Nature with nurture: The role of intrinsic genetic and extrinsic environmental factors on eosinophilic esophagitis

Abstract

Purpose for Review

As the rising prevalence and incidence of eosinophilic esophagitis (EoE) has quickly outpaced the rate of esophageal biopsies, particularly in Westernized countries, several studies have suggested a link between intrinsic genetic and extrinsic environmental risk factors and the development, presentation, and diagnosis of EoE. This review aims to critically assess existing studies describing the role of the environment on the development, symptomatic presentation, and diagnosis of this recently recognized chronic immune-mediated disease.

Recent Findings

We present and critically evaluate the working hypotheses and supportive studies thus far on environmental factors on EoE, describe sources of potential bias in diagnosis due to socioeconomic factors and thus undermining studies of EoE etiology, and highlight opportunities for future research.

Summary

As genetics alone do not explain the rapid rise of EoE, we must look to environmental, or extrinsic, factors both in the early life period which shape the development of the gut microbiome, as well as later life contributing to diagnosis of this new disease. Future etiologic studies linking risk factors to EoE development in individual patients are needed.

INTRODUCTION

Eosinophilic esophagitis (EoE) is an immune-mediated chronic inflammatory disease defined by eosinophilic infiltration of the esophagus and symptoms of esophageal dysfunction such as dysphagia, food bolus obstruction, and less commonly heartburn, globus, or poor feeding in children.¹ The management of EoE includes treatment of inflammation and fibrostenotic sequelae with medications (e.g. proton pump inhibitor, topical corticosteroids, novel biologic agents), dietary therapy, and endoscopic dilation.

Since it was first described nearly three decades ago, the prevalence of EoE has climbed to as high as 1 case per 1000 people and more dramatically, the incidence has risen rapidly up to 12.8/100,000 new cases per year.^2,3 This upsurge now outpaces the increase in esophageal biopsies, suggesting that the increase is not only due to increased awareness for this new disease.⁴ This rise has costly complications as the estimated cost of EoE care likely exceeds $1 billion in outpatient visits, endoscopic procedures, emergency room care, and off-label medication use.⁵

Despite working hypotheses, the etiology of EoE is unknown. Given the pathognomonic eosinophilic infiltration, EoE was initially regarded as an allergic disease and frequently associated with other atopic diseases such as allergic rhinitis, asthma, eczema, and food allergies, suggesting common disease mechanisms.⁶ Unlike its close atopic relatives, EoE is not an IgE mediated process,⁷ however similarly, environmental factors are thought to influence the disease onset and severity. We propose considering a “nature versus nurture” framework to consider factors that determine disease development; these factors include intrinsic or host characteristics and extrinsic factors both in the early life period and later life exposures. Identifying the factors that affect disease development and potentially progression within this framework will lead to an understanding of the multifactorial mechanisms of disease and perhaps a path towards disease mitigation or prevention. In this review, we will present the emerging data on the intrinsic biologic (“nature”) and environmental experiences (“nurture”) on EoE, and highlight areas of uncertainty and emerging research.

Nature: Intrinsic factors contributing to EoE

There are a number of intrinsic factors such as genetic or innate host characteristics, which may predispose an individual to developing EoE. To date, several genetic susceptibility variants related to epithelial barrier dysfunction (FLP, DSG1, CPN14, SPINK5, SPINK7) and T-helper type 2 mediated immunity (CCL26, POSTN, TSLP), have been linked to EoE in candidate and genome-wide association studies.⁸ However, due to multiple pathways and complex modifying interactions with environmental factors on the inheritance of EoE, disease development cannot be solely attributed to genetic factors and inheritance is not Mendelian. In a recent study of 37 EoE patients and their 237 first-degree adult relatives, 14.6% of were found to have esophageal eosinophilia, particularly in male first-degree relatives (p = 0.027) and associated with hay fever, allergic eye symptoms, and food allergies, suggesting increased risk of esophageal eosinophilia among atopic family members of EoE patients.⁹ However, in another study examining family clustering among twins, EoE was more frequent in dizygotic twins (22.0%) compared to non-twin siblings (2.5%, p < 0.001), suggesting that the majority of the effect was due to shared family environment, particularly early life factors, rather than genetics.¹⁰ In reality, however, the pathogenesis of EoE is likely more complicated, with genetic factors interacting with environment factors (also termed “GxE interaction). In the first study specifically examining this concept of genetic variants and early-life exposures, breastfeeding conferred a protective effect to developing EoE in infants with a single nucleotide polymorphism (SNP) in the susceptibility gene CAPN14 (aOR 0.08, 95% CI 0.01–0.59) compared to those who were not breastfed. When this SNP was not present, however, there was no impact on EoE risk regardless of breastfeeding status. This supports the hypothesis that genetically susceptible individuals are primed by environmental factors to later develop EoE.¹¹ As such, future research to identify and explain these interactions are needed.

The microbiome and EoE

A complex and diverse community of microorganisms, the gut microbiome, plays crucial roles in metabolic function influencing the development of the immune system. As such, dysbiosis or microbial imbalance, plays a key role in the pathogenesis of gastrointestinal and atopic diseases, and factors that disrupt this balance (e.g. medications, breastfeeding, environmental exposures) are thought to be associated with inflammation and ultimately disease progression.¹² Early microbiome studies in EoE demonstrated differences in the esophageal microbiome between EoE versus non-EoE controls, increased esophageal bacterial load (but not diversity) with inflammation in EoE versus controls, less microbial diversity from stool in EoE, and altered salivary microbiome in children with EoE.^13–17 In contrast, recent work by Johnson et al. found no esophageal microbiome differences between newly diagnosed EoE adult cases and non-EoE controls, nor within EoE cases based on clinical features.¹⁸ Data on alterations of the esophageal microbiome in response to treatment remains inconclusive with two studies of patients with active untreated EoE, EoE in remission, and control subjects showing no relationship between the level of eosinophilia and esophageal microbiome measures,¹⁵ and no global microbiome changes in response to dietary interventions,¹⁴ and normalization of microbial load after elimination diet and change in the esophageal microbiome after treatment with swallowed topical steroids.^19,20 However, as these studies are limited to prevalent EoE cases, the relationship between this imbalance and inflammation, or the significance of the esophageal-specific microbiome is unclear and inconclusive. Current knowledge gaps and areas for future research include determining if alterations in the microbiome result in inflammation or are an effect of the disease state, as well as identifying and targeting bacteria that play a role in the inflammatory response as a potential disease target.

Nurture: Extrinsic factors and the development of EoE

As genetics alone do not explain the rising incidence and prevalence of EoE, considering extrinsic factors such as environmental exposures and experiences, is key to understanding disease development. It is these factors, particularly in the first three years of life, that influence the development and colonization of the gut microbiome, promote dysbiosis, and inform the developing immune system. Shaped by early life exposures, the gut microbiome is important in mucosal barrier function and regulating immune homeostasis, and has also been implicated in the development of atopic disease such as asthma, food allergies, and atopic dermatitis.²¹ As environmental exposures are implicated in the development of these allergic diseases, much work has also focused on identifying more global environmental exposures beyond the early life period such as aeroallergens, air and water quality, and population density.

Early life factors

As EoE is most commonly diagnosed later in childhood and adulthood, the largest body of work on environmental roles in the development of EoE investigates early life factors, or those impacting fetal development in utero and exposures within the first three years of life. Of these, cesarean delivery, preterm delivery and neonatal intensive care unit (NICU) admission, infant formula feeding, and antibiotic exposure in the first year of life are the most well described. In the first study that described this relationship, using pediatric EoE cases in North Carolina, Jensen et al. found that antibiotic use in infancy was strongly associated with having EoE (OR 6.0, %95 CI 1.7–20.8) with cesarean delivery (OR 2.2, 95% CI 0.8–6.4), preterm birth (OR 4.2, 95% CI 0.7–43.4), and non-exclusive breastfeeding (OR 3.5, 95% CI 0.6–19.5) also trending towards an increased odds of EoE.²² These findings were further supported and other factors identified by case-control studies in two separate cohorts of pediatric EoE in Ohio (prenatal- maternal fever, preterm labor; intrapartum- cesarean delivery; infancy- antibiotic use, use of acid suppressant) and Massachusetts (cesarean delivery, antibiotic use in the first year of life).^23,24 In contrast, only one study to date, a case-control single-center study from Canada, described that breastfeeding (52.8% EoE vs 42.9% control were exclusively breastfed, p > 0.05) and exposure to smoking (OR 0.96, 95% CI 0.58–1.59) were not associated with pediatric EoE compared to controls.²⁵ One study assessed these early life exposures in relation to the development of EoE in adulthood and also found positive associations between antibiotics in infancy (OR 4.64, 95% CI = 1.63–13.2), cesarean delivery (OR = 3.08, 95% CI = 0.75–12.6), preterm delivery (OR 2.92, 95% CI = 0.71–12.0), and neonatal ICU admission (OR 4.00, 95% CI = 1.01–15.9).²⁶ Several of these studies are not without limitation as the data were collected from questionnaires from single centers and thus at risk of recall and selection biases. However similar findings are echoed in recent larger database studies. Using claims data, Witmer et al. echoed findings that antibiotic use in early infancy and premature birth were associated with EoE diagnosis.²⁷ Moreover, in a recent population-based study in Denmark that took advantage of the linked healthcare registries in that country which eliminated the possibility of recall bias, preterm delivery, NICU admission, and maternal and infant use of antibiotics were associated with increased risk of developing EoE, with increased risk with greater frequency of use and timing close to delivery.^28,29

Another category of early life exposures that has recently been identified relates to acid suppressant medications. As up to half of individuals with EoE are able to attain disease remission on proton pump inhibitors (PPIs), this class of medications is commonly used as first-line treatment for EoE. However, similar to antibiotics, exposure to acid suppressants in the early life period has been shown to lead to dysbiosis and alterations in the gut microbiome, which could be a potential mechanism in EoE.³⁰ This relationship between the development of disease and use of acid suppressive mediations in early life, including both PPIs and histamine-2 receptor antagonists (H2RA) is supported by observational studies in other allergic diseases such as atopic dermatitis, asthma, allergic rhinitis, and food allergy.^31,32 More specifically, both PPIs and H2RA in infancy has been associated with EoE in two single center case-control and one large database study to date.^23,24,27 These findings have been recently supported by a large Danish population-based study of pediatric EoE cases showing an increased risk of developing EoE with maternal (aOR for 3+ prescriptions 5.1, 95% CI 1.8–14.9) and infant (aOR for 1 prescription 11.4, 95% CI 5.1–25.6, aOR for 3+ prescriptions 23.5, 95% CI 9.1–60.58) acid suppression use in a dose response manner.³³ One issue to consider, however, in assessing these observation data is how to ensure the PPI was truly used prior to disease onset (and would therefore potentially be causal) as opposed to being used for symptoms from unrecognized disease (which would be an association attributable to protopathic bias).

Aeroallergens and Seasonality

As current hypotheses on environmental risk factors for EoE stem from those associated with other atopic conditions; aeroallergen exposure was one of the first to be scrutinized. Several early studies supported a relationship between seasonal pollen count or aeroallergen exposures and EoE.^34–36 Subsequent studies on seasonal trends of food bolus obstructions, worsening symptoms, and new diagnosis of EoE during peaks of spring and summer months also implicated the role of seasonality and aeroallergens, however aeroallergens do not universally exacerbate EoE, and the evidence of any association is mixed and thus inconclusive.^37–43 For example, in a meta-analysis summarizing studies reporting the month or season of diagnosis or disease recrudescence (defined as esophageal food impaction) and histologic evaluation, Lucendo et al. observed no significant seasonal variation in disease diagnosis or symptom exacerbation.⁴⁴ Although there were numerically fewer cases diagnosed in winter months, there were no statistical differences in the seasonal distribution of new EoE diagnoses with 27.1% diagnosed in spring months and 21.5% in the winter (p = 0.13). Among studies reporting disease recrudescence, EoE recurred throughout the year with no statistical differences in seasonal preference (p = 0.70). In a retrospective study of children with EoE, 160 had suspected aeroallergen triggers and 20% (32) were found to have histologically active EoE with seasonal variation.⁴⁵ The majority had comorbid atopic disease (100% allergic rhinitis, 75% asthma), and disease worsening was observed in the spring. More recent work by Reed et al. found that among 782 patients previously in disease remission and adherent on EoE-specific therapies, only 2% (10 adults, 3 children) had seasonal exacerbations defined by histologic activity and worsening endoscopic findings.⁴⁶ Among these, flare events occurred more in the summer (43%) and fall (29%), and all patients carried a diagnosis of allergic rhinitis. In a cross-sectional study of a large number of esophageal biopsies from a US national pathology database, esophageal eosinophilia was most prevalent in cold (OR 1.39, 95% CI 1.34–1.47) and arid (OR 1.27, 95% CI 1.19–1.36) climate zones compared to tropical zones (OR = 0.87, 95% CI 0.71–10.8), suggesting that additional investigation on climate and geographic features may also provide clues into environmental risk factors.⁴⁷

Potential reasons for these heterogenous findings are, in part, due to using the month or season of presentation or diagnosis as a surrogate for disease activity. Well documented diagnostic delay in EoE is multifactorial – due to patients’ minimizing symptoms, delayed reporting, and adaptive eating behaviors, physician clinical suspicion, access to subspecialty care, wait time to undergo endoscopy, and the influence of concurrent atopic disease activity – and may confound the relationship between disease activity with seasonal variations and aeroallergen exposure.^48–50 Furthermore, limitations in the studies to date include retrospective designs, defining seasonal exposure time, and potential geographic variations in aeroallergens.

Population density

The impact of population density is also posited as a contributor to the development of EoE, however the evidence to date is heterogenous. In a study of 14,381 patients with EoE in national pathology database, Jensen et al. observed that esophageal eosinophilia was inversely associated with population density as a surrogate measure of urban or rural living settings.⁵¹ Controlling for confounders such as distance traveled to endoscopy center (as a proxy for access to care), patients in the lowest population dense or rural areas had greater odds of EoE compared to those living in high population dense areas (aOR 1.59, 95% CI 1.45–1.76). Alternatively in a nationwide provider survey of allergists and gastroenterologists with potential diagnostic bias, higher EoE patient burden was observed in urban (0.58 patients per week) and suburban (0.44) settings when compared to rural (0.36, p < 0.001).⁵² In contrast, no residential variation among EoE patients non-responsive to PPI (50.9% rural vs 49.1% urban) was described in a cohort in Iowa.⁵³ To further examine the likely multifactorial effects of population density on the development or diagnosis of EoE, McGowan et al. conducted a cross-sectional study of over 18,000,000 children enrolled in Medicaid.⁵⁴ They found that although living in a rural environment was protective against EoE diagnosis, this effect was due to disparities in access to pediatric gastroenterology providers and the underdiagnosis of EoE. In a follow-up paper, and after adjusting for urban vs rural status and neighborhood-level poverty, they observed that living farther away from a gastroenterologist was associated with less EoE diagnosis (26–50km aOR 0.84, 95% CI 0.75–0.93; 51–100km aOR 0.83, 95% CI 0.74–0.94; 101–200km aOR 0.68, 95% CI 0.59–0.79; >201km aOR 0.50, 95% CI 0.40–0.61).⁴⁸ Additionally, those in higher poverty areas had decreased odds of EoE diagnosis regardless of distance to provider, suggesting that there are unknown factors (e.g. social determinants of health) and barriers to diagnosis outside of proximity to a gastroenterologist that may explain differences in risk attributable to urban versus rural residence.

Environment quality

A different line of research on the role of environmental exposures on EoE has focused on the influence of air and water quality. In a recent case-crossover study, emergency room visits for chest pain, dysphagia, and food impaction among patients with EoE in Utah were associated with high particulate pollution levels above EPA air quality standards.⁵⁵ In contrast, a recent case control study of a large US-wide database of over 29,000 cases, EoE was more common in rural populations where particulate matter concentrations were lower, echoing prior studies demonstrating associations between EoE and low population density.⁵⁶ Expanding on these findings to all domains of environmental quality (air, land, water, built, and sociodemographic) measured by the Environmental Quality Index (EQI), a large analysis of a national US pathology database demonstrated highest odds of EoE in the worst EQI quintile (OR 1.25, 95% 1.04–1.50), driven by poor water quality.⁵⁷ The odds of EoE were reduced in regions with worse air (OR 0.87, 95% 0.74–1.03) and land (OR 0.87, 95% CI 0.76–0.99) quality. Given these findings, investigation of water quality on EoE in the US revealed that metal contaminants in drinking water (inorganic mercury OR 1.22, 95% CI 1.15–1.28; beryllium OR 1.35, 95% CI 1.30–1.39; thallium OR 1.25, 95% CI 1.21–1.29) were associated with having EoE.⁵⁸ Prior assessment of environmental risks in the state of North Carolina revealed that EoE was associated with living within one mile of swine farming operations (aOR 2.56, 95% CI 1.33–4.95).⁵⁹ Additionally, housing components such as brick exterior (aOR 1.83, 95% CI 1.11–3.02), gas heating (14% EoE vs 8% controls, p=0.06), and forced air (57% EoE vs 45% controls, p=0.009) were found to be associated with EoE in a cohort of patient living in their homes for an average of 7.2 years (SD 5.9) prior to diagnosis.⁶⁰ These early studies support the hypothesis that a broad range of environmental exposures may contribute to the development or disease activity in EoE.

Detergents

Given the hypothesis that agents (e.g. drinking water) that come in direct contact with the esophageal mucosa may impact epithelial barrier function, detergents and other household products may play a role in the development of EoE. In vitro exposure to the common household detergent, sodium dodecyl sulfate (SDS), demonstrated impaired esophageal epithelial barrier dysfunction and epithelial differentiation and a corresponding in vivo study in mice treated with SDS showed marked esophageal eosinophilia compared to controls.⁶¹ It is yet to be established whether this mechanism may be at work in humans, and if it could explain a portion of the changing trends in EoE epidemiology.

CONCLUSION

Through the lens of viewing intrinsic and extrinsic factors in disease development, EoE is a result of complex interactions between environmental exposures in genetically susceptible individuals (Figure 1). As such, viewing “nature with nurture” instead of as competing forces, is a more appropriate and holistic framework to consider and investigate contributors of disease. The most studied are factors that potentially alter colonization of the microbiome and may alter immune tolerance, primarily during the early life period, including the use of antibiotics and acid suppressants, and cesarean section. A range of subsequent environmental exposures are also posited to be associated with EoE, and this is an area of emerging research. Despite a wide range of epidemiologic studies shedding light on these factors, our understanding of the pathogenesis and mechanisms that drive EoE onset in an individual remains naïve. Future prospective studies with comprehensive exposure assessment, in the context of underlying genetics are needed to assess disease risk and outcomes through a lens of gene (or epigenetic)-environment interactions. Limitations of current work include biases due to retrospective data collection or recall, incomplete data from large databases, the use of surrogate markers of environmental exposures, and disentangling confounding factors which impact making a diagnosis both in the early life period and in later life. In addition, our understanding of potential risk factors for developing EoE hinge on making an appropriate diagnosis, which may be limited by multilevel barriers at the patient (e.g. access to healthcare, education and literacy, economic stability, individual health priorities, behavioral), provider (e.g. expertise, communication, resources), and organizational or systems (e.g. availability of services, lost to follow-up, insurance) levels. As the diagnosis of EoE is complex, often delayed, and likely does not reflect the timing of initial disease onset, challenges for future research include clearly defining exposures that precede disease development and reducing disparities in diagnosis. Increased risk of EoE in rural communities, high poverty neighborhoods, and areas of poor air or water quality are prime examples of not only environmental vulnerabilities, but potential clues about disparities in EoE care and access that may limit our understanding of environmental risk factors for disease development, particularly for factors distributed differently across socioeconomic groups. As our view of the role of the environment on EoE continues to expand, current studies should inform future goals of not only identifying and reducing modifiable exposures of EoE, but also identifying and addressing disparities in the diagnosis of EoE.

FIGURE 1.

Intrinsic and extrinsic risk factors for the development of EoE.