Quality improvement of spray-dried, protein-loaded D,L-PLA microspheres by appropriate polymer solvent selection

The aim was to study the effect of the type of polymer solvent on characteristics of microspheres produced by spray drying. The water-soluble model protein, bovine serum albumin (BSA) was microencapsulated into biodegradable poly(D,L-lactic acid) using the following 10 different polymer solvents: acetaldehyde dimethyl acetal, acetone, dichloromethane, dioxane, ethyl acetate, ethyl vinyl ether, nitromethane, tetrahydrofuran, 1,1,1-trichloroethane, and 1,1,2-trichloroethylene. These solvents having similar toxicity levels differ greatly in their physico-chemical characteristics such as boiling point, vapour pressure, miscibility and interfacial tension with an aqueous phase, and solubility parameter. The effect of these solvents on microsphere morphology was studied by SEM-micrographs. Regular particle morphology was obtained when dichloromethane, ethyl acetate, or nitromethane was used as the polymer solvent, whereas the trichlorinated solvents, tetrahydrofuran, and dioxane produced a substantial number of coalesced particles. The results are interpreted in terms of boiling point, vapour pressure, and polymer-solvent affinity. Further, BSA-loading and -integrity in the microspheres, and burst release were analysed. The theoretical loading of 2.9% was attained with dichloromethane, ethyl acetate and nitromethane, in agreement with observations of particle morphology. HPLC- and SDS-PAGE analysis of the microencapsulated BSA did not show any protein degradation or dimerization, whereas solid-phase ELISA clearly revealed that the in vitro protein antigenicity was substantially reduced (50%), particularly by water miscible solvents. Dichloromethane and ethyl acetate did not show any detrimental effect on protein antigenicity. Finally, burst release could be related again to particle morphology, with dichloromethane and nitromethane giving a burst release of only 5%. In conclusion, dichloromethane, ethyl acetate and nitromethane proved to be the most suitable solvents for the polymer-protein system studied.