Abstract
Blood-based diagnostic biomarkers for amyotrophic lateral sclerosis will improve patient outcomes and positively impact novel drug development. Critical to the development of such biomarkers is robust method validation, optimization and replication with adequate sample sizes and neurological disease comparative blood samples. We sought to test an amyotrophic lateral sclerosis biomarker derived from diverse samples to determine if it is disease specific. Extracellular vesicles were extracted from blood plasma obtained from individuals diagnosed with amyotrophic lateral sclerosis, primary lateral sclerosis, Parkinson's disease and healthy controls. Immunoaffinity purification was used to create a neural-enriched extracellular vesicle fraction. MicroRNAs were measured across sample cohorts using real-time polymerase chain reaction. A Kruskal-Wallis test was used to assess differences in plasma microRNAs followed by post hoc Mann-Whitney tests to compare disease groups. Diagnostic accuracy was determined using a machine learning algorithm and a logistic regression model. We identified an eight-microRNA diagnostic signature for blood samples from amyotrophic lateral sclerosis patients with high sensitivity and specificity and an area under the curve calculation of 98% with clear statistical separation from neurological controls. The eight identified microRNAs represent disease-related biological processes consistent with amyotrophic lateral sclerosis. The direction and magnitude of gene fold regulation are consistent across four separate patient cohorts with real-time polymerase chain reaction analyses conducted in two laboratories from diverse samples and sample collection procedures. We propose that this diagnostic signature could be an aid to neurologists to supplement current clinical metrics used to diagnose amyotrophic lateral sclerosis.