In vitro transformation systems in experimental carcinogenesis: Towards an integrated approach

Abstract

The development of cancer in man is a multistep process. Estimates of overall carcinogenic risk from exposure to specific carcinogens can be made from epidemiologic or experimental studies. However, these estimates reflect cumulative risks for all of the individual changes involved in cancer development. Although the overall involvement of a specific agent in carcinogenesis can be determined from in vivo studies, its role in the induction or development of the individual changes in this multistep process can not generally be determined. In contrast, assays that measure the transformation of cells in vitro measure individual (phenotypic) steps involved at different stages ofmultistep carcinogenesis (e.g., preneoplastic transformation of normal cells, neoplastic transformation of preneoplastic cells, malignant transformation of neoplastic cells). While these assays can provide information on the risk or mechanism for specific steps in the carcinogenic process, each endpoint must be integrated with observations made in vivo to be useful in understanding the risk or mechanism for tumor development. A number of in vitro and in vivo systems have been used to estimate the carcinogenic risk from exposure to carcinogens and to provide information on the mechanisms involved in tumor development. However, as noted above, these systems are limited in the amount and type of information they can provide. Experimental approaches that take advantage of the strengths of both in vivo and in vitro systems are needed. These approaches would bridge the gap between the extremes of whole-animal carcinogenesis and in vitro transformation. Several "in vivo/in vitro" systems that bridge this gap, involving the use of mammary epithelial cells [-2, 7], skin keratinocytes [8, 10], or tracheal epithelial cells [-5, 9], have been described. In these systems, carcinogenesis or transformation studies can be carried out in vivo, yielding benign lesions or malignant tumors or in vitro, yielding preneoplastic or neoplastic cells. Two additional properties of these experimental systems make them particularly useful for bridging the gap between whole-animal and single-cell studies. First, cells that have been characterized in vitro can be examined in a nearnormal in vivo environment through the use of repopulation assays in which