Inflammatory lung disease is associated with increased epithelial permeability, but it is unclear how inflammatory cells alter epithelial permeability. Neutrophils have azurophilic granules containing elastase, cathepsin G, and defensins which are released at sites of inflammation. Experiments using whole animals and cultured cells suggest that neutrophil elastase contributes to increased epithelial permeability. Using Madin-Darby canine kidney epithelial (MDCK) monolayers, a well-described epithelial model, we asked whether neutrophil elastase directly affects epithelial permeability independent of cell death or cell detachment from the substratum. We measured permeability using 3H-mannitol. We found that neutrophil elastase increased epithelial permeability in a time- and concentration-dependent fashion. Increased permeability required prolonged (> or = 6 h) exposure to elastase, but was not associated with cytolytic injury or cell detachment. These findings are potentially relevant to the lung because we found a similar time- and concentration-dependent effect when we added elastase to cultured human bronchial epithelial cells. In MDCK cells, permeability increased without alterations in cell actin at the light microscopic level. Interestingly, elastase-induced permeability was both prevented and reversed by serum, but not by serum albumin. Complete reversal occurred if serum was added up to 16 h after adding elastase. Proteolytic activity is important in HNE-induced epithelial permeability because soy bean trypsin inhibitor completely blocks the effect and alpha 1 proteinase inhibitor (alpha 1 PI) partially blocks the effect. Charge interactions also appear to be important because the polyanions heparin and sulfated dextran completely blocked increased permeability following elastase but only partially blocked elastolytic activity in isotonic solutions.(ABSTRACT TRUNCATED AT 250 WORDS)