Geldanamycin (GM; NSC 122750), a naturally occurring benzoquinoid ansamycin, was approved for preclinical evaluation by the Decision Network Committee of the NCI's Division of Cancer Treatment on the basis of notable in vitro antiproliferative activity against human cancer cells, including prostatic carcinoma. Previously, we demonstrated that plasma levels of GM well above 0.2 micro M, the 50% growth inhibitory concentration against the NCI in vitro cancer screen, were achieved in mice and dogs treated by rapid intravenous (iv) injection with doses of 20 and 2 mg/kg, respectively. However, the short systemic duration and lack of significant oral bioavailability suggested that continuous iv (civ) infusion may be required for optimal therapeutic effects. In addition, the agent appeared to be a potent hepatotoxin in the dog, producing profound elevations in serum levels of the transaminases and other indicators of liver function. Failure to develop a formulation suitable for civ administration, due to the limited solubility of GM, precluded an evaluation of the relationship between the pattern and duration of systemic exposure to the drug and the degree of hepatotoxicity. Efforts to identify soluble ansamycin derivatives related in structure to GM with retained in vitro antitumor activity were undertaken. Two active compounds with the desired physicochemical properties, NSC 655480D and NSC 658514D, were designed and prepared. When administered by bolus iv injection to dogs at doses which provided similar plasma concentration- time profiles, NSC 655480D (4.6 mg/kg) effected greater elevations of the hepatic enzyme levels in serum than GM (2.0 mg/kg), whereas there was no evidence of toxicity with NSC 658514D (5.5 mg/kg). Therefore, NSC 658514D was selected for additional studies in which it was administered by 24 and 72 hr civ infusion. The compound exhibited no indications of toxicity when infused for 24 hr at a rate (0.41 mg/kg/hr) that achieved a steady state plasma concentration of 0.35 micro m. However, maintaining a similar infusion rate of 0.5 mg/kg/hr for 72 hr, or increasing the amount infused during 24 hr by a factor of 5, both proved to be lethal. In addition, toxicity of increasing severity developed when the 24 hr infusion of 0.4 mg/kg/hr was repeated at 7 day intervals. This was associated with a decrease in the apparent CL of the drug with each successive dose, from a value of 47.9 to 19.5 mL/min/kg during three courses of treatment. In summary, these studies have demonstrated that GM and related compounds are highly toxic in dogs exposed to plasma concentrations similar to those necessary for in vitro antitumor effects.
