P897 Longitudinal multi-tissue transcriptomic study reveals patient-specific drug response determinants in Inflammatory Bowel Disease patients.
Cervera Seco, L.M.(1)*;Sánchez Mayor, M.(1);Corraliza, A.M.(2,3);Salas, A.(2,3);Panés, J.(2,3);M. Marigorta, U.(1,4);
(1)Centre for Cooperative Research in BioSciences CIC bioGUNE- Basque Research and Technology Alliance BRTA- Bizkaia Technology Park, Integrative Genomics Lab, Derio, Spain;(2)Institut d'Investigacions Biomèdiques August Pi i Sunyer IDIBAPS- Hospital Clínic, Department of Gastroenterology, Barcelona, Spain;(3)Centro de Investigación Biomédica en Red de Enfermedades Hepática y Digestivas, CIBERehd, Barcelona, Spain;(4)IKERBASQUE, Basque Foundation for Science, Bilbao, Spain;
Despite the lack of understanding of Inflammatory Bowel Disease (IBD) pathogenesis, association studies have made inroads into the elucidation of the genetic component of disease diagnosis. The corresponding determinants of drug response, in turn, remain obscure. Primary response is low, secondary loss of response is prevalent and remission rates are low (30% of patients roughly). Along with the expanding repertoire of drugs, it is critical to develop a strategy to reveal the factors contibuting to drug response at the patient level. Our goal is to understand drug response in IBD looking to establish a precision medicine strategy to assign patients a specific treatment based on their transcriptomic profile.
We carried out a longitudinal follow-up of 147 IBD patients, from whom we sampled 264 intestinal biopsies and 119 blood extracts. Patients were exclusively treated with one of the three main available biologics: 248 with infliximab (anti-TNF), 47 with ustekinumab (anti-IL12/IL23) and 88 with vedolizumab (anti-α4β7). RNA-seq profiles were obtained at three different times (baseline, 14 weeks and 46 weeks after treatment initiation), resulting in 232, 92 and 59 samples, respectively. Differential expression analysis and pathway enrichment analysis were performed to get drug-specific genes and response signatures.
The initial exploratory analysis confirms that tissue is the main determinant of expression variability among patients. As for longitudinal response, anti-TNF-specific features are quintessentially inflammatory, especially involved in translation and immune pathways, while response to ustekinumab is associated with extracellular recognition. Of note, vedolizumab-specific profiles strikingly differ between biopsy and blood samples. We also revealed that drug response is determined during first weeks of treatment, as up- and down-regulated genes are shared across time points. This may entail huge consequences in the clinic since the determination of response could be assessed earlier. Using clinical activity to define responders, we identified around 200 differentially expressed genes that could predict response to infliximab at time zero. We also improved our understanding regarding individual response by carrying out trend analyses based on connectivity scores that identify individual-specific regulated pathways.
While we identified common patterns across time in enriched pathways, we could discern specific drug response profiles. Besides providing a renewed understanding of the molecular effects of biological drugs, we propose that personalised models predicting response are feasible, and a promising venue to develop therapeutic guidelines of biologics in IBD.