Formulation of GM-CSF into biodegradable polymeric microspheres

Abstract

The use of spray drying to prepare protein-containing biodegradable polylactide-co-glycolide microspheres suitable for targeting and prolonged release was investigated. A therapeutic protein, granulocyte-macrophage colony-stimulating factor (GM-CSF) was compared to a relatively stable model protein, bovine serum albumin (BSA). A bioassay was developed for analyzing the activity of GM-CSF which was quantified using an enzyme-linked immunosorbent assay (ELISA). The effect of formulation and processing factors on GM-CSF stability was evaluated. The spray drying process parameters were screened using a Plackett-Burman experimental design. The microspheres were characterized for size and protein release and compared to those prepared by solvent extraction. BSA and Tween 80 were found to maximize the recovery of active GM-CSF in solution and during processing indicating that the loss of GM-CSF was mainly due to adsorption or aggregation. Lyophilization of the formulations resulted in greater recovery of active GM-CSF. Significant parameters for the spray drying of microspheres were nitrogen flow rate, polymer solvent, nozzle tip position and solution feed rate. The spray dried microspheres had an average size of about 3 {dollar}\mu\rm m{dollar} with more than 90% of the particles below 5 {dollar}\mu\rm m{dollar} which is suitable for targeting applications. About 41% of GM-CSF activity was retained in the microspheres; 60% of this was released in about 10 hours, after which small amounts were detected up to two days. The absence of further release was attributed to protein-polymer interactions or instability of released protein in the medium. In the case of BSA, about 55-60% was released rapidly in one day and the remaining slowly in four days. The rapid release could be due to improper coating of protein by spray drying or uneven distribution of lyophilized protein in the polymer. Changing the polymer molecular weight, or incorporating BSA in a spray dried form, did not change the release profile significantly. The microspheres prepared by solvent extraction had an average particle size of 3.7 {dollar}\mu\rm m{dollar} with 75% below 5 {dollar}\mu\rm m.{dollar} The activity of GM-CSF retained in this product was 47%. The release profile for GM-CSF from these microspheres was superimposable over that for the spray dried product indicating that the release was independent of the method used for microsphere preparation. The results demonstrate that spray drying is a suitable technique for preparation of microspheres for targeting. The release of proteins from this delivery system is affected by protein-polymer interactions. The instability of the released protein in the medium may preclude the detection of release.