A possible role in chemical carcinogenesis for epigenetic, heritable changes in gene expression.

Although genetic changes are clearly important in the initiation of carcinogenesis, there is reason to think that epigenetic changes may also play a role in the process. A key feature of carcinogenesis is the long latency between exposure to carcinogenic insults and the appearance of malignancy. Thus, if epigenetic changes are to be involved, they must somehow be inherited at each cell division without the continued presence of the carcinogen. I propose that self-perpetuating changes in patterns of gene expression are a plausible mechanism for an epigenetic component of carcinogenesis. Networks of transcription factors that regulate each other's and their own expression are known to control important developmental processes, particularly the determination of entire cell lineages. An inherent property of many such autoregulatory networks is the existence of two very distinct, stable steady-states, defined in terms of the concentration of each transcription factor in the network. In this report, I present a model in which an acute carcinogen exposure is postulated to shift such a network from one steady-state to the other, effectively turning on or off the expression of at least one of the genes. Because of the autoregulatory nature of the network, this new steady-state is stably inherited at each cell division. Such changes in gene expression may ultimately contribute to the malignant phenotype if the regulatory network affects genes important in cell-cycle checkpoints, maintenance of genome stability, signal transduction, or other processes that are altered in tumor cells.