Lamina Propria Phagocyte Profiling Reveals Targetable Signaling Pathways in Refractory Inflammatory Bowel Disease

Abstract

Background and aims: Lamina propria phagocytes are key mediators of inflammatory bowel disease (IBD). We aimed to understand the transcriptomic and functional differences in these cells based on location, disease type, inflammation state, and medication use in patients with IBD.

Methods: Phagocytic immune cells in the lamina propria, as defined by the marker CD11b, were isolated from 54 unique patients (n = 111 gut mucosal biopsies). We performed flow cytometry for cell phenotyping (n = 30) and RNA sequencing with differential gene expression analysis (n = 58). We further cultured these cells in vitro and exposed them to janus kinase inhibitors to measure cytokine output (n = 27). Finally, we matched patient genomic data to our RNA sequencing data to perform candidate gene expression quantitative trait locus analysis (n = 34).

Results: We found distinct differences in gene expression between CD11b⁺ cells from the colon vs ileum, as well as in different inflammatory states and, to a lesser degree, IBD types (Crohn's disease or ulcerative colitis). These genes mapped to targetable immune pathways and metabolic and cancer pathways. We further explored the janus kinase-signal transducer and activator of transcription pathway, which was upregulated across many comparisons including in biopsies from anti-tumor necrosis factor refractory patients. We found that isolated CD11b⁺ cells treated with janus kinase inhibitors had decreased secretion of cytokines tumor necrosis factora and interleukin-8 (P ≤ .05). We also found 3 genetic variants acting as expression quantitative trait loci (P ≤ .1) within our CD11b⁺ data set.

Conclusions: Lamina propria phagocytes from IBD mucosa provide pathogenetic clues on the nature of treatment refractoriness and inform new targets for therapy.

Keywords: Anti-TNF; IBD; JAK-STAT; Phagocytes; RNA Sequencing.

Figure 1

Characterization of CD11b⁺-enriched cells shows an increased overall cell count and granulocyte percentage in inflamed biopsies. (A) After magnetic column sorting, lamina propria CD11b⁺-enriched cells from histologically uninflamed (n = 54) and inflamed biopsies (n = 33) were counted via a hemocytometer. Uninflamed biopsies yielded an average of approximately 400,000 cells and inflamed biopsies, approximately 750,000 cells (Mann-Whitney U-test P = .0043). (B) Flow cytometry gating of CD11b⁺ cells in a representative unenriched sample (all lamina propria cells) and a CD11b⁺-enriched sample, after gating on live, single, CD45⁺ cells. (C) CD11b⁺-enriched cells from histologically uninflamed (n = 20) and inflamed biopsies (n = 10) were phenotyped via flow cytometry. Mean percentages of each cell type are shown.

Figure 2

CD11b⁺ cell transcriptional profiles show largest differences based on location. (A) Venn diagram of the number of significant (P ≤ .05) DE genes in CD11b⁺ cells across 3 global comparisons: colon vs ileum, UC vs CD, and inflamed vs uninflamed. The one differentially expressed gene common among all comparisons (center) was NTS. (B) Principal component analysis (PCA) of all samples (n = 58) based on the transcriptional profiles showed separation primarily by location. Ileum samples showed additional grouping by inflammation and diagnosis.

Figure 3

CD11b⁺ cell differential gene expression analysis across multiple sample comparisons. CD11b⁺ cells showed differential gene expression between (A) all samples, ileum vs colon (n = 58), (B) inflamed colon, CD vs UC (n = 15), (C) uninflamed ileum, CD vs UC (n = 21), (D) colon, uninflamed vs inflamed (n = 32), (E) ileum, uninflamed vs inflamed (n = 26), (F) anti-TNF colon, uninflamed vs inflamed (n = 8), and (G) anti-TNF ileum, uninflamed vs inflamed (n = 8). Colored genes were significantly differentially expressed (P ≤ .05).

Figure 4

Targetable signaling pathways are upregulated in inflamed and anti-TNF refractory samples. (A) Heatmap of enriched pathways from IPA sorted by the pathway activation z-score for the colon vs the ileum with the border bolded for significant (adjusted P ≤ .05) enrichment. The title of each column lists the 2 compared groups. A positive (red) z-score indicates pathway upregulation in the first group, whereas a negative (blue) z-score indicates pathway upregulation in the second group. Gray indicates undirected pathway enrichment (the z-score could not be calculated). (B) Simplified upstream regulatory network of the STAT3 pathway in anti-TNF–treated ileum samples. The red color intensity of each gene correlates with a positive z-score (upregulation) in inflamed anti-TNF–treated ileum samples.

Figure 5

JAK inhibitors decrease inflammatory cytokine secretion from lamina propria CD11b⁺-enriched cell cultures. Lamina propria CD11b⁺-enriched cells were cultured in the presence of a JAK inhibitor, ruxolitinib or tofacitinib, and stimulated with LPS for 24 hours. Supernatant cytokine levels were measured via Luminex. Cytokine fold change for each condition was compared with unstimulated, untreated baseline average. Mean and 95% confidence intervals plus P-values from relevant 2-way ANOVA comparisons are shown (P ≤ .05 or ns = not significant). (A) TNF-α (n = 22), (B) CXCL8/IL-8 (n = 18), (C) IL-6 (n = 22), and (D) IL-1β (n = 18). ANOVA, analysis of variance.