Molecular analysis of duodenal eosinophilia.

BACKGROUND: Eosinophilic duodenitis (EoD), characterized by nonspecific gastrointestinal symptoms and increased numbers of duodenal eosinophils, may be in the eosinophilic gastrointestinal disease spectrum. However, diagnostic thresholds and pathogenic processes of duodenal tissue eosinophilia are inadequately characterized. OBJECTIVE: We aimed to define an EoD transcriptome and pathologic pathways. METHODS: RNA sequencing and histologic features of human duodenal biopsy samples were analyzed as a function of duodenal eosinophils levels. For analyses, we defined EoD as more than 52 peak eosinophils/hpf (n = 8), duodenal eosinophilia as 30 to 52 eosinophils/hpf (n = 11), and normal controls as fewer than 30 eosinophils/hpf (n = 8). Associations between gene expression and histologic features were analyzed with Spearman correlation. RESULTS: We identified 382 differentially expressed genes (EoD transcriptome) between EoD and normal controls (>2-fold change [adjusted P < .05]). The EoD transcriptome distinguished EoD from controls (duodenal eosinophilia and normal controls). The duodenal eosinophil count was correlated with a distinct EoD transcriptome when 50 to 60 peak eosinophils/hpf were present. The EoD transcriptome was enriched in genes involved in IL-4/IL-13 signaling, mast cells, and myeloid progenitor cells. Among duodenal histologic features, lamina propria eosinophil sheets was the most associated with transcriptomic changes (r = 0.66; P < .01). EoD gene signatures were shared with eosinophilic esophagitis and eosinophilic gastritis but not with eosinophilic colitis or celiac disease. CONCLUSION: We have identified an EoD transcriptomic signature that emerges at 50 to 60 peak eosinophils/hpf and established EoD as part of a spectrum of upper eosinophilic gastrointestinal disorder associated with type 2 immunity and distinct from eosinophilic colitis and celiac disease. These findings provide a basis for improving diagnosis and treatment.

Harnessing artificial intelligence to infer novel spatial biomarkers for the diagnosis of eosinophilic esophagitis.

Eosinophilic esophagitis (EoE) is a chronic allergic inflammatory condition of the esophagus associated with elevated esophageal eosinophils. Second only to gastroesophageal reflux disease, EoE is one of the leading causes of chronic refractory dysphagia in adults and children. EoE is a clinicopathologic disorder and the histological portion of the diagnosis requires enumerating the density of esophageal eosinophils in esophageal biopsies, and evaluating additional features such as basal zone hyperplasia is helpful. However, this task requires time-consuming, somewhat subjective manual analysis, thus reducing the ability to process the complex tissue structure and infer its relationship with the patient's clinical status. Previous artificial intelligence (AI) approaches that aimed to improve histology-based diagnosis focused on recapitulating identification and quantification of the area of maximal eosinophil density, the gold standard manual metric for determining EoE disease activity. However, this metric does not account for the distribution of eosinophils or other histological features, over the whole slide image. Here, we developed an artificial intelligence platform that infers local and spatial biomarkers based on semantic segmentation of intact eosinophils and basal zone distributions. Besides the maximal density of eosinophils [referred to as Peak Eosinophil Count (PEC)] and a maximal basal zone fraction, we identify the value of two additional metrics that reflect the distribution of eosinophils and basal zone fractions. This approach enables a decision support system that predicts EoE activity and potentially classifies the histological severity of EoE patients. We utilized a cohort that includes 1,066 biopsy slides from 400 subjects to validate the system's performance and achieved a histological severity classification accuracy of 86.70%, sensitivity of 84.50%, and specificity of 90.09%. Our approach highlights the importance of systematically analyzing the distribution of biopsy features over the entire slide and paves the way toward a personalized decision support system that will assist not only in counting cells but can also potentially improve diagnosis and provide treatment prediction.

A Deep Multi-Label Segmentation Network For Eosinophilic Esophagitis Whole Slide Biopsy Diagnostics.

Eosinophilic esophagitis (EoE) is an allergic inflammatory condition of the esophagus associated with elevated numbers of eosinophils. Disease diagnosis and monitoring require determining the concentration of eosinophils in esophageal biopsies, a time-consuming, tedious and somewhat subjective task currently performed by pathologists. Here, we developed a machine learning pipeline to identify, quantitate and diagnose EoE patients' at the whole slide image level. We propose a platform that combines multi-label segmentation deep network decision support system with dynamics convolution that is able to process whole biopsy slide. Our network is able to segment both intact and not-intact eosinophils with a mean intersection over union (mIoU) of 0.93. This segmentation enables the local quantification of intact eosinophils with a mean absolute error of 0.611 eosinophils. We examined a cohort of 1066 whole slide images from 400 patients derived from multiple institutions. Using this set, our model achieved a global accuracy of 94.75%, sensitivity of 94.13%, and specificity of 95.25% in reporting EoE disease activity. Our work provides state-of-the-art performances on the largest EoE cohort to date, and successfully addresses two of the main challenges in EoE diagnostics and digital pathology, the need to detect several types of small features simultaneously, and the ability to analyze whole slides efficiently. Our results pave the way for an automated diagnosis of EoE and can be utilized for other conditions with similar challenges.

Evaluating Eosinophilic Colitis as a Unique Disease Using Colonic Molecular Profiles: A Multi-Site Study.

BACKGROUND & AIMS: Colonic eosinophilia, an enigmatic finding often referred to as eosinophilic colitis (EoC), is a poorly understood condition. Whether EoC is a distinct disease or a colonic manifestation of eosinophilic gastrointestinal diseases (EGIDs) or inflammatory bowel disease (IBD) is undetermined. METHODS: Subjects with EoC (n = 27) and controls (normal [NL, n = 20], Crohn's disease [CD, n = 14]) were enrolled across sites associated with the Consortium of Eosinophilic Gastrointestinal Disease Researchers. EoC was diagnosed as colonic eosinophilia (ascending ≥100, descending ≥85, sigmoid ≥65 eosinophils/high-power field) with related symptoms. Colon biopsies were subjected to RNA sequencing. Associations between gene expression and histologic features were analyzed with Spearman correlation; operational pathways and cellular constituents were computationally derived. RESULTS: We identified 987 differentially expressed genes (EoC transcriptome) between EoC and NL (>1.5-fold change, P < .05). Colonic eosinophil count correlated with 31% of EoC transcriptome, most notably with CCL11 and CLC (r = 0.78 and 0.77, P < .0001). Among EoC and other EGIDs, there was minimal transcriptomic overlap and minimal evidence of a strong allergic type 2 immune response in EoC compared with other EGIDs. Decreased cell cycle and increased apoptosis in EoC compared with NL were identified by functional enrichment analysis and immunostaining using Ki-67 and cleaved caspase-3. Pericryptal circumferential eosinophil collars were associated with the EoC transcriptome (P < .001). EoC transcriptome-based scores were reversible with disease remission and differentiated EoC from IBD, even after controlling for colonic eosinophil levels (P < .0001). CONCLUSIONS: We established EoC transcriptomic profiles, identified mechanistic pathways, and integrated findings with parallel IBD and EGID data. These findings establish EoC as a distinct disease compared with other EGIDs and IBD, thereby providing a basis for improving diagnosis and treatment.

Loss of Endothelial TSPAN12 Promotes Fibrostenotic Eosinophilic Esophagitis via Endothelial Cell-Fibroblast Crosstalk.

BACKGROUND & AIMS: Eosinophilic esophagitis (EoE) can progress to fibrostenosis by unclear mechanisms. Herein, we investigated gene dysregulation in fibrostenotic EoE, its association with clinical parameters and specific pathways, and the functional consequences. METHODS: Esophageal biopsies from subjects with EoE were collected across 11 Consortium of Eosinophilic Gastrointestinal Disease Researchers sites (n = 311) and 2 independent replication cohorts (n = 83). Inclusion criteria for fibrostenotic EoE were endoscopic rings, stricture, and/or a history of dilation. Endoscopic, histologic, and molecular features were assessed by the EoE Endoscopic Reference Score, EoE Histology Scoring System, EoE Diagnostic Panel, and RNA sequencing. Esophageal endothelial TSPAN12 expression and functional effects on barrier integrity and gene expression were analyzed in vitro. RESULTS: TSPAN12 was the gene most correlated with fibrostenosis (r = -0.40, P < .001). TSPAN12 was lower in fibrostenotic EoE and correlated with EoE Endoscopic Reference Score, EoE Diagnostic Panel, and EoE Histology Scoring System (r = 0.34-0.47, P < .001). Lower TSPAN12 associated with smaller esophageal diameter (r = 0.44, P = .03), increased lamina propria fibrosis (r = -0.41, P < .001), and genes enriched in cell cycle-related pathways. Interleukin (IL)-13 reduced TSPAN12 expression in endothelial cells. Conversely, anti-IL-13 therapy increased TSPAN12 expression. TSPAN12 gene silencing increased endothelial cell permeability and dysregulated genes associated with extracellular matrix pathways. Endothelial cell-fibroblast crosstalk induced extracellular matrix changes relevant to esophageal remodeling. CONCLUSIONS: Patients with fibrostenotic EoE express decreased levels of endothelial TSPAN12. We propose that IL-13 decreases TSPAN12, likely contributing to the chronicity of EoE by promoting tissue remodeling through fibroblast-endothelial cell crosstalk.

Desmoplakin and periplakin genetically and functionally contribute to eosinophilic esophagitis.

Eosinophilic esophagitis (EoE) is a chronic allergic inflammatory disease with a complex underlying genetic etiology. Herein, we conduct whole-exome sequencing of a multigeneration EoE pedigree (discovery set) and 61 additional multiplex families with EoE (replication set). A series of rare, heterozygous, missense variants are identified in the genes encoding the desmosome-associated proteins DSP and PPL in 21% of the multiplex families. Esophageal biopsies from patients with these variants retain dilated intercellular spaces and decrease DSP and PPL expression even during disease remission. These variants affect barrier integrity, cell motility and RhoGTPase activity in esophageal epithelial cells and have increased susceptibility to calpain-14-mediated degradation. An acquired loss of esophageal DSP and PPL is present in non-familial EoE. Taken together, herein, we uncover a pathogenic role for desmosomal dysfunction in EoE, providing a deeper mechanistic understanding of tissue-specific allergic responses.

Machine Learning Approach for Biopsy-Based Identification of Eosinophilic Esophagitis Reveals Importance of Global features.

GOAL: Eosinophilic esophagitis (EoE) is an allergic inflammatory condition characterized by eosinophil accumulation in the esophageal mucosa. EoE diagnosis includes a manual assessment of eosinophil levels in mucosal biopsies-a time-consuming, laborious task that is difficult to standardize. One of the main challenges in automating this process, like many other biopsy-based diagnostics, is detecting features that are small relative to the size of the biopsy. RESULTS: In this work, we utilized hematoxylin- and eosin-stained slides from esophageal biopsies from patients with active EoE and control subjects to develop a platform based on a deep convolutional neural network (DCNN) that can classify esophageal biopsies with an accuracy of 85%, sensitivity of 82.5%, and specificity of 87%. Moreover, by combining several downscaling and cropping strategies, we show that some of the features contributing to the correct classification are global rather than specific, local features. CONCLUSIONS: We report the ability of artificial intelligence to identify EoE using computer vision analysis of esophageal biopsy slides. Further, the DCNN features associated with EoE are based on not only local eosinophils but also global histologic changes. Our approach can be used for other conditions that rely on biopsy-based histologic diagnostics.

Broad transcriptional response of the human esophageal epithelium to proton pump inhibitors.

BACKGROUND: Proton pump inhibitors (PPIs) have been recognized as a primary treatment of eosinophilic esophagitis (EoE), an allergic inflammatory disease of the esophageal mucosa. The mechanisms underlying esophageal epithelial responses to PPIs remain poorly understood. OBJECTIVE: We hypothesized that PPIs can counteract IL-13-mediated esophageal epithelial responses that are germane for EoE pathogenesis. METHODS: Transcriptional responses of human esophageal cells to IL-13 and the PPIs omeprazole and esomeprazole were assessed by RT-PCR and RNA sequencing. Cytokine secretion was measured by multiplex analysis and ELISA. RESULTS: Human esophageal epithelial cells robustly responded to PPI stimulation by inducing a set of 479 core genes common between omeprazole and esomeprazole treatments. The transcriptional response to PPIs was partially mediated through the aryl hydrocarbon receptor signaling pathway, as the aryl hydrocarbon receptor antagonist GNF-351 modified approximately 200 genes, particularly those enriched in metabolic processes and regulation of cell death. PPI treatment reversed approximately 20% of the IL-13 transcriptome. Functional analysis of the PPI-responsive, upregulated genes revealed enrichment in metabolic and oxidation processes, and the unfolded protein response. In contrast, downregulated genes were overrepresented in functional terms related to cell division and cytoskeletal organization, which were also enriched for the genes in the EoE transcriptome reversed by PPIs. Furthermore, PPI treatment decreased the IL-13-induced proliferative response of esophageal epithelial cells. CONCLUSIONS: These results demonstrate broad effects of PPIs on esophageal epithelium, including their ability to curtail transcriptomic processes involved in cellular proliferation and IL-13-induced responses, and they highlight the importance of AHR signaling in mediating these responses.

Functional role of kallikrein 5 and proteinase-activated receptor 2 in eosinophilic esophagitis.

Eosinophilic esophagitis (EoE) is a chronic, food antigen-driven, inflammatory disease of the esophagus and is associated with impaired barrier function. Evidence is emerging that loss of esophageal expression of the serine peptidase inhibitor, kazal type 7 (SPINK7), is an upstream event in EoE pathogenesis. Here, we provide evidence that loss of SPINK7 mediates its pro-EoE effects via kallikrein 5 (KLK5) and its substrate, protease-activated receptor 2 (PAR2). Overexpression of KLK5 in differentiated esophageal epithelial cells recapitulated the effect of SPINK7 gene silencing, including barrier impairment and loss of desmoglein-1 expression. Conversely, KLK5 deficiency attenuated allergen-induced esophageal protease activity, modified commensal microbiome composition, and attenuated eosinophilia in a murine model of EoE. Inhibition of PAR2 blunted the cytokine production associated with loss of SPINK7 in epithelial cells and attenuated the allergen-induced esophageal eosinophilia in vivo. Clinical samples substantiated dysregulated PAR2 expression in the esophagus of patients with EoE, and delivery of the clinically approved drug α1 antitrypsin (A1AT, a protease inhibitor) inhibited experimental EoE. These findings demonstrate a role for the balance between KLK5 and protease inhibitors in the esophagus and highlight EoE as a protease-mediated disease. We suggest that antagonizing KLK5 and/or PAR2 has potential to be therapeutic for EoE.

Molecular, endoscopic, histologic, and circulating biomarker-based diagnosis of eosinophilic gastritis: Multi-site study.

BACKGROUND: Eosinophilic gastritis (EG) is a clinicopathologic disorder with marked gastric eosinophilia and clinical symptoms. There is an unmet need among patients with EG for more precise diagnostic tools. OBJECTIVE: We aimed to develop tissue- and blood-based diagnostic platforms for EG. METHODS: Patients with EG and control subjects without EG were enrolled across 9 Consortium of Eosinophilic Gastrointestinal Disease Researchers-associated sites. An EG Diagnostic Panel (EGDP; gastric transcript subset) and EG blood biomarker panel (protein multiplex array) were analyzed. EGDP18 scores were derived from the expression of 18 highly dysregulated genes, and blood EG scores were derived from dysregulated cytokine/chemokine levels. RESULTS: Gastric biopsy specimens and blood samples from 185 subjects (patients with EG, n = 74; control subjects without EG, n = 111) were analyzed. The EGDP (1) identified patients with active EG (P < .0001, area under the curve ≥ 0.95), (2) effectively monitored disease activity in longitudinal samples (P = .0078), (3) highly correlated in same-patient samples (antrum vs body, r = 0.85, P < .0001), and (4) inversely correlated with gastric peak eosinophil levels (r = -0.83, P < .0001), periglandular circumferential collars (r = -0.73, P < .0001), and endoscopic nodularity (r = -0.45, P < .0001). For blood-based platforms, eotaxin-3, thymus and activation-regulated chemokine, IL-5, and thymic stromal lymphopoietin levels were significantly increased. Blood EG scores (1) distinguished patients with EG from control subjects without EG (P < .0001, area under the curve ≥ 0.91), (2) correlated with gastric eosinophil levels (plasma: r = 0.72, P = .0002; serum: r = 0.54, P = .0015), and (3) inversely correlated with EGDP18 scores (plasma: r = -0.64, P = .0015; serum: r = -0.46, P = .0084). Plasma eotaxin-3 levels strongly associated with gastric CCL26 expression (r = 0.81, P < .0001). CONCLUSION: We developed tissue- and blood-based platforms for assessment of EG and uncovered robust associations between specific gastric molecular profiles and histologic and endoscopic features, providing insight and clinical readiness tools for this emerging rare disease.