Eosinophilic Esophagitis (EoE) is a food allergy-associated inflammatory disease affecting both children and adults [12]. EoE is characterized by persistent esophageal inflammation due to an immune-allergic pathogenesis with clinical manifestations relating to esophageal dysfunction. Eosinophil accumulation and infiltration of esophageal mucosa results in histological changes leading to dysphasia, stricture and food impaction [12].

Diagnoses of EoE have grown rapidly in recent 5 years. In Western countries, the prevalence of EoE has increased dramatically from 10 to 50 for every 100,000 general population [3]. Meanwhile, albeit published reports of EoE from Asian countries are limited, systematic literature review of EoE in Asian countries by Kinoshita et al. [4] reveals comparable EoE prevalence amongst Western and Asian patients. As a matter of fact, EoE is now becoming a serious disease not only in Western countries but also in Asia [4]. In industrialized countries, the incidence of EoE has increased dramatically in the past 30 years, resulting in a considerable public health and economic burden [2]. In many patients, the disease can be controlled by adherence to an elemental diet eliminating foods that trigger EoE [2]. However, this approach involves disruptive changes in lifestyle and eating habits [2].

Generally, the types of EoE have been characterized as food allergies. Around 70% of EoE patients have current or past food allergy history or positive skin pricks test particularly to a range of foods. The basis for this understanding commenced approximately 20 years ago, when Kelly et al. [5] reported that children with EoE, treated with a nonallergenic amino acid-based formula diet for 6 to 8 weeks, had a complete clinical resolution. However, reintroduction of the food proteins was associated with a clinical relapse [5]. Markowitz et al. [6] replicated these results in 2003 with a larger cohort. Recent clinical trials with practical elimination diet [7] and skin testing directed elimination diet [8] provided additional evidence of EoE as a food allergen driven condition. Previous study found that children with EoE were mainly sensitized to egg, dairy milk, soy, beef, wheat, and peanuts [9]. In a pediatric EoE cohort, sensitizations to food allergens have been identified in 75% of the study population [10], supporting the role of food allergy in EoE. Noteworthy is the observation of local immunoglobulin E (IgE) in esophageal mucosa of pediatric EoE patients [11]. In adult and pediatric patients with EoE, specific IgE to food and inhalant allergen epitopes have been recognized in 91% of the subpopulation [12], further supporting a role for food allergic mechanisms in EoE.

EoE also has been recognized as a connection between food allergy and other atopic diseases. EoE is strongly associated with atopic disease and frequently occurs together with asthma, eczema, allergic rhinitis, and anaphylactic food hypersensitivity, and a strong atopic family history [1013]. Allergy-prone children can begin with allergic sensitization, often with eczema in the first year of life, later developing allergic rhinitis and can progress to asthma [14]. This stepwise increment of conditions has been well described as the 'atopic march' [14]. In particular, food allergic children with multiple allergies, for example, asthma, eczema, or rhinitis at a young age are more likely to develop EoE [13]. Recent study has discovered that up to half of food allergy patients who have other atopic illnesses may progress to EoE later in life [10]. Interestingly, while these allergic diseases can follow the "atopic march," EoE shows a wide incident age range, and thus appears not to follow this established concept. Nonetheless, the strong clinical connection between EoE and other well characterized allergic diseases appears to suggest shared pathogenetic mechanisms with mutual impact on severity. As the result, EoE is therefore increasingly seen as a syndrome rather than a single disease [15]. There is also extensive clinical overlap of EoE with asthma, the immunological basis of which has been reviewed [16].

As with many chronic inflammatory diseases, clinicians now realize that the division of EoE into IgE-mediated and non-IgE-mediated food allergy has been an oversimplification. The pathogenesis differs in terms of genetic susceptibility, environmental risk factors, age of onset, clinical presentation, prognosis and response to standard and new therapies [17]. Therefore, in this review, we will focus on the underlying immunological basis of the various EoE pathogenetic mechanisms, specifically about how food allergens induced EoE, integrating results from human studies focusing on specific pathways with results from animal studies, where much molecular data are available.

A number of clinical studies report that food allergy may be the cause of EoE pathogenesis [678910]. Most patients with EoE respond to an elemental diet with resolution of symptoms and normalization of biopsies [12]; in many cases, reintroduction of foods routinely causes relapse of esophageal eosinophilia [9]. The most widely known food allergies identified are linked to eggs, soy, chicken, beef, wheat, rice, barley, peas, corn, and peanuts [9]. It still remains unclear why these foods are more prone to trigger EoE than others, however there is an evidence that specific foods possess essential immunological properties that can induce intrinsic immune responses. For example, milk sphingomyelin can activate invariant natural killer T (iNKT) cells promoting Th2-response [18], and peanut allergen Ara h1 can bind to CD209 on dendritic cells (DCs) [19]. In addition, clinical reports demonstrate that peanuts or other nuts are attributed to 80% of fatal food allergy reactions [20], and the combined skin prick and patch-test-positive dataset of the Cincinnati Center for Eosinophilic Disorders specifies that patients with EoE most generally have peanut, corn, and sensitivities (unpublished data, stated in [21]). In humans, elimination and reintroduction diets imply the causal nature of foods in EoE [9], as does intragastric ovalbumin promoting EoE in mice [22]. This diet involves eliminating the 6 most common foods that have been recognized in allergic diseases, which are egg, milk, soy, wheat, nuts, and seafoods [9]. Elimination diet is initially suggested by a gastroenterologist as a result of referral to allergists to perform a food allergy evaluation for EoE. Eighty-eight percent resolution is observed in children with EoE after eliminating these foods from their diets [923]. Significant changes are required in daily diet and lifestyle and supplementation with an amino acid-based formula, which has been shown to produce clinical improvement in EoE patients [5]. The restricted number of foods that can promote EoE suggests there are certain essential immunological properties of these food proteins that can promote allergenicity [24], although this remains controversial. Taken together, the essential properties of certain food proteins, barrier dysfunction, and immune dysregulation susceptibility affected by early life events disposing to immune dysregulation, all contribute to sensitization.

Innate immune cells are important mediators of EoE pathogenesis. These cells include various type of T cells particularly Th2 cells, NKT cells, innate lymphoid cell (ILC) cells, and other inflammatory cells. They are individually discussed below including how they work together to promote food allergy-induced pathogenesis of EoE.

Within the field of chronic allergic inflammation, eosinophilic esophagitis is viewed as the trademark Th2 disorder of esophagus [25]. This is supported by the early findings that inflammation in esophageal biopsies of EoE patients is frequently eosinophilic in nature [568]. Initial studies demonstrated elevated numbers of CD4 positive cells producing interleukin (IL)-4, IL-5, and IL-13 in peripheral blood mononuclear cells (PBMC) [26] and esophageal mucosa biopsies of EoE patients [1227] correlating with the amount of esophageal eosinophilia [27]. Subsequent unbiased clustering algorithms have demonstrated that the various EoE phenotypes fall into Th2-prevalent immune response as the cause of the elevation of the cytokines IL-4, IL-5, and IL-13 and eosinophils in blood and tissue [122527]. Current studies indicate that there are food-specific Th2 cells in the peripheral blood of adult and pediatric patients with EoE [28]. Consistent with the IL-5 dependent and IgE-independent disease mechanisms for eosinophilic gastrointestinal syndromes [29], EoE subjects have food-specific Th2 cells that produce IL-5 as opposed to subjects with food anaphylaxis in whom food-specific Th2 cells generate IL-4, an interleukin which is vital for IgE class switching [28]. When cultured, IL-5 positive cells need longer time and chromatin remodeling to develop [30], in line with the chronic nature of the EoE disease process. If demonstrated in the periphery of EoE subjects, these cells could offer understanding into disease pathogenesis and offer a potential alternate allergic sensitization marker. In addition, there may be better alternate biomarkers of the Th2 high endotype of EoE, such as increased mRNA levels of eotaxin-3, IL-5, IL-5 receptor α-chain, and IL-13 that are associated with increased numbers of esophageal mucosa tissue eosinophilia [122627]. The case for CD4 positive Th2 cells as controllers of the EoE disease was therefore confirmed (Fig. 1).

The mice that lack the key Th2 cytokines (IL-4, IL-5, or IL-13) all have considerable reduction in esophageal eosinophilic features in the EoE model [31]. By analyzing mice overexpressing IL-5, either by hereditary manipulation or by IL-5 administration, researchers established the specific role of IL-5 in recruiting circulating eosinophils to the esophagus [3132]. Noti et al. [33] demonstrated mRNA levels of IL-4, IL-5, IL-13 in esophageal biopsies of EoE mice supporting that the EoE model depends strongly on Th2.

In addition, in the peanut oral sensitization induced EoE disease mouse model, characterized by esophageal inflammation, there is increased generation of peanut-specific IgE, and elevated peanut protein stimulated secretion of IL-4, IL-5, and IL-13 in splenocyte cultures [21]. Moreover, CD8-deficient mice are prone to experimental EoE whereas CD4-deficient mice are partly protected from esophageal inflammation [31].

There is randomized controlled trial using an antibody to the α-chain of the receptor for IL-4 (dupilumab; Regeneron Pharmaceuticals Inc., Eastview, NY, USA), which effectively inhibits downstream signaling via the receptors for IL-4 and IL-13 to study the effect of this treatment in adults with active EoE (https://clinicaltrials.gov/ct2/show/NCT02379052). Inhibitors of IL-13, such as the anti-IL-13 antibodies lebrikizumab or QAX576, may be a potential therapeutic choice [34]. In a phase 2 trial, blocking IL-13 with lebrikizumab (Genentech Inc., South San Francisco, CA, USA) show promising results, in reducing intraepithelial esophageal eosinophil counts and improving clinical symptoms in adults with EoE [34]. However, notwithstanding the results of an early safety trial of mepolizumab administration demonstrating a symptomatic improvement and a dramatic reduction in peripheral and esophageal eosinophilia in 4 patients with EoE [35], subsequent placebo-controlled and/or double-blind studies have failed to exhibit clinical effectiveness of mepolizumab or reslizumab in EoE pediatric patients, in spite of decrease in blood and esophageal eosinophilia [3536]. Regardless of whether the failure of clinical response reflects inadequate depletion of tissue eosinophils, irreversible structural changes, or involvement of other cell lineages or variables unaffected by anti-IL-5 therapy is uncertain. A randomized, double-blind, placebo-controlled, trial of an antibody to the receptor for IL-5 (benralizumab; MedImmune, a subsidiary of AstraZeneca, Gaithersburg, MD, USA) that effects depletion of eosinophils for months after a single injection is still ongoing (http://www.smartanalyst.com/SMART_Immunology_Newsletter.pdf).

In human studies, IL-4 or IL-5 receptor blockade produces a positive clinical response in patients with esophageal eosinophilia irrespective of atopic status. In a knockout mouse model of EoE, esophageal eosinophilia is abrogated in RAG1 deficient (lacking B and T cells) or FOXN1 deficient (lacking T cells) mice, but not in IgH6 deficient (lacking B cells) mice [31]. These murine models support the view that Th2 cytokines and eosinophilia can be produced independent of the adaptive immune system.

ILCs were initially categorized as either non-T or non-B effector cells in various disease models of Th2 immunity [37]. Their nomenclature has been further delineated according to their cytokine production profile, into group 1 ILCs (interferon-γ producing which include traditional NKT and ILC1 cells), group 2 ILCs (Th2 cytokines producing, formerly known as ‘nuocytes’ or ‘natural helper cells,’ now named ‘ILC2 cells’), and group 3 ILCs (IL-17 and/or IL-22 producing, or are involved in the formation of lymphoid tissues, which include ILC3 cells and lymphoid tissue–inducer cells) [37]. ILC2 cells resemble Th2 cells in various ways as shown in Fig. 2. ILC2 cells lack antigen-specific receptors, but similar to Th2 cells, they respond to the epithelial derived cytokines IL-25, IL-33 and thymic stromal lymphopoietin (TSLP) [253337]. Study indicates that ILC2 cells development relies on IL-7 and IL-33, which is produced mainly from common lymphoid progenitors [37].

Activation of ILC2 cells is further stimulated by TSLP, IL-25, and IL-33, produced mainly by epithelial cells, as a response to injury and stimulation via pattern-recognition receptors [3238]. Culture studies with lymphoid aggregate cells from EoE human donor esophageal tissue in the presence of TSLP or IL-33 cytokine induce ILC2 proliferation. Thus, ILC2 cells are detected in the esophageal mucosa patients with EoE and respond to cytokines recognized to stimulate ILC2 proliferation [3940]. Clinical studies also demonstrate elevated levels of IL-25, IL-33, TSLP in active EoE as compared to normal esophageal tissue [3840]. In addition, ILC2 cells producing Th2 cytokines (IL-13, IL-5, and IL-9) were initially described in the human esophageal EoE tissue, where they contribute to esophageal tissue eosinophilia and play a fundamental role in inflammation [40]. Although ILC2s lack surface markers for B, T, or NKT cells, they do express homologous chemo-attractant receptor molecules commonly expressed on Th2 lymphocytes (CRTH2) [39]. In line with the rapid increase in knowledge about ILC2 cells, including experimental models of Th2 allergic disease and clinical studies of ILC2 cells in pediatric EoE [38], the contributions of ILC2 cells relative to that of Th2 cells in EoE have emerged.

ILC2 cells may have an important role in a subset of EoE patients with lack of or partial response to steroids [41]. IL-5 and IL-13 secretion are not downregulated by steroids in ILC2 cells as they are in CD4 positive T cells. TSLP can trigger steroid resistance in ILC2 cells by up-regulating the pro-survival factor Bcl-xL and promoting phosphorylation of the transcription factor STAT5 [42]. Hence, if ILC2 cells are pathogenic in this subset of patients, this would help explain steroid resistance.