Biomarkers for Reflux Disease: A Literature Review

Abstract

Background and Aims

Gastroesophageal reflux disease (GERD) encompasses an array of disorders unified by the reflux of gastric contents. Owing to the multitude of potential disease manifestations, both esophageal and extra-esophageal, no single biomarker can capture the disease spectrum, making it more plausible that there be a set of GERD biomarkers, each quantifying specific aspects of GERD-related pathology. This review aimed to comprehensively search the literature on biomarkers of GERD, specifically in relation to endoscopically negative esophageal disease and excluding conventional pH-impedance monitoring.

Methods

We searched PubMed and Embase databases from January 1^st, 1996 to June 10th, 2015 for biomarkers of GERD and abstracts from recent international gastroenterology meetings.

Results

Of 1937 citations retrieved, 72 were included. Histopathologic biomarkers, baseline impedance, and serologic assays are some of the candidate biomarkers reviewed. The most unifying concept was of manifestations of impaired esophageal mucosal integrity, evident by increased ionic and molecular permeability, and/or destruction of tight junctions.

Conclusions

Impaired mucosal integrity quantified by baseline mucosal impedance, proteolytic fragments of junctional proteins, or histopathological features, has emerged as a promising GERD biomarker.

INTRODUCTION

The Montreal definition of gastroesophageal reflux disease (GERD) was craftily devised to encompass an array of disorders unified by the common umbrella of being caused by the reflux of gastric contents into the esophagus¹. Hence, a GERD diagnosis can be based on either tissue damage or ‘troublesome’ symptoms attributable to reflux. While this definition may be intellectually satisfying, it also presents a challenging task to the clinician because multiple, potentially interacting variables are in play. Conceptually, these variables can be grouped as pertaining to histopathological injury, esophagogastric junction (EGJ) competence, afferent sensitivity, neuromuscular function, or symptom perception. Furthermore, thanks to the craftiness of the Montreal Consensus, the defining features of GERD for an individual can reside in any of these domains, thereby sabotaging the prospect for a single ‘gold standard’ diagnostic test. No wonder that GERD or ‘refractory GERD’ has become the most common diagnosis in referral gastroenterology practices².

So, faced with the patient with a sore throat, cough, or chest pain who believes that they have “refractory GERD”, what is a clinician to do? Although potentially related to GERD, none of these symptoms are specific for reflux disease, making diagnostic error a real possibility. Certainly, utilizing a biomarker measuring a parameter of reflux relevant to the disease entity in question would be valuable. As background, the National Institutes of Health Biomarkers Definitions Working Group defines a biomarker as “a characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention”³. The World Health Organization defines a biomarker as “any substance, structure, or process that can be measured in the body or its products and influence or predict the incidence of outcome or disease”⁴. Hence, by these definitions, a biomarker is independent of patient-reported symptoms or outcomes. In the case of GERD, owing to the multitude of potential disease manifestations, no single biomarker could capture the spectrum of the disease making it more plausible that there be a set of GERD biomarkers indicating GERD-related pathology. The least controversial example would be endoscopy using the Los Angeles and Prague Classifications for the stratification of esophagitis and Barrett’s metaplasia^{5, 6}. Although not perfect, fine-tuning of these classifications is a much smaller task than the broader topic of biomarkers for the majority of potential GERD cases, excluding the presence of esophagitis or esophageal metaplasia. The hope is that identification of such biomarkers can become clinical tools in identifying and managing patients with non-erosive reflux disease (NERD), functional heartburn (FH), extraesophageal syndromes and negative reflux monitoring, and persistent reflux symptoms despite medical or surgical therapy. That will be the focus of this review.

METHODS

A systematic search of English-language articles from January 1^st, 1996 to June 10^th, 2015 was done by reviewing PubMed and Embase databases, as delineated in Figure 1. Abstracts from recent congresses (2012 – 2015) (Digestive Diseases Week, United European Gastroenterology Week) were also reviewed. We restricted our analysis to the esophageal and extra-esophageal syndromes with an established association with reflux per the Montreal consensus. Review articles, letters, reports and duplicates were excluded. The remaining studies were screened based on titles and abstracts and full articles were then reviewed. A total of 72 studies were identified using this methodology. Due to heterogeneity in study design and populations, quantitative comparisons and measures of diagnostic accuracy (sensitivity, specificity, positive/negative predictive values) could not be used to compare the identified biomarkers. The absence of a ‘gold-standard’ also limited the ability to draw comparisons across studies. Therefore, although a systematic search strategy was used to review the literature, the topic did not lend itself to a ‘systematic’ review from a statistical standpoint. We therefore present a narrative review of the data, limited to an analysis of the esophageal GERD syndromes due to length constraints.

Figure 1.

Search strategy and yield for GERD biomarkers.

HISTOPATHOLOGIC BIOMARKERS

Cell-to-Cell adhesion molecules

In the absence of overt mucosal damage manifest by reflux esophagitis, Barrett’s metaplasia, adenocarcinoma or peptic strictures, biomarkers of mucosal injury have been studied to identify patients with NERD. Historically, histologic evaluations of biopsy samples from normal appearing mucosa identified basal cell hyperplasia, papillary elongation, intraepithelial eosinophilia, neutrophilia, mononuclear cell proliferation, and dilated intercellular spaces (DIS) as microscopic features of GERD, even in non-erosive cases^7–9. Mechanistically, the integrity of the esophageal epithelium is vital in maintaining a barrier to noxious constituents of refluxate (acid, pepsin, bile acids, etc.). The relative impermeability of the esophageal epithelium is a function of cell-to-cell connections (tight junctions, adherens junctions, and desmosomes) providing a structural framework. Hence, manifestations of the synthesis or degradation of these tight junctions have been studied as GERD biomarkers. Animal and in vitro models of GERD have shown changes in expression and localization of Claudin-3 and -4^{10, 11}. In human studies, expression of Claudin-1, -2 and -4, ZO-1, filaggrin and occludin are reportedly altered in GERD, but no correlation has been found with histopathological parameters such as DIS and basal cell hyperplasia^12–14. NERD patients exhibit no change in the expression of these proteins¹². Furthermore, direct correlation between acid exposure and pathology is lacking. A recent study by Pardon et al evaluated the effect of luminal acid on the functional integrity and expression of these proteins¹⁵. Although they were able to induce impaired mucosal integrity after perfusion with acidic and weakly acidic solutions as measured by trans-epithelial electrical resistance, diffusion of paracellular tracers and immunofluorescence, they did not observe corresponding changes in expression of these adhesion molecules. Hence, they suggest that there may be a molecular “redistribution” rather than a change in expression leading to compromise of the mucosal barrier. Oshima et al provide further support for this hypothesis, reporting that esophageal acid exposure caused a “delocalization” of Claudin-4 from the superficial layer of the mucosa¹¹. Interestingly, there are also data suggesting that duodenal contents, including bile acids (deoxycholic acid) and trypsin, rather than gastric acid, regulate the function of tight junctions¹⁶.

Unlike the case for tight junctional proteins, evidence has been found for up-regulation of genes for desmosomal proteins. In a study comparing 95 GERD patients (51 erosive esophagitis) to 27 asymptomatic controls, Wex et al reported up-regulation by between 1.7 and 8.1-fold of genes for the desmosomal components of the esophageal epithelium barrier membrane in the GERD patients. These patients also exhibited histologic changes of DIS, basal cell hyperplasia, and elongation of the papillae suggesting that desmosomal proteins may be better than junctional proteins as markers for mucosal damage¹⁷.

Despite the inconsistent data on differentiating features of cell adhesion molecule expression, a recent report on markers of proteolytic cleavage of a junctional protein is promising. In both a mouse knockout model and tissue from GERD patients, Jovov et al found that proteolytic cleavage of (or absence of) e-cadherin was a critical event in the induction of junctional permeability of the esophageal epithelium and that this could be detected by a corresponding quantitative increase in soluble N-terminal fragment of the molecule in serum. Based on their comparison of 20 GERD patients (with and without a history of esophagitis) and 23 asymptomatic control subjects, they hypothesized that the serologic identification of these cleaved fragments may become a valuable biomarker for GERD¹⁸. Although intriguing, this hypothesis will need to be tested in larger numbers of subjects and in varied pathologies.

Dilated Intercellular Spaces (DIS)

Maintaining a tight epithelial barrier protects the subepithelial space from injury. Seminal studies by Tobey et al identified DIS as a key histologic feature among patients with both erosive and non-erosive GERD^{19, 20}, distinct from control subjects and patients with FH¹⁹. DIS was further noted to be a marker of epithelial permeability or a “shunt leak” marked by rearrangement of intercellular glycoconjugates in Ussing chamber studies, thereby allowing transepithelial passage of macromolecules up to 20 kD in size, water and hydrogen ions. The increased permeability then leads to edema and, hypothetically, activation of previously shielded esophageal chemosensitive nociceptors involved in pain perception^{20, 21}. Mechanistically, this seems a plausible explanation for hypersensitivity to reflux in patients with non-erosive disease and several studies have supported that hypothesis^{22, 23}. One such study evaluated patients with heartburn and found that the presence of microscopic esophagitis (including DIS) accurately discriminated between patients with objective reflux syndromes (non-erosive disease and hypersensitive esophagus) and those with functional heartburn, who were histopathologically indistinguishable from HV²⁴. Moreover, DIS has been consistently reproduced in experiments exposing mucosa to acidic conditions and, to a lesser degree, after exposure to bile acids and weakly acidic refluxate (i.e. duodeno-gastroesophageal reflux)^{25, 26}.

From a therapeutic standpoint, proton-pump inhibitor (PPI) therapy has been shown to reliably reverse DIS in patients with erosive esophagitis, although less so in patients with non-erosive disease^{27, 28}. Farré has suggested that PPI resistance in these NERD cases may be due to the perpetuation of DIS from non-acidic refluxate²⁵. Others have noted that DIS (and acid exposure) extends more proximally in the esophagus in NERD patients, which could be responsible for enhanced symptom perception²⁹.

Despite the promise of DIS as a discriminant between GERD and FH, the lack of specificity of DIS for GERD is a significant limitation as a biomarker. As already alluded to, DIS is not specific to acid-induced injury, as it can also be seen with weakly acidic refluxate^{25, 30, 31}. There is also no consensus on where along the esophagus biopsies should be taken to assess for DIS. More importantly, DIS has been described in conditions of emotional stress³⁰, eosinophilic esophagitis³², exposure to aspirin³⁰ and Candida infections²³. Furthermore, although studies exposing asymptomatic controls to acid have reliably reproduced DIS, these alterations have not correlated with symptom perception, thereby invoking the involvement of additional factors in the genesis of hypersensitivity in NERD patients³³. Other investigators have drawn similar conclusions, noting correlation between symptoms and DIS in patients with erosive esophagitis, but not reliably in those with NERD³⁴. Perhaps its best utility is in its negative predictive value; the absence of DIS can point toward a FH diagnosis. From a technical standpoint, as originally described, DIS measurements have been obtained using transmission electron microscopy with good sensitivity and specificity, albeit expensive, not widely available, and time-consuming^{29, 35}. Alternatively, light microscopy has been evaluated for the detection of DIS and several studies have validated it against electron microscopy^{23, 36–38}.

Immunohistochemical markers

In the absence of specific histopathologic markers for GERD, several immunohistochemical studies have sought a more specific biomarker. One such potential proinflammatory biomarker is Proteinase-Activated Receptor-2 (PAR-2), which has been thought to play a role in altering transepithelial resistance and mediation of visceral hypersensitivity. El-Rehim et al found PAR-2 to be overexpressed in patients with GERD compared with controls, without a significant difference between erosive and non-erosive cases³⁹. Kandulski et al have also shown upregulation of PAR-2 gene expression in patients with GERD, correlating with interleukin-8 expression and histopathologic alternations of GERD⁴⁰. Further studies are needed to validate these findings. Capsaicin-sensitive transient receptor potential cation channel subfamily V member 1 (TRPV1) and Protein Gene Product (PGP) 9.5 (neuronal and hypersensitivity markers) have also been evaluated in GERD patients, particularly in those with NERD. Bhat et al found that total acid exposure time in NERD patients correlated with density of PGP 9.5 and expression of TRPV1, suggesting that acid plays a role in upregulating these markers of hypersensitivity⁴¹. Interestingly, Miwa et al found that the absence of PGP 9.5 on esophageal biopsies was associated with lack of response to PPI therapy, an observation that may be useful in distinguishing NERD patients from FH⁴². Lyso-PAF acetyltransferase and bcl-2 have also been studied, though they have yielded negative results^{43, 44}. A recent study by Calebrese et al evaluated macroscopically normal appearing mucosal biopsy specimens from NERD and esophagitis patients using shotgun proteomics and immunohistochemical analyses. They detected distinct proteins in NERD not present in erosive esophagitis. These included Transitional Endoplasmic Reticulum ATPase (TER ATPase), GAPDH, Alpha 1 Acid Glycoprotein 1, Annexin A1, Calmodulin and 14-3-3. The authors concluded that NERD and erosive reflux disease may be distinct entities and that proteins involved in cellular proliferation, keratinization and stress responses may serve as differentiating biomarkers⁴⁵.

BASELINE IMPEDANCE

Conventional characterization and quantification of refluxate via catheter-based 24-hour pH-impedance monitoring has suffered from substantial variability in sensitivity and specificity for GERD, particularly in NERD. Its use is also limited by patient discomfort, compliance, and reliance on symptom reporting, which by definition, excludes it as a potential biomarker⁴⁶. At a mechanistic level, conventional impedance measurements can discriminate between liquid and gas in the esophageal lumen and determine the direction of content flow, thereby making it the most accurate way of detecting and characterizing reflux⁴⁷. However, this has not translated to reliably discriminating among GERD, NERD, FH, esophageal hypersensitivity and healthy volunteers (HV).

An alternative utilization of impedance technology is the determination of baseline impedance (BI), which measures the ‘resting’ impedance of the esophageal mucosa in the absence of swallowing or reflux events. Conceptually, it is reflective of mucosal integrity, analogous to an Ussing chamber measurement of trans-epithelial resistance and it has emerged as a potential biomarker of mucosal integrity, independent of patient symptom reporting. Seminal work by Farré et al first proposed that BI correlated with transepithelial resistance, a known marker of esophageal mucosal integrity, based on in-vivo and in-vitro studies of acid perfusion in animals and humans⁴⁸. They noted that GERD patients (both erosive and non-erosive) had lower BI values compared with HV, and suggested that this might serve as a reliable biomarker for GERD. Zhong et al further studied the relationship between BI and conventional markers of GERD, reporting that reflux events with prolonged acid exposure times and DIS were associated with low BI⁴⁹. Subsequently, Savarino et al confirmed the correlation between BI and histopathologic markers of GERD, noting lower BI in erosive and non-erosive GERD patients compared to HV or patients with FH, who were indistinguishable from one another⁵⁰. Furthermore, a recent study found that lower BI values correlated with esophageal acid exposure on pH monitoring and were lower in patients with a positive Bernstein test⁵¹.

Distinguishing NERD from FH has been clinically challenging. Kandulski et al recently reported on BI measurements in 52 patients with NERD, esophagitis and FH, and found that a cut-off value of <2,100Ω distinguished GERD patients (erosive and non-erosive) from FH with 78% sensitivity and 71% specificity⁵². They also found an inverse association between DIS and BI in the distal esophagus. Martinucci et al studied a group of patients with FH, and noted that lower BI levels were seen in those with greater than 50% symptom relief after PPI therapy, when compared to those with minimal relief and HV⁵³. Furthermore, De Bortoli et al found that patients with FH responsive to PPI therapy had lower impedance values compared to non-responders and HV, suggesting a role for BI in predicting treatment response⁵⁴.

Overall, investigators have consistently found a strong correlation between treatment response and improvement in BI. Kessing et al reported that patients with refractory GERD off PPI with low BI measurements prior to treatment had a significant increase, but not normalization, of BI values^{55, 56}. Loots et al studied a population of infants with GERD on PPI versus placebo, and noted that BI improved with PPI compared to placebo, although this was not associated with symptomatic response^{57, 58}. A recent study evaluating the effect of endoscopic fundoplication and PPI therapy on BI and heartburn severity found that both treatment methods led to partial recovery in BI when compared to HV, though as noted by Loots, there was no correlation with heartburn improvement⁵⁹. Mauritz et al also confirmed a favorable treatment response in children undergoing laparoscopic fundoplication, noting increased BI levels post-operatively correlating with fewer reflux episodes⁶⁰. On the other hand, studies in NERD patients by Ribolsi et al failed to find a difference in BI in PPI-responsive versus non-responsive patients⁶¹.

Baseline Impedance and extra-esophageal syndromes

Extra-esophageal manifestations of GERD (chronic cough, asthma and laryngitis) continue to pose a diagnostic and therapeutic challenge for gastroenterologists. Rightly or not, these chronic symptoms are often attributed to GERD without concomitant typical GERD symptoms of heartburn and regurgitation. Clearly, a reliable biomarker could be of great value in that circumstance, especially since conventional pH-impedance testing has proven of limited value. Given the encouraging data for esophageal syndromes, BI has also been evaluated in the context of these extra-esophageal syndromes. From a mechanistic standpoint, Woodland et al evaluated BI in the proximal and distal esophagus of HV. As expected, they noted lower BI distally, but interestingly, they found that mucosal calcitonin gene-related peptide (CGRP), a marker or neural sensitivity, was more superficially located in the proximal esophagus, perhaps explaining the heightened sensitivity of this area⁶². It would be of interest to apply these findings to patients with hypersensitivity, chronic cough, and laryngitis. Ribolsi et al recently studied 156 chronic cough patients (43.5% responsive to PPI, 56.5% non-responsive) and found a greater probability of PPI response in those with pathologic acid exposure time or low BI⁶³. Pauwels et al compared BI in patients with chronic cough, typical GERD symptoms, and HV. They reported lower proximal and distal BI in the cough patients compared to the typical GERD patients and lower values in both groups compared to controls. However, the typical reflux patients actually had more impedance detected reflux and proximal reflux than did the cough patients leading them to conclude that the low BI in the cough patients was likely unrelated to reflux⁶⁴. In laryngopharyngeal reflux (LPR), a study by Lee et al found an inverse correlation between BI in the distal esophagus and distal acid exposure time, but no correlation between the two in proximal measurements, questioning the role of proximal mucosal integrity in LPR⁶⁵. Finally, Choi et al evaluated BI in 77 patients with non-cardiac chest pain (NCCP). They found no difference in BI between NCCP and controls 3 cm above the lower sphincter, but noted that proximal BI was lower in the NCCP patients 17 cm above the lower sphincter⁶⁶.

Alternative methodologies for baseline impedance measurements

Despite the encouraging data for BI as a GERD biomarker, there are several technical limitations to consider. Conventional BI measurements are made from catheter-based pH-impedance studies, which are cumbersome and uncomfortable for patients. There is no uniform agreement on how, where, or when to measure BI in the 24-hr tracing. Hence, investigators have endeavored to devise specialized instrumentation specifically for measuring BI.

Electrical tissue impedance spectroscopy (ETIS) is an alternative, through-the-scope methodology, designed to quickly measure mucosal integrity during endoscopy. Initial animal studies in 2011 yielded encouraging results, and subsequent studies by Weijenborg et al comparing patients with erosive disease and HV, found good correlation between impedance measurements by ETIS (taken in areas of normal appearing mucosa) and traditional transepithelial resistance and permeability measurements obtained in Ussing chamber studies^{35, 67, 68}. However, data on patients with NERD and correlations with conventional BI measurements are not yet available. Manambe et al recently described a variation on impedance measurement, the bioelectrical admittance method, which measures admittance (the numerical reciprocal of impedance) endoscopically. Using this method, they found a negative correlation between admittance and tissue resistance evaluated by Ussing chambers⁶⁹. Finally, Yuksel et al have devised a minimally invasive technique to obtain impedance measurements, using a single-channel, mucosal impedance catheter with sensors that traverse the working channel of the upper endoscope. Their findings mimicked those of conventional catheter based BI measurements, i.e., mucosal impedance was lower in patients with GERD compared to HV⁷⁰. A recent abstract presented by the same group at Digestive Diseases Week in Washington, DC (May, 2015) compared conventional BI measurements to mucosal impedance measurements, and found that there was less variability with the latter technique⁷¹. A follow-up study by Ates et al applied this minimally invasive technology in the evaluation of 61 patients with erosive esophagitis, 81 patients with NERD, HV, and disease controls (achalasia and eosinophilic esophagitis). They found that mucosal impedance measurements were lower in patients with GERD when compared to controls and achalasia and could reliably distinguish between the groups. Interestingly, patients with eosinophilic esophagitis also demonstrated low impedance measurements, but were distinguished from GERD in that they were uniformly low throughout the esophagus compared to GERD patients, who had increasing impedance proximally. Moreover, values for the GERD patients normalized with acid suppressive therapy⁷².

SUMMARY

We used a systematic methodology to review biomarkers for GERD, independent of defining endoscopic features and conventional pH-impedance monitoring. Keeping with the National Institutes of Health Biomarkers Definitions Working Group³, the biomarkers discussed are characteristics that are “objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention” and independent of patient-reported symptoms or outcomes. A major conclusion of this review, and one that was somewhat predictable, was that there can be no single biomarker for GERD because of the heterogeneity of the condition. We also found a wealth of emerging data.

Most encouraging, is progress made in the domain of defining esophageal mucosal integrity and its compromise consequent from reflux injury. Dilated intercellular spaces and the associated reduced trans-epithelial resistance were first described in animal experiments and have now been translated to clinical relevance^7–9. Impaired mucosal integrity has been a consistent observation among reflux populations, defined by physiological abnormalities and/or enhanced esophageal sensitivity and differentiating them from ‘FH’ where the pathology lies elsewhere. Reflux injury mediated by acid, pepsin, and bile acids degrades tight junctions in the esophageal mucosa, altering its ionic and molecular permeability and, in the process, changing its electrical characteristics as determined by BI measurements and releasing fragments of junctional proteins into the circulation as potential fingerprints of that process¹⁸. Although the methodologies in both domains needs to be refined, recent reports are encouraging and the development of an instrument specifically designed to measure esophageal mucosal impedance during endoscopy are promising⁷².

In conclusion, this review summarizes a comprehensive literature search on GERD biomarkers, acknowledging that there is unavoidable selection bias due to the nature of the tropic reviewed. However, we present the first review of its kind that attempts to evaluate biomarkers in this widely heterogenous disease entity. It is, of course, overly simplistic to think that there would be a single biomarker for a disease of such varied presentations. Rather, the hope is for biomarkers to emerge for distinct parameters of GERD pathology. The concept of impaired mucosal integrity quantifiable by measures of mucosal impedance has emerged as the most promising domain of investigation, with several novel endoscopic and histopathologic tools currently under study that will likely play a larger role in the diagnosis of GERD in years to come.

Abbreviations

BI

baseline impedance

CGRP

calcitonin gene-related peptide

DIS

dilated intercellular spaces

EGJ

esophagogastric junction

GERD

gastroesophageal reflux disease

LPR

laryngopharyngeal reflux

NERD

non-erosive reflux disease

OP

oropharyngeal

PAR-2

Proteinase-Activated Receptor-2

TRPV1

transient receptor potential cation channel subfamily V member 1

FH

functional heartburn

HV

healthy volunteers