Previous studies have indicated that milrinone, a specific type III phosphodiesterase inhibitor, may be able to induce chloride secretion in cystic fibrosis (CF) tissues. We have now assessed the effect of this agent in vivo on the nasal epithelium of CF mutant mice and also in the nose and lungs of human subjects with CF. Wild-type mice showed a small hyperpolarization of the nasal potential difference (PD) in response to milrinone (100 microM, 1.6 +/- 0.6 mV, n = 8, P < 0.05). In contrast, CF mice carrying either the most common human mutation of the gene for the CF transmembrane regulator (CFTR), DeltaF508 (protein mislocalized), or the G551D mutation (protein normally localized) failed to demonstrate this response. Milrinone perfused alone had no significant effect on the baseline nasal PD of human subjects without CF (14.7 +/- 4.0 mV preperfusion; 15.3 +/- 4.6 mV postperfusion), but significantly (P < 0.05) augmented the hyperpolarization induced by a subsequently perfused low-chloride solution (with milrinone, 36.8 +/- 3.0 mV, n = 6; without milrinone, 18.1 +/- 2.2 mV, n = 19). In contrast, in human subjects with CF (n = 6), milrinone alone significantly (P < 0. 05) altered the nasal baseline PD (52.2 +/- 3.3 mV preperfusion; 57. 4 +/- 4.2 mV, postperfusion) but not the subsequent responses to the low-chloride solution (with milrinone, 1.1 +/- 2.2 mV, n = 4; without milrinone, 0.6 +/- 0.5 mV, n = 28) or to isoproterenol (100 microM). In a separate study in subjects (n = 6) with the DeltaF508 mutation, nasal coadministration of milrinone with isoproterenol produced no effect in the presence of amiloride and a low-chloride solution (-0.8 +/- 0.5 mV). This was also the case in the nasal epithelium of CF subjects (n = 4) carrying at least one G551D allele (-0.3 +/- 0.8 mV). Similarly, milrinone did not hyperpolarize the PD of either the tracheal (n = 6) or segmental (n = 6) airways of CF subjects (DeltaF508) when applied topically in vivo in the presence of amiloride, isoproterenol, or adenosine triphosphate (all 100 microM) in a low-chloride solution. These data do not support the use of milrinone to induce chloride secretion in CF airways in vivo.