208 Preliminary Results of Novel Method for Identifying Ischemic Stroke Vesicular Protein Signature in Cardiac Surgery Patients

Abstract

INTRODUCTION: Acute ischemic stroke (AIS) treatment depends on timely diagnosis. Identifying an efficient, blood-based protein signature for AIS has been limited by the lack of a pre-ischemia control group. Here we report the preliminary results of a model for AIS protein signature discovery using cardiac surgery patients and novel mass spectrometry techniques. METHODS: Twenty patients undergoing aortic arch repair with deep hypothermic circulatory arrest (n = 17) or coronary artery bypass graft (n = 3) were enrolled. Blood samples were obtained immediately pre- and postoperatively. Magnetic resonance imaging (MRI) of the brain was obtained within 72 hours of surgery. Presence of infarct and total ischemic volume was determined. A novel protein enrichment and mass spectrometry analysis technique (MAG-Net) was used to characterize the pre- and post-surgical proteomes present in extracellular vesicles. Analysis of variance (ANOVA) (false discovery rate < 0.05) was used to analyze protein signatures associated with infarction. RESULTS: Post-surgical infarct occurred in 15 patients (75%), with a median volume of 108mm3. A total of 5,376 proteins were identified, 1,125 of which showed a significant difference between paired pre- and post-operative concentrations. Of these 1,125 proteins, 524 proteins were unique to patients without infarction, 340 proteins had a significant change in both populations, and 261 proteins were unique to patients with infarcts. There were no significant proteins for the interaction term comparing pre- and post-operative differences between the infarct/no infarct groups. CONCLUSIONS: This clinical model for AIS protein signature development relies on the high rate of ischemia during cardiac surgery to allow for pre-stroke sample collection. The MAG-Net technique enables efficient processing of biologically relevant extracellular vesicle proteins and analysis of an unbiased protein signature. The preliminary results demonstrate the feasibility of combining these innovations for AIS protein signature discovery.