Gynecologic Manifestations of Hereditary Nonpolyposis Colorectal Cancer
Gynecologic Manifestations of Hereditary Nonpolyposis Colorectal Cancer
Hereditary nonpolyposis colorectal cancer (HNPCC) is an autosomal dominant cancer susceptibility syndrome associated with inherited defects in the DNA mismatch repair system. HNPCC family members are at high risk for developing colorectal, endometrial, and ovarian cancers. Studies of HNPCC families have helped define the important role that mismatch repair genes play in the molecular pathogenesis of endometrial and ovarian cancers. This review will describe some of the important clinical and molecular features of HNPCC-related endometrial and ovarian cancer and describe how genetic susceptibility can be identified in patients with sporadic endometrial and ovarian cancers. It is important to identify patients with HNPCC, as families of mutation carriers may benefit from genetic counseling, testing, and intensified cancer surveillance.
Hereditary nonpolyposis colorectal cancer (HNPCC) is an autosomal dominant cancer susceptibility syndrome associated with inherited defects in the DNA mismatch repair system. Initially described by Warthin in 1913 as the cancer family syndrome, HNPCC is characterized by genetic susceptibility to multiple malignancies, most commonly of the colon and endometrium (Table 1).[2,3] HNPCC is believed to account for approximately 1% to 5% of all colorectal cancers and an estimated 0.5% to 1.4% of all endometrial cancers.[5,6] The clinical features of HNPCC include earlyonset cancers, synchronous and/or metachronous malignancies, and a nearly complete (80%) penetrance of the disease. Lifetime risk for cancer in patients with inherited DNA mismatch repair mutations has been estimated as being as high as 91% for males and 69% for females. As shown in Table 2, the cumulative incidence of colorectal, endometrial, and ovarian cancer in HNPCC is significantly higher than in the general population. The majority (70%-80%) of HNPCC cases result from germline mutations in the DNA mismatch repair genes MLH1 and MSH2.[8,9] Germline mutations in other mismatch repair genes MSH6, PMS1, PMS2 are found less frequently. Loss of the normal DNA mismatch repair leads to genomic instability and confers a mutator phenotype.[10,11] The mutator phenotype is an essential feature of tumorigenesis. Mismatch repair- deficient cells rapidly accumulate somatic mutations that can affect a variety of target genes regulating cell growth and/or apoptosis. Mutations in DNA mismatch repair- deficient cells are typically single- base mismatches and insertion/ deletions that frequently occur within repeated DNA sequences. Tumors deficient in mismatch repair often exhibit instability in microsatellite DNA repeats. This tumor phenotype is called microsatellite instability (MSI). The MSI phenotype is observed in a high proportion (80%-90%) of colorectal cancers in patients with inherited (germline) mutations in MLH1 or MSH2. Since endometrial and ovarian cancer frequently occur in HNPCC families,[3,14-17] it is not surprising that defective mismatch repair plays a significant role in sporadic endometrial and ovarian tumorigenesis. Clinical diagnosis of HNPCC relies on family history, and multiple criteria have been employed to identify putative carriers of mismatch repair gene defects (Table 3). The criteria have been expanded over time to include extracolonic malignancies as well as other risk factors for genetic disease such as synchronous or metachronous tumors. Identification of HNPCC is important because families of mutation carriers may benefit from genetic counseling, testing, and intensified cancer surveillance. This review will discuss the molecular basis of HNPCC-related malignancies and will specifically focus on the clinical relevance of mismatch repair deficiency in sporadic and inherited forms of endometrial and ovarian cancer. Molecular Biology of Mismatch Repair Mismatch repair genes function to ensure fidelity of DNA replication in cell division. Replication errors can occur through mismatches produced by physical or chemical alteration of nucleotides, misincorporation of nu cleotides during replication, and genetic recombination events. There are now seven mismatch repair genes that have been investigated in cancers: MLH1, MSH2, MSH3, MSH6, PMS1, PMS2, and MLH3. Figure 1 illustrates how these proteins normally function as complexes to repair single-base mismatches or insertion/deletion type mutations. The fact that the majority of MSI-positive tumors have MLH1 or MSH2 defects supports the unique functionality of MLH1 and MSH2. The MSH2/MSH6 and MSH2/MSH3 heterodimers function as sensors, recognizing mismatched DNA, whereas the MLH1/PMS2 heterodimer initiates correction. Studies of mismatch repair gene knockout mice have provided evidence for the importance of an intact DNA repair system in protecting against cancer. Colon tumors form in MSH2 null (both alleles deleted) mice and demonstrate MSI. By contrast, MSH6 null mice tend to develop tumors later in life, and these tumors do not exhibit MSI. Partial functional redundancy between the MSH2/MSH6 heterodimer and the MSH2/MSH3 heterodimer explains the low level of MSI found in MSH3 and MSH6 mutants. Accumulation of point mutations and insertion/deletion (frameshift) mutations in repeated DNA sequences is detectable as MSI. Tumors demonstrating MSI (mismatch repair deficiency) show different polymerase chain reaction (PCR) products at multiple microsatellite sequences when compared to the normal (nontumor) DNA from the same patient (Figure 2). A consensus panel of five microsatellite markers has been established (D2S123, D5S346, D17S250, Bat 25, and Bat 26) to facilitate and standardize identification of tumors that demonstrate MSI. Mechanisms of Cancer Formation
Mismatch repair deficiency contributes to cancer formation through two distinct pathways (Figure 3, upper pathway). These pathways are likely related to each other; however, a well-defined molecular link between the two mechanisms has yet to be defined. In mismatch repair-deficient endometrial cancers, tumor-suppressor genes that contain coding sequence repeats such as PTEN (phosphatase and tensin homolog), BAX (BCL2- associated X protein), and IGF2R (insulin- like growth factor 2 receptor) may preferentially acquire mutations and ultimately lead to tumor formation.[ 22-24] With the exception of PTEN (which has been shown to be mutated in approximately 85% of hereditary endometrial cancers), mutations in tumor suppressors and cell-cycle regulatory proteins in MSIpositive endometrial cancers are infrequent.[ 22,24,26] The alternative mechanism for how loss of DNA repair contributes to tumorigenesis (Figure 3, lower pathway) is that cells deficient in mismatch repair fail to activate apoptotic pathways in spite of overwhelming DNA damage. Cell line studies provide evidence for the role of mismatch repair proteins in the cell death pathway. MSH2 null mouse embryonic fibroblasts do not undergo apoptosis in response to the DNA-damaging agent MNNG (N-methyl-N'-nitro-Nnitrosoguanidine), and overexpression of MSH2 or MLH1 can induce (rescue) apoptosis in MSI-positive or MSI-negative cells. Tumor cell selection may therefore be initiated by a failure of defective mismatch repair proteins (MLH1 and MSH2) to stimulate existing apoptotic pathways.[ 27] As noted, MSI in hereditary cancer is associated with an underlying defect in one of the DNA mismatch repair genes. Sporadic (nonfamilial) endometrial tumors, however, rarely demonstrate mutations in these genes.[29-32] In sporadic endometrial cancers, it is much more common for MLH1 to be inactivated epigenetically through promoter hypermethylation. An estimated 70% to 90% of endometrial cancers with MSI demonstrate aberrant methylation of the MLH1 promoter.[33,34] Promoter hypermethylation in endometrial cancer is associated with loss of MLH1 expression. The elevated cancer risk associated with the presence of both germline (HNPCC) coupled with the frequent somatic (epigenetic methylation) inactivation of the mismatch repair genes (MLH1) highlights the significance of these genes in tumorigenesis. Endometrial Cancer Genetic Susceptibility
Approximately 5% of all endometrial cancers result from inherited cancer susceptibility. Of inherited endometrial cancers, an estimated 0.5% to 1.4%[5,6] are HNPCC-related. HNPCC is the best-understood inherited endometrial cancer susceptibility syndrome. Endometrial cancer is the most frequent extracolonic tumor in HNPCC,[2,3,17] and the cumulative lifetime risk of endometrial cancer in HNPCC families is high- approximately 40% to 60%.[3,14] Although the majority of germline mutations in HNPCC families are in MLH1 and MSH2, families that carry germline mutations in MSH6 appear to have a preponderance of endometrial cancers regardless of whether or not they meet strict (Amsterdam) clinical criteria.[37,38] An analysis of 20 families with 146 MSH6 mutation carriers demonstrated an overall lower risk for all HNPCCrelated tumors but a significantly higher risk for endometrial cancer compared to MLH1 and MSH2 mutation carriers. Clinical criteria for diagnosing HNPCC have broadened over time (Table 3) to reflect the understanding that there is genetic risk in patients who do not meet strict (Amsterdam) clinical diagnostic criteria. There are at least three risk factors for having germline (inherited) defects in one of the mismatch repair genes. Genetic susceptibility is found in patients who develop endometrial cancer at a young age (< 60 years old),[36,40] who present with synchronous or metachronous malignancies, or whose tumors demonstrate MSI without epigenetic silencing of MLH1 (MLH1- unmethylated tumors). Women who do not meet strict clinical (Amsterdam) criteria for HNPCC but have early-onset (< 60 years old) disease appear to have increased familial clustering of malignancies. In a large phone survey of 455 women with early-onset endometrial cancer (20-54 years old) compared with over 3,000 age-matched controls, the odds ratio for endometrial cancer in a firstdegree relative was 2.8 (95% confidence interval [CI] = 1.9-4.2), and the odds ratio for colorectal cancer within the same family was 1.9 (95% CI = 1.1-3.3). Another study examined 291 early-onset (< 60 years old) endometrial cancer cases with available parental data. Nine kindreds (3.1%) met clinical criteria for HNPCC, and an additional 3% had familial clustering of malignancies. Patients with double primary malignancies of the colon and endometrium are another subgroup at-risk for inherited cancer susceptibility. In a study of 40 unrelated women with colorectal cancer and endometrial cancer, 7 patients had MLH1 or MSH2 mutations and 6 of the 7 had family histories suggestive of HNPCC. The relative risk of colorectal cancer in first-degree relatives of patients who carried germline mutations in MLH1 or MSH2 was 8.1 (95% CI = 3.5- 15.9) and the relative risk of endometrial cancer was 23.8 (95% CI = 6.4-61.0). First-degree relatives of patients with double primary malignancies without MLH1 or MSH2 defects were also at increased risk for colorectal cancer or endometrial cancer, with a relative risk of 2.8 (95% CI = 1.7-4.5) and 5.4 (95% CI = 2.0- 11.7), respectively. Another study of women diagnosed with both colon and endometrial cancer (N = 80) revealed that the relative risk of colorectal cancer before age 55 in first-degree relatives of probands who had both cancers before age 55 was 30.5 (95% CI = 18.8-46.6). In addition to clinical factors (early- onset disease and double primary malignancies), loss of mismatch repair in endometrial tumors can be helpful in identifying women with genetic susceptibility. A molecular phenotype (MSI-positive, MLH1-unmethylated) has been identified that appears to highlight carriers of germline mismatch repair mutations. Women with MSI-positive endometrial cancer without MLH1 methylation demonstrate familial clustering of malignancies, develop cancer almost 10 years earlier, and have an excess of HNPCC-associated metachronous malignancies. Clinical Significance of Mismatch Repair Deficiency
Identification of HNPCC patients is important because family members have been shown to benefit significantly from enrollment in colorectal cancer surveillance programs. Whether the loss of mismatch repair in HNPCC-related or sporadic endometrial cancers is related to tumor behavior or overall survival, however, is controversial. In HNPCC-related endometrial cancers, no difference has been seen in 5-year cumulative survival (all stages) or in histologic subtype compared to age- and stage-matched sporadic endometrial cancers. In a small study (109 tumor samples) that compared MSI-positive (N = 10 tumors with high-level MSI) to MSI-negative sporadic endometrial cancers, MSI positivity was associated with higher tumor grade and poor clinical outcome. A larger retrospective study of sporadic MSI-positive and MSI-negative endometrial cancers, however, found no difference in 5-year survival, 5-year recurrencefree survival, stage, grade, or histologic subtype. Another retrospective case/control study of 29 MSI-positive endometrial cancer patients demonstrated statistically significant improvement in 5- year survival (77% vs 48%) in a cohort of patients weighted heavily towards advanced-stage disease (> 50% of the patients had International Federation of Gynecology and Obstetrics [FIGO] stage III or IV disease). This benefit was seen in univariate and multivariate analysis. An analysis of 70 MSI-positive vs 159 MSI-negative endometrial cancers (81% completely surgically staged) found that MSIpositive tumors tend to present with early-stage disease and have a less aggressive histology. There were also no differences seen in recurrence or survival, but the median follow-up for both groups was only 24 months. Overall, it is important to recognize patients who may have inherited susceptibility to endometrial cancer, as family members will benefit from colorectal surveillance and may similarly benefit from gynecologic surveillance. The relationship between MSI and endometrial cancer biology and overall survival in HNPCC-related or sporadic endometrial cancer is not yet fully understood. A larger analysis including prospectively acquired clinical data in surgically staged patients with known molecular phenotypes would be helpful to better understand differences in clinical outcomes in these patients.