DOP53 In-depth characterisation of the serum antibody epitope repertoire in Inflammatory Bowel Disease by high-throughput phage-displayed immunoprecipitation sequencing

• Published in: Journal of Crohn's and colitis Vol. 16; no. Supplement_1; pp. i100 - i102
• Main Authors: Bourgonje, A R, Andreu-Sánchez, S, Vogl, T, Hu, S, Vich Vila, A, Leviatan, S, Kurilshikov, A, Klompus, S, Kalka, I N, van Dullemen, H M, Weinberger, A, Visschedijk, M C, Festen, E A M, Faber, K N, Wijmenga, C, Dijkstra, G, Segal, E, Fu, J, Zhernakova, A, Weersma, R K
• Format: Journal Article
• Language: English
• Published: 21-01-2022
• ISSN: 1873-9946; 1876-4479
• DOI: 10.1093/ecco-jcc/jjab232.092

Abstract
Background
Patients with IBD show distinct antibody responses, particularly against microbiota. However, a comprehensive overview of the antibody epitope repertoire in IBD is lacking. Here, we characterized serum antibody responses in patients with IBD and population controls using a high-throughput phage-displayed immunoprecipitation sequencing (PhIP-seq) workflow and associated these to disease phenotypes and the faecal microbiome.
Methods
PhIP-seq was leveraged to characterise antibody responses against 344,000 rationally selected peptide antigens in 497 patients with IBD which were compared with 1,326 individuals from a population-based cohort (Fig. 1A-B). Antibody profiles were linked to 23 IBD-specific clinical features such as disease location and surgical history and to faecal microbiota composition (Fig. 1C).
Results
Patients with IBD demonstrated distinct antibody epitope repertoires compared with individuals from the general population, with 373 differentially abundant antibody-bound peptides (202 overrepresented, 171 underrepresented) belonging to bacterial flagellins (69), virulence factors (102), other antigens of both commensal and pathogenic bacteria (90) as well as viruses (67) and food proteins (24) (Figure 2). In particular, antibody responses against bacterial flagellins, many of which belong to Lachnospiraceae bacteria (e.g. Roseburia spp.), but also Eubacterium spp. and pathogens (e.g. Legionella, Clostridium, Burkholderia) dominated in patients with Crohn’s disease (CD), and were associated with ileal disease involvement and more complicated disease behaviour (e.g. fibrostenotic disease, surgical history) as well as anti-Saccharomyces cerevisiae antibody positivity. Furthermore, many other antigens were newly identified, e.g. decreased responses to E. coli virulence factors and genome polyproteins of enteroviruses, and increased responses to food antigens (wheat, barley) and autoantigens (particularly collagen type I and VI). Antibody epitope repertoires were able to accurately discriminate CD from population controls (area under the curve [AUC]=0.88, test set evaluation), showing very high discriminative performance (positive and negative predictive value of 72% and 93%, respectively, representing predicted classes in test set) (Fig. 3A-C), which was less accurate for ulcerative colitis (UC) (Fig. 3D-F).
Conclusion
This study demonstrates the size, diversity and complexity of systemic antibody epitope repertoires in patients with IBD compared to controls, showing that distinct clinical phenotypes of IBD are characterized by unique antibody signatures. PhIP-seq is a powerful tool for identifying systemic immune-based biomarkers and exposing novel immunological targets in immune-mediated inflammatory diseases like IBD.