MLH1 Promoter Methylation Could Be the Second Hit in Lynch Syndrome Carcinogenesis

Abstract

(1) Background: MLH1 hypermethylation is an epigenetic alteration in the tumorigenesis of colorectal cancer (CRC) and endometrial cancer (EC), causing gene silencing, and, as a consequence, microsatellite instability. Commonly, MLH1 hypermethylation is considered a somatic and sporadic event in cancer, and its detection is recognized as a useful tool to distinguish sporadic from inherited conditions (such as, Lynch syndrome (LS)). However, MLH1 hypermethylation has been described in rare cases of CRC and EC in LS patients. (2) Methods: A total of 61 cancers (31 CRCs, 27 ECs, 2 ovarian cancers, and 1 stomach cancer) from 56 patients referred to cancer genetic counselling were selected for loss of MLH1 protein expression and microsatellite instability. All cases were investigated for MLH1 promoter methylation and MLH1/PMS2 germline variants. (3) Results: Somatic MLH1 promoter hypermethylation was identified in 16.7% of CRC and in 40% of EC carriers of MLH1 germline pathogenic variants. In two families, primary and secondary MLH1 epimutations were demonstrated. (4) Conclusions: MLH1 hypermethylation should not be exclusively considered as a sporadic cancer mechanism, as a non-negligible number of LS-related cancers are MLH1 hypermethylated. Current flow charts for universal LS screening, which include MLH1 methylation, should be applied, paying attention to a patient’s family and personal history.

1. Introduction

MLH1 promoter methylation is a well-known epigenetic alteration in the tumorigenesis of colorectal cancer (CRC) and endometrial cancer (EC), the two most recurrent tumors found in Lynch syndrome (LS) patients. Lynch syndrome, also known as hereditary nonpolyposis colon cancer (HNPCC) syndrome, is an inherited syndrome (OMIM #120435, https://omim.org/entry/120435, accessed on 2 November 2023) responsible for about 3% of all colorectal cancers. It is caused by germline mutations in the mismatch repair (MMR) genes MSH2, MSH6, MLH1, and PMS2 and is associated with high-level microsatellite instability (MSI-H) in tumor tissues. Patients with LS have an elevated risk of developing CRC and endometrial cancer (EC) (up to 80% and 60%, respectively) and are at a moderate risk for developing other types of cancer such as stomach, small bowel, bladder, other urothelial, and ovarian cancers [1]. Pathogenetic variants of the MSH2 and MLH1 genes account for the most part of LS, with a frequency of 40–60% and 40–50%, respectively, while minor percentages have been found for MSH6 (10–20%) and PMS2 (2%) [1]. Very recently, Mallorca Group suggested that MMR genes cause four dominantly inherited cancer syndromes with different penetrance and expressivity [2].

Carcinogenesis related to MMR defects is not an exclusive mechanism of LS; in fact, about 15% of sporadic CRC and 30% of sporadic EC are MMR-defective [3]. MLH1 promoter hypermethylation is a well-known epigenetic alteration in the tumorigenesis of sporadic CRC and EC. In fact, non-inherited MMR-defective CRC is usually characterized by a BRAF V600E somatic mutation and/or MLH1 promoter hypermethylation [4]. As well, non-inherited MMR-defective ECs are caused by MLH1 somatic hypermethylation.

MLH1 promoter hypermethylation causes gene silencing and, as a consequence, the loss of MLH1 protein expression and microsatellite instability (MSI) in tumor tissue. Usually, MLH1 methylation is considered a somatic epigenetic mechanism that characterizes sporadic cancers, and it is used in the diagnostic flow chart based on universal tumor screening for mismatch repair defects to exclude LS [5,6]. In detail, when immunohistochemical (IHC) loss of MLH1 and PMS2 proteins is observed in a tumor, both MLH1 germline variants and epigenetic silencing should be considered. MLH1 promoter methylation and germline MLH1 variants are commonly considered two mutually exclusive mechanisms in the carcinogenesis of CRC and EC [7,8]. In detail, when somatic MLH1 hypermethylation and/or a BRAF mutation are identified in CRC, or only MLH1 hypermethylation is identified in EC, patients are excluded from the LS germinal test, as reported by Tibiletti et al. [5].

Notably, the identification of patients and their relatives affected by LS is demonstrated by several studies as a cost-effective strategy for CRC and EC prevention [9]. It is well known that both intensive surveillance and risk-reducing surgery improve the long-term survival of LS patients [1]. For this purpose, MLH1 promoter methylation and BRAF analyses are crucial to discriminate LS with respect to sporadic MLH1-defective CRCs and ECs. Nevertheless, MLH1 hypermethylation has been described in rare cases of CRC and EC in LS patients who are carriers of pathogenic MLH1 germline variants [10,11,12] or are carriers of a primary (de novo) or secondary (inherited) MLH1 epimutation [13,14,15]. Primary epimutation corresponds to pure epigenetic events, and secondary epimutation corresponds to the secondary epigenetic effect of cis-acting genetic alterations transmitted following a Mendelian inheritance pattern [16].

The co-occurrence of MLH1 methylation with germline pathogenetic variants has been recently highlighted by Moreira et al., who reported an important proportion, greater than 15%, of MLH1 hypermethylated CRC from LS patients [17]. According to these data, somatic MLH1 hypermethylation can be considered as a second hit of MLH1 silencing in LS, and, in association with MLH1 germline variants, as leading to the loss of function of MMR mechanisms and to the accumulation of errors during DNA replication.

Taking into account the pivotal role of MLH1 methylation in the current flow charts for LS identification [5,8,18,19,20,21], in this study, a series of 61 cancers including 31 CRCs, 27 ECs, 2 ovarian cancers and 1 stomach cancer were studied in order to investigate the involvement of MLH1 promoter hypermethylation in cancers of patients suspected for Lynch syndrome.

2. Patients and Methods

2.1. Patients Cohort

We retrospectively studied 61 cancers including 31 CRCs, 27 ECs, 2 ovarian cancers and 1 stomach cancer from 56 patients. As summarized in Table S1, the series included 41 women and 15 men and; the mean age of patients at diagnosis was 51.6 years (range 21–84 years). All patients were referred to the cancer genetic counselling service of ASST Sette Laghi in Varese from 2008 to 2020 to ascertain the presence of a cancer predisposition syndrome, according to the flow chart reported by the Italian Associazione Italiana Famigliarità ed Ereditarietà dei tumori Gastroenterici (AIFEG) consensus [1]. In detail, the investigated cases in this study were selected from all of the CRCs and the ECs diagnosed by the Department of Pathology at ASST Sette Laghi, and were subjected to the universal immunohistochemical analysis of the four MMR proteins and from cancers of patients referred to genetic counselling for suspected LS. Written informed consent was obtained prior to analysis for all patients who agreed to genetic testing.

The 61 cancers were selected for their absence of MLH1 expression (d-MLH1) and, only for CRCs, BRAF wild-type status. All 61 cancers had been surgically removed and were evaluated by an expert pathologist (FS). As shown in the supplementary material (Table S1), the histological types of the series included 16 not otherwise specified adenocarcinomas (ADK), 13 mucinous (MUC), and 1 signet ring cell carcinoma (SRCC) for CRCs; 18 endometrioid, 1 papillary squamous cell carcinoma (PSCC), 1 squamous carcinoma (SCC), and 5 not otherwise specified adenocarcinomas (ADK) for ECs; 1 endometrioid ovarian cancer; and 1 intestinal type stomach cancer. Histological type was not available (na) for the remaining 4 cancers.

2.2. DNA Extraction and MSI Analysis

Tumor DNA was obtained from formalin-fixed and paraffin-embedded (FFPE) tissue using three representative 8 µm sections of tumor samples. In each section, neoplastic areas were selected by an expert pathologist (FS) and manually micro-dissected to minimize contamination by normal cells. DNA was extracted using a QIAamps DNA FFPE Tissue kit (Qiagen, Hilden, Germany) or a Maxwell^® DNA FFPE Kit and an automatized Maxwell 16 system (Promega, Madison, WI, USA), according to the manufacturers’ protocols.

MSI analysis was possible in 52 out of 61 cancers. The MSI analysis was performed on DNA using a pentaplex PCR panel of mononucleotide repeats (NR-21, NR-22, NR-24, BAT25, and BAT26), which are semi-monomorphic in the Caucasian population, as reported by Suraweera et al. [22]. The amplified fragments were subjected to electrophoresis using a SeqStudio genetic analyzer (Thermo Fisher scientific, Waltham, MA, USA) and were analyzed by two independent molecular biologists (GC and NS) using GeneMapper software version 5 (Thermo Fisher scientific, Waltham, MA, USA). Microsatellite instability was scored as a high level of instability (MSI-H) when at least 2 out of 5 of the analyzed microsatellites were unstable. A low level of microsatellite instability (MSI-L) was defined when only 1 microsatellite was unstable. When no microsatellite instability was identified for either locus, the sample was scored as microsatellite-stable (MSS).

2.3. MLH1 Methylation Analysis

The analysis of MLH1 promoter methylation was performed using methylation-specific multiplex ligation-dependent probe amplification (MS-MLPA), using a SALSA MS-MLPA ME011 MMR kit (MRC-Holland, Amsterdam, The Netherlands) on the same tumoral DNA analyzed for MSI. MS-MLPA was performed according to the manufacturer’s instructions (www.mrc-holland.com, accessed on 2 November 2023). In brief, about 100 ng of DNA was hybridized for 16 h at 60 °C with methylation-specific probes, which contained an HhaI methylation-sensitive digestion site. The probes were subsequently digested with an HhaI enzyme, which digests only GCGC unmethylated sequences, and were amplified using PCR with universal FAM-labeled primers. PCR products were run on a SeqStudio genetic analyzer (Thermo Fisher scientific, Waltham, MA, USA) and checked with GeneMapper software version 5 (Thermo Fisher scientific, Waltham, MA, USA). Data analysis was carried out with Coffalyser.net software v.220513.1739 (MRC-Holland, Amsterdam, The Netherlands). According to the protocol, a sample was classified as methylated when CpG sites in the MLH1 promoter region exhibited methylation with a level of methylation higher than a 0.2 ratio. This value corresponds to the limit of the blank for each probe. All data obtained through MS-MLPA analysis were confirmed by bisulfite pyrosequencing of the Deng-C region. In detail, MLH1 bisulfite pyrosequencing addressed the methylation level of 5 CpG within the Deng-C region and was addressed using PCR amplification on bisulfite-converted DNA, followed by pyrosequencing (Qiagen, Hilden, Germany). A sample was classified as methylated when the mean of all five cytosines was higher than 10% of methylation. The 10% cut-off level was set by analyzing artificial control samples at different percentages of DNA methylation (0%, 10%, 50%, and 100%), which were prepared by mixing commercial fully methylated DNA and fully unmethylated DNA (Human WGA Methylated and Non-methylated DNA Set; Zymo Research).

Germline MLH1 methylation analysis was performed on DNA extracted from blood using an MS-MLPA ME011 MMR assay (MRC-Holland, Amsterdam, The Netherlands).

2.4. Germline MMR Analysis

Germline analysis of the MMR genes including MSH2, MSH6, MLH1, PMS2, and EPCAM was performed on blood-derived DNA as per standard procedures using Sanger sequencing, targeted NGS panel sequencing, and MLPA testing, as previously reported [23]. In detail, this approach identified both point variants together with large deletions and duplications. The identified genetic variants were divided into five classes according to the International Agency for Research on Cancer recommendations and were classified in accordance with the guidelines from Insight Classification (InSiGHT Variant Interpretation Committee: Mismatch repair Gene Variant Classification Criteria, 2018; www.insight-group.org, accessed on 2 November 2023 [24]). Class 4 and 5 variants were considered pathogenic, while class 1 and 2 variants were considered benign and not reported. Class 3 variants were considered of uncertain clinical significance (VUS).

2.5. Statistical Analysis

Statistical analysis was performed using GraphPad Prism software (version 5, San Diego, CA, USA). Qualitative variables, such as presence or absence of MLH1 methylation among cancer groups, were analyzed with the chi-square test with a 95% confidence interval (CI). p-values were considered significant when lower than 0.05.

3. Results

3.1. Somatic and Germline Analysis

Microsatellite instability (MSI) analysis was available for 52 out of 61 cancers. A high level of microsatellite instability (MSI-H) was identified in 96.1% (25/26) of the investigated CRCs and in 70.8% (17/24) of the ECs; only one EC revealed a low level of MSI (MSI-L). The two ovarian cancers revealed MSI-H patterns. MSI analysis was not possible for the stomach sample due to poor tumoral DNA quality.

Methylation analysis was performed on all samples and MLH1 hypermethylation was observed in 25 out of 61 (41.0%) d-MLH1 cancers, including 7/31 (22.6%) CRCs and 15/27 (55.5%) ECs. The two ovarian and one stomach cancers also revealed MLH1 hypermethylation. Hypermethylated tumors showed a methylation level ranging from 0.2 to 0.6, with four cases (T-09 of patient P-09; T-10 and T-11 of patient P-10 from family F-5; T-43 of patient P-38 from family F-6) showing very high levels of methylation (greater than 0.9). For these four latter cases, MLH1 methylation analysis was performed on blood DNA, and a constitutive MLH1 hypermethylation was identified in each patient.

All patients affected by d-MLH1 cancers were tested for the MLH1 and PMS2 genes in germinal setting, and pathogenic variants (class 4 and 5) of the MLH1 gene were observed in 23 patients. Of these, five patients developed multiple cancers: P-10, P-13, P-16, and P-17 had both CRC and EC, while P-18 was affected by two CRCs. Overall, 28 tumors, respectively (18 CRCs and 10 ECs), belonging to 23 LS patients were considered. Only one patient (P-38 from Family F-6) affected by CRC carried the PMS2 c.2380C>T p.(Pro794Ser) class 3 missense variant. Interestingly, the CRC of this patient showed a high level of MLH1 methylation and the analysis of the blood sample demonstrated constitutional primary MLH1 hypermethylation (supplementary Figure—Figure S1).

In Supplementary Table S1, we report all somatic and germline data for the 61 investigated cancers.

3.2. Correlation between Somatic Methylation and Germline Variants of MLH1 Gene

Figure 1 shows the distribution of somatic methylation in the 58 d-MLH1 tumors from LS patients (n = 28, Figure 1a) and from sporadic patients (n = 30, Figure 1b). Remarkably, somatic MLH1 promoter hypermethylation was identified in 3/18 (16.7%) CRCs and in 4/10 (40%) ECs of MLH1 germline pathogenic variant carriers (Figure 1a). All seven cancers showing concomitant MLH1 somatic hypermethylation and germline MLH1 variants displayed immunohistochemical loss of expression of both MLH1 and PMS2, presented with a high level of MSI, and belonged to five families (Figure 2, Figure 3 and Figure 4a). Although statistical analysis did not reveal a positive association between MLH1 hypermethylation and LS (chi-square p-value = 0.0622), a relevant proportion of CRCs and ECs in LS patients showed MLH1 hypermethylation.

Notably, different methylation patterns were observed in different cancers from the same family with the same constitutional MLH1 pathogenic variant. In detail, as described in Figure 2a, two carriers of the MLH1 c.458_462del p.(Glu153Alafs*17) variant from the same family (family F-1) had MLH1 hypermethylated EC (P-02), while her son (P-18) had two MLH1 unmethylated CRCs.

Analogously, patient P-13 (carrier of the MLH1 c.1852_1854del p.(Lys618del) variant) from family F-2 (Figure 2b) developed multiple cancers, of which two were CRCs without methylation at the MLH1 promoter and one was an MLH1 hypermethylated EC. The different patterns of methylation in patients carrying the same germline variant suggest that somatic MLH1 methylation is not linked to germinal condition in these patients, but rather that it occurs sporadically in cancer cells. Also, in family F-3 (Figure 3a), the concomitance of a germinal MLH1 variant and somatic MLH1 hypermethylation was ascertained in the index case (P-05). Overall, these constitutive and somatic conditions demonstrated that hypermethylation of MLH1 could also occur sporadically as a second hit of an inherited pattern of MLH1. MLH1 is a tumor suppressor gene, and it is well known that the loss of its function occurs as a consequence of a second hit involving MLH1, such as deletion, point mutation, or chromosome aneuploidy. Our data demonstrated that MLH1 hypermethylation could also be a second hit occurring in LS cancers, causing loss of MLH1 gene function.

Interestingly, in family F-4 (Figure 3b), the index case (P-03) carrier of the MLH1 c.1558+1G>A pathogenic variant was affected by CRC, with clonal MLH1 hypermethylation in those areas of tumors showing IHC clonal loss of MLH1 expression.

On the contrary, in family F-5, three cancers (two CRC and one EC, Figure 4a) from two members (P-09 and P-10) carrying the c.168_c.116+713del MLH1 variant revealed the same somatic methylation pattern. This family was previously described, and it was demonstrated that a large deletion of the promoter region of MLH1 caused epigenetic silencing of this gene and was defined as secondary epimutation [23].

Finally, in family F-6 (Figure 4b), a young woman (P-38) who was a carrier of a variant of uncertain significance (VUS) in PMS2 c.2380C>T p.(Pro794Ser), was affected by a CRC, with a high level of MLH1 hypermethylation that was also confirmed in the blood sample. The presence of MLH1 hypermethylation was also in the normal sample, which led us to define this condition as a constitutional MLH1 hypermethylation (MLH1 primary epimutation). Interestingly, three years later, this patient developed a complex MLH1-defective endometrial hyperplasia.

4. Discussion

It is well known that MLH1 promoter methylation, together with the presence of a BRAF V600E somatic variant and a germline pathogenic MLH1 variant, is generally considered to be a mutually exclusive mechanism in CRC carcinogenesis and that the epigenetic event is the principal mechanism of MLH1 silencing in sporadic carcinogenesis [4,7,8]. On these bases, MLH1 promoter methylation analysis is recommended in several guidelines and in universal CRC and EC screening in order to distinguish non-heritable from germinal origin of tumors [5,8,18,19,20,21]. However, the consolidated knowledge about these two mutually exclusive mechanisms was gained from the study of several CRC cohorts [7,8], whereas no data are available on ECs.

Recently, Yokoyama et al. [10] described an interesting case in which both an MLH1 germline variant and MLH1 promoter hypermethylation were observed in endometrial cancer in the same patient and concluded that LS cannot be excluded even if MLH1 promoter hypermethylation on tumoral tissue is observed. The concomitance of the MLH1 germline pathogenic variant and somatic epigenetic silencing of the MLH1 gene was also described in very rare cases of constitutional primary and secondary epimutations [14]. Thus, differentiating “true sporadic” cases with somatic MLH1 hypermethylation from seemingly sporadic cases with constitutional MLH1 methylation or somatic methylation co-occurring with a deleterious germline MLH1 variant poses a clinical and molecular diagnostic challenge.

In order to explore the relevance of the presence of MLH1 promoter methylation in LS-related tumors, we investigated 61 cases including CRC and EC in relation to their MLH1 germline condition. Interestingly, MLH1 somatic hypermethylation was observed in concomitance with germline MLH1 pathogenetic variants in CRC (16% of cases), in agreement with data reported by Moreira et al. [17], and more frequently in EC (40% of cases). The more common involvement of MLH1 hypermethylation in EC with respect to CRC is highlighted in this work for the first time, and it is supported with evidence obtained from sporadic EC analyses [25,26]. In fact, as described by Borrego et al., in endometrial cancer, which is a hormone-sensitive tissue, non-genetic factors, such as body mass index and hormonal conditions, may play a role in the instauration of somatic MLH1 methylation [25]. Nevertheless, a larger cohort of LS hypermethylated cancers is needed to validate these findings.

Pedigree analyses and methylation studies of the somatic and germinal status in the same family demonstrated that MLH1 hypermethylation may represent a constitutional alteration co-occurring with a germline in-cis genetic variant (secondary epimutation, family F-5) or could be unrelated to any variants (P-38, family F-6). This latter case (P-38) showed a high level of MLH1 methylation and was diagnosed as a primary epimutation, a germline condition of high cancer risk also in absence of Amsterdam criteria in this family [27]. Finally, MLH1 methylation in LS-associated cancers could occur as the second hit in patients carrying a pathogenic germline MLH1 variant. In fact, family F-2 showed that MLH1 methylation is present only in EC and absent in both CRCs of the same patient (P-13). In family F-1, two carriers of the same MLH1 germline variant, who were P-18 (the son) and P-02 (the mother), had two unmethylated CRCs (P-18) and a hypermethylated EC (P-02), respectively.

Moreover, even though our results were obtained from a very select cohort of patients, our findings are also confirmed by the recent literature [10,17], indicating that the MLH1 epigenetic mechanism could be involved in LS carcinogenesis as a second hit. Altogether, the co-occurrence of MLH1 hypermethylation and an MLH1 germline variant suggests that MLH1 hypermethylation should not be used to exclude LS.

Future studies are necessary to determine if the concomitant genetic and epigenetic MLH1 inactivation is more frequent in EC in comparison with CRC and to understand if high levels of tumor MLH1 promoter methylation may represent a real and useful marker of an underlying germline MLH1 methylation.

5. Conclusions

In conclusion, our data confirmed that MLH1 hypermethylation is not an exclusive mechanism of non-inherited cancers but also plays a non-negligible role in LS-related cancers, especially in EC. The co-occurrence of an MLH1 germline mutation and somatic MLH1 promoter hypermethylation was observed in different cancers of the same patient and in cancers of different patients from the same family. In these families, LS cannot be excluded even if MLH1 promoter hypermethylation is observed. Thus, current flow-charts for universal LS screening, which include MLH1 methylation testing to rule out LS, should be applied, also paying attention to a patient’s family and personal history even when somatic results are suggestive for sporadic cancer.