We have developed a rapid, sensitive and selective method for the determination of the cyclosporin analog PSC 833 in human and mouse plasma using cyclosporin A as internal standard. The assay uses liquid-liquid extraction with diethyl ether for sample clean-up followed by reversed-phase high-performance liquid chromatography with UV detection at 210 nm. Good peak shapes were obtained using a NovaPak Phenyl column operating at 72 degrees C. Good selectivity from endogenous compounds was achieved using a mobile phase composed of methanol-acetonitrile-water (34:34:32). The retention times of cyclosporin A and PSC 833 were approximately 7.8 and 11.7 min, respectively, with two major endogenous peaks at 9.2 and 16.7 min. Selective decreasing of the retention times of cyclosporin A and PSC 833 relative to these interferences occurring upon aging of the column was balanced by increasing the percentage of methanol relative to acetonitrile. No other late eluting peaks were present, resulting in a total analysis time of 20 min per sample. The assay performance in human plasma was good. The absolute recovery of PSC 833 after the sample clean-up step was 48+/-6%. The lower limit of quantitation was 0.05 microM using 500 microl of sample. Within the linear dynamic range of the assay (0.10-5.0 microM) the accuracy was close to 100% and within-day and between-day variation less than 7%. Because of the limited availability of blank mouse plasma, the concentration in samples from mice were determined using calibration curves constructed in human plasma. The lower limit of quantitation in mouse was 0.25 microM using 200 microl of sample. Overall, the performance of the assay in mouse plasma was somewhat less than in human plasma but accuracy and precision were within the ranges that are considered acceptable for bio-analytical assays.