Molecular analysis of duodenal eosinophilia

Abstract

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.

Keywords: Duodenitis; celiac disease; eosinophil; eosinophilic duodenitis; transcriptome.

FIG 1.

A conserved pattern of gene expression from duodenal tissue of subjects with EoD. A, Clustering analysis based on 382 differentially expressed genes (EoD transcriptome). Heatmap (upregulated [red] vs downregulated [blue]) of expression profiles of differentially dysregulated genes between normal controls (NLs) and subjects with EoD (FDR P < .05; ≥1.5-fold change). Each column represents an individual subject (NL, n = 8; with EoD, n = 8). B, Duodenal transcriptome data on NLs (yellow) and subjects with EoD (red) reduced to 3-dimensional presentation by multidimensional scaling analysis for visual presentation of the expression distance between samples. A and B, Arrows and bold circles represent samples having at least 30 eosinophils/hpf in at least 3 hpf. C, The number of genes that exhibit the indicated minimum fold difference in subjects with EoD versus in NLs subjects is graphed (left). Genes with 5-fold or greater change (n = 18) are listed (right).

FIG 2.

The EoD transcriptome is associated with DE. A, Peak duodenal eosinophil count and duodenal CCL26 and CLC correlation. B, Individual EoD transcriptome expression and peak duodenal eosinophil count (subjects numbered 1–27 by increasing eosinophilia [upper panel]); dashed lines represent thresholds (30 eosinophils/hpf [black] vs 52 eosinophils/hpf[red]). Heatmap (red, upregulated; blue, downregulated) of 382 differentially expressed genes (EoD transcriptome [lower panel]); columns represent individual subjects (normal control [NL], n = 8; with DE, n = 11; with EoD, n = 8) matched vertically to upper panel. C, NL (yellow), DE (blue), and EoD (red) duodenal transcriptome data as a 3-dimensional presentation. B and C, Arrows and bold circles represent samples having at least 30 eosinophils/hpf in at least 3 hpf. D, Conserved duodenal cytokine gene expression pattern of EoD versus NL and/or DE. E, EoD score generation schematic summary distinguishing those with EoD versus NLs and quantifying EoD disease severity. F, Peak duodenal eosinophil count (left) and EoD score (right). G, Peak duodenal eosinophil count and EoD score correlation. A-D and F-G, Circles represent individual subjects. D and F, Mean ± SEM. P values determined by using the Kruskal-Wallis test followed by the Dunn multiple-comparison test.*P < .05; **P < .01; ***P < .001; ****P < .0001. TPM, Transcripts per kilobase million.

FIG 3.

Functions and cell types enriched in the EoD transcriptome. A and B, Functional annotation enrichment analyses of 256 upregulated (left) and 126 downregulated (right) genes of the EoD transcriptome using CluGO overview charts (A) and showing the 7 most significant terms in biologic process by ToppGene (for the full list, see Table E3) (B). C, Specific increase in gene expression–estimated proportion of cell types in EoD. Data presented as means ± SEMs. Each circle represents an individual subject or control. P values determined by using the Kruskal-Wallis test followed by the Dunn multiple-comparison test. *P < .05; **P < .01. CMP, Common myeloid progenitor; GMP, granulocyte-monocyte progenitor; MEP, megakaryocyte-erythrocyte progenitor; NL, normal control.

FIG 4.

Duodenal histologic features and associations with duodenal transcripts. A, Representative hematoxylin and eosin–stained duodenal EoD biopsy specimens (original magnification, ×20). B, Histologic feature clustering in duodenal biopsy samples with features arranged to ensure that members of the same cluster are adjacent in the correlation plot and in the same order as in the cluster members. Color map shows correlations among histologic features, with darker red shades indicating stronger positive correlations. C, Comparison of histologic features among the normal control (NL), DE, and EoD groups. Data are means ± SEMs. Each circle represents an individual subject or control. P values determined by using Kruskal-Wallis test followed by Dunn multiple-comparison test. D, Hierarchic relationships between duodenal histologic features from EoD transcriptome gene expression profile correlations. Spearman r–based heat diagram for gene-level correlations (top left). Darker shades indicate stronger correlations, red for positive and blue for negative. Associations between the EoD score and individual histologic features (top right). Red coloration indicates statistical significance at P < .05 with a Spearman r correlation. The top 10 genes significantly associated with each histologic feature in EoD (bottom). *P < .05; **P < .01; ***P < .001.

FIG 5.

Comparison of the EoD transcriptome with other EGID transcriptomes. A, Venn diagram analysis of shared genes for upregulated (left) and downregulated (right) expression among EGID transcriptomes (EoD, EoE, EoG, and EoC). B, The 50 most upregulated (left) and downregulated (right) transcripts in the EoD transcriptome relative to the normal control (NL) are given with the corresponding relative fold change in expression in tissue from subjects with EoE, EoG, and EoC relative to the NLs. C, Enrichment analysis of tissue-specific genes in the tissue-matched EGID transcriptome (ie, esophagus-specific genes in EoE transcriptome). Each pie chart indicates the proportion of tissue-specific genes in the upregulated and downregulated portions of each tissue-matched transcriptome. P values determined by using the Fisher exact test. D, Comparisons of gene expression by quantitative RT-PCR in samples from subjects with active EGID (esophagus (EoE, n = 147; NL, n = 48), stomach (EoG, n = 13; NL, n = 33), duodenum (EoD, n = 8; NL, n = 8), and colon (EoC, n = 12; NL, n = 16]) relative to the NL. Data are means 6 SEMs. P values determined by using the Mann-Whitney U test. *P < .05.

FIG 6.

Comparison of the EoD transcriptome with celiac disease gene expression. A, Venn diagram analysis of shared genes for upregulated (left) and downregulated (right) expression between EoD and celiac disease [EoD from this study; celiac disease from 403 genes differentially expressed in at least 2 of 3 recent RNAseq transcriptomic studies comparing active celiac disease and controls (PMID reference numbers: 31700112, 30097691, and 30998704)]. B, Comparison of top 20 gene ontology terms related to downregulated expression between EoD and celiac disease. Red bar indicates common terms between EoD and celiac disease.