Back

A dysregulated stromal remodelling programme characterises prior anti-TNF failure in ulcerative colitis

Thomas, J. P.; Wooldridge, T.; Cozzetto, D.; Lambie, N.; Kudo, H.; Saifuddin, A.; Gul, L.; Modos, D.; Goldin, R.; Matthews, N.; Korcsmaros, T.; Powell, N.

2026-08-12 systems biology
10.64898/2026.08.11.744139 bioRxiv
Show abstract

Prior anti-tumour necrosis factor (TNF) failure is associated with reduced efficacy of subsequent advanced therapies in ulcerative colitis (UC), but the biological basis of this treatment-refractory state remains unclear. We integrated clinical outcomes and baseline colonic transcriptomic data from UC patients in the UNIFI phase III trial programme with regulatory and signalling network inference, connectivity mapping, and single-cell-resolution spatial transcriptomics. Colonic transcriptomic analyses identified coordinated enrichment of extracellular matrix organisation, collagen remodelling and integrin-associated programmes, increased stromal cell representation and elevated inferred MAPK/EGFR activity in UC patients with prior anti-TNF failure. Causal network inference prioritised MAPK3 as a candidate regulator of this state, while connectivity mapping identified MEK/EGFR inhibitors as candidate perturbagens. MEK inhibition suppressed stromal pathways and reduced inferred MAPK/EGFR activity ex vivo. Spatial profiling of active UC and non-IBD colonic tissues localised these programmes to UC-enriched stromal niches. Ligand-receptor inference further identified reciprocal stromal-myeloid communication within these niches. Collectively, these findings define a stromal remodelling programme associated with prior anti-TNF failure and nominate MAPK/EGFR signalling as a potentially tractable component of treatment-refractory UC.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.