P827 The clinical phenotype of collagenous colitis is associated with T-cell-related genetic variants
G. Roda1, D. Piovani2, J. Prantesh3, I. Gordon3, R. Asselta4, S. Duga4, S. Vetrano1, S. Bonovas2, N. Harpaz5, S. Danese1, I. Peter6, J.F. Colombel7, F. Rieder3
1Humanitas Research Hospital, IBD Center, Milano, Italy, 2Humanitas Research Hospital, IBD Center/Department of Biomedical Sciences, Milano, Italy, 3Cleveland Clinic, Department of Gastroenterology, Cleveland, USA, 4Humanitas Research Hospital, Dept. Biomed. Sciences, Milano, Italy, 5Icahn School of Medicine at Sinai, Department of Pathology, New York, USA, 6Icahn School of Medicine at Sinai, Division of Genetics and Genomic Sciences, New York, USA, 7Icahn School of Medicine at Sinai, The Henry D. Janowitz Division of Gastroenterology, New York, USA
Background
Collagenous colitis (CC) is a common cause of chronic diarrhoea and highly associated with immune-mediated diseases. The aetiology of CC may involve a dysregulated immune response in genetically predisposed individuals. Our preliminary data identified ancestral HLA haplotype 8.1 and non-HLA CC risk loci being associated with CC. We here carried out the first genotype/clinical phenotype association study in a large cohort of CC patients.
Methods
This study combined Immunochip data from 300 patients with centrally read histologically confirmed CC with retrospectively collected clinical data in a quarternary care centre. We tested 21 SNPs previously associated with CC (Roda et al. submitted) for their genetic association with selected phenotypic variables (histology, autoimmune diseases, relapse rate, concomitant medications, number of bowel movements at diagnosis, sex and response to therapy). We performed univariable linear, logistic and Poisson regression analyses for associations with the outcome variables.
Results
300 CC patients [median age 63 (IQR: 54–63); 80% female] were included after quality control. The genetic variants associated with clinical phenotypes can be found in Table 1. We identified a link of several genes involved in T cell function with CC phenotypes. IL2A was associated with risk of additional autoimmune disorders reinforcing the possible role of T cells in both diseases. RUNX3 involved in Th17 cells and TGF-beta pathways was a risk factor for IEL infiltrates. Notably, we identified two protective factors: one near C5orf30, previously implicated in rheumatoid arthritis and collagen-induced arthritis, was linked with a reduced rate of flares; ZNF365, involved in uric acid excretion and associated with ileal Crohn’s disease, was found protective in relation to collagen band thickness. When assessing non-genetic variables, number of bowel movements correlated with collagen band thickness (
| | | | | | |
| 2 | rs35230759 | 0.012 | ZNF804A | uric acid excretion | Autoimmunity |
| 3 | rs671046 | 0.034 | IL2A | immune tolerance | |
| 1 | rs11249221 | 0.025 | RUNX3 | Th17 cell/TGF-beta | IELs infiltrate |
| 2 | rs1444543 | 0.018 | CTNNA2 | - | |
| 5 | rs2600825 | 0.026 | SLCO4C1 | - | |
| 2 | rs35230759 | <0.001 | ZNF804A | - | Relapse rates |
| 5 | rs115530159 | 0.026 | C5orf30 | rheumatoid arthritis and collagen-induced arthritis | |
| 6 | rs3129716 | 0.036 | HLA | Immune responses | |
| 10 | rs871985 | 0.017 | ZNF365 | - | Collagen band thickness |
Conclusion
We found a significant association of T-cell-related genes and CC clinical phenotypes. Additionally, our study identified current smoking and increased collagen band thickness as risk factors for disease severity.
- Posted in: Poster presentations: Genetics 2020