Molecular Identification of Tumor-Derived Extracellular Vesicles Using Thermophoresis-Mediated DNA Computation

Abstract

Molecular profiling of tumor-derived extracellular vesicles (tEVs) holds great promise for non-invasive cancer diagnosis. However, sensitive and accurate identification of tEVs is challenged by the heterogeneity of EV phenotypes which reflect different cell origins. Here we present a DNA computation device mediated by thermophoresis for detection of tEVs. The strategy leverages the aptamer-based logic gate using multiple protein biomarkers on single EVs as the input and thermophoretic accumulation to amplify the output signals for highly sensitive and specific profiling of tEVs. Employing this platform, we demonstrate a high accuracy of 97% for discrimination of breast cancer (BC) patients and healthy donors in a clinical cohort (n = 30). Furthermore, molecular phenotyping assessed by tEVs is in concordance with the results from tissue biopsy in BC patients. The thermophoresis-mediated molecular computation on EVs thus provides new opportunities for accurate detection and classification of cancers.