Methylation Analysis of DNA Mismatch Repair Genes Using DNA Derived from the Peripheral Blood of Patients with Endometrial Cancer: Epimutation in Endometrial Carcinogenesis

Germline mutation of DNA mismatch repair (MMR) genes is a cause of Lynch syndrome. Methylation of MutL homolog 1 (MLH1) and MutS homolog 2 (MSH2) has been detected in peripheral blood cells of patients with colorectal cancer. This methylation is referred to as epimutation. Methylation of these genes has not been studied in an unselected series of endometrial cancer cases. Therefore, we examined methylation of MLH1, MSH2, and MSH6 promoter regions of peripheral blood cells in 206 patients with endometrial cancer using a methylation-specific polymerase chain reaction (MSP). Germline mutation of MMR genes, microsatellite instability (MSI), and immunohistochemistry (IHC) were also analyzed in each case with epimutation. MLH1 epimutation was detected in a single patient out of a total of 206 (0.49%)—1 out of 58 (1.72%) with an onset age of less than 50 years. The patient with MLH1 epimutation showed high level MSI (MSI-H), loss of MLH1 expression and had developed endometrial cancer at 46 years old, complicated with colorectal cancer. No case had epimutation of MSH2 or MSH6. The MLH1 epimutation detected in a patient with endometrial cancer may be a cause of endometrial carcinogenesis. This result indicates that it is important to check epimutation in patients with endometrial cancer without a germline mutation of MMR genes.


Introduction
Lynch syndrome is an autosomal dominant syndrome characterized by early development of colorectal and endometrial cancers, with an elevated risk of ovarian and other cancers [1]. Lynch syndrome accounts for 2%-6% of cases of endometrial cancer and is caused by germline mutations in DNA mismatch repair (MMR) genes (MLH1, MSH2, MSH6, and PMS2) [2]. Genomic rearrangements in the epithelial cell adhesion molecule (EPCAM) gene can also lead to a silencing of the closely linked MSH2 gene and cause Lynch-associated cancers. These genes encode components of the mismatch repair system, which comprises two main protein complexes: the MutL homologue (MLH1, PMS2) and the MutS homologue (MSH2, MSH6). An aberration in one of these genes, and loss of these proteins, prevent appropriate repair of base mismatches produced during DNA replication.

Methylation Analysis
The methylation status of CpG islands in the MLH1, MSH2, and MSH6 promoters was analyzed by MSP. This method detects methylation of bases by a bisulfite reaction and subsequent detection by polymerase chain reaction (PCR) with primers specific to bisulfite converted methylated and unmethylated sequences. We referred to the primers and PCR conditions for MSP analysis as previously reported [21,22]. MSP analyses for MLH1, MSH2, and MSH6 are summarized in Table 2. CpGenome Universal Methylated DNA and Unmethylated DNA (Millipore, Temecula, CA, USA) were used as the respective positive controls for methylated and unmethylated PCR. The MSP products were subsequently sequenced to confirm the presence of methylation in the methylated MSP.

Germline Mutation Analysis
Mutation screening of MMR genes was performed by direct sequencing, as previously reported [4,23].

Microsatellite Instability Analysis
To detect the loss of function in MMR, five satellite markers (BAT25, BAT26, D2S123, D5S346, and D17S250) were evaluated, as recommended by the National Cancer Institute (NCI) [24]. MSI analysis was performed as previous reported [4]. PCR primers for these regions were purchased from Research Genetics (Huntsville, AL, USA). The antisense primers contained a fluorescent marker, Cy5 amidite (indodicarbocyanine), at the 5 end. AmpliTaq polymerase and AmpliTaq buffer (Perkin-Elmer, Boston, MA, USA) were used in the PCR, in which 1 µL of sample DNA (template, 0.1 µg/µL) was added to 24 µL of premixture (distilled water, 16.125 µL; 1.25 mmol/L deoxyribonucleotide triphosphate (dNTP), 4 µL; 10× BPCR buffer, 2.5 µL; optical density (OD) 2.2 forward primer, 0.625 µL; OD 2.2 reverse primer, 0.625 µL; and Taq polymerase, 0.125 µL) in a total reaction volume of 25 µL. Forward and reverse primers specific for D2S123, D5S376, D17S250, BAT25, and BAT26 regions were used. DNA denaturation at 95 • C for 5 min was followed by 40 cycles for 30 s at 95 • C, 40 s at 55 • C, and 40 s at 72 • C. The reaction mixture was then heated to 72 • C for 7 min, cooled, and stored at 4 • C. PCR products were combined with size markers, denatured for 5 min at 80 • C, and then electrophoresed in a 6% Long Ranger 7 M urea denaturing gel on an AFL red DNA sequencer (Amersham Pharmacia Biotech, Tokyo, Japan). DNA fragment sizes were analyzed using gene scanning software (Allele Links; Amersham Pharmacia Biotech). Tumors showing an allelic shift of two or more markers were referred to as high level MSI (MSI-H), those with an allelic shift of one marker as low level MSI (MSI-L), and those with no allelic shift at any marker as microsatellite stable (MSS).

Results
One of the 206 patients (0.49%) was found to have MLH1 epimutation. Figure 1 shows the MSP of this patient and the MLH1 methylation in peripheral blood lymphocytes. Direct sequencing of methylated and unmethylated MSP products from blood lymphocytes confirmed methylation in the MLH1 promoter region in this case ( Figure 2). This patient had endometrial cancer at 46 years old: endometrioid adenocarcinoma G1 with post-surgical stage IA (FIGO 2008). She had a history of colorectal cancer at 45 years old and met the revised Bethesda guidelines because of the early onset of colorectal cancer and metachronous Lynch-associated cancers (Table 3). Based on the familial clustering of Lynch syndrome-associated cancers and the age of onset, this family met the Amsterdam II criteria for Lynch syndrome. We could not definitively rule out the possibility of familial adenomatous polyposis (FAP) because the medical records of her colorectal cancer and her relatives' cancers were unavailable to us, since they were treated at other hospitals. Nevertheless, FAP seems unlikely, given the proband was not advised of any diagnosis of polyposis following her surgery for colorectal cancer. Germline genetic testing showed no mutation of MMR genes (MLH1, MSH2, and MSH6). Her family tree is shown in Figure 3. Methylation was also found in uterine cervix tissue (non-cancer region) and in a peripheral blood sample, while the cancerous region of the uterine endometrium had methylated and unmethylated alleles ( Figure 4). MSI tests with five satellite markers showed MSI-H, which indicated the possibility of deficiency in MMR ( Figure 5). IHC of MMR proteins ( Figure 6) clearly showed decreased MLH1 expression in the adenocarcinoma region. As indicated above, this case met the Amsterdam II criteria and six cases fulfilled the revised Bethesda guidelines. In patients under 50 years old, the rate of epimutation was 1 in 58 patients (1.72%). Twenty-four patients, including this epimutation case, in 206 patients were clinically recommended to be checked for MSI; seven patients showed MSI, and another 17 patients showed MSS.
Genes 2016, 7, 86 6 of 14 markers showed MSI-H, which indicated the possibility of deficiency in MMR ( Figure 5). IHC of MMR proteins ( Figure 6) clearly showed decreased MLH1 expression in the adenocarcinoma region. As indicated above, this case met the Amsterdam II criteria and six cases fulfilled the revised Bethesda guidelines. In patients under 50 years old, the rate of epimutation was 1 in 58 patients (1.72%). Twenty-four patients, including this epimutation case, in 206 patients were clinically recommended to be checked for MSI; seven patients showed MSI, and another 17 patients showed MSS.

Amsterdam II Criteria
There should be at least three relatives with a hereditary nonpolyposis colorectal cancer (HNPCC)associated cancer [colorectum, endometrium, small bowl, ureter, renal pelvis] All of the following criteria should be present: 1. One should be a first-degree relative of the other two.
2. At least two successive generations should be affected.
3. At least one should be diagnosed at <50 years of age.
4. FAP should be excluded in the colorectal cancer cases, if any.
5. Tumors should be verified by pathological examination.  Figure 5). IHC of MMR proteins ( Figure 6) clearly showed decreased MLH1 expression in the adenocarcinoma region. As indicated above, this case met the Amsterdam II criteria and six cases fulfilled the revised Bethesda guidelines. In patients under 50 years old, the rate of epimutation was 1 in 58 patients (1.72%). Twenty-four patients, including this epimutation case, in 206 patients were clinically recommended to be checked for MSI; seven patients showed MSI, and another 17 patients showed MSS.

Amsterdam II Criteria
There should be at least three relatives with a hereditary nonpolyposis colorectal cancer (HNPCC)associated cancer [colorectum, endometrium, small bowl, ureter, renal pelvis] All of the following criteria should be present: 1. One should be a first-degree relative of the other two.
2. At least two successive generations should be affected.
3. At least one should be diagnosed at <50 years of age.
4. FAP should be excluded in the colorectal cancer cases, if any.
5. Tumors should be verified by pathological examination.

Revised Bethesda Guidelines
Tumors from individuals should be tested for MSI in the following situations: 1. Colorectal cancer diagnosed in a patient who is less than 50 years of age.
2. Presence of synchronous, metachronous colorectal, or other Lynch-associated tumors * regardless of age.
3. Colorectal cancer with microsatellite instability-high (MSI-H) histology ** diagnosed in a patient who is less than 60 years of age.
4. Colorectal cancer diagnosed in one or more first-degree relatives with a Lynch-related tumor, with one of the cancers being diagnosed under age 50 years.
5. Colorectal cancer diagnosed in two or more first-or second-degree relatives with Lynch-related tumors, regardless of age.

Amsterdam II Criteria
There should be at least three relatives with a hereditary nonpolyposis colorectal cancer (HNPCC)-associated cancer [colorectum, endometrium, small bowl, ureter, renal pelvis] All of the following criteria should be present: 1. One should be a first-degree relative of the other two. 2. At least two successive generations should be affected.
3. At least one should be diagnosed at <50 years of age. 4. FAP should be excluded in the colorectal cancer cases, if any.
5. Tumors should be verified by pathological examination.

Revised Bethesda Guidelines
Tumors from individuals should be tested for MSI in the following situations: 1. Colorectal cancer diagnosed in a patient who is less than 50 years of age. 2. Presence of synchronous, metachronous colorectal, or other Lynch-associated tumors * regardless of age.
3. Colorectal cancer with microsatellite instability-high (MSI-H) histology ** diagnosed in a patient who is less than 60 years of age. 4. Colorectal cancer diagnosed in one or more first-degree relatives with a Lynch-related tumor, with one of the cancers being diagnosed under age 50 years. 5. Colorectal cancer diagnosed in two or more first-or second-degree relatives with Lynch-related tumors, regardless of age. 3. Colorectal cancer with microsatellite instability-high (MSI-H) histology ** diagnosed in a patient who is less than 60 years of age.
4. Colorectal cancer diagnosed in one or more first-degree relatives with a Lynch-related tumor, with one of the cancers being diagnosed under age 50 years.
5. Colorectal cancer diagnosed in two or more first-or second-degree relatives with Lynch-related tumors, regardless of age.

Discussion
This is the first study on the relationship of the epimutation of MMR genes in unselected endometrial cancer cases. Epimutation was first reported in 2002 in a case of colorectal cancer by Gazzoli et al., based on methylation in a single allele of the MLH1 promoter in peripheral blood DNA [9]. Multiple cases of epimutation in MMR genes have subsequently been described [10][11][12][13][14][15][16][17][18][19]. Based on these cases, the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) suggest additional tests for epimutation in some cases of colorectal cancers [26]. However, the best approach for endometrial cancer is unclear. Our results suggest that the epimutation of MMR genes should be considered in the evaluation of cases of endometrial cancers.
Screening and testing for Lynch syndrome in the NCCN Guidelines and the Japanese Clinical Practice Guidelines for Hereditary Colorectal Cancer are shown in Figures 7 and 8 [26,27]. MSI tests, or IHC, are recommended for patients with clinical information on family history, young age at onset, and Lynch-associated cancer, and patients who fulfill the Amsterdam II criteria or the revised Bethesda guidelines. In cases with MSI-H or negative staining for one or more MMR proteins, germline genetic tests should then be considered. The NCCN guidelines indicate that patients with endometrial cancer at ages less than 50 years old should be considered for testing for Lynch syndrome. In the current study, the patient with MLH1 epimutation was 46 years old, had a history of colorectal cancer (Lynch-associated cancer), had a family history of Lynch-associated cancers, and met the revised Bethesda guidelines. Additionally, she had MSI-H and IHC showed a loss of expression of MLH1. The screening strategy recommends germline genetic testing of MMR genes; however, an epimutation results in normal DNA sequencing and cannot be revealed without the use of a methylation detection technique, e.g., MSP.

Discussion
This is the first study on the relationship of the epimutation of MMR genes in unselected endometrial cancer cases. Epimutation was first reported in 2002 in a case of colorectal cancer by Gazzoli et al., based on methylation in a single allele of the MLH1 promoter in peripheral blood DNA [9]. Multiple cases of epimutation in MMR genes have subsequently been described [10][11][12][13][14][15][16][17][18][19]. Based on these cases, the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) suggest additional tests for epimutation in some cases of colorectal cancers [26]. However, the best approach for endometrial cancer is unclear. Our results suggest that the epimutation of MMR genes should be considered in the evaluation of cases of endometrial cancers.
Screening and testing for Lynch syndrome in the NCCN Guidelines and the Japanese Clinical Practice Guidelines for Hereditary Colorectal Cancer are shown in Figures 7 and 8 [26,27]. MSI tests, or IHC, are recommended for patients with clinical information on family history, young age at onset, and Lynch-associated cancer, and patients who fulfill the Amsterdam II criteria or the revised Bethesda guidelines. In cases with MSI-H or negative staining for one or more MMR proteins, germline genetic tests should then be considered. The NCCN guidelines indicate that patients with endometrial cancer at ages less than 50 years old should be considered for testing for Lynch syndrome. In the current study, the patient with MLH1 epimutation was 46 years old, had a history of colorectal cancer (Lynch-associated cancer), had a family history of Lynch-associated cancers, and met the revised Bethesda guidelines. Additionally, she had MSI-H and IHC showed a loss of expression of MLH1. The screening strategy recommends germline genetic testing of MMR genes; however, an epimutation results in normal DNA sequencing and cannot be revealed without the use of a methylation detection technique, e.g., MSP.
Our case may explain why some patients do not have germline mutations in MMR genes, despite fulfilling the Amsterdam II criteria or the revised Bethesda guidelines, and having MSI or loss of MMR proteins. Mangold et al. found germline mutations of MLH1 or MSH2 in 62% of patients who fulfilled the revised Bethesda guidelines and had MSI-H [28]. Colorectal cancer patients with suspected Lynch syndrome, who were selected for epimutation tests because of negative germline tests for MMR genes, had MSI and IHC loss of MLH1 due to epimutation at rates of 3%-10% [11,13,14,16,17,[29][30][31][32][33]. The current study suggests that similar findings may occur in endometrial cancer. MSI or IHC were not performed for all patients because of the limitation of the setting of this study and the unavailability of cancer tissue. In 24 of 206 patients, MSI was checked, as clinically recommended; seven patients showed MSI, and another 17 patients showed MSS. MLH1 epimutation was detected only from these seven patients. Further research for detecting the epimutation of MMR genes in selected groups (e.g., the MSI-H group, IHC loss of the MMR group, or both) is needed. These previous reports and our findings may support the routine evaluation of the epimutation of MMR genes in cases of suspected Lynch syndrome fulfilling diagnostic criteria, showing MSI and lacking expression of MMR proteins.  Our case may explain why some patients do not have germline mutations in MMR genes, despite fulfilling the Amsterdam II criteria or the revised Bethesda guidelines, and having MSI or loss of MMR proteins. Mangold et al. found germline mutations of MLH1 or MSH2 in 62% of patients who fulfilled the revised Bethesda guidelines and had MSI-H [28]. Colorectal cancer patients with suspected Lynch syndrome, who were selected for epimutation tests because of negative germline tests for MMR genes, had MSI and IHC loss of MLH1 due to epimutation at rates of 3%-10% [11,13,14,16,17,[29][30][31][32][33]. The current study suggests that similar findings may occur in endometrial Our case may explain why some patients do not have germline mutations in MMR genes, despite fulfilling the Amsterdam II criteria or the revised Bethesda guidelines, and having MSI or loss of MMR proteins. Mangold et al. found germline mutations of MLH1 or MSH2 in 62% of patients who fulfilled the revised Bethesda guidelines and had MSI-H [28]. Colorectal cancer patients with suspected Lynch syndrome, who were selected for epimutation tests because of negative germline tests for MMR genes, had MSI and IHC loss of MLH1 due to epimutation at rates of 3%-10% [11,13,14,16,17,[29][30][31][32][33]. The current study suggests that similar findings may occur in endometrial We also checked MSH2 and MSH6 for epimutation. No case of MSH6 epimutation has ever been described, and also there was no MSH6 epimutation in our study. No MSH2 epimutation was detected in our study either; however, methylation screening of MSH2 is no longer an appropriate analysis for detecting MSH2 epimutation. This is because MSH2 epimutation is a secondary consequence of germline deletion of the EPCAM gene, and the standard test is multiplex ligation probe amplification (MLPA) analysis to detect deletion of the 3 end terminal of the EPCAM gene [34]. Additionally, even among cases with an EPCAM deletion, hypermethylation and silencing of MSH2 is restricted to cells expressing EPCAM. This means MSH2 epimutation caused by EPCAM deletion is not always detected in blood cells because they are not epithelial cells. Additionally, carriers of EPCAM deletions show mosaic patterns of MSH2 inactivation, and these features may lead to false-negative results. While expression levels of EPCAM vary in the tissue, EPCAM deletion with consequent MSH2 epimutation has a low risk for endometrial cancer development against a high risk for colorectal cancer [35]. We could not check for EPCAM mutation and MLPA analysis for MSH2, and further investigation is needed.
Studies of mature spermatozoa from male MLH1 epimutation carriers have shown that MLH1 methylation is absent in spermatozoa, while methylation is present soma-wide [13,31,36,37]. Constitutional epimutation is the term describing the epigenetic change found in somatic tissues, but is absent or erased in the germline. Additionally, it is identified by finding hemiallelic methylation of the promoter lesion in all three embryonic germ cell lineages: endodermal (buccal and colonic epithelium), mesodermal (blood lymphocytes), and ectodermal (hair follicles) [38][39][40]. In the current study, we found the MLH1 methylation in blood lymphocytes and uterine cervix for normal tissue; however, they are both mesodermal and it would be better to check for methylation in ectoderm and endoderm to clearly determine MLH1 epimutation. Furthermore, validation of the presence and level of methylation should be performed using at least two independent methods, such as bisulfite pyrosequencing and clonal bisulfite sequencing. We used MSP to detect methylation for screening and confirmed with bisulfite sequencing; however, to clarify whether the epimutation occurred in hemiallelic or mosaic forms, bisulfite pyrosequencing should also be performed for detecting the ratio of methylation. Although most MLH1 epimutation are characterized by monoallelic methylation, in this case, it would be expected to be due to monoallelic methylation [13,16,41].
Most cases identified with a constitutional MLH1 epimutation have a personal history of Lynch syndrome-like tumors, but no family history. This is because MLH1 epimutation tends to arise de novo on the maternal allele [11,13,15,31,42]. Detecting epimutation of MLH1 may be useful for patients themselves, and for their families. The lifetime risk of cancers in patients with epimutation of MMR genes has not been examined, in contrast to Lynch syndrome. However, since MLH1 epimutation has the same effects as MLH1 germline mutation for the patients themselves, following patients with epimutation of MMR genes may be helpful for surveillance and early detection of Lynch-associated cancers. Additionally, some patients and their first-degree relatives who are suspected for Lynch syndrome with an uninformative genetic test result are needed to undergo clinical surveillance. As such, detecting epimutation of MLH1 of patients can prevent the unnecessary clinical surveillance for their families because of a low incidence of inheritance of MLH1 epimutation. In this study, the patient with MLH1 epimutation almost met the Amsterdam II criteria; she clearly had familial accumulation, indicating the possibility of inheritable constitutional epimutation. We were unable to collect DNA samples from the patient's family members; therefore, we do not know whether this epimutation was inheritable or not. This autosomal dominant transmission of MLH1 epimutation was recently described as a secondary MLH1 epimutation. This inheritable secondary MLH1 epimutation is thought to link to a germline promoter c.-27C>A or c.85G>T single nucleotide variant (SNV), or both, in European ancestry, as previously reported [36,43]. This epimutation was erased in the spermatozoa, the same as the constitutional epimutation, even though the epimutation will be apparent soma-wide in the offspring. This indicated that the MLH1 epimutation is erased in the germline and reset in each member of the next generation who inherits the c.-27C>A or c.85G>T genetic alleles, or both. It is possible the proband described here also has an underlying genetic alteration causing the epimutation, given her family history of cancer. Although sequencing of the MLH1 gene did not identify any coding mutation, it remains possible that another genetic alteration around MLH1 exists in this family, but this was not investigated in this study due to the lack of available samples and consent.
The novelty of this study is that the subjects were unselected patients with endometrial cancer, and epimutation has not previously been reported on a patient series ascertained on the basis of the development of endometrial cancer. However, a limitation of this study was that the cancers from this patient series were not systematically screened for MSI, according to universal testing schemes for Lynch syndrome. Therefore, we are unable to infer the frequency of MLH1 epimutation among cases whose cancer exhibited MSI. In patients with colorectal cancer, Castillejo et al. [17] found MLH1 constitutional epimutation in five of 847 patients who met the revised Bethesda guidelines (0.59%), but in zero of 2123 patients who did not meet the guidelines. For comparison, MLH1 germline mutation was detected in 18 of the cases that met the guidelines (2.13%). Among 106 patients who met the guidelines and showed a loss of MLH1 on IHC, five had MLH1 constitutional epimutation (4.7%) [17]. This result indicates that constitutional MLH1 epimutation may be found more frequently in a subgroup showing MSI-H or IHC loss of MMR, meeting the guidelines with endometrial cancer, or all three. Further cases and subgroup analysis are required.

Conclusions
MLH1 epimutation was detected in a single patient out of 206 (0.49%) with endometrial cancer. This finding indicates that the epimutation of MMR genes should be evaluated in cases of suspected Lynch syndrome that fulfill the revised Bethesda guidelines, show MSI, and lack expression of MMR proteins.