs. non-EoE subjects, suggesting its potential role as a non-invasive biomarker
Strisciuglio et al, 2021 (45)	20Children and adolescents	Case-control study	Mucosal DNAmethylationprofile	Analyses revealed striking disease-associated differences in mucosal DNA methylation profiles in children diagnosed with EoE compared to controls
Jensen et al, 2020 (46)	36Children and adults	Case-control study	DNAmethylation profile	EoE patients that respond versus do not respond to treatment have differences in their methylation profile

Other epigenetic mechanisms, different from miRNA and involved in EoE pathogenesis or response to therapies, have been recently assessed. For example, pediatric patients with EoE showed differences in mucosal DNA methylation profiles compared to controls(45). Moreover, DNA methylation differences have also been found in responder and non-responder patients(46)

Exhaled nitric oxide

Fractional exhaled nitric oxide (FeNO) is a bio-marker of eosinophilic asthma (47). However, considering the common atopic etiology, FeNO was also measured in a prospective study of 11 non-asthmatic subjects with active esophagitis before and after treatment, without any supporting role in the management of EoE (Table 6) (48). Moreover, FeNO did not help distinguish EoE from GERD (48). Therefore, no studies have shown a potential role of FeNO in EoE diagnosis and monitoring (49).

Table 6.

Other non-invasive biomarkers

Author, year	Population	Study	Biomarkers	Outcome
Leung et al, 2013 (48)	11Children and adults	Prospective study	FeNO	No supporting role for FeNO determination in the management of EoE
Lanz et al, 2012 (49)	55Children and adolescents	Case-control study	FeNO	Measurement of FeNO does not help identify EoE from GERD
Moye et al, 2019 (50)	24Children and adolescents	Prospectivecase-controlstudy	Plasmametabolomicsprofile	Notable candidate biomarkers include dimethylarginine, putrescine, and N-acetylputrescine
Cunnion et al, 2016 (51)	75Children and adults	Case-control study	Urinary 3-BT	Median normalized 3-BT levels were increased 93-fold in patients with EoE compared to controls

Metabolomics

Only one study assessed the metabolomic profile in patients with EoE. However, Moye et al. showed that plasma urea cycle metabolites (dimethylarginine, putrescine, and N-acetylputrescine) are elevated in children with EoE, and their levels are modified by proton pump inhibitor treatment (50).

3-Bromotyrosine (3-BT) is a chemical marker of eosinophil activation and is high in patients with asthma. Cunnion et al. found that 3-BT levels were increased 93-fold in patients with EoE compared to controls, providing proof of concept testing urine by a mass spectrometry method (Eosinophil Quantitated Urine Kinetic, EoQUIK) can provide a non-invasive tool to evaluate eosinophil degranulation in EoE (51).

Genetic risk loci

Eosinophilic esophagitis is a multifactorial disease. Although recent evidence suggested a fundamental pathogenetic role of the environmental factors, several studies have also reported that genetic predisposition is a significant risk factor in the development of EoE (52). Different studies, including candidate-gene identification and genome-wide association studies (GWAS), have identified gene loci that have been associated explicitly with EoE (53). These gene loci are categorized into four major groups: 1) genes involved in Type 2 (T2) inflammation, 2) epithelial barrier dysfunction, 3) enhanced fibrosis, and 4) altered immune response (54). The main genes are TSLP, calpain 14 (CAPN14), CCL26, EMSY, LRRC32, STAT6, and ANKRD27 Table 7. Additional studies founded mutations within the filaggrin gene and the promoter region of TGFB1 (55, 56). TSLP is released by activated epithelial cells and plays a fundamental role in promoting T2 differentiation (57). Levels of TLSP are increased in patients with atopic diseases, including EoE (58). CAPN14 is a cysteine protease and plays a fundamental role in the integrity of the esophageal epithelial barrier. Furthermore, its expression is only limited to the esophageal mucosa (59). However, CAPN14 expression was almost 4-fold increased in EoE patients compared to controls. Higher levels of CAPN14 expression are associated with the downregulation of DSG-1, filaggrin, and zonulin, which are pivotal proteins of the epithelial barrier (59) . CCL26 gene, which encodes for EOT3, is the most highly overexpressed esophageal transcript in patients with EoE and is critical in disease pathogenesis (60) . STAT6 is essential for T2 development and is a signaling intermediate for IL-4 and IL-13 post-IL-4 receptor alpha (IL-4Ra) engagement (53). LRRC32 is a TGF-beta binding protein, and EMSY is involved in transcriptional regulation (53). In this context, the Cincinnati Children’s Hospital researchers developed a specific diagnostic panel comprising a 96-gene quantitative PCR array to identify patients with EoE, monitor the disease and response to therapy, and improve the diagnosis and treatment (61).

Conclusion

EoE is an emerging disease affecting patients at any age and is currently considered one of the upper GI tract disorders with a relevant burden on patients and the healthcare systems (6). To date, the GI endos-copy is the gold standard for the diagnosis and follow-up of patients with EoE. Therefore, there is a critical need for non-invasive biomarkers to replace such invasive monitoring. Although this review showed promising non-invasive biomarkers, none of these has been incorporated into guideline recommendations. Despite several signs of progress in understanding EoE pathogenesis, we have more to learn as we strive to improve diagnostic modalities, discover more effective and patient-targeted therapeutic strategies, and develop more accurate disease monitoring systems. We are hopeful that the growing number of genetic, molecular expression, and immunologic analyses, in conjunction with increased differentiation of clinical phenotypes and biomarker supported endotypes, will help us explain differing therapeutic responses, predict clinical response, guide individual therapies, and improve patient outcomes. The future of EoE is exciting from both a diagnostic and therapeutic standpoint. Therefore, further research is required to confirm phe-notypes and histological or serological biomarkers to provide a novel endotype classification based on different cytokine or genetic signatures.