DOP072Fecal host transcriptomics: a novel tool to monitor disease activity in patients with inflammatory bowel disease

Background The intestinal mucosa undergoes rapid renewal, with cells continuously shed into the gut lumen. Recent findings indicate that these cells remain viable and can provide valuable information about bowel pathology. Analysis of endoscopic washes obtained during colonoscopies indicates that bulk RNA sequencing of fecal material accurately reflects histologic inflammation in patients with inflammatory bowel disease (IBD). Whether similar signals can be obtained from stool samples remains unknown. We aimed to determine whether stool RNA sequencing can reliably indicate the presenece, severity and localization of intestinal inflammation in patients with IBD. Methods Stool samples from pediatric and adult patients with IBD and from healthy controls were subjected to bulk RNA sequencing after microbial RNA depletion. Inflammation and location scores were generated based on differential expression, dimensionality reduction, and bowel-transcriptomics knowledge. Correlations were then assessed between fecal host transcriptomics (FHT) and presenece, severity and location of inflammation, determined by endoscopy. Results Stool samples were collected from 141 subjects, including 92 patients with Crohn’s disease, 42 patients with ulcerative colitis and 7 healthy controls. The quality of human RNA isolated from fecal content was high, with a median of 3,440 (IQR 2,213-5,807) genes detected per sample. As a proof-of-concept, the correlation between S100A8/A100A9 (calprotectin) RNA and calprotectin protein levels in 107 stool samples was strong (R = 0.72, Figure 1A). Unsupervised hierarchical clustering of 200 genes in the FHT dataset separated samples according to endoscopic activity (Figure 1B), indicating that clinical disease activity is the predominant source of variation in the dataset. Based on a set of several dozen transcripts, an inflammatory score was generated that was highly accurate in distinguishing IBD patients with active disease vs. controls and IBD patients in endoscopic remission (AUC 0.97 compared to 0.72 of fecal calprotectin. Figure 1C-D). Moreover, FHT showed better correlation with SES-CD among patients with Crohn’s disease (R = 0.73), compared with fecal calprotectin (R = 0.64, Figure 1E-F). Finally, a location score was able to differentiate IBD patients with small-bowel restricted disease vs. those with colonic involvement (Figure 1G-H). Conclusion FHT is a promising novel non-invasive tool to measure inflammatory activity in patients with IBD and accurately defines presence of inflammation, severity and location. Extensive validation work is on-going to determine its ability to support therapeutic decisions. Conflict of interest: Segal, Arik: No conflict of interest Weintraub, Yael: No conflict of interest Matar, Manar: No conflict of interest Nevo- Shor, Anat: No conflict of interest Tal-Shifman, Noa: No conflict of interest Kantor, Irena: Lecturing fee - Takeda SAB - Tracells Bahar Halpern, Keren: No conflict of interest Yehuda-Cohen, Avishag: No conflict of interest Refael, Tal: No conflict of interest Rokach, May: No conflict of interest Jana, Tamar: No conflict of interest Shamir, Raanan: No conflict of interest Itzkovitz, Shalev: No conflict of interest Kahan-Hanum, Maya: No conflict of interest Harnik, Yotam: No conflict of interest Shouval, Dror: Lecturing fees - Takeda SAB - Tracells