Abstract P777: Statistical Modeling of Proteomic Signaling During Stroke in Patients Undergoing Thrombectomy

Abstract

Introduction: The previously published Blood and Clot Thrombectomy Registry and Collaboration (BACTRAC) protocol (clinicaltrials.gov NCT03153683) utilizes mechanical thrombectomy to obtain tissue samples for banking. Peripheral blood proximal to the clot and intracranial blood distal from the clot were isolated. Proteomic and statistical analyses revealed normalized (intracranial-systemic) CCL19 expression was a predictor of infarct volume. Statistical modeling analyses were used determine the CCL19-associated proteomic signaling network occurring during ischemic stroke relating to infarct volume. Methods: Arterial intracranial and systemic blood samples underwent analysis for inflammatory proteins using Proximity Extension Assay (PEA) via Olink (Olink Proteomics, Boston, MA). Systemic expression was used as an internal control to normalize expression in the intracranial blood. Bivariate regression was used to examine the relationship between the intracranial normalized CCL19 expression and infarct volume. A backwards stepwise regression was then used to determine a model of predictability of infarct volume by CCL19 and associated inflammatory proteins. Results: 25 subjects (>18 yrs) with a mean infarct volume of 8,172 ± 82,284 mm 3 and mean infarct time of 513 ± 246 minutes were included in this study. Their median age was 64 (24-91) and 10 (40%) were male. 16 subjects (64%) had hypertension, 15 (60%) had BMI > 25, and 6 (24%) had a previous stroke. The stepwise regression model shows normalized expression of 16 proteins correlated with an increase in infarct volume (p<0.005): CCL20, CXCL1, OSM, CD6, OSMR, TGF-alpha, TRANCE, CXCL10, LIF-R, CCL19, CDCP1, Flt3L, CCL23, CD244, TRAIL, NOTCH1. Conclusions: In our model, the expression of these proteins were consistently changed, though the directionality differed. LIF-R, NOTCH1, TRAIL, CD6, CCL23, TGF-alpha, and CCL20 were positively correlated, while the expressions of Flt3L, OSM, OSMR, TRANCE, CD244, CDCP1, CXCL1, CXCL10, and CCL19 were negatively correlated with infarct volume. This model depicts the proteomic signaling occurring during stroke in relationship to infarct volume, which reveals potential biomarkers and therapeutic targets for the early phase of ischemic stroke.