Molecular Genetics of Hereditary Ovarian Cancer
Molecular Genetics of Hereditary Ovarian Cancer
Epithelial ovarian carcinoma is by far the most lethal of all gynecologic cancers. An estimated 26,000 new cases are diagnosed annually, and 14,000 deaths are attributable to this malignancy each year. The low overall 5-year survival rate of 37% is largely a reflection of the difficulty in diagnosing the disease at an early stage; nearly 80% of epithelial ovarian carcinomas have extended beyond the ovary at diagnosis.
Many of the established risk factors for epithelial ovarian carcinoma are related to the general phenomenon of incessant ovulation. These include nulliparity, an increased number of uninterrupted ovulations, early menarche, and late menopause. Conversely, factors that interrupt ovulation, such as oral contraceptive use and multiparity, lessen risk.
Incidence rates for ovarian cancer are highest in the United States and other affluent nations, suggesting the importance of environmental and/or lifestyle factors, such as diet, in the etiology of this tumor. Published evidence on these risk factors is inconclusive and controversial, however. Thus, it is likely that only a fraction of ovarian cancer risk is attributable to established factors.
The greatest insights into the etiology of ovarian cancer are likely to derive from an increased understanding of its molecular genetic features. All cancers are genetic in origin, in the sense that the driving force of tumor development is genetic mutation. A given tumor may arise through the accumulation of mutations that are exclusively somatic in origin, or through the inheritance of a mutation or mutations through the germ-line, followed by the acquisition of additional somatic mutations. These two genetic scenarios distinguish what are colloquially referred to as sporadic and hereditary cancers, respectively.
Although the neoplastic phenotype is derived, in part, from epigenetic alterations in gene expression, the sequential mutation of cancer-related genes, with their subsequent selection and accumulation in a clonal population of cells, determine (1) whether a tumor develops and (2) the time required for its development and progression. The data to support this multistep, multigenic paradigm are extensive, but perhaps the most compelling evidence is that the age-specific incidence rates for most human epithelial tumors increase at roughly the fourth to eighth power of elapsed time. This suggests that a series of four to eight genetic alterations are rate-limiting for cancer development.
To date, genetic alterations in cancer cells have been described in two major families of genes: oncogenes and tumor-suppressor genes. Genes from both classes must be mutated for cancers to occur.
Oncogenes result from gain-of-function mutations in their normal cellular counterpart proto-oncogenes and behave in a dominant fashion at the cellular level; ie, cell proliferation or development of the neoplastic phenotype is stimulated following the mutation of only one allele. The most common types of oncogene mutations are gene amplification, point mutation, and translocation, and these are nearly always acquired somatically during tumorigenesis. Two known exceptions are the RET and MET genes, inherited mutations of which predispose individuals to multiple endocrine neoplasia type II and papillary renal carcinoma, respectively.
The protein products of tumor-suppressor genes normally function to inhibit cell proliferation and are inactivated through loss-of-function mutations. Knudsons two-hit model established the paradigm for tumor-suppressor gene recessivity at the cellular level, wherein both alleles must typically be inactivated to elicit a phenotypic effect. The most common mutations observed in tumor-suppressor genes are point mutations, either missense or nonsense; microdeletions or insertions of one or several nucleotides causing frameshifts; and large deletions. A mutation in one allele, whether germ-line or somatic, is then revealed following somatic inactivation of the homologous wild-type allele.
In theory, the same spectrum of mutational events could contribute to inactivation of the second allele, but what is typically observed in tumors is homozygosity or hemizygosity for the first mutation, indicating loss of the wild-type allele. As originally demonstrated for the retinoblastoma susceptibility gene, loss of the second allele may occur through mitotic nondisjunction or recombination mechanisms or through large deletions.
This so-called loss of heterozygosity has become recognized as the hallmark of tumor-suppressor gene inactivation at a particular genomic locus. Most hereditary cancer syndromes, including those described in this review, are associated with inherited mutations in tumor-suppressor genes.
Of all of the known risk factors for ovarian history, other than age, a positive family history confers the greatest risk for developing the disease. Consistent with this observation are epidemiology-based estimates that about 10% of all epithelial ovarian carcinoma cases result from a hereditary predisposition, with the germ-line inheritance of a mutant gene conferring autosomal-dominant susceptibility with variable penetrance.
Extraordinary progress has been made recently in the identification of the molecular basis for essentially all of these manifestations of ovarian carcinoma (Table 1), allowing hereditary ovarian cancer incidence to be estimated directly (Table 2). These estimates are somewhat higher than those based on genetic epidemiologic studies of familial clustering of ovarian cancer, because the molecular analyses of unselected ovarian cancer cases allow for the detection of apparently low-penetrance hereditary cancer cases that are not associated with family histories of cancer.
It should also be stressed that other genetic variants likely exist that confer predisposition with low penetrance, in a Mendelian-recessive fashion, or through interactions with other susceptibility loci. However, at present, little is known about these types of genes. Thus, it is likely that an even larger fraction of ovarian cancers will eventually be regarded as hereditary in origin. Nevertheless, for the purposes of this discussion, hereditary cancer will refer to those cases associated with Mendelian-dominant, highly penetrant loci.
Epidemiologic studies and detailed analyses of familial ovarian cancer pedigrees have consistently confirmed the existence of two distinct manifestations of hereditary ovarian cancer: (1) the breast and ovarian cancer syndrome, in which both cancers are seen in excess and in some cases manifest in the same individual; and (2) ovarian cancers associated with an excess of colorectal and endometrial cancers that define the hereditary nonpolyposis colorectal cancer (HNPCC) syndrome.
It has been hypothesized that a site-specific form of ovarian cancer may represent a third distinct manifestation of hereditary disease. This hypothesis is based on the description of families that contain multiple cases of ovarian cancer and no apparent excess of breast cancer. However, genetic linkage analyses have failed to demonstrate linkage of these families to any locus other than the breast and ovarian cancer susceptibility gene BRCA1 on chromosome 17q12-21, suggesting that these kindreds likely represent a variant manifestation of the breast and ovarian cancer syndrome in which early-onset breast cancer is rare or has not yet appeared.
Breast and Ovarian Cancer Syndrome
The breast and ovarian cancer syndrome accounts for 85% to 90% of all familial ovarian cancer cases. The probable genetic relationship of these two malignancies in a hereditary context has been demonstrated in population-based, case-control epidemiologic studies. Following the original report of genetic linkage of early-onset breast cancer families to the BRCA1 locus, some breast and ovarian cancer families were found to demonstrate linkage to BRCA1 as well. This finding has been extended such that it is now clear that most (76% to 92%) breast and ovarian cancer families are linked to BRCA1.
The variable estimates of linkage probably result from genetic heterogeneity. Lower estimates are obtained if all families, including those with cases of male breast cancer, are considered, while higher estimates are obtained if families with cases of male breast cancer or fewer than two cases of ovarian cancer are excluded.
Most of the breast and ovarian cancer families that are not linked to BRCA1, especially those with cases of male breast cancer, are linked to the more recently described BRCA2 locus on chromosome 13q12-13. The incidence of ovarian cancer compared to breast cancer appears to be lower in BRCA2-linked than in BRCA1-linked families, however, raising questions about the penetrance of this gene for ovarian cancer.
Hereditary Nonpolyposis Colorectal Cancer Syndrome
Epithelial ovarian carcinoma is also recognized as a component of the HNPCC syndrome. Known previously as the cancer family syndrome and Lynch syndromes I and II (depending on the absence or presence of extracolonic malignancies, respectively), HNPCC is an autosomal dominant genetic syndrome characterized by three or more first-degree relatives with colon or endometrial cancer, at least two of whom were diagnosed with colon cancer at age 50 or younger. In addition to cancers of the colon and endometrium, HNPCC family members are at an increased risk for cancers of other gastrointestinal sites, the upper urologic tract, and the ovary.
Limited data on the risk of ovarian cancer in these families indicate a 3.5-fold increase in the number of cases observed over that expected in the general population; HNPCC has been estimated to account for 10% to 15% of all familial ovarian cancer cases. Significant heterogeneity in ovarian cancer frequency is seen among HNPCC families, a finding that suggests genetic heterogeneity. Consistent with this observation are linkage data indicating at least three genetic loci that contribute to the HNPCC phenotype.[11,12]
The cloning and characterization of the genes responsible for HNPCC have led to a better understanding of the etiology of HNPCC-associated tumors and have opened the door to the possibility of genetic screening for this disorder. However, no significant insights into the factors controlling the penetrance and tissue specificity of various mutations have yet emerged.