P004 Untargeted serum metabolome shows Crohn Disease (CD) associated signature, signals that correlate with disease activity and CRP, and signals that may precede disease flare. (30th January 2023)

• Record Type: Journal Article
• Authors: Levhar, N; Hadar, R; Braun, T; Efroni, G; Abramovich, I; Gottlieb, E; Granot, M; Sandra, N; Selinger, L; Picard, O; Yavzori, M; Lahat, A; Eliakim, R; Weiss, B; Kopylov, U; Ben-Horin, S; Amir, A; Haberman Ziv, Y
• Is Part Of: Journal of Crohn's and colitis. Volume 17(2023)Supplement 1
• Page Start: i173
• Page End: i173
• Publication Date: 2023-01-30
• Subjects: Inflammatory bowel diseases -- Periodicals; 616.344005
• DOI: 10.1093/ecco-jcc/jjac190.0134
• Languages: English
• ISSNs: 1873-9946

Abstract

Abstract: Background: The serum metabolome contains potential disease biomarkers, some of which are produced endogenously, derived from the diet, or formed by the gut microbiome. We aimed to characterize metabolomics profiles linked with Crohn Disease (CD), with active CD, and with remitting pre-flare CD. Methods: Serum untargeted metabolomics was performed using liquid chromatography Mass Spectrometer (LC/MS). PERMANOVA on beta diversity was used to define factors affecting serum metabolome. Multivariate analyses in MaAsLin2 were applied to identify differentially abundant metabolites between CD and controls, between active CD and those in remission, and between those that subsequently flared and those that remained in remission within 24 months of follow-up. Results: The cohort included 202 samples from 53 CD patients and 31 healthy controls (156 samples from CD and 46 from controls; Table 1). 545 metabolites were included in our predefined library and were subsequently analyzed. PERMANOVA on the serum metabolomic data indicated that fasting, age, gender, disease, and subject are significant factors contributing to metabolomics variations. Using multivariate analyses and controlling for fasting, age, gender, and subject, we identified 34 specific metabolites that were significantly higher in CD and 26 that were significantly higher in Controls (FDR<0.25; Fig. 1). CD samples were significantly enriched for purine derivatives, simple sugars, and histidine, compared to healthy controls. In contrast, amino-acids, caffeine, and TCA cycle related metabolites were consistently downregulated in CD. 37 metabolites were higher in active CD and 24 in quiescent CD. Saturated fatty acids and acetylated amino-acids were positively correlated with active CD, while histidine and other standard amino-acids were downregulated in active samples compared to remission. Interestingly, acetylated amino acids, asparagine and phenylalanine, were also significantly correlated with CRP. A sub-cohort of 25 CD subjects was enrolled during remission and monitored every 3 months over 2 years. During that prospective follow-up, 11 flared. Focusing on the pre-flare samples, patients who eventually flared showed a significant reduction in N-acetyl glycine and reduction in two ketogenic derivates, acetoacetate and 3-hydroxybutyric acid, in comparison to patients who remained in remission throughout (FDR<0.15). Conclusion: CD, CD activity, and a subsequent flare in CD patients with longitudinal data were associated with distinct serum metabolic signature after controlling for confounders like age, gender, and fasting. Additional validation is needed to test if these can be used as predictive biomarkers.