A Scalable Bioreactor Platform for Reproducible Production and Characterization of Ovarian Cancer-Derived Extracellular Vesicles
Wei Fu, Kalpana Deepa Priya Dorayappan, Colin Hisey, Lakshmi Narasimhan Chakrapani, Sydney Wiggins, Shyam Sundaram, Zachary Lambert, Kim Truc Nguyen, Sudhiksha Anbu Chelian, Eduardo Reategui, Karuppaiyah Selvendiran, Derek J. HansfordExtracellular vesicles (EVs) from ovarian cancer cells are valuable sources for candidate biomarker studies, but conventional static flask culture yields limited material and is difficult to scale reproducibly. We evaluated a serum-free CELLine AD 1000 bioreactor workflow for producing EVs from four ovarian cancer-related (OC-related) cell lines (OVCAR4, CaOV3, PA1, SW626) and human dermal fibroblasts (HDFa) as a non-cancer control. Cells were adapted to CDM-HD serum-free medium and maintained for eight weeks with twice-weekly conditioned-medium collection. EVs were isolated by differential ultracentrifugation followed by size-exclusion chromatography and characterized by nanoparticle tracking analysis, imaging flow cytometry, Western blotting, and transmission and scanning electron microscopy. Across longitudinal harvests, OC-related cultures generally produced higher EV particle concentrations and A280-based bulk protein estimates than HDFa, while individual cell lines showed distinct production profiles and membrane-associated growth patterns. A parallel OVCAR4 T-175 flask, maintained in its original serum-containing medium, provided a contextual reference indicating higher per-collection EV particle recovery with the bioreactor, although this was not a matched culture-format comparison. EV-enriched preparations contained vesicle-like particles, with modal diameters of approximately 96–128 nm. Using imaging flow cytometry, the CD9 signal was higher in OC-related EVs and CD63 was most prominent in HDFa; CD9 and CD63 were also detected in OC-related EV lysates by Western blotting. Because one bioreactor was operated per cell line, these findings should be interpreted as preliminary and descriptive rather than statistically comparative. Overall, this study provides a practical serum-free CELLine AD 1000 workflow for generating characterized OC-related EV material for downstream analytical studies.