The Rb gene and the negative regulation of cell growth.

I N RECENT YEARS, several lines of evidence have converged to support the idea that loss of genetic information is a critical step in the pathogenesis of many types of tumors. According to this model, cells shed certain critical growth regulatory genes during their evolution from normalcy to malignancy. Such growth regulators may normally act to constrain cellular proliferation; once lost, normally existing barriers to cell growth are removed, leading in turn to the unregulated proliferation of neoplasia. Such carcinogenic mechanisms involving loss of genetic information stand in stark contrast to the lessons taught over the past decade by the field ofoncogene research. More than a dozen oncogenes have been associated with human tumors. Without exception, these oncogenes derive from pre-existing normal cellular genes, the proto-oncogenes, which have been implicated in regulation of normal cell growth and differentiation.’ The genetic changes that convert normal protooncogenes into oncogenes invariably involve alterations that potentiate the activity of the normal genes, creating deregulated, hyperactive alleles. It is these hyperactive genes that then drive malignant proliferation. Somatic cell genetics has provided the first body of evidence indicating the importance of loss of genetic information from tumor cell genomes. The results all stem from a simple experimental plan involving the fusion of tumor cells with more normal, nontumorigenic cells. The resulting cell hybrids, which carry chromosomes from both parents, are generally nontumorigenic. This has led to the conclusion that the normal partners in the fusion are supplying genes that reimpose normal growth control on the malignant cell.2 Various control experiments have been performed to demonstrate that this loss of tumorigenicity is not attributable to trivial mechanisms deriving from tissue or species incompatibility between the two parent cells. Ultimately the most persuasive experiments are those showing that reversion from tumorigenicity correlates with the presence in the hybrid cell of a specific chromosome orginating from the normal parent cell. It would seem that this normal chromosome carries a specific gene (or genes) that imposes normal growth control on the malignant cell. Such genes are presumed to replace similarly acting genes that were lost previously during the genetic events that led initially to the creation of the malignant partner in the cell fusion. This point is made most directly by experiments in which single chromosomes are transferred individually from a normal cell to a malignant recipient. Such an experiment represents a more focused experimental strategy than earlier manipulations in which an entire complement of normal chromosomes are comingled with those of the tumor cell. For example, such experiments have shown that a normal chromosome 1 1 causes loss of tumorigenicity when introduced into cells of a Wilms’ tumor line.3 Experiments using other single-chromosome transfers are in progress. A second line of evidence pointing toward genetic loss during tumorigenesis derives from karyotypic studies of specific human tumors. A decade ago, study of chromosomes prepared from retinoblastoma tumor cells revealed interstitial deletions affecting the q14 band of chromosome 13. Analogous experiments on Wilms’ tumor demonstrated a specific loss ofgenetic material associated with the p13 band of chromosome I I In the case of retinoblastoma, this karyotypic abnormality suggested the loss of a hypothetical gene, termed Rb, associated with the 13q14 chromosomal band. Insight into the origins of retinoblastomas was extended by others who addressed the fate of the surviving intact Rb allele on the paired chromosome I 3. A series of indirect but elegant genetic analyses made it clear that this allele is also lost during the formation of a retinoblastoma.6’7 Taken together, this body of work strongly suggested that both copies of the I 3q I 4-associated Rb gene need be lost or inactivated in order to trigger tumorigenesis. The observed loss/inactivation of both copies of the Rb gene during tumorigenesis carried a number of implications about the function of the normal Rb alleles. Most important, this result indicated that either copy of the Rb gene can serve effectively to control normal cell proliferation. However,