Repeat administration of gene therapy for cystic fibrosis is likely to be essential for long-term clinical efficacy. This may be minimized by the use of slow-release gene transfer preparations with more prolonged expression and longer dosing intervals for the patient. Poly(D-L-lactide-co-glycolide) (PLG) is a biodegradable and biocompatible polymer that has been used to encapsulate plasmid DNA. PLG-DNA microspheres were generated and characterized with respect to morphology, size (80% of particles <5.2 microm), and encapsulation efficiency (50.7+/-2.3%, n=6). Gel electrophoresis of DNA re-extracted from the microspheres confirmed that despite a decrease in the proportion of supercoiled conformation, it had not been degraded by the preparation process. Gene transfer efficiency was tested using microspheres encapsulating the reporter gene beta-galactosidase in vitro on Cos 7 cells and a CF airway epithelial line (CFTEo approximately ) and ex vivo in a sheep tracheal (s.t.) model. In both cases, transgene expression was significantly (P<0.01) lower at the first time point tested (24 h in vitro, 48 h ex vivo) compared to lipid-#67-mediated gene transfer. However, PLG-mediated expression in vitro was sustained at 48 h, while lipid #67-mediated expression levels had dropped significantly (P<0.05) to 50.3+/-13.7 and 38.2+/-2.7% (Cos 7 and CFTEo approximately cells, respectively) of the 24-h level. This pattern was also seen in the s.t. model where at 72 h, PLG-mediated expression was 125.4+/-7.2% of the 48-h level demonstrating significantly (P<0.05) better retention of transfection efficiency than lipid #67, where levels had fallen to approximately half the 48 h level. By 96 h, expression was still retained in the PLG-transfected group (87.3+/-12.5% of 48 h expression) but was undetectable in the lipid -#67-transfected s.t. Finally, PLG microspheres, encapsulating the reporter gene chloramphenicol transferase (CAT, 80 microg) were instilled intranasally into Balb/C mice. Compared to lipid-#67-mediated delivery, where whole lung CAT expression was highest at 48 h (13.7 x 10(3)+/-0.05 CAT U/microg protein, n=6) and then not detectable at further time points, CAT expression was not detectable in PLG-transfected mice at 48 h, but was detectable at 7, 14 and 21 days after transfection. These data demonstrate that PLG-mediated gene transfer can produce prolonged gene expression in airway epithelia. However, gene transfer efficiency still requires significant improvement.