Integrated nanoscale deterministic lateral displacement arrays for separation of extracellular vesicles from clinically-relevant volumes of biological samples

Extracellular vesicles (EVs) offer many opportunities in early-stage disease diagnosis, treatment monitoring, and precision therapy owing to their high abundance in bodily fluids, accessibility from liquid biopsy, and presence of nucleic acid and protein cargo from their cell of origin. Despite their growing promise, isolation of EVs for analysis remains a labor-intensive and time-consuming challenge given their nanoscale dimensions (30-200 nm) and low buoyant density. Here, we report a simple, size-based EV separation technology that integrates 1024 nanoscale deterministic lateral displacement (nanoDLD) arrays on a single chip capable of parallel processing sample fluids at rates of up to 900 μL h-1. Benchmarking the nanoDLD chip against commonly used EV isolation technologies, including ultracentrifugation (UC), UC plus density gradient, qEV size-exclusion chromatography (Izon Science), and the exoEasy Maxi Kit (QIAGEN), we demonstrate a superior yield of ∼50% for both serum and urine samples, representing the ability to use smaller input volumes to achieve the same number of isolated EVs, and a concentration factor enhancement of up to ∼3× for both sample types, adjustable to ∼60× for urine through judicious design. Further, RNA sequencing was carried out on nanoDLD- and UC-isolated EVs from prostate cancer (PCa) patient serum samples, resulting in a higher gene expression correlation between replicates for nanoDLD-isolated EVs with enriched miRNA, decreased rRNA, and the ability to detect previously reported RNA indicators of aggressive PCa. Taken together, these results suggest nanoDLD as a promising alternative technology for fast, reproducible, and automatable EV-isolation.