On the clonal origin of tumors: a review of experimental models

A number of current models of carcinogenesis postulate that rare events are critical in the formation of malignant neoplasms. The most fundamental prediction of the hypothesis that these events are rare is the clonality of neoplasms. Evidence from spontaneous neoplasms supports the contention that such neoplasms are clonal. However, there are a number of reasons to believe that spontaneous neoplasms are derived from large numbers of cells. The issue can be resolved with experimental cancer models. In order to determine whether experimentally induced neoplasms are derived from the clonal expansion of single cells, a variety of benign and malignant tumors can be induced in mosaic rodents. These animals comprise 2 genetically distinguishable cell lineages. If the neoplastic tissues obtained from mosaic animals are composed entirely of progeny of one or the other of the 2 cell lineages, it may be concluded that they are clonal. If, on the other hand, the neoplasms began from the proliferation of many cells, then neoplastic masses would be expected to contain cells of both lineages. The results from a number of these experiments have led to the conclusion that chemically induced neoplasms are clonal. Furthermore, malignant neoplasms are generally believed to develop in a stepwise manner. If they were derived from a single cell, then each of the stages leading to the formation of the cancer should be clonal. A variety of stages thought to be necessary precursors of cancer have been analyzed in mosaic rodents. These preneoplastic lesions have been determined to be clonal in origin. Thus, theories of carcinogenesis must account for the rarity of the events critical to the formation of cancer.