The solubility, in human urine, of the major hydroxylated metabolite (M1) of an experimental cognition enhancer was characterized through a series of in vitro experiments in an effort to estimate the probability of crystalluria occurring following oral administration of the parent compound. The aim of these experiments was to determine if a safety margin existed between clinically observed urine concentrations and the solubility of M1. The mean urine concentrations of M1 in young and elderly subjects following oral administration of the parent compound at the highest doses tested, were 4865 +/- 2368 ng/mL and 2764 +/- 791 ng/mL, respectively. In vitro solubility experiments with M1 were conducted in drug-free human urine (37 degrees C) from four male and four female healthy subjects under conditions of high and low urine osmolality. Mean concentrations (n = 16) of M1 in human urine to which solid M1 was added, were 3656 +/- 621 ng/mL, 4678 +/- 1169 ng/mL and 5378 +/- 2474 ng/mL after stirring for 24, 48 and 72 h, respectively, indicating that the ex vivo mean solubility of M1 in human urine is no greater then approximately 5 microg/mL. Addition of solid M1 to urine from human subjects dosed with the parent compound resulted in mean urine M1 concentrations 23.5% greater than those observed in vivo. The results from both experiments indicated a significant overlap between urine concentrations of M1 in vivo following the highest oral administration of the parent drug and M1 solubility measured in vitro, suggesting a high potential for in vivo saturation of urine with M1 with subsequent precipitation, crystalluria, and nephrotoxicity. Consequently, the results of these studies have placed restrictions on the dose that could be administered during clinical development of this compound.