Eosinophil progenitor levels are increased in active pediatric eosinophilic esophagitis

To the Editor

Eosinophilic esophagitis (EoE) is a chronic, food antigen–driven gastrointestinal disease characterized by marked esophageal eosinophilia. CCL26 (eotaxin-3) levels mediate eosinophil tissue recruitment [1, 2], and eosinophil tissue counts serve as the basis for diagnosis [3]. Currently, EoE treatment efficacy is often determined via invasive serial endoscopies with esophageal biopsies; thus, there is a compelling need to develop novel and less-invasive blood-based biomarkers to evaluate EoE disease activity. Though peripheral blood eosinophil counts moderately correlate with esophageal tissue eosinophil levels, their relationship is not sufficient to serve as a clinically useful biomarker [4]. Similar to the mature eosinophil, the lineage-committed eosinophil progenitor (EoP) is mobilized during allergic disease [5]. EoP frequency is increased in the peripheral blood of adult patients with asthma in response to allergen challenge [6, 7]. Also, EoP levels are increased in the lung tissue and sputum of allergen-challenged patients with asthma, suggesting a direct contribution to tissue eosinophilia via in situ differentiation after migration from the peripheral blood [6, 8]. Notably, peripheral blood EoP levels correlate with the severity of asthma in adults, advancing the concept of the EoP as a clinically useful biomarker [6]. No clinical studies to date have measured EoP levels in EoE. We hypothesized that EoP levels would be increased in the peripheral blood of pediatric patients with active EoE disease and would correlate sufficiently with the level of mature eosinophils in the esophagus to serve as a relevant biomarker for EoE disease activity.

To test our hypothesis, peripheral blood samples for EoP enumeration were obtained from consented pediatric patients (Table 1) undergoing endoscopy for evaluation of EoE at the Cincinnati Center for Eosinophilic Disorders (CCED, Cincinnati Children’s Hospital Medical Center [CCHMC]). Included subjects were ages 1-18 years and had a diagnosis of EoE. Patients met the consensus criteria for EoE and failed proton pump inhibitor (PPI) therapy, which serves as part of the diagnostic criteria of EoE [3]. Patients were excluded if other inflammatory gastrointestinal diseases, such as Crohn’s disease or other eosinophilic gastrointestinal diseases, were present or if they had used systemic steroids, such as prednisone or other immune modulators, in the previous two months. Demographic and medical record information for each participant was maintained in a CCED research database and reviewed to obtain relevant patient information. Peripheral blood mononuclear cell (PBMC) isolation and flow staining were performed as detailed in the online repository. EoPs were identified as a subpopulation of live PBMCs using the gating strategy in Figure E1 in the online repository. Subsequently, patients were divided into two groups on the basis of their esophageal eosinophil levels, active EoE (n = 17; median age 14.0 [interquartile range 5.5-16.5] years, peak esophageal eosinophils ≥15 eosinophils/high-power microscopic field [hpf], median 52.0 [interquartile range, 36.0-86.0] eosinophils/hpf) and inactive EoE (n = 14; median age 7.5 [3.0-12.5] years, peak esophageal eosinophils <15 eosinophils/hpf, median 0.5 [0.0-3.0] eosinophils/hpf). Patient age between the inactive and active disease groups was not significantly different (p = 0.08), and median EoP levels did not vary by age (Figure E2 in the online repository).

Table I.

Patient Characteristics

Sex	Agea (years)	Atopic Phenotype b	Therapyc	PPIb	Peak Esophageal Eosinophil Count (per hpf)	Blood Eosinophil Progenitor Count (per mL blood)
Inactive EoE (n = 14)
F	8	AR+ AD− As−	Food Trials	Yes	0	0
F	18	AR+ AD+ As+	Food Trials	Yes	0	6
F	9	AR+ AD− As−	Food Trials	Yes	0	11
M	3	AR+ AD+ As+	Food Trials	Yes	0	22
M	3	AR− AD− As−	Food Trials	Yes	1	18
F	6	AR+ AD+ As+	Food Trials	Yes	3	10
M	2	AR+ AD+ As−	Food Trials	Yes	2	4
M	3	AR− AD− As−	Food Trials	Yes	0	15
F	3	AR− AD+ As−	Food Trials	No	0	31
M	11	AR+ AD− As−	Food Trials	Yes	3	12
M	14	AR− AD− As−	Food Trials	Yes	1	42
F	12	AR+ AD+ As−	Food Trials	Yes	0	41
F	7	AR+ AD+ As+	Food Trials	No	4	30
M	15	AR+ AD− As−	Food Trials and Steroids	Yes	8	0
Active EoE (n = 17)
M	3	AR+ AD+ As−	Food Trial	Yes	16	16
M	15	AR− AD− As−	Food Trial	No	56	35
M	7	AR+ AD+ As+	Food Trial	Yes	110	27
F	4	AR− AD+ As+	Food Trial	Yes	23	52
F	6	AR+ AD+ As+	Food Trial	No	42	17
M	17	AR+ AD− As−	Unrestricted Diet	Yes	56	45
F	17	AR+ AD+ As−	Steroids	Yes	49	18
M	5	AR− AD+ As−	Unrestricted Diet	Yes	219	81
F	10	AR+ AD− As−	Unrestricted Diet	No	36	46
M	14	AR+ AD+ As+	Food Trial + Steroids	Yes	89	21
M	5	AR+ AD− As−	Food Trial	Yes	34	20
M	11	AR+ AD− As−	Unrestricted Diet	Yes	36	41
M	17	AR+AD− As+	Unrestricted Diet	No	83	87
M	17	AR− AD− As−	Unrestricted Diet	Yes	120	44
F	16	AR+ AD− As−	Food Trial	Yes	81	63
F	16	AR− AD− As−	Unrestricted Diet	Yes	45	37
M	16	AR+ AD− As−	Unrestricted Diet	No	52	37

^aAge at time of endoscopy

^bPer medical records; AR, allergic rhinitis; AD, atopic dermatitis; As, asthma; PPI, proton pump inhibitor

^cTherapy at time of endoscopy

First, we ascertained whether there were differences in blood EoP levels in patients with active EoE disease versus inactive disease. The median absolute EoP level per mL of blood in patients with active EoE was increased more than two fold over the EoP level in patients with inactive EoE (37.0 [interquartile range, 20.5-49.0] versus 13.5 [5.5-30.3] EoPs per mL of blood, p < 0.001 Mann-Whitney, Figure 1A). Absolute EoP levels also significantly correlated with the esophageal eosinophil level (Spearman r = 0.56, p = 0.0011, Figure 1B). We next investigated the ability of EoP levels to serve as a biomarker for EoE disease activity. A receiver operator curve showed an area under the curve of 0.84 (95% CI, 0.70 to 0.98, p = 0.0015, Figure 1C). A cutoff value of less than 15.5 absolute EoPs per mL accurately excluded active disease in 100% of patients. Concurrently, levels equal to or greater than 15.5 absolute EoPs per mL reliably predicted active disease in 74% of patients (sensitivity = 100%, specificity = 57%, PPV = 74%, NPV = 100%, LR⁺ = 2.3, LR⁻ = 0). Using leave-one-out cross-validation and a threshold of 15.5, we found that the mean specificity was 57 ± 3 % while retaining 100% sensitivity. These data support the utility of EoP as an excellent disease activity biomarker in patients with EoE.

Figure 1.

A. Median absolute EoP levels in patients with active versus inactive EoE are shown. (For all box plots, Box = 25-75^th percentile, whiskers = minimum to maximum values, all points shown, each point represents an individual patient) B. Median absolute EoPs correlate with peak esophageal eosinophils (Spearman). C. Receiver operating characteristic analysis of the ability of EoP levels to discriminate between patients with and without active EoE. The solid diagonal line indicates values that have no discriminatory value. D. Comparison of median EoP levels in active EoE divided into two groups, restricted diet (food trials) versus unrestricted diet, is shown. E. Median values of peak esophageal eosinophils in active EoE divided into two groups, restricted diet (food trials) versus unrestricted diet, are shown.

As peripheral blood EoP levels performed well as a disease activity biomarker, we assessed whether differences in EoP levels existed between different therapeutic groups. Both swallowed steroids and elimination diets, in which disease-triggering foods are avoided, are effective treatments in many pediatric patients with EoE [3, 9]. Most of our patient cohort was on dietary therapy at the time of their endoscopy (Table 1). To investigate the effect of dietary restrictions on blood EoP levels, we subdivided the patients with active EoE disease into two groups, unrestricted diet or restricted diet, at the time of endoscopy. We observed that the median absolute EoP level was significantly increased amongst patients with active EoE on an unrestricted diet versus those on a restricted diet (44.5 [interquartile range, 38.0-72.3] versus 24.0 [17.8-47.8] EoPs per mL of blood, p = 0.04, Figure 1D). Interestingly, esophageal eosinophils levels were similar between patients with active EoE on restricted diet and those on an unrestricted diet (p = 0.4, Figure 1E), suggesting that EoP levels are sensitive to dietary restrictions in patients with active EoE disease. Future studies, including serial measurements in patients, are needed to establish the relationship between EoP levels and EoE therapies.

As EoP levels are elevated in the peripheral blood of adult patients with atopic diseases [5], we investigated whether a history of other allergic diseases was associated with increased blood EoP levels in our population. A clinical history of an atopic disorder (atopic dermatitis, allergic rhinitis, or asthma) did not correlate with EoP elevations and esophageal disease activity (Figure E3 in the online repository). Additionally, peripheral blood eosinophil counts (n = 10) did not correlate with peripheral blood EoP levels (Spearman r = 0.16, p = 0.65, Figure E4 in the online repository). We acknowledge that larger studies to evaluate the affects of atopy are needed.

This pilot study suggests that EoP levels do show promise as a disease-monitoring biomarker in patients previously diagnosed with EoE. The current disease monitoring standard dictates that all patients undergo invasive endoscopy with tissue biopsies due to poor correlation between clinical symptoms and tissue inflammation [10]. Any reduction of invasive procedures through the use of less-invasive blood biomarkers would be valuable. Thus, in our study, 8/14 patients (57%) would have avoided endoscopy due to correctly classified inactive disease. Such a biomarker of “disease inactivity” would be clinically beneficial for this population, as it may alleviate many of the burdens associated with repeated invasive esophageal biopsies.

Capsule summary: Eosinophil progenitor (EoP) levels are increased in the blood of pediatric patients with active eosinophilic esophagitis (EoE). Notably, blood EoP levels show promise as a novel, blood-based biomarker to monitor EoE disease activity.

Abbreviations

EoP

eosinophil lineage–committed progenitor

EoE

eosinophilic esophagitis

hpf

high-power microscopic field

PBMCs

peripheral blood mononuclear cells

PPI

proton pump inhibitor