Recombinant adeno-associated virus (rAAV) has emerged as a primary vector for commercial in vivo gene therapies. However, current manufacturing processes for rAAV-based gene therapies face significant inefficiencies, leading to shortages for clinical trials and approved therapies, as well as to prohibitively high production costs (more than $300,000 per dose for certain therapies). Timely quantification of total and full particle titers during rAAV production is crucial for product quality assurance and process optimization. Conventional analytical tools for rAAV titer quantification, such as polymerase chain reaction (PCR) and enzyme-linked immunosorbent assay (ELISA), are labor-intensive offline methodologies, characterized by high costs. This study introduces a novel approach for real-time quantification of full and total rAAV titers during production via transient transfection, based on a suspended microchannel resonator (SMR) technology combined with machine learning. The SMR measures single-cell mass, density, and volume of rAAV-producing cells, which closely reflect changes in cellular states and specific productivity during transfection. Subsequently, a machine learning model (ML) predicts full and total rAAV titers by using only SMR-derived single-cell biophysical measurements. The ML model is successfully validated on a test set, demonstrating its capability to predict full and total rAAV titers using only SMR-derived single-cell measurements. The SMR/ML platform has the potential to provide cost-effective, online measurements of rAAV titers during production in mammalian cultures, offering a promising tool for enhancing quality control and process efficiency throughout the development and manufacturing stages.
