Dichloromethane metabolism to formaldehyde and reaction of formaldehyde with nucleic acids in hepatocytes of rodents and humans with and without glutathione S-transferase T1 and M1 genes.

Metabolism of dichloromethane (DCM) to formaldehyde (HCHO) via a glutathione S-transferase (GST) pathway is thought to be required for its carcinogenic effects in B6C3F1 mice. In humans, this reaction is catalyzed primarily by the protein product of the gene GSTT1, a member of the Theta class of GST, and perhaps to a small extent by the protein product of the gene GSTM1. Humans are polymorphic with respect to both genes. Since HCHO may bind to both DNA and RNA forming DNA-protein crosslinks (DPX) and RNA-formaldehyde adducts (RFA), respectively, these products were determined in isolated hepatocytes from B6C3F1 mice, F344 rats, Syrian golden hamsters, and humans to compare species with respect to the production of HCHO from DCM and its reaction with nucleic acids. Only mouse hepatocytes formed detectable amounts of DPX, the quantities of which corresponded well with quantities of DPX formed in the livers of mice exposed to DCM in vivo [Casanova, M., Conolly, R.B., and Heck, H. d'A. (1996). Fundam. Appl. Toxicol. 31, 103-116]. Hepatocytes from all rodent species and from humans with functional GSTT1 and GSTM1 genes formed RFA. No RFA were detected in human cells lacking these genes. Yields of RFA in hepatocytes of mice were 4-fold higher than in those of rats, 7-fold higher than in those of humans, and 14-fold higher than in those of hamsters. The RFA:DPX ratio in mouse hepatocytes incubated with DCM was approximately 9.0 +/- 1.4, but it was 1.1 +/- 0.3 when HCHO was added directly to the medium, indicating that HCHO generated internally from DCM is not equivalent to that added externally to cells and that it may occupy separate pools. DPX were not detected in human hepatocytes even at concentrations equivalent to an in vivo exposure of 10,000 ppm; however, the possibility that very small amounts of DPX were produced from DCM cannot be excluded, since HCHO was formed in human cells. Maximal amounts of DPXliver that might be formed in humans were predicted from the amounts in mice and the relative amounts of RFA in hepatocytes of both species. With predicted DPXliver as the dosimeter, the unit risk, the upper 95% confidence limit on the cancer risk, and the margin of exposure were calculated at several concentrations using the linearized multistage and benchmark dose methods. Since the actual delivered dose is smaller than that predicted, the results suggest that DCM poses at most a very low risk of liver cancer to humans.