The use of cationic lipids for gene transfer to airway epithelia has shown promise in in vitro and in vivo studies. However, previous studies have used a wide variety of different lipid preparations and different formulations. Few studies have been designed to optimize the variables involved in transfection, and none have been focused on airway epithelia. Therefore we examined variables that affect cationic lipid-mediated transfection of HeLa cells and of airway epithelial cells grown on permeable filter supports at the air-liquid interface. To quantitate expression of cDNA, we assayed expression of luciferase. We found that the ratio of DNA to lipid was an important variable that determined transfection efficiency. In both HeLa and airway epithelial cells, the optimum charge ratio of cationic lipid to anionic DNA was similar to 1.25, consistent with the notion that a positively charged complex facilitates interaction with the negatively charged cell membrane. After testing a series of readily available cationic lipids, we found that 1,2-dimyristyloxypropyl-N,N-dimethyl hydroxyethyl ammonium bromide (DMRIE)I dioleoyl phosphatidylethandamine (DOPE) appeared to show good efficacy. The concentration of DNA and cationic lipid also played an important role: in HeLa cells the optimum concentration of cationic lipid was similar to 5 mu M and in airway epithelial cells was similar to 15 mu M. Higher concentrations appeared to increase toxicity in HeLa cells, but in confluent airway epithelia the DNA-lipid complex had minimal effects on total cellular protein or on transepithelial resistance. Expression increased as the duration of exposure to the DNA-lipid complex increased, with prolonged times (24 h) providing greater expression than 4-6 h. These studies using an in vitro model of airway epithelium may help guide attempts to develop in vivo gene transfer for airway disease associated with a number of diseases including cystic fibrosis, surfactant protein B deficiency, and alpha(1)-antitrypsin deficiency.