Summary of the First International Gastrointestinal Eosinophil Research Symposium

INTRODUCTION

Chris A. Liacouras, MD, and Glenn T. Furuta, MD

The last decade raised eosinophilic esophagitis (EE) from a clinical curiosity to a recognized disease. The care of the patient with EE demands an extraordinary attention to detail; the diagnosis must be made in the proper fashion, and then consistent collaboration between multiple subspecialists is required to provide outstanding longitudinal care. Eosinophilic esophagitis is no longer an emerging diagnosis, but rather a distinct clinicopathological disease. Its natural history, pathophysiology, and treatments are undergoing discovery and refinement that have been reviewed in a number of recent articles.

In January 2006 the North American Society of Pediatric Gastroenterology convened a steering committee to develop an international conference focusing on EE. The overall goals of the symposium were to present new data regarding clinical and basic elements of the disease, engage new investigators to enter into research focusing on EE, and initiate an international research consortium. To that end, the First International Gastrointestinal Eosinophil Researcher Symposium was organized and held October 17-18, 2006, in Orlando, FL. Young investigators presented abstracts focusing on EE. More than 300 pediatric and adult pathologists, allergists/immunologists, and gastroenterologists shared experiences and data derived from their area of expertise.

During the first day of the symposium, 5 subcommittees (clinical symptoms, endoscopy, histology, allergy, and treatment) presented their review of the world's literature. This work is being synthesized into a consensus statement. The next day, the symposium was divided into 3 sections: EE natural history, pathogenesis, and treatments. The following summaries provide the current state of knowledge as it was presented at the symposium.

I. NATURAL HISTORY OF EE EPIDEMIOLOGY

Peter Bonis, MD

The epidemiology of EE is incompletely understood. Available data suggest that the disease is new (perhaps first arising in the 1970s) and the worldwide incidence is rising in both adults and children, although the extent to which increased recognition accounts for the rising incidence is uncertain. A more complete understanding of its incidence, geographic variability, and associated burden of disease is still evolving.

Few population-based studies have evaluated the epidemiology of EE. Available reports suggest that the incidence has been increasing in recent years, although in some regions it may have stabilized. The prevalence in children living in Western Australia was estimated to have increased from 0.5 cases to 0.89 cases per 10,000 children between 1995 and 2004 ⁽¹⁾. The incidence was estimated to be 0.6/10,000 population in adults in a separate report from Australia, ⁽²⁾ and 0.15/10,000 population in a study from Olten County, Switzerland ⁽³⁾. In all of these reports, the authors noted a marked increase in recent years, which did not appear to be related to changes in referral patterns or increased recognition.

A report from the United States of children and adolescents (up to ages 19) estimated an incidence of 1.3/10,000 population in 2003 ⁽⁴⁾, which the authors point out exceeded that of inflammatory bowel disease (IBD) in children (0.7/10,000 population in a report from Wisconsin) ⁽⁵⁾. The prevalence was estimated to be 4.3/10,000 population in 2003. However, in contrast to the studies above, the authors note that the incidence has remained stable in recent years (M.E. Rothenberg, personal communication, October 2006).

All of these reports are limited by the potential for ascertainment bias because the estimates were based upon patients who were recognized clinically. A population-based study of adults from Sweden attempted to define the prevalence of esophageal eosinophilia in the general population ⁽⁶⁾. Eosinophils are not normally present in the esophagus, and thus any esophageal eosinophilia is abnormal. Interestingly, some degree of esophageal eosinophilia was present in about 5% of the population, whereas 1% had more severe esophageal eosinophilia with a density consistent with EE. The study did not perform objective testing for gastroesophageal reflux disease (GERD), however, thus the degree to which these findings may have been attributed to GERD (or other causes) is unclear.

The geographic distribution of the disease has not been well established, although reports have emerged from many regions of the world. It is possible that a better understanding of the disease distribution could help uncover etiologic associations. Such an understanding will require the establishment of systematic surveillance and consensus on its definition.

Several reports have established familial clustering of the disease ^(3,4,7); however, the degree to which these represent a genetic predisposition versus common environmental conditions is unclear.

Morbidity and costs related to the disease (and its treatment) are incompletely understood. Presentation has been sufficiently severe to warrant an elimination or elemental diet in children, which may have a substantial impact on quality of life for patients and their families. By contrast, many adults with EE do not have severe symptoms, even without treatment ⁽⁸⁾. Emerging data, however, suggest that EE is an important cause of dysphagia and food impaction in adults ⁽⁹⁾.

An improved understanding of the natural history of the disease and endpoints of treatment will help to define morbidity related to the disease. For example, it remains unclear whether EE in children persists into adulthood. Whether persistent esophageal eosinophilia in patients who have minimal or no symptoms is associated with adverse health outcomes also is unknown.

NATURAL HISTORY IN CHILDREN AND ADULTS

Phil E. Putnam, MD, and Alex Straumann, MD

Eosinophilic esophagitis is a young disorder. The first comprehensive descriptions of the adult form of this inflammatory esophageal disease were published in the early 1990s ^(10-12). We have come to recognize that EE is a clearly defined clinicopathological entity, characterized by esophagus-related symptoms in combination with a dense esophageal eosinophilia, both of which persist despite treatment with prolonged proton pump inhibitors ⁽¹³⁾. Initially thought to be rare, EE has emerged as 1 of the most common causes of dysphagia and esophageal food impaction in adults ⁽⁹⁾. Despite enormous research activity, the natural history of EE remains poorly understood.

The natural history of a disease describes the expected course followed by the given disease over time, its characteristic pattern, and its time-intensity gradient. This, in turn, supplies answers for research purposes, to lay the foundation for the evaluation of any therapeutic measure; patient care, to provide a prognosis to an affected individual; and socioeconomic considerations, to estimate the burden of illness.

The natural history of a disease can be studied by observing the course of the illness if left untreated. Various outcomes can occur, including spontaneous permanent resolution, remission with relapse, relentless progression, or a progression to a fixed stable state. With regard to EE in children, there is little published data regarding the long-term outcome of this disease when it has its onset in childhood. In addition, the literature that has been reported has generally included information regarding treated individuals who are observed only for short periods. Because of these brief periods of observation, it is still not known whether children with EE grow up to become adults with EE or whether pediatric EE is the same disease as adult EE.

With regard to EE in infants, children, and adolescents, there have been several studies that have provided information regarding clinical symptoms seen at various ages. Studies by Liacouras et al ⁽¹⁴⁾ and Noel et al ⁽¹⁵⁾ demonstrate that infants with EE generally present with symptoms of reflux, feeding disorder, or irritability. School-age and older children generally continue to complain of symptoms similar to gastroesophageal reflux. In addition, they also exhibit signs of vomiting, regurgitation, and abdominal pain. Older children and adolescents present more often with signs and symptoms of dysphagia, swallowing difficulties, and possible food impaction. At this point, investigators are unsure whether these various symptoms represent different subtypes or phenotypes of EE, or whether they manifest a progression of the disease. It is interesting to note that the majority of the pediatric EE literature relates to children who present to their physician with clinical symptoms. These chronic symptoms eventually prompt evaluation, including upper endoscopy with biopsy leading to the diagnosis of EE.

Frequently, the symptoms of EE can be intermittent or sporadic. Clinicians are now histologically identifying many asymptomatic children who have severe EE. These patients are typically either found incidentally during evaluation for another disorder or found at endoscopic follow-up after therapy for EE has been instituted. Because they have minimal or no symptoms, many of these children are either lost to follow-up or discharged from care. Thus, the question exists as to what happens to the disease in these children over time. The treatment for asymptomatic children with histological EE is controversial.

With regard to children, EE appears to be a chronic disease at best and a progressive disease at worst. Spontaneous remission and resolution of the disease seem unlikely. Histological relapse is extremely common when treatment such as dietary therapy or corticosteroids is withdrawn. In addition, the absence of clinical symptoms does not predict an absence of histological inflammation. In individuals who have been treated, the question remains whether treatment with either dietary therapy or medical therapy will result in permanent resolution or control of the inflammatory process. In general, effectively treated individuals have not been observed to develop dysphagia or fibrosis. Moreover, will aggressive therapy alter the natural history of esophageal fibrosis?

Based on previously published studies ⁽⁸⁾, it has been established that in adults, EE is a chronic disease with persistence of symptoms and eosinophilic-tissue infiltration for years; unequivocally restricted to the esophagus without a risk of extension to eosinophilic gastroenteritis; a localized disease, given that no cases with development of hypereosinophilic syndrome have been observed; likely a benign disease because so far no EE patient has experienced an esophageal malignancy; a nonfatal disease, given that no EE-related deaths have been reported; and a disease that substantially impairs the quality of life.

Recent publications, as well as clinical observations, suggest that the inflammatory activity of EE may be exogenously influenced by aeroallergens, such as pollen ⁽¹⁶⁾. In contrast, EE may fluctuate spontaneously, independently of aero- or food allergens. EE may persist for years in a stand-by mode. In contrast, chronic EE may lead to a remodeling of the esophagus, may predispose to viral and/or fungal superinfections, or may lead to an emesis-induced perforation of the esophagus. EE also has been shown to be a risk factor for procedure-induced perforation with the potential for a fatal outcome, and may impair the function of the lower esophagus sphincter and consecutively lead to gastroesophageal reflux.

There are no specific markers or presenting factors that can predict the outcome of EE; however, features such as the age of presentation, response to steroids, and the response to an elemental or elimination diet may be important criteria that may predict the outcome of EE in the future. Additionally, current research suggests that there may be specific phenotypes that may allow for gene chips or biomarkers to be developed. We need better instruments to assess the natural history of EE. We propose that a prospective, long-term, multicenter trial should be initiated, one that includes a substantial number of patients. In this study clinical findings, laboratory analyses, endoscopic findings, and histological as well as immunological and genetic markers must be assessed according to a study protocol using clearly defined criteria.

EOSINOPHILS IN THE ESOPHAGUS: HOW MANY ARE TOO MANY?

Eduardo Ruchelli, MD

Except in the esophagus, eosinophils are present normally in variable numbers in the mucosa of the entire gastrointestinal (GI) tract ^(17,18). The squamous epithelium of the normal esophagus is completely devoid of eosinophils. When present, intraepithelial eosinophils are considered a sign of esophagitis. Their presence, however, is not diagnostic of any particular etiology. Several types of esophagitis frequently include eosinophils as part of the inflammatory changes. In children, the most common causes of esophagitis include gastroesophageal reflux, allergic response, Crohn disease, and infections such as those caused by Candida species and herpesviruses. In general, infections tend to be associated with a predominantly neutrophilic response, whereas the inflammation in gastroesophageal reflux disease (GERD) and allergen exposure is primarily eosinophilic in nature.

Eosinophilic esophagitis has emerged in the last 10 years as a distinct type of esophagitis ^(4,13,19-22). It is characterized not only by its association with food or airborne allergens but also by the large number of eosinophils that are usually present in the esophagus of these patients (>15-20 eosinophils per high power field). Although biopsies from patients with GERD usually have a lower number of eosinophils, the distinction between these 2 types of esophagitis on histological grounds alone is frequently difficult. In some biopsies the eosinophil count is within a borderline range. Because the inflammation is often patchy, the number of eosinophils varies significantly, even among samples taken from relatively close areas. Therefore, the eosinophil count may be affected by sampling. Because of this variability from 1 area to another, eosinophils should be counted in the most intensely inflamed high power field (peak count) and multiple biopsies should be obtained. Although infrequent, there are well-documented cases of severe eosinophilia with complete response to antireflux medication ⁽²³⁾, and conversely, there are some patients who have typical clinical and endoscopic features of EE in whom the biopsy fails to demonstrate a large number of eosinophils. Despite these limitations, numerous studies and our own experience indicate that patients with EE tend to have a larger number of eosinophils than patients with GERD, and at a minimum, a peak eosinophil count >15 eosinophils per high power field should raise the possibility of EE. However, a pathognomonic eosinophil count identifying EE does not exist. The distinction between EE and GERD rests on both histological and clinical findings.

In addition to the large number of eosinophils, other histopathological features have been observed in patients with EE. These features include a superficial distribution of eosinophils along the luminal surface of the mucosa (superficial layering), superficial aggregates of eosinophils (eosinophilic microabscesses), and detached fragments of squamous cells admixed with numerous eosinophils ^(9,24,25). These pathological changes explain the endoscopic appearance of the esophagus in some patients with EE, in whom white plaques are observed. Other features include prominent basal layer hyperplasia and fibrosis of the lamina propria. It is unclear whether prominent basal layer hyperplasia is merely secondary to an inflammatory process involving a large number of eosinophils, or whether it is driven by other mediators, particularly when in some instances the degree of basal layer hyperplasia is out of proportion to the number of eosinophils. Only a few reports have described the presence of lamina propria fibrosis in patients with EE ⁽⁸⁾. The rarity of this finding may be due to the fact that most esophageal biopsies do not include the lamina propria. Despite the limited number of well-documented cases, fibrosis may explain the narrowing of the esophagus or motility abnormalities that cause dysphagia or food impaction in patients with EE, particularly older children and adults. It is also possible that eosinophils infiltrate deeper layers of the esophagus leading to edema, fibrosis, and/or muscle thickening, which contribute to altered motility. Documentation of such a finding, however, is extremely rare in the literature ^(26,90).

In summary, mucosal biopsies should be obtained from all patients undergoing evaluation for GERD-like symptoms that do not respond to acid blockade, or who present with dysphagia or food impaction. Biopsies should be obtained regardless of the gross appearance of the mucosa, and multiple biopsies should be obtained. Biopsies should also be obtained from stomach and duodenum to rule out other diseases such as eosinophilic gastroenteritis and IBD. Currently, in the proper clinical context (lack of response to proton pump inhibition or normal pH monitoring of the distal esophagus), the number of intraepithelial eosinophils is the primary histological feature distinguishing EE from GERD. The number of eosinophils should be determined in the most intensely inflamed high power field. Additional features such as eosinophilic microabscesses, surface layering of eosinophils, basal layer hyperplasia, and lamina propria fibrosis and inflammation should be assessed in all biopsies. Esophageal biopsy histology is critical for the diagnosis, but ultimately the diagnosis rests on both histological and clinical findings.

NONINVASIVE MARKERS

Sandeep K. Gupta, MD

EE is a chronic condition with periods of exacerbation and remission. Esophagogastroduodenoscopy and histological examination of esophageal mucosal biopsies are required to establish the diagnosis, assess response to therapy, document disease remission, and evaluate symptom recurrence. Because of the invasiveness of esophagogastroduodenoscopy and the costs associated with repeated procedures, there is an acute need to identify noninvasive biomarkers that correlate with disease presence, remission, severity, and response to therapy. Biomarkers may be broadly divided into 3 groups (Table 1) ⁽²⁷⁾. The following lists the characteristics of an ideal biomarker for EE:

Table 1

Correlates with EE state

Connects with EE severity

Reflects changes due to therapy

High sensitivity

High specificity

Reproducible

Performed on specimens which are noninvasively obtained and relatively easy to obtain

Simple methodology

Cost effective

Identification of an EE Biomarker

Although there are gaps in our understanding of EE, the underlying pathogenesis is thought to be related to a Th2 response that is controlled by cytokines including interleukin (IL)-5, IL-13, and eotaxin ⁽¹³⁾. In addition, a number of studies have identified the pluripotent potential of eosinophils and have documented the fact that they contain a number of cytokines, chemokines, and granular proteins ⁽²⁸⁾. These biologically active mediators are released upon activation, and likely participate in the inflammatory cascade. Thus, future identification of novel EE biomarkers likely will be related to mediators dictating the influx of eosinophils into the esophageal mucosa or those mediators that are released by the eosinophils themselves. These biomarkers could be obtained from a number of different sources including sputum, breath, blood, stool, and urine. This section focuses on biomarkers studied in a number of eosinophilic diseases, including EE. Well-designed studies that take into account severity of concomitant atopic diseases, proper controls, treatment interventions, and the influence of preservative, fixation, and storage methods will be critical to final interpretation of these studies.

Eosinophil Cell Count

Peripheral eosinophil counts have not been studied in a rigorous manner to determine utility as a biomarker. Studies to date demonstrate that peripheral eosinophil counts may be elevated in as many as 75% to 100% of patients, but others show no correlation. Eosinophils in the sputum have been used as a reliable marker of disease activity in patients with asthma, occupational asthma, and eosinophilic bronchitis ⁽²⁹⁾, but have not been examined in patients with EE. Because murine studies demonstrate bronchial eosinophilia in models of EE, sputum eosinophil counts may correlate with esophageal eosinophilia. Intact stool eosinophils have not yet been examined as a marker of EE.

CD23

CD23 is the low-affinity receptor for immunoglobulin E (IgE) ⁽³⁰⁾. This membrane protein is present on a variety of cells, including enterocytes, dendritic cells, eosinophils, and subpopulations of B cells and T cells. CD23 expression is induced by cytokines associated with allergic responses, and it could be a marker of GI allergy. CD23 has been detected in the stools of patients with food allergies, but not in those of controls ⁽³¹⁾. The utility of this marker in a mixed (IgE/non-IgE) process, such as EE, is unclear but worthy of further study.

Mast Cell Products

Several reports suggested that mast cells and their products may participate in the pathogenesis of EE, although the molecular events are poorly defined ^(32-34). Leukotrienes, a mast cell product, may play a role in muscle contraction and 1 clinical report found that supraphysiological doses of the leukotriene receptor antagonist montelukast was effective in reducing symptoms but not tissue eosinophilia. In a recent study leukotriene protein levels were measured in esophageal mucosal biopsies of controls and children with EE, but were found to be similar ⁽³⁵⁾. Interestingly, the levels were elevated in esophageal biopsies of patients with eosinophilic inflammation of the GI tract that extended beyond the esophagus compared with controls. Thus, although leukotrienes may not be involved in the pathogenesis of EE, these mast cell proteins could conceivably serve as mediators of more extensive GI eosinophilic inflammation. Additional studies into the role(s) of leukotrienes in eosinophilic inflammation could include examining other body specimens and measuring leukotriene mRNA levels.

Because of its association with other allergic diseases and its capacity to induce smooth-muscle contraction, histamine is another mast cell product deserving consideration. It has a stable metabolite, N-methylhistamine, which can be measured in urine, as previously shown in patients with IBD but not in patients with EE ⁽³⁶⁾.

Cytokines and Chemokines

A multitude of cytokines and chemokines serve as inflammatory mediators. Some of these are specific to particular inflammatory states, whereas others are general mediators of inflammation. Eotaxin-3, an eosinophil-specific chemokine, was examined in esophageal biopsies of a cohort of children (controls, EE, and chronic/reflux esophagitis) ⁽³⁷⁾. The eotaxin-3 gene, mRNA, and protein were upregulated in esophageal biopsies of children with EE compared with controls. Higher levels of eotaxin-1 and eotaxin-2 mRNA, but not protein, also were noted in the biopsies of children with EE compared with controls. Eotaxin-3 protein level in plasma was twice as high in children with EE compared with controls; eotaxin-1 and eotaxin-2 protein levels in plasma were not performed. These interesting findings raise the possibility of eotaxin-3 differentiating patients with EE from controls.

In another study mRNA of several cytokines was measured on esophageal biopsies from children with EE and controls ⁽³⁸⁾. Levels of IL-5, eotaxin-1, and RANTES (regulated upon activation, normal T cell expressed, and secreted protein) were higher in children with EE compared with controls, whereas levels of IL-4, IL-13, and eotaxin-3 were similar between EE and controls. Interestingly, eotaxin-2 levels were higher in controls compared with children with EE. In an adult study, eotaxin-1 protein levels in esophageal biopsies were similar between controls and EE patients ⁽³⁹⁾.

Eosinophil-derived Neurotoxin

Eosinophil-derived neurtoxin (EDN) is 1 of several toxic proteins contained within the cytoplasmic granules of eosinophils ⁽²⁸⁾. Other granular proteins include major basic protein, eosinophil peroxidase, and eosinophil cationic protein. These proteins are secreted on eosinophil stimulation and actively participate in the subsequent inflammatory response. Elevated levels of some of these granular proteins have been reported in the serum and urine of children with asthma and in the stool of children with IBD ⁽⁴⁰⁾. In a longitudinal study of children with EE, baseline EDN levels in blood and stool decreased following corticosteroid therapy ⁽⁴¹⁾. A similar trend in serum IL-5 levels also was noted, and it is plausible that these biomarkers could serve as markers of disease activity in lieu of repeated endoscopies.

Biomarkers as a Group

A recently published novel study examined a group of laboratory tests as markers of EE activity ⁽⁴²⁾. Children with active EE, inactive EE, and controls were enrolled in this cross-sectional study. Plasma eosinophil counts, EDN levels, and eotaxin-3 levels were higher in active EE compared with controls. Levels of plasma IL-5, eotaxin-1, eotaxin-2, and stool EDN were similar in active EE versus controls. Additional studies on this concept should be pursued and be of longitudinal design.

Conclusions

In an attempt to reduce the necessity of repeated endoscopy and aid in the diagnosis of patients with unclear diagnoses, the identification of a role for these potential biomarkers is acutely needed. Future studies must focus on longitudinal measurement, durability of the testing technique, and the reproducibility of these tests in different laboratories. In addition, these data will need to be weighed against the associated costs, the practicality of methodology, and patient burdens associated with specimen collections.

II. PATHOGENESIS OF EE MOUSE MODELS OF EOSINOPHIL-ASSOCIATED GI DISEASES

James J. Lee, PhD

Mouse models of human disease are becoming invaluable to our understanding of the underlying mechanisms contributing to inflammatory responses that occur in the digestive tract. Specifically, the conserved character of mammalian physiology and biochemistry, and the underlying molecular mechanisms that mediate inflammatory responses, are allowing investigators using the mouse not only to test ideas derived from patient studies but also to develop novel hypotheses to improve understanding of human GI diseases. In particular, the use of mice has been beneficial to studies of the rare and enigmatic eosinophil, leading to a rethinking of the classical paradigm describing these leukocytes as uniquely destructive effector cells that contribute only to pathology ⁽⁴³⁾. In addition, studies of eosinophils in the mouse have provided a plethora of reagents and model systems that investigators in the research community now are using in strategies to define mechanisms of eosinophil effector function and the role(s) of these leukocytes in disease.

The development and use of gene transfer technologies in the mouse have led to the creation of 2 broad categories of mouse models, transgenic mice and gene knockout animals. These model systems have each contributed significant insights regarding the activities of eosinophils:

Transgenic mouse models modulating eosinophil numbers: These mouse models result from the integration and expression of DNA introduced into the genome of the mouse. This strategy is the oldest of the gene-transfer technologies used in the mouse ⁽⁴⁴⁾ and has matured into a strategy capable of directing both a spatial (ie, cell) and temporal-specific pattern of gene expression.

Our laboratory has used this technology during the past decade to generate mouse models that have become instrumental in studies of eosinophil effector functions. For example, the development of a hypereosinophilic mouse model was achieved through the creation of a transgenic line of mice (NJ.1638) expressing the eosinophilopoietic cytokine IL-5 in all mature circulating T cells ⁽⁴⁵⁾. Surprisingly, the importance and long-term utility of this model has not been through insights gained regarding the production of eosinophils. Instead, this model has become an invaluable source of eosinophils and eosinophil gene products for ex vivo studies on isolated cells ⁽⁴⁶⁾, in vivo studies using adoptive cell transfer ⁽⁴⁷⁾, and the production of eosinophil-specific antibodies ⁽⁴⁸⁾.

The production of eosinophil-specific antibodies has been particularly important to the mouse research community at large; however, a reassessment of our mouse-specific eosinophil granule protein antibodies has recently shown that many of these reagents display cross-reactivity against human eosinophil granule proteins and are capable of detecting the presence of eosinophils and eosinophil degranulation in clinical biopsies from patients. In particular, an anti-eosinophil peroxidase (EPO) mouse monoclonal antibody was identified from these assessments for its utility in immunohistochemistry with formalin-fixed paraffin-embedded tissue samples. This antibody represents a new and novel reagent capable of detecting the presence of eosinophil activation/degranulation in tissue sections. Preliminary studies with anti-EPO mouse monoclonal antibody already have demonstrated the unique character of this reagent and its potentially invaluable use as an eosinophil-specific reagent in the diagnosis of patients with eosinophil-associated diseases. We also have used transgenic technology to express a cytocidal protein uniquely in eosinophil lineage-committed marrow progenitor cells, generating a transgenic line of mice (PHIL) congenitally deficient of eosinophils with no additional effects on other white blood cell types ⁽⁴⁹⁾. These initial studies using PHIL in mouse models of asthma already have shown that eosinophils are required for the histopathologies and lung dysfunction associated with allergic respiratory inflammation. The availability of this model to the research community studying GI diseases invariably will provide the insights required to define the necessity and importance of this cell type in the development of the disease pathologies.

Gene knockout animals deficient of eosinophil-specific granule protein-encoding genes: Gene knockout animals are genetically engineered mouse models with deficiencies in the expression of a specific gene of interest. Briefly, embryonic stem cells and gene transfer via homologous recombination are used to produce mutations, initially in the genome of the embryonic stem cells, before the introduction of these genetically manipulated cells into recipient embryos and passage of the induced mutation through the germline. Thus, this gene transfer technology represents a strategy to silence (ie, knock out) the expression of specific genes through targeted mutagenesis. We have used this technology to target and/or knock out the expression of the genes encoding the abundant eosinophil secondary granule proteins EPO ⁽⁵⁰⁾ and major basic protein-1 ⁽⁵¹⁾ in the mouse. These knockout animals have been the focus of several GI studies that have demonstrated a link between the expression of these eosinophil-specific genes and the pathologies associated with IBD ^(52,53).

In summary, mouse models of eosinophil-associated human inflammatory diseases are critical components of ongoing investigations of eosinophil activities in GI diseases. In particular, the creation and characterization of these animal models have led to the development of reagents with clinical diagnostic applicability. More important, these models uniquely have allowed investigators to test prevailing hypotheses extending from clinical studies of patients as well as to develop novel insights regarding the mechanisms of eosinophil effector functions.

HOW DO EOSINOPHILS CONTRIBUTE TO INFLAMMATION? MURINE MODELS OF INTESTINAL EOSINOPHILIA

Simon P. Hogan, PhD

Eosinophil-associated GI disorders (EGID) are characterized by a rich eosinophilic inflammation of the GI tract in the absence of known causes for eosinophilia or other GI disorders. These disorders include EE, eosinophilic gastritis, eosinophilic gastroenteritis, eosinophilic enteritis, and eosinophilic colitis, and are being recognized with increasing frequency. Clinical studies suggest that eosinophils have a pathogenic role in EGID; however, the function of eosinophils in these disorders remains an enigma. Recently, several groups have developed experimental models of EGID to delineate the role of eosinophils in the pathophysiological features of EGID. This review briefly summarizes these recent experimental findings examining the contribution of eosinophils to disease pathogenesis.

EGID

Eosinophil accumulation in the GI tract is a common feature of numerous GI disorders including classic IgE-mediated food allergy, ⁽⁵⁴⁾ eosinophilic gastroenteritis, ⁽⁵⁵⁾ allergic colitis, ⁽⁵⁶⁾ EE ^(57,58), IBD ⁽⁵⁹⁾, and GERD ^(60,61). Patients with EGID suffer from a variety of problems, including failure to thrive, abdominal pain, irritability, gastric dysmotility, vomiting, diarrhea, microcytic anemia, hypoproteinemia, and dysphagia ⁽⁶²⁾. Studies suggest that the clinical symptoms are a direct consequence of eosinophil tissue involvement. Mucosal eosinophil involvement manifests with diarrhea, vomiting, bleeding, or malabsorption. Muscular involvement often is associated with obstructive symptoms, and serosal eosinophil involvement presents with ascites ⁽⁶³⁾. Furthermore, the level of eosinophil infiltration has been shown to correlate with clinical symptoms and disease severity.

Experimental Eosinophilic Gastroenteritis

The infiltration of eosinophils into the GI tract of humans, such as in patients with eosinophilic gastroenteritis, is frequently associated with several characteristic pathological changes, including weight loss and gastric dysmotility ⁽⁶⁴⁾. To elucidate the etiology of disease, the authors developed an experimental model oral antigen-induced eosinophilic GI inflammation that mimics eosinophilic gastroenteritis ⁽⁶⁵⁾. Oral administration of antigen (ovalbumin) to ovalbumin-sensitized mice induced a pronounced eosinophilic inflammation of the small intestine (duodenum, jejunum, and ileum). The oral antigen-challenged mice suffered from variable levels of reduced activity, increased respiratory rate, pilar erecti, and failure to thrive (cachexia). Postmortem GI examination of these mice revealed the presence of gastromegaly and evidence of gastric dysmotility. Using an in vivo gastric retention assay, we demonstrated impaired gastric emptying in oral allergen-challenged mice. In addition, morphometric analysis revealed a significant decrease in the villus/crypt ratio in the small intestine of oral allergen-challenged mice compared with control-challenged animals. Similarly, patients with a variety of inflammatory GI disorders also present histologically with reduction in the intestinal villus/crypt ratio ⁽⁶⁶⁾. Collectively, these results establish that experimental oral antigen challenge induces several pathological processes that are commonly observed in patients with EGID, including gastric dysmotility, cachexia, and reduced villus/crypt proportions.

To help explain the pathogenesis of the oral antigen-induced cachexia and dysmotility, we performed electron microscopic analysis of the jejunum from control- and oral antigen-challenged mice. Notably, eosinophils in the jejenum of oral antigen-challenged mice were in close proximity to damaged enteric nerves ⁽⁶⁵⁾. The enteric nerves contained swollen axonal chambers with variable loss of internal organelles, including the dense core granules of Schwann cells. Interestingly, these features, indicative of axonal necrosis, have been observed in EGID in humans ⁽⁶⁷⁾. Using eotaxin-1-deficient mice, we demonstrated that oral antigen challenge-induced cachexia and gastric dysmotility were dependent on eosinophils, suggesting a pathogenic role for eosinophils ⁽⁶⁵⁾. Collectively, eosinophil infiltration and degranulation into the GI tract was associated with marked GI pathology including villus/crypt shortening, gastric dysmotility, gastromegaly, and failure to thrive.

Experimental EE

Using an experimental model of EE, we have examined the contribution of eosinophils in EE. Nine doses of intranasal Aspergillus fumigatus antigen induced a concomitant eosinophilic inflammation of the lung and esophagus. Experimental EE was characterized by marked levels of esophageal eosinophils, free eosinophil granules, and epithelial cell hyperplasia, features that mimic the human disorders. Interestingly, exposure of mice to oral or intragastric allergen does not promote EE, indicating that hypersensitivity in the esophagus occurs with simultaneous development of pulmonary inflammation. In the absence of IL-5, eosinophil accumulation in the esophagus following oral allergen challenge was ablated ⁽⁶⁸⁾. The reduction in eosinophil numbers in the esophagus of these mice was associated with a reduction in epithelial cell hyperplasia, suggesting a pathophysiological connection between eosinophils and the development of EE.

Experimental Eosinophilic Colitis

Using a well-established model of colonic injury (2.5% dextran sulfate sodium [DSS] model), we have dissected out the molecular and cellular mechanisms involved in eosinophil recruitment and the potential contribution of eosinophils to the pathogenesis of eosinophilic colitis ^(52,69). Administration of DSS induced colonic eosinophilic inflammation and GI dysfunction (diarrhea with blood, and shortening of the colon), which resembles ulcerative colitis in patients. Eosinophils in DSS-treated mice appeared to be undergoing cytolytic eosinophilic degranulation, as evidenced by nuclear chromatolysis, disruption of plasma membrane, and the presence of free eosinophilic granules in the extracellular spaces adjacent to these eosinophils ⁽⁵²⁾. We also demonstrated that eosinophil degranulation was associated with elevated levels of colonic luminal EPO activity in mice with colonic injury. Moreover, EPO activity in DSS-treated mice was 1000-fold higher than that observed in control-treated animals. To examine the role of EPO in the pathogenesis of colonic injury, we used EPO-deficient mice ⁽⁵²⁾. DSS-induced colitis was significantly attenuated in EPO^-/- mice compared with strain-matched wild-type mice. This attenuation was associated with the loss of EPO activity in DSS-treated EPO^-/- mice and occurred despite no reduction in colonic eosinophil levels. To confirm a role for EPO in pathogenesis of colonic injury, we used an EPO inhibitor, resorcinol, and demonstrated that drug inhibition of EPO also attenuated DSS-induced colonic injury ⁽⁵²⁾. Thus, these studies demonstrate that eosinophil-derived EPO plays an important role in the immunopathogenesis of colonic injury.

In conclusion, EGIDs are becoming more prevalent in the Western world. A variety of clinical and experimental models have revealed that eosinophils promote potent proinflammatory effects mediated by their ability to release their cytotoxic secondary granule constituents and a variety of lipid mediators and cytokines. Although much progress has been made concerning the GI eosinophil and EGID, there is still a paucity of knowledge compared with other cell types and GI diseases that may be even less common (eg, IBD). It is hoped that a better understanding of the pathogenesis and treatment of EGID will emerge by combining holistic clinical and research approaches involving experts in the fields of allergy, gastroenterology, nutrition, and pathology.

CONSIDERING THE ROLE OF MAST CELLS IN EE

Barry K. Wershil, MD, and Glenn T. Furuta, MD

Mast cells are located in virtually all vascularized tissues. Their ability to produce proinflammatory mediators, both preformed and newly synthesized, has fueled speculation that mast cells play a prominent role in a variety of allergic and nonallergic inflammatory conditions. Here, the authors review evidence and theory that mast cells participate in the pathophysiology of EE.

Mast Cells and the Esophagus

Mast cells are normally found in the mucosa and submuscosa of the esophagus, but rarely in the muscularis ⁽⁷⁰⁾. Mast cells contain a panel of preformed mediators within their cytoplasmic granules, including histamine, neutral proteases, tumor necrosis factor (TNF)-α, and others ⁽⁷¹⁾. In addition, mast cells can synthesize numerous biologically potent mediators, such as leukotrienes, platelet-activating factor, and a host of cytokines and chemokines ⁽⁷¹⁾. The production of these mediators by mast cells can have a number of direct and/or indirect effects. As an example, leukotriene production by mast cells could have direct effects in inducing smooth-muscle contraction and edema by increasing vascular permeability. Also, leukotrienes can recruit inflammatory cells from the circulation, which would be an indirect mast cell pathway affecting the inflammatory response.

Mast cell activation to release mediators can occur through several mechanisms. The most extensively studied is antigen cross-linking of IgE antibodies on the surface of the mast cell, leading to degranulation and mediator release. Although this may be relevant in some cases of EE, it is probably a minority. There are many alternative pathways that can induce mast cell mediator independent of IgE, however. Particularly relevant to EE is the ability of eosinophil-derived proteins to induce mast cell degranulation (see below). Neuropeptides contained in nerve terminals in the GI tract also can induce mast cell degranulation, and there are substantial data implicating substance P in nerve/mast cell interactions, although this is still an area of controversy in human systems ⁽⁷²⁾. Other mechanisms that can induce mast cell mediator release in the GI tract include acid reflux ⁽⁷³⁾ and bile acids ⁽⁷⁴⁾. Thus, a number of classic and alternative mechanisms may be involved in the activation of esophageal mast cells during inflammation.

Interaction of Mast Cells With Eosinophils

Mast cells also may have direct and indirect interactions with eosinophils. As previously mentioned, some of the cationic proteins contained by eosinophils, particularly eosinophil-derived major basic protein, can induce mast cell degranulation and TNF-α production ⁽⁷⁵⁾. This represents a potential positive feedback loop augmenting the inflammatory response. Mast cells also can produce a number of relevant cytokines and chemokines that influence eosinophil function, recruitment, and survival, such as IL-4, IL-5, GM-CSF, IL-13, and eotaxin-1 ⁽⁷¹⁾.

Mast Cells and EE

Much of the data implicating mast cells in the pathogenesis of EE is indirect and inferential. An increased number of mast cells have been noted in esophageal biopsies from patients with EE and the mast cell counts correlate with esophageal eosinophilia ^(37,39,76). Elevated levels of mast cell-specific or associated genes also were noted ⁽⁷⁶⁾; however, studies differ in being able to identify evidence of mast cell degranulation ^(37,76). This may not be critical because with some stimuli such as bacterial cell wall products, mast cells can release newly synthesized mediators without overt degranulation ⁽⁷⁷⁾. These findings certainly suggest the participation of mast cells in the pathogenesis of EE, but it is important to consider that these findings may represent an epiphenomenon.

Assuming that mast cells do participate in the disease leads to several interesting speculations, however. The release of mediators such as histamine, leukotrienes, TNF-α, and eosinophil chemoattractants can act to augment edema, inflammatory cell recruitment, and smooth-muscle contractility. Mast cells are also a source of transforming growth factor (TGF)-β, which has been implicated in various conditions associated with tissue fibrosis, including EE. The net effect of mast cell participation could be proinflammatory effects on esophageal motility and stricture formation. There is the possibility, however, that some aspects of mast cell involvement in EE may be beneficial. For example, in patients with esophageal achalasia, a primary motility disorder, mast cell numbers expand in the esophagus and the mast cells seem to associate with the interstitial cells of Cajal ⁽⁷⁸⁾, which are important in motor activity in the GI tract. This association appears to correlate with survival of the interstitial cells of Cajal ⁽⁷⁸⁾. Thus, it is possible that mast cells could play a protective role in some conditions.

Conclusions

Current evidence suggests a role for mast cells in the pathogenesis of EE, and the authors have put forth several potential mechanisms for how they may participate. Further research will be necessary to more clearly define their role, either positive or negative, and the mediators involved.

EOSINOPHIL TRAFFICKING TO THE ESOPHAGUS: ROLE OF CHEMOKINES IN EE

Marc E. Rothenberg, MD, PhD

Multiple lines of evidence suggest that EE is associated with Th2-type immune responses. In particular, elevated levels of eosinophil-active Th2 cytokines (eg, IL-4, IL-5, IL-13) as well as mast cells are present in the esophagus of patients with EE ^(19,37). In addition, experimental models of EE can be induced in mice by allergen exposure, especially in the respiratory tract following mucosal or epicutaneous sensitization, as well as by overexpression of Th2 cytokines (IL-5 and IL-13) ^(68,79-81). Collectively, these experimental systems demonstrate an intimate connection between the development of eosinophilic inflammation in the respiratory tract and esophagus, not only in response to external allergic triggers but also in response to intrinsic Th2 cytokines.

Using whole genome microarray expression profile analysis, the dysregulated expression of ∼1% of the human genome in the esophagus of patients with EE was recently described ⁽³⁷⁾. Interestingly, the EE transcript profile markedly distinguishes EE from chronic esophagitis (typical of reflux); the latter mainly expresses genes comparable to normal individuals. Of the entire dysregulated genome, the gene with the greatest overexpression and correlation with esophageal eosinophilia is eotaxin-3, induced 50- to 100-fold compared with control individuals ⁽³⁷⁾. Comparison of allergic and nonallergic EE patients revealed that the gene transcript signature was conserved markedly across these 2 major patient phenotypes.

Experimental modeling in mice has established that Th2 signaling is required for induction of experimental EE. In particular, mice with the targeted deletion of STAT6 are protected from allergen and IL-13-induced experimental EE ⁽⁸¹⁾. Notably, IL-13-deficient mice have impaired allergen-induced EE ⁽⁸¹⁾. Because IL-4 and IL-13 are potent inducers of the eotaxins, it is interesting to speculate that IL-13 may induce EE by upregulating the production of eotaxin-3. Of note, IL-13 selectively induces eotaxin-3 in keratinocytes from the skin ⁽⁸²⁾. Notably, the esophageal tissue from patients with EE markedly overexpresses eotaxin-3, but not eotaxin-1 or eotaxin-2, suggesting that IL-13-induced signaling in esophageal keratinocytes may specifically induce eotaxin-3 ⁽³⁷⁾.

During the past several years, evidence has accumulated that EE has a strong familial association. In the authors' experience, nearly 10% of parents of patients with EE have a history of esophageal strictures and ∼8% have biopsy-proven EE, suggesting that EE is a genetic disorder ⁽⁴⁾. Indeed, genetic analysis of a single nucleotide polymorphism (+2496T>G, rs2302009) in the eotaxin-3 gene has shown association with EE by both population-based case-control comparison and family-based transmission disequilibrium test ⁽³⁷⁾. Of note, this single nucleotide polymorphism, located in the 3′UTR of the eotaxin-3 mRNA, may affect mRNA stability. The induction of inflammatory cytokines often is controlled at the level of mRNA stability. Moreover, this single nucleotide polymorphism may disrupt a putative AU-rich element sequence and thus modify the mRNA levels and/or responsiveness to glucocorticoids. Interestingly, eotaxin-3 (+2496T>G) is in strong linkage disequilibrium with upstream genetic variants, making a ∼7-kb haplotype block in the white population ⁽⁸³⁾. This genetic variant (+2496T>G) is likely to contribute to EE susceptibility, as are any in strong linkage disequilibrium with it.

Advances in understanding the molecular pathogenesis of EE started only 4 years ago. Because of the lack of data and agreement, the cutoff value of esophageal eosinophil concentrations has not been agreed upon. Although an empiric threshold level of eosinophils may prove useful for disease classification, the authors recommend that the EE gene transcript signature including overexpression of eotaxin-3 be considered in disease definition. Further studies should focus on the molecular mechanisms involved in EE pathogenesis, its differentiation from the pathogenesis of atopy in general, and the involvement of inflammatory cells and inflammatory mediators. We envision that molecular diagnostics, similar to the approach being taken to classify cancer, will be applied to EE and become useful for diagnosis and differentiation from reflux esophagitis and prediction of therapeutic responsiveness and prognosis. It is important to note that the mechanisms by which any therapeutic intervention improves EE have not been established, highlighting the value of translational research aiming to develop optimal EE therapy. The authors anticipate that mechanism-based therapeutic intervention (eg, anti-eotaxin-3, eotaxin-1 receptor antagonists, anti-IgE, anti-IL-5, anti-IL-13) will prove to be successful therapies for EE.

CONTRIBUTION OF EOSINOPHILS TO FIBROSIS

Steven J. Ackerman, PhD, and Ignatius Gomes, PhD

Eosinophils as Effector Cells of Tissue Remodeling and Fibrosis

Tissue fibrosis is characterized by the pathological accumulation of extracellular matrix (ECM) in response to tissue damage, a variety of infectious processes, and dysregulated host inflammatory responses. Eosinophils are now thought to play a significant role in mediating tissue remodeling and fibrosis ⁽⁸⁴⁾ in a variety of eosinophil-associated diseases, including asthma ^(85,86), eosinophil myalgia syndrome ⁽⁸⁷⁾, eosinophilic endomyocardial fibrosis ⁽⁸⁸⁾, idiopathic pulmonary fibrosis ⁽⁸⁹⁾, scleroderma ⁽⁸⁷⁾, and EE ^(62,90). Eosinophils are implicated in fibrogenesis through their clinical disease associations ⁽⁸⁴⁾, their elaboration of fibrogenic growth factors such as TGF-β ^(91,92) and platelet-derived growth factor BB ⁽⁹³⁾, and secretion of their granule cationic proteins such as major basic protein (MBP) ⁽⁹⁴⁾ and EPO ⁽⁹⁵⁾, or matrix metalloproteinases (MMPs) such as MMP-9 ⁽⁹⁶⁾. The topographic association of degranulating eosinophils and deposition of MBP in tissues with pathological fibrosis is a recurrent finding in a broad group of eosinophilic illnesses, and eosinophils have been identified as the major TGF-β-producing cell in the lungs of asthmatics ⁽⁸⁶⁾.

Recent human and animal model studies provide compelling evidence for eosinophils as effector cells in tissue remodeling and fibrosis (Fig. 1). For example, a reduction in bronchial mucosal eosinophils induced by treatment of asthmatics with anti-IL-5 antibody (Mepolizumab) significantly decreases the expression of a number of ECM proteins including tenascin, lumican, and type III collagen in the bronchial reticular basement membrane ⁽⁹⁷⁾. Likewise, anti-IL-5 treatment significantly decreases tissue eosinophils and deposition of ECM proteins in the skin in allergen-induced late-phase reactions in atopic subjects ⁽⁹⁸⁾. Direct evidence for the role of the eosinophil in the pathogenesis of tissue remodeling and fibrosis comes from recent studies in which eosinophil-deficient mice were used to demonstrate essential roles for eosinophils in the development of airway hyperresponsiveness and remodeling, including mucus (goblet) cell metaplasia, smooth-muscle cell hyperplasia, and subepithelial fibrosis ^(49,99,100).

Fig. 1

Eosinophil Expression of Fibrogenic Factors

Multiple growth factors and cytokines expressed by eosinophils ^(101,102) have been implicated in the development of tissue remodeling and fibrosis (Fig. 1). Among these, TGF-β is the most widely studied and regarded as the most potently fibrogenic. TGF-β regulates the expression of the profibrogenic cytokine IL-6, the myofibroblast marker α-smooth muscle actin, and other extracellular matrix proteins such as the collagens. TGF-β expression is correlated with bronchial airway fibrosis and severity of asthma ⁽¹⁰³⁾. Induction of pulmonary fibrosis by TGF-β and other cytokines ^(92,104-110) provides some of the best evidence to date for the involvement of these factors in inducing tissue remodeling and fibrosis in the lung.

Mechanisms of Eosinophil-induced Fibrogenesis in Asthma and EE

Eosinophil-fibroblast interactions have been implicated in the generation of subepithelial fibrosis and airway remodeling characteristic of human asthma in murine allergic asthma models ^(96,111). Mechanistic assessments of eosinophil-fibroblast interactions that may lead to fibrosis have been described in only a limited number of studies, however. Eosinophils are 1 of the major inflammatory cells recruited to the lung in response to allergic pulmonary inflammation, where they reside for an appreciable amount of time (likely due to GM-CSF-mediated survival) and could interact with neighboring fibroblasts within the parenchyma. The significance of these eosinophil-fibroblast interactions remains unclear, as do their contributions to physiological tissue repair and/or the development of pathological tissue fibrosis. The authors previously reported that eosinophil granule MBP synergizes with TGF-β-primed or IL-1-primed human lung fibroblasts to induce significant increases in gene transcription and secretion of members of the IL-6 family of inflammatory cytokines, including IL-6 and IL-11 ⁽⁹⁴⁾. TGF-β-induced fibroblast secretion of IL-6 has been implicated in the overproduction of collagens, tissue inhibitor of metalloproteinases, and glycosaminoglycans in fibrogenesis ^(112,113).

Eosinophil-lung fibroblast co-culture in the presence of IL-5 induces fibroblast to myofibroblast transdifferentiation with increased expression of α-smooth muscle actin and ECM ⁽¹¹⁴⁾. Eosinophils may have an indirect impact on fibroblast phenotype and fibrogenesis through activation of the epithelial-mesenchymal trophic unit ⁽¹¹⁵⁾; for example, through secretion of TGF-β, or MBP and EPO (Fig. 1) ⁽⁹⁵⁾. Alternatively, eosinophils may induce fibrogenesis through TGF-β induction of the epithelial to mesenchymal transition as shown to occur in kidney fibrosis ⁽¹¹⁶⁾ and, more recently, the lung ^(117-119). Although correlative relationships have been established between eosinophils, fibroblasts, and their interactions in fibrogenesis, the specific mechanisms by which eosinophil-expressed cytokines and granule cationic proteins regulate or alter normal fibroblast function and phenotype leading to pathological tissue fibrosis require further study.

Subepithelial fibrosis, a major component of airway remodeling in the pathogenesis of bronchial asthma, is believed to be initiated by insults that include Th2-mediated allergic responses ^(120-122). Eosinophilic inflammation is believed to drive the differentiation of airway fibroblasts to myofibroblasts as characterized by the expression of myofibroblast-specific markers such as α-smooth muscle actin and by the deposition of ECM proteins such as collagens, fibronectin, and other ECM constituents, such as tenascin and lumican ^(97,115). Eosinophils recruited to the lung in asthma likely interact with fibroblasts beneath the reticular basement membrane and become activated to release fibrogenic growth factors such as TGF-β, driving fibroblasts to differentiate into myofibroblasts, which then deposit pathological amounts of collagens and other ECM proteins contributing to airway subepithelial fibrosis ⁽¹²³⁾. Activated eosinophils in asthmatic lung secrete their granule contents through processes termed piecemeal degranulation and cytolysis ⁽¹²⁴⁾. This property of activated eosinophils may allow them to selectively secrete profibrogenic growth factors such as TGF-β and granule cationic proteins such as MBP1 that in turn prime or induce the fibrogenic phenotype of resident pulmonary fibroblasts (Fig. 1) ^(84,125). A recent report showing correlations between pulmonary expression of eotaxin-1, expression of eotaxin-1 receptor, TGF-β₁, and pulmonary fibrosis in a bleomycin mouse model support this general mechanism ⁽¹²⁶⁾.

Studies of tissue remodeling and fibrosis in EE are extremely limited to date, principally because of the difficulties inherent in obtaining sufficient amounts of esophageal biopsy tissue below the hyperplastic epithelium. For this reason, evidence for progressive remodeling and fibrosis of the esophagus derives principally from endoscopic, radiological, and histological features of the disease ⁽⁸⁾. These features include epithelial hyperplasia, furrowing, trachealization, decreased esophageal caliber, strictures ⁽¹²⁷⁾, and increased thickness of the muscularis propria and total esophageal wall ⁽⁹⁰⁾, with a few studies showing increased subepithelial collagen (ie, ECM) deposition ⁽⁸⁾. Of note, no studies demonstrate whether esophageal eosinophils are the principal source of factors involved in epithelial hyperplasia or subepithelial fibrosis in EE or what other remodeling and fibrogenic factors are expressed; however, a recent study of esophageal biopsies from pediatric patients with EE showed increased levels of subepithelial fibrosis and increased expression of TGF-β₁ by eosinophils and its signaling molecule phospho-SMAD2/3 compared with patients with GERD and normal controls ⁽¹²⁸⁾. Beyond this report, esophageal remodeling and fibrosis have not been systematically studied to define the changes in epithelial cell and fibroblast phenotype, extracellular matrix constituents, and smooth-muscle hyperplasia that may contribute to remodeling of the esophagus in chronic EE. The recent genome-wide expression profiling studies of EE esophageal tissue that identified increased expression of eotaxin-3 as the likely candidate responsible for eosinophil recruitment ⁽³⁷⁾ interestingly did not identify many genes known to participate in tissue remodeling and fibrosis, perhaps because the biopsy specimens analyzed were sufficiently superficial to include mainly hyperplastic epithelium and not subepithelial fibrotic tissue.

Eosinophils infiltrating the esophagus in patients with EE have been shown to express both IL-4 and IL-13 ⁽¹²⁹⁾, cytokines clearly involved in the development of airway remodeling and subepithelial fibrosis in asthma. More important, in terms of the natural history of EE, the time frame is not well defined from disease onset to the development of epithelial hyperplasia, thickening of the muscularis propria and esophageal wall, and subepithelial deposition of collagens and other ECM constituents that contribute to esophageal remodeling and fibrosis. Also unclear are the relationships of esophageal remodeling to disease severity and duration, and to what extent esophageal remodeling and fibrosis are reversible with treatments that significantly reduce tissue eosinophils in the esophagus. Of interest, a recent open-label phase I/II safety/efficacy study of 4 adult EE patients treated with anti-IL-5 (Mepolizumab) suggests that decreasing esophageal eosinophils ∼6- to 9-fold leads to reversal of esophageal remodeling in terms of decreased epithelial hyperplasia ⁽¹³⁰⁾.

Summary

The role of the eosinophil in the development of esophageal remodeling and fibrosis in EE has not been adequately defined and warrants further study. Better understanding of the mechanisms by which eosinophils promote tissue remodeling and fibrogenesis may ultimately lead to the development of novel therapeutic approaches for blocking and/or reversing the debilitating and sometimes life-threatening fibrosis seen in many eosinophil-associated allergic diseases, including EE.

III. TREATMENT OF EE ENDPOINTS FOR TREATMENT OF EE

Chris A. Liacouras, MD

During the past 10 years, EE has become increasingly recognize. There are several accepted methods of treatment that provide symptomatic relief. Unfortunately for most of these treatment plans, when the treatment is removed, the disease often recurs. Additionally, symptomatic relief does not always correlate with histological improvement. Because of these issues, controversy exists about whether patients should be treated for symptom relief or to achieve histological normalcy.

Physicians must consider several factors when treating a disease: Physicians endeavor to approve patient symptoms. Along with the improvement of symptoms, physicians often attempt to improve the patient's quality of life. When dealing with diseases such as EE, symptoms often occur as a result of severe tissue inflammation. Thus, in most cases, the improvement of patient symptoms only occurs when tissue inflammation is improved. Finally, when therapies are introduced, physicians attempt to minimize complications. In fact, do no harm should always be considered when applying any treatment.

There are several factors that make the treatment of EE controversial. Because the disease was first identified only 10 years ago, the natural history is not well understood. Additionally, patients present in a variety of ways, ranging from daily, severe symptoms (vomiting, abdominal pain) to intermittent symptoms (dysphagia) occurring only once every 6 weeks to 8 weeks. Several modes of therapy have been shown to be effective; however, these therapies have advantages and disadvantages. Because there is no single, standardized accepted method for treatment, doctors are presented with many different options regarding the proper therapeutic approach.

When deciding about the endpoint of treatment, several questions must be asked. These questions include: Is it satisfactory to simply treat a patient's symptoms rather than treat the underlying histology? Is it acceptable to use medical therapy intermittently that would allow the disease to wax and wane? Should patients be kept on chronic daily therapy to prevent all histological inflammation? Would the complications of the disease occur less often if the disease was treated more aggressively?

To answer these questions, the disease and its potential complications must be clearly defined. Patients with EE currently present in a variety of ways. Typically, young children generally present similarly to patients who have GERD. Their symptoms include chronic vomiting, regurgitation, irritability, nausea, and abdominal pain. Older children and adults may complain of heartburn, waterbrash, and nausea; however, these older patients typically begin to experience episodes of dysphagia. As children progress into adolescence and early adulthood, acute food impactions and dysphagia often occur. The disease predominantly affects males, and patients often have other associated allergic conditions such as eczema, asthma, and allergic rhinitis. Finally, a strong family history of allergic disorders frequently exists.

Several therapies have been shown to be successful in treating EE. In cases in which dysphagia is associated with esophageal strictures, esophageal dilatation has been effective in improving symptoms ⁽¹³¹⁾. Systemic and topical corticosteroids improve not only the patient's symptoms but also the underlying tissue pathology ^(132,133). In fact, when high doses are used, the abnormal esophageal tissue often completely normalizes ^(15,134-136). Other medications, such as leukotriene receptor antagonists, have been shown to improve symptoms without any impact on tissue eosinophilia ⁽¹³⁷⁾. Unfortunately, in most cases, when these medications are withdrawn, tissue inflammation and clinical symptoms reappear; however, chronic treatment with systemic corticosteroids has been related to significant side effects, which include poor growth, bone abnormalities, cushingoid effects, and mood disturbances ⁽¹³⁸⁾. To date, severe complications associated with the use of topical steroids have not been reported; however, problems such as esophageal candidiasis, dry mouth, and epistaxis have been reported. Additionally, there continues to be concern that adrenal axis suppression may develop when topical steroids are used chronically at high doses.

Apart from medical therapy, dietary therapy has been shown to be effective. The removal of food antigens has been shown to significantly improve both the patient's symptoms and the esophageal eosinophilia ^(139-141). The advantages of using dietary therapy include minimal to no side effects, removal of the antigenic cause of the disease, and clinical improvement with tissue normalization. The disadvantage of restricted diets include the issue of compliance-these diets can be difficult to administer, especially to older children-the use of enteral tube feeding (creating a need for nasogastric or gastrostomy tubes), the cost required for the administration of specialized formulas, and issues surrounding quality of life. Recently, a small number of patients with EE have developed food tolerance, after 4 years to 5 years of dietary avoidance, to foods that they previously were unable to ingest.

It is also important to understand the possible complications of EE. There have been a large number of reports of older children and adults who have developed esophageal strictures requiring dilatation ⁽¹⁴²⁾. Often, these patients present emergently and require esophageal dilatation. Although dilation has been shown to be successful in the acute setting for the majority of patients, there have been reports of severe esophageal laceration, chest pain, hospitalization, and perforation when performing endoscopy or esophageal dilatation in these patients ⁽¹⁴³⁾. Other complications associated with EE include the development of a small-caliber esophagus or a sliding hiatal hernia. Recently, otolaryngologic and airway abnormalities such as subglottic stenosis, obstructive apnea, tonsillar hypertrophy, tracheal inflammation, and intraarytenoid mucosal edema also have been reported ⁽¹⁴⁴⁾. There have also been reports of esophageal superinfections that have developed in patients who have had chronic EE.

The major reason why it has been difficult to determine the endpoint of therapy relates to the natural history of the disease. There is no supportive literature demonstrating that patients with an untreated, severe esophageal eosinophilia always develop disease progression. If it were known that severe esophageal eosinophilia generally progressed to a narrowed or strictured esophagus, then it would be hard to argue that aggressive treatment should not be undertaken to prevent these complications. There have been several reports suggesting that patients can be observed and followed for several years without a significant change in their clinical symptoms or anatomic abnormalities. These studies also indicate that the disease does not simply disappear. Patients continue to have significant tissue eosinophilia and clinical symptoms ⁽⁸⁾. Finally, in an ongoing study at the Children's Hospital of Philadelphia, 24 patients were identified with EE over a 9-year period. All of these patients refused medical or dietary therapy. On average, 6 years later, they returned because of increasing clinical symptoms. Repeat endoscopy continued to show significant esophageal eosinophilia. Interestingly, all of the patients' initial symptoms were related to GERD; however, upon return, their symptoms changed to dysphagia and difficulty in swallowing. Finally, 5% to 10% of patients with EE who refused treatment developed a sliding hiatal hernia on repeat endoscopy.

Another argument concerning whether to treat these patients often relates to issues surrounding quality of life. Many investigators argue that several therapies may have a profound effect on the patient's quality of life. For example, when a strict dietary therapy is introduced, such as the use of an amino acid-based formula, patients are informed that they must avoid all other foods; moreover, these formulas typically require the insertion of a nasogastric tube or gastrostomy tube to ensure adequate energy intake ⁽¹⁴⁾. Additionally, the cost of using formulas and the social impact relating to the children not being able to eat the same foods as their peers is of great concern to many physicians. Yet, patients who present with EE typically have severe symptoms that cause profound changes to their quality of life. These patients frequently cannot eat normal foods because of severe dysphagia and often cannot participate in routine social activities because of their underlying disease.

Finally, concern always exists when significant tissue inflammation or abnormal mucosal lesions are left untreated. Disorders such as colonic polyposis, GERD, and celiac disease may progress to colonic cancer, Barrett esophagus, and lymphoma without proper treatment, and thus histological normalcy has been the goal. In other chronic diseases such as IBD, the mucosa may not appear histologically normal after treatment; typically, symptom relief has been the desired treatment goal. With regard to EE, the degree of basal cell hyperplasia and severe tissue inflammation raises concerns about future complications such as esophageal narrowing, stricture development, and the development of Barrett esophagus. To date, fibrotic narrowing is a documented risk, whereas neoplasia has not been identified. Where the treatment standard will be set is not certain, but these are the factors that must be considered. Once the disease has been treated successfully, an attempt should be made to minimize medical therapy whenever possible. Because clinical symptoms can occur intermittently, despite severe mucosal eosinophilic inflammation, and until more data are known regarding the long-term risks of chronic esophageal eosinophilia, upper endoscopy with biopsies should be strongly considered after withdrawal of therapy to assess recurrence of tissue inflammation.

At this point, how should we treat the disease? It is hard to argue that symptomatic patients should not be initially treated; however, the duration of treatment remains controversial. Patients who remain symptomatic should continue to receive therapy. In these patients treatment should be geared toward both clinical and histological resolution. If dietary therapy is used and patients can maintain dietary compliance, then this type of therapy should be continued for several years. Eventually, those foods known to cause EE can be reintroduced followed by repeat endoscopy with biopsy to determine whether the esophageal eosinophilia has recurred. With regard to medical therapy, treatment should again be geared toward treating both the symptoms and abnormal histology.

The more difficult issue is what to do with patients who have minimal or no clinical symptoms, but who continue to have significant EE. Some investigators recommend observation without treatment in clinically asymptomatic patients who continue to have severe esophageal inflammation. Because the disease does not spontaneously resolve, many of the above arguments suggest that treating abnormal histology may prevent the patients' future development of complications related to EE. Although disagreement exists regarding the need to treat asymptomatic esophageal inflammation, most physicians agree that in all cases patients continue to receive close follow-up and repeat upper endoscopies at routine intervals to assess changes to tissue inflammation.

IS THERE A MOLECULAR BASIS FOR ENTERAL TREATMENTS FOR INTESTINAL INFLAMMATORY DISEASES?

Simon Murch, PhD, FRCP, FRCPCH

Enteral treatments are used in a variety of intestinal inflammatory disorders. Dietary therapies may include removal of specific individual antigens, or a more stringent exclusion process in which hypoallergenic formulas are used as exclusive nutritional intake. One question is whether improvement so obtained is simply due to the absence of antigens that are capable of triggering IgE-mediated or non-IgE-mediated immune responses or whether a broader anti-anti-inflammatory process may be induced. Mechanisms suggested for such effects include promotion of epithelial integrity or modulation of chemokine production, or induced changes in dendritic cell function affecting Th1/Th2 balance or generation of regulatory lymphocytes. In addition, the effects of dietary constituents may be mediated indirectly by alteration in the gut flora, which has powerful immunomodulatory effects and specific interactions with the epithelium. Probiotic organisms represent a further rapidly developing therapeutic modality, which may be applicable to a range of inflammatory disorders. Studies of mechanism have uncovered a variety of potential means by which enteral therapies may affect mucosal immunopathologies, and further work is needed to refine and optimize these therapies, which have the great benefit of a low potential for adverse effects.

Enteral Nutrition Therapy in IBD

Enteral nutrition therapy, based on polymeric or elemental formulas given as exclusive intake while excluding other foods, has proven efficacy in achieving primary remission in Crohn disease. The important caveat is that compliance with therapy is important, and thus palatability and motivation are important considerations. Several mechanisms have been suggested, including promotion of epithelial restitution through trophic constituents such as glutamine, the inclusion in formulas of anti-inflammatory mediators such as TGF-β or omega-3 fatty acids, or possible direct modulation of nuclear factor-κB responses by dietary nucleic acids. In addition, there is emerging evidence that enteral therapy may alter the mucosal flora. Although direct evidence is thus far lacking in human studies, 1 murine study clearly identified induced change in the flora as critical in inhibiting the development of colitis in transgenic animals ⁽¹⁴⁵⁾.

Induced alteration in the flora by diet may affect mucosal immune responses in several ways, including production of short-chain fatty acids such as butyrate, which both provides epithelial nutrition and modulates chemokine responses ⁽¹⁴⁶⁾, and by inducing mucosal IgA and regulatory lymphocyte generation. Specific pattern-recognition receptors such as Toll-like receptors in turn have an impact on the host immune response to the flora, and provide a critical bridge between innate and adaptive immune responses.

Dietary Therapy in Allergic and Eosinophilic Gut Diseases

Evidence for a specific anti-inflammatory effect of dietary therapies in allergic disease, beyond exclusion of antigen, is still relatively scarce. Evidence, however, exists for opposing effects of IgA and IgE, each transported by specific mechanisms across the epithelium, in respectively retarding or promoting transcellular allergen intake ^(147,148). The secondary consequence of such antigen delivery is of induced degranulation of subepithelial mast cells, leading to a state of hyperpermeability that favors further antigen ingress ⁽¹⁴⁷⁾. Interaction between eosinophils and mast cells is of particular importance in inducing the disturbance of visceral nociceptive pathways that appears characteristic of many mucosal allergic and eosinophilic conditions in which visceral hyperalgesia is a prominent clinical feature ⁽¹⁴⁹⁾. Recent evidence identifies allergen-induced mucosal mast cell degranulation as particularly important in establishing a chronic state of hyperalgesia, which may persist for some time after antigen withdrawal ^(150,151).

As in IBD, the flora again represent a distinct pathway in which mucosal inflammatory events may be modulated in allergic gut disease. One important coexisting factor in allergic gut disease is low-grade enteropathy, leading to modulation of the flora through carbohydrate malabsorption, and to inhibition of regulatory lymphocyte generation through deficiency in micronutrients such as zinc and vitamin D.

Probiotics

The use of probiotic organisms represents an attractive future therapeutic option, with early evidence of efficacy in a variety of intestinal inflammatory disorders, including IBD, allergies, and neonatal necrotizing enterocolitis ^(152,153). The field is complex and clearly requires much further study of basic mechanism; however, in addition to a variety of studies in humans, experimental work confirms potential efficacy in restoration of the epithelial barrier, reversing Th2 skewing, and induction of regulatory immune responses ⁽¹⁵³⁾. The potential use of genetically modified organisms, which deliver high concentrations of regulatory cytokines such as IL-10 to the mucosa, represents an adventurous and potentially highly important therapeutic stratagem for the future ⁽¹⁵⁴⁾.

MOLECULAR MECHANISMS FOR ANTICYTOKINE TREATMENTS FOR EE

Anil Mishra, PhD

The accumulation of eosinophils in the esophagus is a commonly observed medical problem in patients with diverse diseases, including GERD, EE, eosinophilic gastroenteritis, and parasitic infections ^{(57,79,132,140,155,156)}. Esophageal eosinophilia is associated with allergic responses; for example, patients with EE have a high rate of atopy (and asthma) and their clinical symptoms and eosinophilic infiltrations are ameliorated by an elemental diet or anti-inflammatory therapy (cromoglycate or glucocorticoids) ^(132,140). Although the role of allergens in the induction of eosinophilia in the esophagus has been debated ⁽⁵⁷⁾, our experimental studies have established a link between aeroallergens and esophageal eosinophilia ⁽⁷⁹⁾. Exposure of anesthetized mice with repeated intranasal or intratracheal challenges of aeroallergen (eg, extracts of Aspergillus fumigatus) using a protocol to induce allergic airway inflammation promote marked EE (Fig. 2). This allergen-induced esophageal eosinophilia is accompanied by intraepithelial eosinophils, extracellular granule deposition, and epithelial cell hyperplasia, features that mimic the pathophysiological changes observed in individuals with various forms of EE ^(57,79). More important, the eosinophilic inflammation occurs in the lungs and esophagus but not the stomach or intestine, demonstrating an intimate immunological connection between Th2-associated allergic responses in the lung and esophagus. Of the cytokines produced by Th2 cells, IL-5 is the most specific for eosinophils. IL-5 induces eosinophil growth, differentiation, activation, and survival, and it primes eosinophils to respond to chemoattractants such as eotaxin, an eosinophil selective CC chemokine ^(155-157). Recent experiments have demonstrated that IL-5 is overexpressed in the esophagus of patients with EE ⁽³⁹⁾. Additionally, systemic overexpression of IL-5 (via pharmacological or transgenic approaches) promotes eosinophil trafficking to the esophagus in mice ⁽¹⁵⁸⁾ and neutralizing anti-IL-5 treatment in a murine model of EE restricts eosinophil trafficking to the lung and esophagus (Fig. 3) ⁽⁷⁹⁾. In addition to producing IL-5, Th2 cells are thought to induce eosinophil-associated inflammation through the secretion of an array of cytokines (IL-4, -5, -6, -9, -10, -13, and -25) that activate inflammatory and residential effector pathways both directly and indirectly. In particular, IL-4 and IL-13 are produced at elevated levels in the asthmatic lung and are thought to be central regulators of many of the hallmark features of disease. IL-13 appears to be particularly important because it is produced in high quantities by Th2 cells and regulates multiple features of allergic disease (IgE production, mucus overproduction, eosinophil recruitment and survival, airway hyperreactivity, and the expression of CD23, adhesion molecules, and chemokines [eg, eotaxin]) ⁽¹⁵⁹⁾. Dysregulation of IL-13 has been reported in a variety of allergic conditions including asthma ⁽¹⁶⁰⁾, atopic dermatitis ^(161-163), and allergic rhinitis ^(164-166). Overexpression of IL-13, by pharmacological administration or transgenic approaches, induces multiple features of asthma, including eosinophilia, mucus overproduction, and airway hyperresponsiveness ^(167,168). Neutralization of IL-13 activity by treatment with a soluble form of the IL-13 receptor α₂ chain reverses many of the characteristic features of experimental asthma ^(169,170).

Fig. 2

Fig. 3

Based on the importance of IL-13 in asthma and the high concordance between asthma and EE in humans and in murine models of asthma, the authors hypothesized that overexpression of IL-13 in the lung may be associated with the development of EE. To test this hypothesis, we delivered IL-13 to the lung by intratracheal administration. Data established that pulmonary inflammation, triggered by IL-13, is associated with the development of EE (Fig. 4). Human IL-13-induced lung and esophageal eosinophilia was shown to be reduced by pretreatment with human neutralizing IL-13 (CAT354) antibody ⁽¹⁷¹⁾; however, mice genetically deficient in IL-13 were not protected from the development of experimental EE by intranasal allergen exposure (unpublished data), but epidermal allergen exposure showed partial protection in the induction of esophageal eosinophilia ⁽⁸¹⁾. Therefore, it remains unclear that IL-13 has any direct role in EE induction. Based on these observations, the authors strongly recommend a multicenter clinical trial for anti-IL-5 therapy in EE patients. Furthermore, to explore other possible anticytokine therapies, we next focused our attention on the role of B and T cells in the pathogenesis of EE. Because the experimental EE (induced by respiratory allergen exposure) induces both esophageal and pulmonary inflammation, we compared lung and esophageal responses in B and T cell-deficient mice. B cell-deficient mice induce partial reduction in lung inflammation but no protection in esophageal inflammation, whereas T cell-deficient mice were protected from the lung and esophageal eosinophilic inflammation. These data indicated no role of allergen-induced antibodies in experimental EE, and T cell adaptive immunity is required for EE. Furthermore, the authors observed CD8⁺ T cells had no role, but CD4⁺ T cells partially contribute to the induction of experimental EE. In contrast, the role of CD4⁺ T cells in regulating lung eosinophilia is relatively more critical. These findings indicate, first, that EE is not associated with eosinophilic lung inflammation, and second, that both CD4⁺ and CD4^- T cells have a critical role in EE pathogenesis. These data further prompted us to target CD4⁺ and CD4^- specific T cell populations for the treatment strategy for EE. We are now focusing our studies on activated lymphocyte subpopulations, such as CD25⁺ T cells because these lymphocyte subpopulations are present in both CD4⁺ and CD4^- subsets in the esophagus. Emerging data from our laboratory indicates that both of these populations are increased in the esophagus of allergen-induced experimental EE, which indicates their possible role in EE pathogenesis. Therefore, we propose testing neutralizing antibodies against anti-CD25 in experimental EE.

Fig. 4

IV. SUMMARY AND CONCLUSIONS

The recent explosion of literature and increasing clinical experiences have established EE as a chronic disease. Few patients, if any, outgrow their illness. As such, a number of clinical management issues and basic research questions have emerged. Although the First International Gastrointestinal Eosinophil Researcher Symposium provided much in the way of clinical advice and new knowledge, it also raised important questions that will take time to answer:

Etiology and pathogenesis

Maintenance treatment strategies

Endpoints for treatment

Natural history and complications

Optimal method to identify food antigens/aeroallergens

Methods to differentiate eosinophilia associated with GERD and EE

Noninvasive technologies to assess disease activity

Evaluation and management of asymptomatic patient with esophageal eosinophilia

The most pressing clinical issues focus on the natural history of the disease and the necessity and determination of appropriate maintenance treatments. These areas are undeniably linked; if patients are free from long-term morbidity, then the importance of chronic treatments is diminished. This scenario does not appear to be the case for most patients who experience a waxing and waning course that can affect daily life in an untoward manner. As more physicians recognize the chronicity of the disease, a larger body of experience dictates that complications and long-term problems may arise. Although strictures and narrow-caliber esophagus are the best-known sequelae in adults, EE is certainly associated with a less well-defined but clinically important impact on a child's eating behavior.

Along these same lines, the best treatment endpoint has yet to be determined. To date, research protocols have focused primarily on tissue eosinophilia as an endpoint because many clinicians are somewhat reluctant to leave the esophagus with any degree of inflammation. Using clinical phenotypes (allergic vs nonallergic) along with noninvasive or mucosal biomarker profiles (ie, eotaxin-3, IL-5) may hold promise for future identification of patients who will benefit, or not, from more targeted therapeutic intervention and monitoring. For example, the suggestion that there may be allergic and nonallergic phenotypes may guide future treatment into different approaches (nutritional vs medical). Each of these areas demands careful documentation and basic studies to know the mechanisms by which eosinophils affect the esophageal mucosa.

Eight years ago, Dr David Landon made the following observation: My suspicion is that some EEs evolve into the fibrotic variety. Only one of these has reversed on fluticasone. The problem would seem to be that once fibrosis becomes truly established, the disease seems irreversible. Whether other therapy, or a more basic attack, would change this outcome, I don't know.' As collaborative basic and clinical research projects continue to address these unresolved problems, we will gain increasing clarity as to the best manner in which to provide immediate and long-term care for our patients.

Acknowledgments

We are grateful for the efficient and gracious administrative support of Margaret Stallings, Sandy Fasold, and Kim Rose.

Authors' Affiliations

Dr Liacouras is with the Division of Gastroenterology, Hepatology, and Nutrition, University of Pennsylvania School of Medicine, The Children's Hospital of Philadelphia; Dr Bonis is with Tufts University School of Medicine; Drs Putnam and Mishra are with the Department of Pediatrics, Cincinnati Children's Hospital Medical Center, University of Cincinnati College of Medicine; Dr Straumann is with the Department of Gastroenterology, Kantonsspital Olten, Switzerland; Dr Ruchelli is with the Department of Pathology, University of Pennsylvania School of Medicine, The Children's Hospital of Philadelphia; Dr Gupta is with the Department of Clinical Pediatrics, Indiana University School of Medicine; Dr Lee is with the Department of Biochemistry and Molecular Biology and Division of Pulmonary Medicine, Mayo Clinic (Scottsdale); Dr Hogan is with the Division of Allergy and Immunology, Cincinnati Children's Hospital Medical Center, University of Cincinnati College of Medicine; Dr Wershil is with the Department of Pediatrics, Albert Einstein College of Medicine, and the Division of Pediatric GI and Nutrition, The Children's Hospital at Montefiore; Dr Rothenberg is with the the Division of Allergy and Immunology, Cincinnati Children's Hospital Medical Center, University of Cincinnati College of Medicine; Dr Ackerman is with the Department of Biochemistry and Molecular Genetics, University of Illinois at Chicago College of Medicine; Dr Gomes is with the Department of Pediatrics, University of Illinois at Chicago College of Medicine; Dr Murch is with the Department of Paediatrics and Child Health, Warwick Medical School; and Dr Furuta is with the Department of Pediatrics, Division of Pediatric Gastroenterology, Hepatology, and Nutrition, University of Colorado Health Science Center.