Search in the Abstract Database

Abstracts Search 2019

OP19 Corticosteroid response rectal gene signature and associated microbial variation in treatment naïve ulcerative colitis

Y. Haberman*1,2, R. Karns2, P. Dexheimer2, M. Schirmer3, T. Braun1, M. Collins2, A. Mo4, M. Rosen2, N. Gotman5, PROTECT Study group, S. Kugathasan6, T. D. Walters7, G. Gibson4, S. Davis Thomas5, C. Huttenhower8, R. J. Xavier9, J. S. Hyams10, L. A. Denson2

1Sheba Medical Center, Tel Hashomer, Israel, 2Cincinnati Children Hospital Medical Center, Cincinnati, USA, 3Broad Institute of MIT and Harvard University, Cambridge, USA, 4Georgia Institute of Technology, Atlanta, USA, 5University of North Carolina, Chapel Hill, USA, 6Emory University, Atlanta, USA, 7Hospital For Sick Children, Toronto, Canada, 8Harvard School of Public Health, Boston, USA, 9Broad Institute of MIT and Harvard University, Boston, USA, 10Connecticut Children’s Medical Center, Hartford, USA

Background

Molecular mechanisms driving disease course and response to initial therapy in ulcerative colitis (UC) are poorly understood. In the full PROTECT cohort, the strongest predictor of corticosteroids (CS)-free remission by Weeks 12 or 52 was Week 4 (WK4) remission. We used pre-treatment rectal biopsies in new-onset UC, and defined key pathways linked to WK4 response to standardised induction with CS in the largest prospective paediatric UC cohort to date.

Methods

PROTECT enrolled 428 newly diagnosed paediatric UC patients at 29 North American sites. mRNA-Seq and 16S rRNA defined pre-treatment rectal gene expression and microbial communities in 206 participants. Independent group of 50 participants were used to validate the CS response gene signature. WK4 remission was defined as PUCAI < 10 without additional therapy/colectomy.

Results

Moderate–severe UC patients (152/206) from the discovery cohort and all 50 from the validation cohort received standardised induction therapy with CS. WK4 remission was achieved in 75/152 (49%) and 21/50 (42%) of the discovery and validation groups respectively. 115 genes were differentially expressed (FDR<0.05 and FC ≥1.5) between moderate–severe UC patients who did or did not achieve WK4 remission in the discovery cohort. The corticosteroid response gene signature is highly associated with CXCR chemokines (p < 7.12E−12), innate myeloid immune signatures (p < 1.62E−15), and response to bacteria (p < 2.16E−13). Principle component analyses (PCA) PC1 that summarise the variation of the CS response signature, was significantly different between those that achieved WK4 clinical remission (p < 0.001) and mucosal healing (p = 0.002) defined as faecal calprotectin < 250 μg/g in the discovery cohort. This was replicated in the independent validation cohort, and was also associated with response to anti-TNFα and anti-α4β7 integrin induction in adults. The response gene signature was associated with shifts in microbes previously implicated in mucosal homeostasis; positive association with taxa such as Campylobacter, Veillonella, and Enterococcus implicated in mucosal inflammation, and a negative association with taxa from the Clostridiales order that are considered beneficial. Finally, the addition of the pre-treatment rectal gene signature PC1 [OR = 0.4 (0.2–0.8 95% CI)] improved WK4 clinical prediction model of remission with CS [AUC = 77.7 (70.0–85.4 95% CI)].

Conclusion

We identified a gene signature linked to WK4 CS response, which was validated in independent UC patients, and showed associations with response to anti-TNFα and anti-α4β7 integrin in adults, and with specific microbial taxa. Our data may prioritise future therapies for non-responders to current approaches.