Bioassay and relative cytotoxic potency of cyclophosphamide metabolites generated in vitro and in vivo.

Cytotoxic potencies of cyclophosphamide, 4-ketocyclophosphamide, carboxyphosphamide, mechlorethamine (HN2), bis(2-chloroethyl)amine (nor-HN2), chlorambucil, acrolein, and cyclophosphamide metabolites generated in vivo and in vitro were determined via bioassay. Our bioassay procedure was to incubate the potential cytotoxic agent with Walker 256 ascites cells in vitro, inject these cells into host rats, record survival times, estimate the number of viable cells which must have been injected to account for the observed survival time, and calculate percentage cell kill from these estimates. A log-linear relationship between tumor cell kill and exposure time or drug concentration was observed. Cyclophosphamide and 4-ketocyclophosphamide were noncytotoxic, and incuba tion of the latter with a microsomal or 105,000 X g supernatant fraction did not activate it. Acrolein and carboxyphosphamide were only minimally cytotoxic. HN2, chlorambucil, and nor-HN2 all were cytotoxic; HN2 was most and nor-HN2 was least potent. The cytotoxic potency of cyclophosphamide metabolites generated in vitro and in vivo was expressed as the concentration of metabolite(s) in nor-HN2 or formaldehyde equivalents that was required to kill 90% of the tumor cells. For total cyclophosphamide metabo lites obtained after the incubation of cyclophosphamide with hepatic microsomes or with a 9000 X g supernatant fraction and from blood or urine after cyclophosphamide injection, the concentrations of drug required to kill 90% of the tumor cells were 0.087, 0.42, 0.66, and 4.5 /uM, respectively, when expressed in nor-HN2 equivalents, and were 0.035, 0.037, 0.041, and 0.17 ¿uM, respectively, when expressed in formalde hyde equivalents. Cyclophosphamide, 4-ketocyclophospha mide, carboxyphosphamide, nor-HN2, and acrolein could not account for the cytotoxic activity of the cyclophosphamide metabolite generated in vitro by hepatic microsomal mixedfunction oxidase action, since the latter was a more potent cytotoxic agent than any of these compounds, although it was less potent than HN2. Trapping of the microsome-generated metabolite with semicarbazide reduced its cytotoxic potency. These data support the contention that cyclophosphamide is activated (to a cytotoxic metabolite) primarily by mixedfunction oxidase action of the hepatic endoplasmic reticulum which oxidizes it to aldophosphamide, and that aldophosphamide is inactivated (as a cytotoxic agent) by aldehyde oxidase (EC 1.2.3.1) and/or aldehyde dehydrogenase (EC 1.2.1.3), which oxidize(s) aldophosphamide to carboxyphosphamide. In addition, the data provide estimates of the potency of aldophosphamide as a cytotoxic agent relative to that of other alkylating agents.