Negative selection in hepatic tumor promotion in relation to cancer risk assessment.

Mechanistic studies with phenobarbital (PB), 2,3,7,8,-tetrachlorodibenzo-p-dioxin (TCDD) and other liver tumor promoters support a general model of promotion involving negative selection where specifically-mutated cells derive a growth advantage in the presence of persistent mitosuppression. Exposure to these liver tumor promoters appears to transiently enhance hepatocyte replication, presumably via transcriptional activation of growth regulatory genes, leading to a homeostatic increase in mitoinhibitory growth factors in the liver to constrain proliferation. Transforming growth factor beta 1 (TGF-beta), a potent mitoinhibitory growth factor for hepatocytes, has been associated with the mitosuppression caused by PB and certain peroxisomal proliferators. Escape from TGF-beta mitosuppression may involve loss or alteration of function of the mannose 6-phosphate/insulin-like growth factor II (M6P/IGFII) receptor, which is required for TGF-beta 1 activation, or alterations of the TGF-beta types I, II and III signal transduction receptors. A risk assessment based on a negative selection mechanism could be conducted for tumor promotion endpoints with TCDD and compared with current approaches that implicitly regard TCDD as an initiator. Benchmark dose calculation using centrilobular induction of cytochromes P450 1A1 and 1A2 as a surrogate for periportal growth stimulation would provide a rational starting point for application of conventional safety factor approaches, similar to those used with non-cancer effects. In the future, tissue and plasma concentrations of specific growth factors, e.g. TGF-beta or hepatocyte growth factor, HGF, might be considered as more direct dose surrogates for tumor-promoting effects of xenobiotics. Uncertainty factor adjustments to a TCDD benchmark dose calculation should eventually rely on direct knowledge of regulation of specific growth regulatory genes and their receptors in relevant species and on species differences in TCDD pharmacokinetics, instead of application of default animal-to-human and interindividual uncertainty factors.