Food Antigen Consumption and Disease Activity Affect Food-Specific IgG4 Levels in Patients with Eosinophilic Esophagitis (EoE)

Abstract

Introduction:

High levels of serum food-specific IgG4 (sIgG4) have been reported in patients with EoE. The objective of this study was to examine whether serum sIgG4 levels to foods and aeroallergens are higher in EoE patients than allergic controls and to investigate the association between sIgG4 and EoE clinical characteristics.

Methods:

This was a case-control study nested in a prospective EoE Cohort. EoE cases were defined per consensus guidelines, and controls were individuals with symptoms who were confirmed to be EoE-negative on upper endoscopy. Demographic and clinical information was prospectively collected. Serum IgE and sIgG4 were measured to foods and aeroallergens by ImmunoCAP. Mean levels of sIgG4 were compared between cases and controls, and logistic regression models were used to examine predictors of elevated milk sIgG4 levels.

Results:

The analysis included 123 individuals (EoE n=93, control n=30) with a similar distribution of allergic disease between EoE patients and controls (86% v. 93%; p=0.30). EoE patients had significantly higher sIgG4 levels to all allergens evaluated, with the exception of birch (p=0.24). Milk sIgG4 levels were independently associated with milk consumption (OR 4.95; p=0.01) and the presence of sIgE to milk (OR 4.23; p=0.008).

Conclusion:

Serum sIgG4 levels to food and aeroallergen proteins were higher in patients with EoE than non-EoE controls, and higher levels of milk sIgG4 were independently associated with milk consumption and the presence of sIgE to milk proteins. Whether sIgG4 plays a pathogenic role in EoE or could be used as an EoE biomarker remains unknown and warrants further study.

Introduction

Over past 20 years, eosinophilic esophagitis (EoE) has rapidly transformed from a case reportable disease to a major cause of upper gastrointestinal morbidity.(1) EoE is characterized by type 2 (T2) allergic inflammation, including tissue infiltration of eosinophils and mast cells, and increased levels of T2 cytokines.(2) In 2014, Clayton et al demonstrated that patients with EoE have higher levels of tissue and serum immunoglobulin G4 (IgG4) than controls.(3) Since that time, this finding has been replicated and extended in multiple studies. IgG4-containing plasma cells are found in the esophageal tissue of patients with EoE,(3) and IgG4 deposits can distinguish patients with EoE from gastroesophageal reflux disease (GERD).(4–6) Esophageal IgG4 levels also correlate with eosinophil counts and IL-10 expression in active disease(7) and decrease with disease remission.(8) As a result, there is now compelling evidence that IgG4 is a relevant feature of this disease.

Our group recently reported that serum levels of IgG4 to specific food proteins in milk (Bos d 4, Bos d 5, and Bos d 8) and wheat were higher in children with EoE than unselected controls.(9) This study was limited, however, by the fact that the control population was not confirmed to be EoE-negative and did not have a similar distribution of allergic disease as the cases. Furthermore, this study was limited to the pediatric population and did not assess sIgG4 levels to aeroallergens, which may drive EoE in certain individuals.(10) Finally, the association between serum sIgG4 to food proteins and clinical characteristics remains unknown. The objective of the present study was to examine whether serum sIgG4 levels to foods and aeroallergens are higher in patients with EoE than non-EoE controls and investigate the association between serum IgG4 levels and clinical characteristics.

Methods

Study Design

The parent EoE Cohort is an ongoing, prospective, observational study designed to investigate immunologic, genetic, and environmental determinants of EoE and eosinophil-associated gastrointestinal disorders (EGIDs). Starting in 2017, children and adults who are either seen in our institution’s multidisciplinary EoE/EGID clinics or undergoing esophago-gastroduodenoscopy (EGD) for evaluation of confirmed or suspected EoE/EGID are invited to enroll. At the enrollment visit, patients are interviewed with standardized questionnaires about the study subject’s demographics, medical history, family history, treatment history, diet, and clinical symptoms of EoE/EGID. At the time of each EGD, gastroenterologists perform the Eosinophilic Esophagitis Endoscopic Reference Score (EREFS), six esophageal biopsies are collected in the proximal, middle, and distal esophagus and evaluated for the presence of eosinophils, and peripheral blood is collected and stored in a −20°C freezer. At follow-up visits, patients are interviewed with a standardized questionnaire about their clinical symptoms, treatment, and current diet. Relevant health information, laboratory results, pathology results, endoscopy scores, treatment, and questionnaire data are recorded in a dedicated REDCap Database, which is supported by the integrated Translational Health Research Institute of Virginia (iTHRIV). The study was approved by the UVA Institutional Review Board, and written informed consent was obtained from all participants.

In order to examine the association between food-specific IgG4 levels and clinical characteristics, we performed a nested case-control study of individuals enrolled in a prospective EoE Cohort. EoE cases were defined, per consensus guidelines, by the following criteria: 1) symptoms of esophageal dysfunction; 2) the presence of ≥15 eosinophils/hpf (eos/hpf) on esophageal biopsy obtained via EGD and 3) exclusion of other potential causes of esophageal eosinophilia.(11) Patients who met these criteria but were histologically and clinically responsive to a PPI were included as cases of EoE, per the most recent consensus guidelines.(11) Only patients with active disease (≥15 eos/hpf) at the time of EGD and serum collection were included as cases (n=93). Controls were defined as individuals who underwent an EGD for symptoms suggestive of EoE but did not have endoscopic or histologic findings consistent with EoE (n=30). Among the patients with EoE, 12 individuals with paired samples during active disease (on a PPI only) and remission on swallowed steroids were also identified for longitudinal analyses.

Serum Food-Specific IgE and IgG4 Measurements

Allergen-specific IgE (sIgE) and IgG4 (sIgG4) to cow’s milk proteins (nBos d 4, nBos d 5, nBos d 8), wheat, egg, soy, cat, timothy grass, birch, ragweed, and dust mite were measured in stored sera using ImmunoCAP^® 250 (Thermo-Fisher/Phadia, Uppsala, Sweden), as previously described.(9) Component-based assays to Bos d 4, Bos d 5 and Bos d 8 were performed due to the potential for non-specific background binding when assaying IgG4 responses using cow’s milk extract. IgE results were expressed in international units per milliliter (IU/mL) where 1 IU = 2.4 ng, and the lower limit of detection was 0.10 IU/mL. IgG4 results were expressed in micrograms per milliliter (μg/mL), and the lower limit of detection was 0.07 μg/mL. Values below the lower limit of detection were defined as one-half of the lower limit of detection for both IgE and IgG4 analyses.

Covariates

Race and ethnicity were defined by self-report and was defined as Non-Hispanic White, Non-Hispanic Black, Hispanic, Asian, Multi-Racial, and Not Otherwise Characterized. For the purposes of analysis, the latter three categories were combined into “Grouped.” Asthma was defined as a positive response to the question, “Has a doctor ever diagnosed you [your child] with asthma?” Eczema was defined as a positive response to “Has a doctor ever diagnosed you [your child] with eczema?” Allergic rhinitis was defined as a positive response to the question, “Has a doctor every diagnosed you [your child] with allergic rhinitis (hay fever or seasonal allergies?).” Food allergy was defined as a positive response to the question, “Have you [your child] ever experienced abdominal pain, itching all over, trouble breathing, flushing, hives, or swelling of your face, throat, hands, or feet within 2 hours of eating something?” Milk consumption was assessed at the enrollment visit using the question, “Do you current eat foods with milk, including baked milk (i.e. baked goods, butter, yogurt, ice cream, cheese)?” Individuals who consumed milk were then asked, “How often do you consume foods with milk” with categorical answers ranging from “less than once a month” to “more than once a day.” Overall milk consumption was again assessed at the time of each EGD. Fibrotic features were assessed at the time of each EGD and the presence of fibrostenosis was defined as the presence of a) a stricture or b) evidence of fibrosis or stromal sclerosis on esophageal biopsy.

Statistical Analysis

Demographic and clinical characteristics were compared between cases and controls using chi-squared, student’s t tests, and Mann-Whitney U tests, as appropriate. The distributions of food and aeroallergen sIgG4 levels were examined by EoE outcomes, and as these values were not normally distributed, geometric mean levels of each sIgG4 were compared between cases and controls using Mann Whitney U tests. Univariate logistic regression models were used to examine predictors of elevated levels of milk sIgG4, which was defined a priori as ≥30 ug/mL for the sum of sIgG4 antibodies to Bos d 4, Bos d 5, and Bos d 8, based on our previous data regarding milk sIgG4 levels in patients with EoE.(9) We then adjusted for potential confounders, including age, sex, race/ethnicity, milk consumption, and milk sIgE using multivariable logistic regression. Longitudinal food sIgG4 levels in the same patient with active disease and remission were compared using the Wilcoxon matched pairs signed rank test. All analyses were performed in Stata SE 14.1 (College Station, TX) or GraphPad Prism (San Diego, CA).

Results

Study Population

A total of 123 individuals were included in this study, of whom 93 were confirmed to have active EoE, and 30 were controls. Compared to controls, patients with EoE were older (median 20 v. 13 years, p=0.03), more likely to be male (69% v. 43%, p=0.01), and less likely to be consuming milk (73% v. 93%, p=0.02). There was no significant difference in the prevalence of allergic disease between cases and controls (Table 1), though patients with EoE were more likely to have sIgE to milk (77% v. 33%; p<0.001). Patients with EoE were more likely to present with abnormal eating patterns (29% v. 10%; p=0.04) and dysphagia (70% v. 30%; p=0.008) and less likely to present with abdominal pain (18% v. 47%; p=0.002) and heartburn (22% v. 43%; p=0.02) than controls. Peripheral blood absolute eosinophil counts were significantly higher in patients with EoE than controls (median 300; IQR 160 – 510 cells/uL versus median 150; IQR 100–250 cells/uL; p<0.001).

Table 1:

Demographic and Clinical Characteristics of Study Population

	Controls (n = 30)	EoE (n = 93)	p value
Age †	13 (3 – 77)	20 (1 – 60)	0.03
Male Sex	13 (43)	63 (68)	0.02
Race/Ethnicity
Non-Hispanic White	24 (80)	82 (88)	0.12
Non-Hispanic Black	0 (0)	2 (2)
Hispanic	3 (10)	1 (1)
Grouped	3 (10)	8 (9)
Allergic History
Overall	27 (93)	79 (86)	0.30
Asthma	11 (38)	44 (47)	0.45
Allergic Rhinitis	24 (83)	66 (71)	0.44
Eczema	10 (34)	36 (39)	0.87
Food Allergy	10 (34)	43 (46)	0.52
Milk sensitization (+sIgE)	10 (33)	72 (77)	<0.001
Milk sIgE‡	0.07 (0.01 – 7.89)	0.32 (0.01 – 55.8)	<0.001
Gastrointestinal Symptoms
Feeding dysfunction	3 (10)	27 (29)	0.04
Poor growth	1 (3)	7 (8)	0.42
Abdominal pain	14 (47)	17 (18)	0.002
Vomiting	10 (33)	23 (25)	0.36
Chest pain	7 (23)	10 (11)	0.08
Heartburn	13 (43)	20 (22)	0.02
Dysphagia	9 (30)	65 (70)	<0.001
Food impaction	9 (30)	38 (41)	0.29
Treatment at EGD
PPI	5 (17)	49 (53)	0.001
Swallowed Steroids	0 (0)	21 (23)	0.004
Food Avoidance§	10 (33)	33 (35)	0.83
Milk Consumption	28 (93)	67 (73)	0.02
EoE Characteristics
EREFs†	0 (0 – 2)	3 (0 – 8)	<0.001
Fibrostenosis	0 (0)	21 (23)	0.004
Peak eosinophils†	0 (0 – 5)	30 (15 – 100)	<0.001
Serum AEC*	150 (40 – 720)	300 (0 – 2350)	<0.001

Values expressed as n (%) unless otherwise defined

^†Median (range)

^‡Geometric mean (range)

^§Control (n=10) patients were avoiding foods due to a history of food allergy.

^¶AEC: absolute eosinophil count (missing data n=16)

At the time of the EGD, 53% of patients with active EoE were on a PPI, 23% were on swallowed steroids, and 35% were on an elimination diet. Twenty-one patients with EoE (23%) had evidence of fibrostenosis and 10 (11%) were classified as PPI-responsive based on previous or subsequent EGDs demonstrating histologic and clinical remission solely on a PPI.

Allergen-Specific IgG4 Levels are Higher in Patients with EoE than non-EoE Controls

IgG4 levels to Bos d 4 (geometric mean 3.0 v. 0.25 ug/mL; p<0.0001), Bos d 5 (6.8 v. 1.0 ug/mL p<0.0001), Bos d 8 (7.0 v. 1.2 ug/mL; p<0.0001), wheat (9.2 v. 1.8; p<0.0001), egg (3.3 v. 1.4 ug/mL; p=0.046) and soy (0.7 v. 0.2 ug/mL; p=0.0002) were higher in patients with EoE than controls (see Figure 1 and eTable 1). Similarly, geometric mean levels were significantly higher to dust mite (geometric mean 0.37 v. 0.19 ug/mL; p= 0.0002), cat dander (0.47 v. 0.18 ug/mL; p=0.0002), timothy grass (0.35 v. 0.16 ug/mL; p=0.008) and ragweed (0.20 v. 0.09 ug/mL; p=0.002) in patients with EoE compared to controls (Figure 2). There was no difference in geometric mean levels of sIgG4 to birch pollen (0.20 v. 0.16 ug/mL; p=0.24). In general, sIgG4 levels to aeroallergens were an order of magnitude lower than sIgG4 levels to food.

Figure 1: Food sIgG4 Levels in Subjects with Active EoE and Controls.

Food-specific IgG4 levels among patients with active EoE (blue) and non-EoE controls (red). Geometric mean levels are represented by the solid line, and the number of individuals with undetectable levels is listed on the bottom of the graph for each assay. Asterisks represent p<0.001 (****) and p<0.05 (**).

Figure 2: Aeroallergen sIgG4 Levels in Subjects with Active EoE and Controls.

Aeroallergen-specific IgG4 levels among patients with active EoE (blue) and controls (red). Geometric mean levels are represented by the solid line, and the number of individuals with undetectable levels is listed on the bottom of the graph for each assay. Asterisks represent p<0.001 (****) and p<0.05 (**).

Milk Consumption and sIgE to Milk Are Associated with High Levels of Milk sIgG4

Among patients with EoE, the sum of sIgG4 levels to Bos d 4, 5, and 8 ranged from 0.1 to 1153 ug/mL (geometric mean 23.61; 95% CI 16.0 – 34.8 μg/mL). Clinical characteristics, including age, sex, race/ethnicity, allergic disease, presenting symptoms, and EREFS criteria were not associated with high-levels of sIgG4 to milk proteins, which was defined as >30 ug/mL to the sum of sIgG4 antibodies to Bos d 4, 5, and 8 (see eTable 2). In contrast, milk consumption and the presence of sIgE to milk proteins were significantly associated with high-levels of sIgG4 to milk proteins ([OR 4.15; 95% CI 1.47 – 11.7] and [OR 4.68; 95% CI 1.75 – 12.5], respectively). This relationship persisted in models adjusting for both milk consumption and the presence of milk sIgE, suggesting an independent effect of both variables on milk sIgG4 levels (Table 2).

Table 2:

Associations Between Clinical Characteristics and Milk sIgG4 >30 μg/mL

	Crude OR	p value	Adjusted OR	p value
Age < 18	1.21 (0.53 – 2.75)	0.65	1.32 (0.45 – 3.85)	0.62
Male Sex	1.55 (0.64 – 3.74)	0.33	2.00 (0.66 – 6.03)	0.22
Race/ethnicity	1.13 (0.80 – 1.58)	0.49	1.19 (0.83 – 1.72)	0.35
Atopic History	0.72 (0.22 – 2.33)	0.58	0.57 (0.14 – 2.35)	0.44
Fibrostenotic Disease	1.30 (0.49 – 3.44)	0.60	0.86 (0.28 – 2.70)	0.80
Peak eos/hpf	0.99 (0.96 – 1.02)	0.56	0.98 (0.95 – 1.01)	0.20
Milk consumption	4.15 (1.47 – 11.7)	0.007	5.03 (1.51 – 16.8)	0.009
Positive Milk sIgE	4.68 (1.75 – 12.5)	0.002	4.83 (1.61 – 14.6)	0.005

We further examined the relationship between IgE and IgG4 for food-specific proteins using stratified analyses. Among patients with EoE, those with sIgE to milk or wheat had higher mean sIgG4 levels to the same protein (Table 3). This relationship was also seen for egg and soy, but it did not reach statistical significance. Interestingly, among controls, this same relationship was seen for wheat proteins; however, there was no difference in mean sIgG4 levels for milk protein components by sIgE status, with the exception of Bos d 8, which was higher in controls without IgE.

Table 3:

Food sIgG4 Levels stratified by Food sIgE to the same allergen in EoE v. Controls

	Control (n=30)			EoE (n=93)
Food sIgG4	(−) sIgE	(+) sIgE	p value	(−) sIgE	(+) sIgE	p value
Milk
Bos d 4	0.3 (0.1 – 0.9)	0.2 (0.04 – 0.9)	0.36	1.3 (0.7 – 2.5)	5.8 (2.7 – 12.4)	0.002
Bos d 5	1.0 (0.4 – 2.7)	1.1 (0.3 – 3.8)	0.78	3.1 (1.7 – 5.7)	12.3 (7.4 – 20.4)	0.0004
Bos d 8	2.3 (1.2 – 4.2)	0.6 (0.2 – 1.7)	0.02	3.9 (2.1 – 7.1)	10.8 (6.6 – 17.7)	0.01
Sum	4.8 (2.3 – 9.8)	2.8 (0.9 – 8.7)	0.45	9.1 (4.5 – 18.4)	37.2 (24.2 – 57.2)	0.0005
Wheat	0.9 (0.5 – 1.6)	5.0 (2.9 – 8.8)	0.0003	2.2 (0.9 – 5.3)	13.3 (9.0 – 19.8)	0.0003
Egg	0.9 (0.3 – 2.5)	3.6 (1.1 – 11.6)	0.07	2.2 (1.2 – 4.1)	4.6 (2.8 – 7.4)	0.06
Soy	0.2 (0.1 – 0.3)	0.3 (0.1 – 0.9)	0.51	0.6 (0.3 – 0.9)	0.9 (0.6 – 1.4)	0.12

Values are expressed as geometric mean (95% CI)

Similarly, we examined the association between milk consumption and milk sIgG4 levels using stratified analyses. Among patients with EoE, mean sIgG4 levels to milk protein components were higher in those consuming milk than those avoiding milk (geometric mean 35.9 μg/mL [95% CI 23.9 – 54.0] vs. 9.2 μg/mL [95% CI 4.3 – 20.1]; p=0.001, Figure 3). There was no difference in milk sIgG4 levels by milk consumption in controls, but this was limited by the small number of control patients who were avoiding milk (n=2). Among those consuming milk, sIgG4 levels to milk protein components were higher in patients with EoE than controls (geometric mean 35.9 μg/mL [95% CI 23.9 – 54.0] vs. 3.9 μg/mL [95% CI 2.0 – 7.8]; p<0.001).

Figure 3: Cumulative Milk sIgG4 Levels Stratified by Milk Consumption.

Milk-specific IgG4 levels (sum of Bos d 4, 5, and 8) among non-EoE control patients (red) and active EoE (blue), stratified by milk consumption. Geometric mean levels are represented by the solid line.

Food sIgG4 Levels Decrease with EoE Disease Remission

In order to assess whether serum sIgG4 levels are associated with EoE disease activity, we measured food sIgG4 levels in 12 adult patients with paired samples during active disease and remission. All 12 patients with active disease were initially treated with a proton pump inhibitor alone, and they subsequently achieved disease remission (<15 eos/hpf) with a swallowed steroid (see eTable 3). There was no recommendation to change consumption of potential trigger foods (milk, wheat, egg, or soy) between the two endoscopies, all patients were actively consuming milk at both EGDs, and no patients were on an elimination diet for treatment of their EoE. Mean levels of sIgG4 decreased with disease remission to milk protein components (geometric mean 33.0 μg/mL [95% CI 13.3 – 82.2] vs. 18.3 μg/mL [95% CI 6.6 – 50.6]; p=0.001), wheat (10.4 μg/mL [95% CI 5.2 – 20.7] vs. 6.0 μg/mL [95% CI 2.5 – 14.2]; p=0.006), and soy (1.0 μg/mL [95% CI 0.5 – 2.1] vs. 0.6 μg/mL [95% CI 0.3 – 1.5]; p=0.02). There was no change in egg sIgG4 levels (Figure 4).

Figure 4: Paired Food sIgG4 Measurements in Active Disease and Remission.

Food-specific IgG4 levels in paired samples collected from EoE patients with active disease and remission. Remission in these patients was achieved by adding a swallowed steroid, and patients were instructed to not change consumption of potential trigger foods (milk, wheat, egg, or soy) between the two endoscopies.

Discussion

In this study, we found that serum sIgG4 levels to food and aeroallergen proteins were significantly higher in patients with EoE than controls with a similar distribution of allergic disease. And, while we did not see any association between milk sIgG4 levels and demographic or clinical characteristics among patients with EoE, higher levels were independently associated with the presence of sIgE to milk proteins and milk consumption. Although previous studies have demonstrated that IgG4 is a relevant feature of EoE,(3–9) this study is the first to demonstrate that these elevated food sIgG4 levels are not merely due to the presence of allergic disease. Furthermore, our findings suggest that while milk consumption may be a driver of the milk sIgG4 response in all individuals, this response is more robust in individuals with EoE, perhaps related to the concomitant presence of sIgE to milk.

In allergic disease, IgG4 is conventionally thought to be a marker of immune tolerance. This paradigm is supported by (i) higher levels of venom sIgG4 in bee-keepers who develop tolerance to stings,(12) (ii) elevations in cat sIgG4 in children living in a house with a cat,(13) and (iii) increases in allergen sIgG4 levels during immunotherapy.(14–16) IgG4 also has unique biological functions among the IgG subclasses to support this paradigm. First, due to weak interactions at the core hinge domain, IgG4 can undergo Fab arm exchange.(17) This process renders IgG4 functionally monovalent and limits its ability to cross-link antigens.(18, 19) IgG4 also has a limited ability to bind C1q and activate complement, and it has a strong affinity for the inhibitory receptor, FcγRIIb.(18) IgG4 is therefore often regarded as a marker of tolerance in individuals with allergic disease.

In this study, however, our control population had a similar prevalence of other allergic conditions as our EoE patients, and we found that serum levels of allergen-specific IgG4 were higher to both foods and inhalants in those with EoE. This finding suggests that there are unique aspects of the inflammation of EoE that drive the production of IgG4, which are not seen in other allergic conditions, including traditional IgE-mediated food allergy.(20) IgG4 is known to be secreted in response to the combination of Th2 cytokines (such as IL-4 and IL-13), IL-10, and chronic antigen exposure. While Th2 inflammation is characteristic of other allergic conditions including atopic dermatitis,(21) asthma,(22) allergic rhinitis,(23) and IgE-mediated food allergy,(23) IL-10 is not a central player in classic allergic inflammation. IL-10, which is produced by regulatory T (Treg) cells,(24) regulatory B cells,(25) macrophages,(26, 27) and eosinophils,(28, 29) is important for maintaining tolerance to exogenous and endogenous antigens. Studies have now demonstrated the presence of IL-10(7, 30) and Treg-like cells(31–33) in the esophageal tissue of patients with active EoE. While an increase in Treg cells is seen in patients developing natural tolerance(34) or undergoing immunotherapy for food allergy(35, 36) and allergic rhinitis,(37) this is otherwise not characteristic of allergic-type inflammation. Whether sIgG4 could therefore be used as a biomarker to diagnose and monitor EoE remains unclear and warrants further study.

We further found that while sIgG4 levels were elevated in patients with EoE to both food and aeroallergens, the sIgG4 levels to aeroallergens were an order of magnitude lower than those to food. We hypothesize that this is related to the quantity of allergen exposure. For example, in the United States, individuals are exposed to median concentration of 2 ug/g dust mite in the house.(38) In contrast, the average American consumes 13 g of cow’s milk-related protein per day.(39) In addition, the majority of aeroallergen exposure occurs through the nasal passages and lungs, and it is likely that a smaller fraction is exposed to the esophagus. That being said, it has been demonstrated that the inflammation in EoE may be triggered or exacerbated by aeroallergen exposure,(10, 40, 41) and dust mite antigen has been identified within the esophageal tissue of patients with EoE.(42) Therefore, this level of exposure may be sufficient in select individuals to induce inflammation.

While serum levels of sIgG4 are higher in patients with EoE than allergic controls, the role of IgG4 in the allergen-induced inflammation of EoE remains unknown. IgG4 deposits and IgG4-containing plasma cells are seen in the esophageal tissue of patients with active disease.(3) Esophageal IgG4 levels correlate with eosinophil counts and IL-10 expression in active disease(7) and decrease with disease remission.(8) Interestingly, recent studies have shown that IgG4 and gliadin proteins co-localize in patients with active EoE, but this is not seen in controls.(43) While IgG4 may be a biomarker of the underlying inflammatory response or increased gut permeability in EoE, it is also possible that it is playing a pathogenic role through the binding of food antigen. Further prospective studies examining the role of IgG4 in the food-induced response of EoE are clearly needed.

This study is limited by the heterogeneity of the EoE population, which was comprised of patients who were on various medication and dietary therapies for EoE at the time of assessment. As we would expect these interventions to decrease sIgG4 levels in EoE patients, it is possible that the true difference in geometric mean sIgG4 levels between cases and controls is actually larger than we observed in this study. In addition, the allergic control population was defined as individuals who were undergoing evaluation for EoE but were found not to have this disease. Therefore, they had a high prevalence of GI symptoms and may not be representative of the general outpatient allergy population. In addition, few of these control patients were avoiding milk, which limits our ability to make conclusions about the role of milk consumption on sIgG4 levels in non-EoE patients. However, control patients consuming milk had significantly lower milk sIgG4 levels than EoE patients consuming milk, which supports our conclusions that this exaggerated sIgG4 response is unique to patients with EoE.

Conclusion

In conclusion, we found that serum sIgG4 levels to food and aeroallergen proteins were significantly higher in patients with EoE than controls with a similar distribution of allergic disease. Furthermore, higher levels of milk sIgG4 were independently associated with milk consumption and the presence of sIgE to milk proteins. Whether sIgG4 plays a pathogenic role in EoE or could be used as an EoE biomarker remains unknown and clearly warrants further study.

Key Messages.

• Serum food and aeroallergen-specific IgG4 levels were higher in EoE patients than allergic controls.

• Higher milk-specific IgG4 levels were associated with milk consumption and sIgE to milk proteins.

• Whether food-specific IgG4 could be used as a biomarker in EoE remains unknown.

Abbreviations:

IgG4

Immunoglobulin G4

IgE

Immunoglobulin E

EoE

Eosinophilic esophagitis

eos/hpf

Eosinophils per high-powered field

EGD

Esophagogastroduodenoscopy

Data availability statement:

The data that support the findings of this study are available from the corresponding author upon reasonable request.