Molecularly Imprinted Nanozyme Intelligent Array Sensing Platform Realizes the Enrichment and Quantitative Detection of Small Extracellular Vesicles

As tumor-associated biomarkers, proteins on small extracellular vesicles (sEVs) play a critical role in revealing tumor subtypes and disease progression. Nevertheless, the rapid isolation of sEVs and the quantitative detection of their surface proteins remain challenging. In this study, we developed a nanozyme (PB-MIP-PEG) array sensing platform based on molecular imprinting technology (MIT), which can efficiently enrich sEVs and simultaneously detect three proteins: CD20, CD19, and PD-L1. Fe₃O₄@TiO₂ magnetic beads interact with phosphate groups in the sEVs phospholipid bilayer and, under an external magnetic field, achieve effective separation from complex samples. Prussian blue (PB) nanozymes with peroxidase (POD)-like activity, coated with a silicon dioxide (SiO₂) imprinting layer, showed enhanced enrichment toward 3,3',5,5'-tetramethylbenzidine (TMB), which contributed to improved catalytic activity and detection sensitivity. Furthermore, density functional theory (DFT) and molecular dynamics (MD) simulations further revealed that hydrogen bonding and π-π stacking are the dominant interactions during the imprinting process. By integration of colorimetric imaging with a position-encoding system and employing a convolutional neural network-long short-term memory (CNN-LSTM) model for automated feature extraction, the concentrations of CD20, CD19, and PD-L1 were rapidly distinguished. This array sensing platform, which integrates molecularly imprinted nanozymes with neural network models, provides a new approach for the intelligent analysis of sEVs surface proteins and demonstrates a potential application value in lymphoma-related monitoring.