DOP017. Dysbiosis in ulcerative colitis: adding the spatial component
A. Lavelle1,2, G. Lennon1,2, A. Balfe1,2, N. Docherty2, J. Hyland1, D. O'Donoghue1, G. Docherty1, O. O'Sullivan3, P. Cotter3, K. Sheahan1, H. Mulcahy1, F. Shanahan3, J.C. Coffey4, D. Winter1,2, P.R. O'Connell1,2, 1St. Vincent's University Hospital, Centre for Colorectal Disease, Dublin, Ireland, 2University College Dublin (UCD), School of Medicine and Medical Science, Dublin, Ireland, 3University College Cork, Alimentary Pharmabiotic Centre, Cork, Ireland, 4Graduate Entry Medical School, University Hospital Limerick, University of Limerick Ireland, Limerick, Ireland
High-throughput sequencing has afforded unprecedented insights into the diversity and function of the colonic microbiota in health and disease. There are however, few data relating to the spatial structuring of the colonic microbiota and how inflammation disrupts the ecological landscape. The aim of this study was to elucidate the biogeography of the healthy microbiota using deep sequencing and to perform a comparative analysis in ulcerative colitis (UC).
4 healthy volunteers undergoing colonoscopy and 4 patients with acute UC undergoing proctocolectomy were sampled at four colorectal levels (caecum, transverse colon, descending colon and rectum). A fifth UC patient was sampled at the caecum and rectum only. Luminal brush samples, whole mucosal biopsies and laser capture microdissected mucus gel layer samples were retrieved at each level. 454 barcoded pyrosequencing of extracted microbial DNA was performed to characterize the microbiota and paired histological specimens were collected to grade regional inflammation at each site.
97 samples were sequenced, producing 3,069,310 reads, of which 3,041,773 reads were assigned to the bacterial kingdom. Samples from patients with UC had significantly lower levels of Bacteroideacea and Akkermanciaceae and significantly higher levels of Ruminococcaceae, Clostridiaceae and Bifidobacteriaceae. Inter-individual variability was the dominant variable, account for 55% of calculated total variance in a between class analysis (BCA) of the combined cohort at the family level. This was similarly evident in a Canonical Analysis of Principal components, using the unweighted UniFrac metric and incorporating inflammation and categorical spatial variables. In health, segregation between the mucus gel layer and the luminal communities was apparent. While this was also apparent in individuals with UC, a BCA suggested that this segregation was less defined. Despite these differences, there was no consistent correlation between local colonic inflammation and bacterial taxa within individuals with UC, suggesting that the dysbiosis in ulcerative colitis is a field effect and not directly correlated with the degree of regional inflammation.
Chronic inflammation in this UC cohort caused a characteristic dysbiosis which modulated an individual's signature microbiota without causing a definitive taxonomic convergence across individuals when compared to health. This underscores the importance of historical contingency and inter-individual variability in the UC microbiota. While segregation between luminal and mucosal communities remains evident in UC, it is less defined than in health.