What Was the Evidence for Tumor Suppressors?

The apparent dominant transmission of cancer traits is paradoxical in light of three observations. First, hybrid cells formed from the experimental fusion of highly malignant tumor cells with normal cells are not usually tumorigenic, suggesting that the normal phenotype is dominant in the presence of tumorigenic mutations. Furthermore, the occasional hybrid cell that regains tumorigenicity in these experiments has lost specific chromosomes originally contributed by the normal cell, implying that it is not gain of a dominant cancer trait but specific chromosomal loss that is responsible for the tumor phenotype. Second, if a single mutation was sufficient in itself to elicit a tumor, then families segregating for autosomal dominant forms of cancer would be expected to have no normal tissue in the diseased organ. This expectation is in direct contrast to the clinical description of these tumors as focal lesions surrounded by normal, functioning tissue of the same organ. Finally, epidemiological analyses of sporadic and familial forms of several human cancers have indicated that the conversion of a normal cell to a tumor cell requires multiple events.

Retinoblastoma: The First Suppressor

 Retinoblastoma is a relatively rare tumor (1 in 20,000 births) of young children and occurs in both a sporadic and autosomal dominant inherited form. Based entirely on statistical data from epidemiology and clinical observations, several remarkable conclusions were made regarding the nature of events, leading to retinoblastoma tumor formation. First, the inherited mutation alone was not sufficient to cause the disease, since there are at least 107 retinoblast cells which are potential targets for retinoblastomas, each carrying the inherited mutation, yet on average, only three independent tumors form per affected individual. This also suggested that at a genetic level, mutations leading to retinoblastoma may be recessive, rather than dominant as suggested by the inheritance pattern. The hereditary tumors were proposed to arise through an initial germline mutation followed by a second mutation in a somatic cell. The rate at which somatic mutations occurred was similar in hereditary and sporadic cases, although sporadic tumors required two somatic mutations, each in the same retinoblast for tumor formation. Entirely consistent with this was the observation that hereditary cases usually occurred at an earlier age, were often bilateral, and had multiple tumors, whereas the sporadic cases were invariably unilateral and single tumors. Because of the small possibility that a second somatic mutation may never occur in hereditary cases, ∼5% of carriers do not develop any tumor. The nature of the two mutational targets in the genome was unknown at the time of these clinical observations, but cytogenetics and molecular genetics eventually led to the answer as well as to a general approach to other human cancers.

Analysis of the chromosome band patterns from hereditary and sporadic retinoblastoma patients revealed a deletion of chromosome 13q14 (chromosome 13, q or long arm, band 1–4), suggesting that the gene for retinoblastoma (Rb) resided somewhere within this region. DNA from hereditary tumors was then analyzed with cloned DNA probes (termed “DNA markers”) that could distinguish the two copies, or alleles, of chromosome 13 within each cell. It was found that, in tumors of affected individuals, the region containing the suspected Rb gene on chromosome 13 was present in a mutant only state. This conversion from a heterozygous state to homozygosity for the mutation was termed loss of heterozygosity (LOH) and constituted the second hit required for tumor formation in hereditary cases. Furthermore, LOH on chromosome 13q14 also occurred in sporadic retinoblastoma. These data lent strong support to the idea that retinoblastoma tumor formation occurs by the unmasking of a recessive genetic defect. The discovery that LOH occurs in other hereditary and most sporadic cancers in humans marked the simultaneous emergence of somatic cell cancer genetics and its coupling to the genetics of hereditary cancer. By identifying the region of chromosome 13q14 with the most consistent LOH in tumor DNA, the gene responsible for retinoblastoma was eventually isolated and its functionality assessed. Most importantly, the gene was shown to be mutationally inactivated in retinoblastoma tumors. When a normal copy of the gene was transferred to tumor cells, their growth and tumorigenic behavior was reduced. Thus, the conjoint application of epidemiology, cytogenetics, molecular genetics, and molecular biology allows the identification of a gene with tumor-suppressing function.

Are There Other Tumor Suppressors and What Cellular Role Do They Normally Play?

Since the first suppressor was isolated, many others have been molecularly identified. As might be expected, these represent genes whose products are involved in many different aspects of cell growth and behavior. These include regulators of the cell cycle, growth and transcriptional regulators, DNA repair enzymes, differentiation factors, elements of cell  motility, and regulators of cellular signaling. Thus, elucidation of the function and nature of tumor suppressors is not only of importance for understanding cancer etiology but also useful for dissecting normal cellular function.

Clinical Relevance