Retinoblastoma: importance of recessive mutations in tumorigenesis

Retinoma. Case studies

The new concept of retinoma, or retinocytoma, brings up the rare 'spontaneous regressions' long-reported in the literature. Systematic investigation of all relatives of children suffering from retinoblastoma is showing up more and more retinoma cases undetected until now. From 1975 to 1990 the authors have identified 11 retinoma cases amongst 103 retinoblastoma patients and their families. The average age of the retinoma patients is 23 years with a mean follow-up of four years. There are four bilateral cases, one of which with phthisis bulbi, three unilateral retinomas associated with retinoblastoma of the fellow-enucleated eye, and four unilateral cases. Of the 11 retinomas, seven patients have a family history of retinoblastoma. Of the seven patients of procreating age, 16 offspring, three abortions and two miscarriages are noted. Of the 16 offspring, 12 developed retinoblastoma, 11 of which bilaterally. One 21-year-old patient presented in the process of malignant transformation. A child, enucleated for retinoblastoma of one eye and showing lesions conforming to retinoma in the fellow eye at two years of age, suffered an osteosarcoma when nine years old. The authors' data show that retinoblastoma and retinoma follow the same genetic changes and consequently require the same investigation and follow-up. This study indicates a frequency of retinoma of 10% amongst retinoblastoma patients and their families which is higher than that usually quoted.

Carcinogenesis revisited

The importance of mutations in carcinogenesis is still unclear. Assuming that mutations of the genetic material are central to this problem, the number needed to give rise to a cancer cell must be established. A one-mutation theory is unsatisfactory for a number of reasons. A four-mutation model fits better and can be calculated in humans assuming that the observed endogenous tumors in 2% of the population equals the frequency of spontaneous carcinogenesis, accepting a mean mutation rate of 2 X 10(-5) mutations per gene per generation, and a production of about 7 X 10(15) cells during our whole lifetime. This model is also consistent with the observed peak incidence of cancer in children, with the hereditary aspects of some pediatric tumors, and with the usually nonhereditary mechanisms of cancer in adults.

Cloning of the esterase D gene: a polymorphic gene probe closely linked to the retinoblastoma locus on chromosome 13

The study of recessive oncogenes such as those responsible for retinoblastoma and Wilms tumor is difficult because the gene products involved are unknown and because the diseases are not associated with unique cellular or molecular phenotypes suitable for genetic manipulation. Since the gene for esterase D (ESD) is known to be tightly linked to the retinoblastoma locus (RB1) in the q14.1 band of chromosome 13, we have cloned the ESD gene from a human cDNA library by using oligonucleotides specific for a partial amino acid sequence of the purified enzyme to provide a genetic marker for further studies on retinoblastoma. The putative ESD gene codes for a message of 1.2 kilobases, which is present in all cell types examined, and maps to 13q14.1, thus confirming that it is the ESD gene. Restriction enzyme analysis reveals a restriction fragment length polymorphism with Apa I; this polymorphism results from the heterozygosity of 32% of the individuals tested and is shown to be useful in identifying carriers of the mutation responsible for retinoblastoma. A preliminary screen of 24 retinoblastoma tumors by Southern blot did not reveal any homologous deletions or rearrangements of the ESD locus.

A detailed analysis of chromosomal changes in heritable and non-heritable retinoblastoma

Full cytogenetic analysis of 27 different retinoblastoma tumors is presented. Gross aneuploidy of chromosome arms 6p and 1q were very common, being observed in 15/27 and 21/27 tumors, respectively. However, we found that chromosome 13 was rarely missing: only 3/27 had a detectable monosomy affecting 13q14. Monosomy of chromosome 13 by small deletion or rearrangement was also not observed in any of 12 retinoblastoma tumor lines analyzed detail at the 300-400 chromosome band level. A novel observation in retinoblastoma was the discovery of non-random translocations at three specific breakpoints, 14q32 (4/12), 17p12 (5/12), and 10q25 (3/12). Genomic rearrangements similar to those described involving C-myc in Burkitt lymphoma 14q+ cells could not be demonstrated in the four 14q+ retinoblastoma lines using molecular techniques, and a probe mapping to the site implicated to have an activating role in lymphoma. These data suggest that there is a target for rearrangement at 14q32 but it is not the same sequence used in some Burkitt lymphomas. Two other breakpoints (2p24 and 8q24) coincided with the mapped position of cellular oncogenes, but also failed to show a molecular rearrangement with the oncogene probes. The breakpoints, 10q25 and 17p12, are constitutional fragile sites which may predispose these regions to act as acceptors of translocations in malignant cells. One line had double minute chromosomes, and was the only one of 16 (6%) tested with the N-myc probe which had an amplification. Different tumors from single patients with multifocal heritable retinoblastoma showed independent karyotype evolution. Unilateral non-heritable tumors exhibited a high level of karyotype stability throughout both in vivo and in vitro growth. The various common patterns of aneuploidy and translocations probably confer an early selective advantage to malignant cells, rather than induce malignant transformation.