Rationale: RNA sequencing studies have previously identified impaired barrier function and the subsequent proinflammatory response as critical to the pathogenies of EoE. However, the investigation into disease initiation is often limited by the inability to capture the disease early. Elimination diets and food reintroductions provide a unique situation to investigate potential early manifestations of EoE. We propose that the recurrence of EoE disease following food reintroductions serves as a surrogate model to study the early processes occurring with antigen re-exposure.

Methods: Ten patients demonstrated recurrent EoE in a patchy endoscopic and histologic manner during trigger food reintroduction, with areas of normal appearing tissue (low eosinophils, <13 eos/HPF) interspersed with regions of exudates/furrows (high eosinophils, ≥15 eos/HPF). Exome RNA sequencing was performed on formalin-fixed, paraffin-embedded esophageal tissue from these areas of high and low eosinophilia. Comparisons were made to esophageal tissues from remission EoE (n=5) and non-EoE control (n=22).

Results: Recurrent EoE (high and low eosinophil tissues) demonstrated transcriptional changes in epithelial and eosinophil genes hallmark to EoE (POSTN, CCL26, CDH26, DSG-1) compared to controls. The recurrent high eosinophil tissue demonstrated alterations in genes associated with tissue remodeling (keratins, ANO1, GATA6), eosinophil infiltrations (IL13, IL5, and IL5RA), and loss of barrier function (SPINK7). Transcriptional changes paralleled severe epithelial injury in these tissues. Looking at the five patients with paired tissue collected during remission tissue and recurrent disease, the progression of EoE is seen transcriptionally. Elevated CCL5 RNA levels differentiated recurring low eosinophil tissue from paired remission. Factors implicated in the regulation, recruitment, and degranulation of eosinophils (GATA2, IL1RL1, IL36G), neutrophils (CXCL6, CXCL8), B-cells (SIGLEC6), and mast cells (SIGLEC6, ADGRE2, HDC, TPSB2) were elevated in the recurrent high eosinophil tissue compared to low eosinophil tissue.

Conclusions: Investigation of patients in whom recurrence of EoE occurs in a patchy manner identifies transcriptional changes involved in the evolution of EoE. Recurrent low eosinophil tissue demonstrated cytokine signaling, suggesting involvement in disease pathogenesis before the onset of robust eosinophilia infiltration. In tissues with elevated eosinophils during recurrent EoE, upregulated inflammatory pathways were accompanied by a dramatic shift of innate and adaptive immune cell activity, substantiating the importance of antigen hypersensitivity as a driver of disease. A food antigen re-exposure model can elucidate early immunological and epithelial changes that likely play a role in the recurrence of EoE.

<b>Figure 1. A)</b> Study design of tissue collection and sequencing performed on the following: (1) baseline remission tissue of elimination diet where the trigger is removed (n=5, ≤10 eos/HPF), (2) patchy disease recurrence (n=10) following food reintroduction. These specimens were categorized into (2a) “recurring high eos tissue” with active disease (≥15 eos/HPF) and (2b) “recurring low eos tissue” (<13 eos/HPF). (3) non-EoE controls (n=22). <b>B)</b> Recurring high eos tissue had greater EoE Histology Scoring System (EoEHSS) grade and stage compared to low eos tissue (p<0.01). Specifically, high eos tissue had more severe and extensive basal zone hyperplasia and dilated intercellular spaces.

Figure 1. A) Study design of tissue collection and sequencing performed on the following: (1) baseline remission tissue of elimination diet where the trigger is removed (n=5, ≤10 eos/HPF), (2) patchy disease recurrence (n=10) following food reintroduction. These specimens were categorized into (2a) “recurring high eos tissue” with active disease (≥15 eos/HPF) and (2b) “recurring low eos tissue” (<13 eos/HPF). (3) non-EoE controls (n=22). B) Recurring high eos tissue had greater EoE Histology Scoring System (EoEHSS) grade and stage compared to low eos tissue (p<0.01). Specifically, high eos tissue had more severe and extensive basal zone hyperplasia and dilated intercellular spaces.

Figure 2. <b>A)</b> Showing differentially expressed genes (DEGs) of potential interest for paired patient category comparison, recurring high eosinophil tissue compared to recurring low eosinophil tissue (fold-change >2, adj p-value <0.05, total 1098 DEGs) and recurring low eosinophil tissue compared to baseline remission tissue (fold-change >1.5, adj p-value <0.05, total 120 DEGs). <b>B)</b> Hallmark GSEA pathway heatmap of paired patient categories: baseline remission and recurring high eosinophil and low eosinophil tissue (n=5), showing normalized enrichment scores for pathways with FDR 10%.

Figure 2. A) Showing differentially expressed genes (DEGs) of potential interest for paired patient category comparison, recurring high eosinophil tissue compared to recurring low eosinophil tissue (fold-change >2, adj p-value <0.05, total 1098 DEGs) and recurring low eosinophil tissue compared to baseline remission tissue (fold-change >1.5, adj p-value <0.05, total 120 DEGs). B) Hallmark GSEA pathway heatmap of paired patient categories: baseline remission and recurring high eosinophil and low eosinophil tissue (n=5), showing normalized enrichment scores for pathways with FDR 10%.

Mark Hazel

Appearances