Background
The inflammatory response to cerebral ischemia is complex and most clinical studies of stroke outcome have focused on relatively few well-investigated proteins. A team from the University of Gothenburg used the Olink® Target 96 Inflammation panel to profile a broad range of inflammation-related proteins, hoping to identify biomarkers associated with ischemic stroke outcome that were independent of established clinical predictors. Plasma proteomics were measured in 534 ischemic stroke cases and logistic regression was used to estimate associations to unfavorable 3-month functional outcome.
Outcome
Twenty proteins showed associations to outcomes in univariate analysis, of which TNFSF14, OSM, SIRT2, STAMBP, and 4E-BP1 were independent of all clinical variables, including stroke severity. Machine learning (LASSO) analysis further identified a 9-protein signature (CCL25, TRAIL, Flt3L, CSF-1, EN-RAGE, HGF, IL-6, OSM, and TNFSF14) that predicted favorable vs unfavorable outcomes with AUC=0.81 – this increased to AUC=0.92 when stroke severity was added to the model.
Pathway enrichment analysis indicated the proteins associated with post-stroke functional outcome were involved in NLRP3 inflammasome regulation and signaling pathways, such as TNF, JAK/STAT, MAPK, and NF-κB. Interestingly, several of these proteins have experimental evidence indicating a role in injury or recovery after stroke, providing biological support for the key findings made. The study suggests that further investigations are strongly warranted to develop these predictive models for stroke outcomes, and if causality can be determined, some of the proteins identified could also open up new therapeutic avenues.
