Brain plasticity and neuroinflammatory protein biomarkers with circulating MicroRNAs as predictors of acute brain injury outcome – A prospective cohort study

Abstract

<h3>Background</h3><p>Brain recovery mechanisms after injuries like aneurysmal subarachnoid hemorrhage (aSAH), ischemic stroke (IS), and traumatic brain injury (TBI) involve brain plasticity, synaptic regeneration, and neuroinflammation. We hypothesized that serum levels of the p75 neurotrophic receptor (p75NTR) and associated signaling proteins, as well as differentially expressed (DE) microRNAs, could predict recovery outcomes irrespective of injury type.</p><h3>Methods</h3><p>A prospective patient cohort with ischemic stroke (IS, <em>n</em> = 30), aneurysmal subarachnoid hemorrhage (aSAH, <em>n</em> = 31), and traumatic brain injury (TBI, <em>n</em> = 13) were evaluated (total <em>n</em> = 74). Serum samples were collected at two post-injury intervals (early: 1–3 days, late: 4–8 days), and outcomes were assessed after three months using the modified Rankin Scale (mRS), categorizing outcomes as favorable (mRS 0–3) or unfavorable (mRS 4–6). Six proteins were measured using ELISAs: p75NTR, NGF, sortilin, IL1β, TNFα, and cyclophilin. DE microRNAs were identified using DESeq2, and their target genes were predicted. Serum molecules between patients with differing outcomes were compared using a Kolmogorov-Smirnov test, 2-tailed <em>t</em>-test and multivariate linear discriminant analysis (LDA).</p><h3>Results</h3><p>Favorable (<em>n</em> = 46) and unfavorable (<em>n</em> = 28) outcome cohorts were balanced with age and sex (<em>p</em> = 0.25 and 0.63). None of the studied proteins correlated with age. Combinatory LDA of the six protein biomarkers indicated strong prognostic value for favorable outcomes (OR 2.09; AUC = 70.3%, <em>p</em> = 0.0058). MicroRNA expression changes over time were identified in the aSAH, TBI, and IS groups (<em>p</em> < 0.05, FDR corrected). Twenty-three microRNAs were commonly DE across all brain injury groups when comparing favorable and unfavorable outcomes (p < 0.05). LDA of four microRNAs targeting the studied proteins showed high prognostic accuracy (OR 11.7; AUC = 94.1%, <em>p</em> = 0.016).</p><h3>Conclusions</h3><p>The combined prognostic microRNA and protein biomarker models demonstrated accurate outcome prognostication across diverse injury types, implying the presence of a common recovery mechanism. DE microRNAs were found to target the studied molecules, suggesting a potential mechanistic role in recovery. Further investigation is warranted to study these molecules in prognostication, as well as therapeutic targets for enhancing recovery.</p>