Background: Extracellular vesicles (EVs), including exosomes, are promising pharmaceutical modalities. They are purified from cell culture supernatant; however, the preparation may contain EVs with the desired therapeutic effects and different types of EVs, lipoproteins, and soluble proteins. Evaluating the composition of particulate impurities and the levels of protein impurities in final preparations is critical for quality control. However, few analytical methods can detect these impurities. Methods: We established and evaluated an analytical method using size-exclusion chromatography–multi-angle light scattering (SEC-MALS) for particle and protein impurity analyses of EV samples. Results: In the particle size distribution analysis of EV samples, SEC-MALS showed higher resolution compared with nanoparticle tracking analysis (NTA) and dynamic light scattering (DLS). MALS showed comparable accuracy and precision to that of other methods for particle size evaluation using polystyrene standard beads with 60, 100, or 200 nm diameter. Coupling SEC-MALS with UV detection quantitatively evaluated soluble protein impurities. Proteomic analysis on the SEC-MALS-fractionated samples identified different EV and lipoprotein marker proteins in different fractions. Conclusions: SEC-MALS can characterize EV preparations obtained from human adipose-derived mesenchymal stem cells, suggesting that it can evaluate the particle component composition in various EV samples and therapeutic exosome preparations.

Temporomandibular joint osteoarthritis (TMJ-OA) involves inflammation and degeneration of mandibular condylar cartilage and subchondral bone. Dedifferentiated fat cells (DFATs) were established from buccal fat pad tissue obtained from a single human donor. We examined the effects of DFAT-derived extracellular vesicle preparations (DFAT-EVs) on inflammatory macrophage responses in vitro and disease-related outcomes in TMJ-OA in vivo. DFAT-EVs were applied to lipopolysaccharide- and interferon-γ-stimulated RAW264.7 macrophages and administered intra-articularly in rats with monosodium iodoacetate-induced TMJ-OA. Fluorescence signals from labeled DFAT-EVs were observed in RAW264.7 cells. DFAT-EVs reduced nitric oxide production and modulated nitric oxide synthase 2, interleukin (IL)-1β, IL-6, IL-10, IL-4, and tumor necrosis factor-α responses, although effects varied among inflammatory mediators. Head withdrawal threshold was significantly higher in the DFAT-EVs group than in the OA group at several time points in the 6-week cohort, but not in the 4-week cohort. Administration of DFAT-EVs attenuated the reduction in mandibular condylar bone volume fraction and was associated with preservation of cartilage architecture and matrix. Osteoarthritis Research Society International scores were significantly lower in the DFAT-EVs group than in the OA group at 4 weeks but not at 6 weeks. These in vitro and in vivo findings support further evaluation of DFAT-EVs for TMJ-OA.