-
Views
-
Cite
Cite
B Verstockt, S Verstockt, P Sudahakar, J Dehairs, H Blevi, G Van Assche, S Vermeire, M Ferrante, DOP38 A vedolizumab specific four-gene colonic signature accurately predicting future endoscopic remission in patients with inflammatory bowel disease, Journal of Crohn's and Colitis, Volume 13, Issue Supplement_1, March 2019, Page S048, https://doi.org/10.1093/ecco-jcc/jjy222.072
Close - Share Icon Share
Abstract
Vedolizumab, a monoclonal antibody targeting α4β7 integrin and mainly inhibiting gut lymphocyte trafficking, has been approved for the treatment of both Crohn’s disease (CD) and ulcerative colitis (UC). Due to the increasing availability of therapeutic compounds in inflammatory bowel disease (IBD), predictive biomarkers are urgently awaited in order to help clinicians decide between anti-TNF, vedolizumab or other therapies.
We obtained inflamed colonic biopsies from 31 patients (20 UC, 11 CD) prior to initiation of vedolizumab. Similarly, inflamed colonic biopsies (15 UC, 9 CD) were collected from 24 patients initiating anti-TNF therapy (Table).
| . | Vedolizumab (n = 31) . | Anti-TNF cohort (n = 24) . |
|---|---|---|
| Diagnosis, n (%) UC / CD | 20 (64.5) / 11 (35.5) | 16 (66.7) / 8 (33.3) |
| Age, years (median, IQR) | 45.3 (29.6–56.3) | 36.0 (22.0–54.9) |
| Gender, n women (%) | 17 (54.8) | 15 (62.5) |
| Disease duration, years (median, IQR) | 8.4 (4.0–15.3) | 0.5 (1.9–7.0) |
| Disease location L1-L2-L3-L4 / E1-E2-E3 | 0-2-9-2 / 3-10-7 | 0-2-6-1 / 0–13-3 |
| Disease behaviour B1-B3-B3 perianal | 6-3-2-5 | 4-3-1-2 |
| Steroid use during induction, n (%) Topical—Systemic | 10 (32.3)–8 (25.8) | 5 (16.1)–7 (22.6) |
| Previous-anti-TNF exposure | 21 | NA |
| Endoscopic remission, number (%) | 19 (61.3) | 8 (33.3) |
| . | Vedolizumab (n = 31) . | Anti-TNF cohort (n = 24) . |
|---|---|---|
| Diagnosis, n (%) UC / CD | 20 (64.5) / 11 (35.5) | 16 (66.7) / 8 (33.3) |
| Age, years (median, IQR) | 45.3 (29.6–56.3) | 36.0 (22.0–54.9) |
| Gender, n women (%) | 17 (54.8) | 15 (62.5) |
| Disease duration, years (median, IQR) | 8.4 (4.0–15.3) | 0.5 (1.9–7.0) |
| Disease location L1-L2-L3-L4 / E1-E2-E3 | 0-2-9-2 / 3-10-7 | 0-2-6-1 / 0–13-3 |
| Disease behaviour B1-B3-B3 perianal | 6-3-2-5 | 4-3-1-2 |
| Steroid use during induction, n (%) Topical—Systemic | 10 (32.3)–8 (25.8) | 5 (16.1)–7 (22.6) |
| Previous-anti-TNF exposure | 21 | NA |
| Endoscopic remission, number (%) | 19 (61.3) | 8 (33.3) |
Clinical characteristics of the vedolizumab and anti-TNF treated cohort
RNA was extracted and single-end RNA sequencing was performed using Illumina HiSeq4000. Normalisation and differential expression was done using DESeq2 R package. Pathways were analysed with Ingenuity Pathway Analysis (IPA). Using randomised generalised linear modelling (RGLM), a predictor for vedolizumab-induced endoscopic remission (absence of ulcerations at month 6 for CD; Mayo endoscopic sub-score ≤1 at Week 14 for UC) was identified in a randomly generated test cohort (n = 20) and validated in 11 independent samples. Through unsupervised consensus clustering, we validated the marker in a publicly available microarray dataset (GSE73661), and studied vedolizumab specificity in the anti-TNF treated cohort.
| . | Vedolizumab (n = 31) . | Anti-TNF cohort (n = 24) . |
|---|---|---|
| Diagnosis, n (%) UC / CD | 20 (64.5) / 11 (35.5) | 16 (66.7) / 8 (33.3) |
| Age, years (median, IQR) | 45.3 (29.6–56.3) | 36.0 (22.0–54.9) |
| Gender, n women (%) | 17 (54.8) | 15 (62.5) |
| Disease duration, years (median, IQR) | 8.4 (4.0–15.3) | 0.5 (1.9–7.0) |
| Disease location L1-L2-L3-L4 / E1-E2-E3 | 0-2-9-2 / 3-10-7 | 0-2-6-1 / 0–13-3 |
| Disease behaviour B1-B3-B3 perianal | 6-3-2-5 | 4-3-1-2 |
| Steroid use during induction, n (%) Topical—Systemic | 10 (32.3)–8 (25.8) | 5 (16.1)–7 (22.6) |
| Previous-anti-TNF exposure | 21 | NA |
| Endoscopic remission, number (%) | 19 (61.3) | 8 (33.3) |
| . | Vedolizumab (n = 31) . | Anti-TNF cohort (n = 24) . |
|---|---|---|
| Diagnosis, n (%) UC / CD | 20 (64.5) / 11 (35.5) | 16 (66.7) / 8 (33.3) |
| Age, years (median, IQR) | 45.3 (29.6–56.3) | 36.0 (22.0–54.9) |
| Gender, n women (%) | 17 (54.8) | 15 (62.5) |
| Disease duration, years (median, IQR) | 8.4 (4.0–15.3) | 0.5 (1.9–7.0) |
| Disease location L1-L2-L3-L4 / E1-E2-E3 | 0-2-9-2 / 3-10-7 | 0-2-6-1 / 0–13-3 |
| Disease behaviour B1-B3-B3 perianal | 6-3-2-5 | 4-3-1-2 |
| Steroid use during induction, n (%) Topical—Systemic | 10 (32.3)–8 (25.8) | 5 (16.1)–7 (22.6) |
| Previous-anti-TNF exposure | 21 | NA |
| Endoscopic remission, number (%) | 19 (61.3) | 8 (33.3) |
Clinical characteristics of the vedolizumab and anti-TNF treated cohort
RNA was extracted and single-end RNA sequencing was performed using Illumina HiSeq4000. Normalisation and differential expression was done using DESeq2 R package. Pathways were analysed with Ingenuity Pathway Analysis (IPA). Using randomised generalised linear modelling (RGLM), a predictor for vedolizumab-induced endoscopic remission (absence of ulcerations at month 6 for CD; Mayo endoscopic sub-score ≤1 at Week 14 for UC) was identified in a randomly generated test cohort (n = 20) and validated in 11 independent samples. Through unsupervised consensus clustering, we validated the marker in a publicly available microarray dataset (GSE73661), and studied vedolizumab specificity in the anti-TNF treated cohort.
Forty-four genes (25 down, 19 up) were significantly differently expressed between future vedolizumab remitters and non- remitters. Involved pathways included glucocorticoid receptor signalling, differential regulation of cytokines in intestinal epithelial cells, granulocyte adhesions and diapedesis. Using these 44 differentially expressed genes as input for the RGLM modelling, we identified a 4-gene signature which could accurately split remitters and non- remitters in both the discovery (accuracy 90.9%, p = 0.02) and validation (100%, p = 0.006) set. Using the same 4-gene signature we could accurately discriminate prospective future remitters from non-remitters in a publicly available microarray data set of 13 open-label vedolizumab treated UC patients (84.6%, p = 0.02). In contrast, this 4-gene signature was not predictive for anti-TNF induced endoscopic remission (62.5%, p = 0.65).
We identified and validated the first, vedolizumab-specific predictive 4-gene expression signature which may guide treatment strategy in IBD patients with colonic involvement.
