In vivo MRI measurements of tumor growth induced by dichloroacetate: implications for mode of action.

Dichloroacetate (DCA) is an important by-product of the chlorination of drinking water that produces liver cancer in rodents. Assessment of the risk that results from concentrations that occur in drinking water will be dependent upon the mode of action held responsible for these tumors. A study by Stauber and Bull [Stauber, A.J. and Bull, R. J (1997) Differences in phenotype and cell replicative behavior of hepatic tumors inducted by dichloroacetate (DCA) and trichloroacetate (TCA). Toxicol. Appl. Pharmacol. 144, 235-246] in mice treated with DCA demonstrated a lesion distribution that was skewed towards many small, altered foci of cells that are assumed to be precursor lesions [EPA, (1996). U.S. Environmental Protection Agency: Proposed Guidelines for carcinogen risk assessment; notice. Fed. Reg. 61, pp. 17960-10811]. The present study was designed to determine the extent to which the tumorigenic effects of DCA could be explained by its effect on tumor growth rates (i.e. tumor promoting activity). In vivo magnetic resonance imaging (MRI) allowed accurate determination of growth rates of individual lesions in mice that had been treated with DCA in drinking water at 2 g/l. Out of thirty treated mice, ten were found to have hepatic tumors detectable by MRI at 48 weeks of treatment. These tumor-bearing animals were assigned to two groups matched on the size of lesions observed by in vivo MR1. Treatment with DCA continued in one group of five mice and was stopped in the other. For both groups, tumor growth rates were determined by measuring changes in size of all lesions greater than 1 mm(3) in volume during a 14-day period. Removal of DCA treatment resulted in growth rates that could not be distinguished from zero across all lesion sizes represented in the sample. These data are in agreement with previous observations of DCAs effects on replication rates within tumors (Stauber and Bull, (1997)). Tumor growth rates observed in animals maintained on treatment decreased with lesion volume in a manner that is consistent with a stochastic Gompertz birth-death process proposed by Tan [Tan, W.Y. (1986) A stochastic Gompertz birth-death process. Stat. Prob. Lett. 4, 25-28]. Parameters of this model obtained by fitting measured growth rates were used to predict the lesion-size distribution expected after one year of DCA treatment. The shape of the predicted lesion-size distribution was similar to that observed by Stauber and Bull (Stauber and Bull, (1997)) in mice sacrificed after 40 weeks of DCA treatment. We conclude that the effects of DCA on the division and/or death rates of spontaneously initiated cells can account for the predominance of small lesions in DCA-treated animals.