Predicting the course of IBD: light at the end of the tunnel?

Prognosis in Crohn's disease (CD) and ulcerative colitis (UC) varies substantially between patients. Accordingly, no single treatment strategy will be suitable for everyone. 'Top-down' strategies inevitably overtreat patients destined to experience indolent disease (and risk unnecessary immunosuppression), whilst 'step-up' strategies expose patients with aggressive disease to avoidable disease-related morbidity. Accordingly, for personalised medicine to succeed in IBD, a reliable method of predicting prognosis is critical. In oncology, gene expression profiling enables reliable prediction of several aspects of tumour behaviour. Unfortunately, this technology has been less successful in other disease areas - often because heterogeneous cell populations are studied, producing gene signatures that simply reflect compositional differences in the starting material. This can be overcome by analysing purified cell populations. Previously, we discovered a prognostic gene expression signature in purified CD8+ T cells from patients with systemic lupus erythematosus and ANCA-associated vasculitis. We therefore investigated whether the same signature might exist in other relapsing-remitting diseases driven by immune responses to antigen, and detected an analogous signature in CD and UC. This signature was detectable from diagnosis, but did not correlate with contemporaneous clinical or biochemical data. A significant association with prognosis was detected in both UC and CD, with patients who were enriched for the signature (subgroup 'IBD1') experiencing a more aggressive disease course, characterised by frequent flares and a higher incidence of treatment failure. This suggests that gene expression profiling could stratify patients with CD or UC at diagnosis according to their future disease course, and represents a step towards personalised therapy.