Detection of allelic imbalance in MLH1 expression by pyrosequencing serves as a tool for the identification of germline defects in Lynch syndrome

Whole genome sequencing completes the molecular genetic testing workflow of patients with Lynch syndrome

Multigene panel tests (MGPTs) revolutionized the diagnosis of Lynch syndrome (LS), however noncoding pathogenic variants (PVs) can only be detected by complementary methods including whole genome sequencing (WGS). Here we present a DNA-, RNA- and tumor tissue-based WGS prioritization workflow for patients with a suspicion of LS where MGPT detected no LS-related PV. Among the 100 enrolled patients, MGPT detected 28 simple PVs and an additional 3 complex PVs. Among the 69 MGPT-negative patients, the lack of somatic MLH1 promoter methylation in a patient with a distinguished MLH1 allelic imbalance selected this sample for WGS. This returned a germline deep intronic MLH1 variant, with further functional studies confirming its' pathogenicity. Interestingly, all three complex PVs and the MLH1 deep intronic PV were found to be recurrent at our center. Our straightforward and cost-effective prioritization workflow can optimally include WGS in the genetic diagnosis of LS.

Altered chromatin landscape and 3D interactions associated with primary constitutional MLH1 epimutations

BACKGROUND

Lynch syndrome (LS), characterised by an increased risk for cancer, is mainly caused by germline pathogenic variants affecting a mismatch repair gene (MLH1, MSH2, MSH6, PMS2). Occasionally, LS may be caused by constitutional MLH1 epimutation (CME) characterised by soma-wide methylation of one allele of the MLH1 promoter. Most of these are "primary" epimutations, arising de novo without any apparent underlying cis-genetic cause, and are reversible between generations. We aimed to characterise genetic and gene regulatory changes associated with primary CME to elucidate possible underlying molecular mechanisms.

METHODS

Four carriers of a primary CME and three non-methylated relatives carrying the same genetic haplotype were included. Genetic alterations were sought using linked-read WGS in blood DNA. Transcriptome (RNA-seq), chromatin landscape (ATAC-seq, H3K27ac CUT&Tag) and 3D chromatin interactions (UMI-4C) were studied in lymphoblastoid cell lines. The MLH1 promoter SNP (c.-93G > A, rs1800734) was used as a reporter in heterozygotes to assess allele-specific chromatin conformation states.

RESULTS

MLH1 epimutant alleles presented a closed chromatin conformation and decreased levels of H3K27ac, as compared to the unmethylated allele. Moreover, the epimutant MLH1 promoter exhibited differential 3D chromatin contacts, including lost and gained interactions with distal regulatory elements. Of note, rare genetic alterations potentially affecting transcription factor binding sites were found in the promoter-contacting region of CME carriers.

CONCLUSIONS

Primary CMEs present allele-specific differential interaction patterns with neighbouring genes and regulatory elements. The role of the identified cis-regulatory regions in the molecular mechanism underlying the origin and maintenance of CME requires further investigation.

Primary constitutional MLH1 epimutations: a focal epigenetic event

Background

Constitutional MLH1 epimutations are characterised by monoallelic methylation of the MLH1 promoter throughout normal tissues, accompanied by allele-specific silencing. The mechanism underlying primary MLH1 epimutations is currently unknown. The aim of this study was to perform an in-depth characterisation of constitutional MLH1 epimutations targeting the aberrantly methylated region around MLH1 and other genomic loci.

Methods

Twelve MLH1 epimutation carriers, 61 Lynch syndrome patients, and 41 healthy controls, were analysed by Infinium 450 K array. Targeted molecular techniques were used to characterise the MLH1 epimutation carriers and their inheritance pattern.

Results

No nucleotide or structural variants were identified in-cis on the epimutated allele in 10 carriers, in which inter-generational methylation erasure was demonstrated in two, suggesting primary type of epimutation. CNVs outside the MLH1 locus were found in two cases. EPM2AIP1-MLH1 CpG island was identified as the sole differentially methylated region in MLH1 epimutation carriers compared to controls.

Conclusion

Primary constitutional MLH1 epimutations arise as a focal epigenetic event at the EPM2AIP1-MLH1 CpG island in the absence of cis-acting genetic variants. Further molecular characterisation is needed to elucidate the mechanistic basis of MLH1 epimutations and their heritability/reversibility.

Elucidating the molecular basis of MSH2-deficient tumors by combined germline and somatic analysis

In a proportion of patients presenting mismatch repair (MMR)-deficient tumors, no germline MMR mutations are identified, the so-called Lynch-like syndrome (LLS). Recently, MMR-deficient tumors have been associated with germline mutations in POLE and MUTYH or double somatic MMR events. Our aim was to elucidate the molecular basis of MSH2-deficient LS-suspected cases using a comprehensive analysis of colorectal cancer (CRC)-associated genes at germline and somatic level. Fifty-eight probands harboring MSH2-deficient tumors were included. Germline mutational analysis of MSH2 (including EPCAM deletions) and MSH6 was performed. Pathogenicity of MSH2 variants was assessed by RNA analysis and multifactorial likelihood calculations. MSH2 cDNA and methylation of MSH2 and MSH6 promoters were studied. Matched blood and tumor DNA were analyzed using a customized next generation sequencing panel. Thirty-five individuals were carriers of pathogenic or probably pathogenic variants in MSH2 and EPCAM. Five patients harbored 4 different MSH2 variants of unknown significance (VUS) and one had 2 novel MSH6 promoter VUS. Pathogenicity assessment allowed the reclassification of the 4 MSH2 VUS and 6 probably pathogenic variants as pathogenic mutations, enabling a total of 40 LS diagnostics. Predicted pathogenic germline variants in BUB1, SETD2, FAN1 and MUTYH were identified in 5 cases. Three patients had double somatic hits in MSH2 or MSH6, and another 2 had somatic alterations in other MMR genes and/or proofreading polymerases. In conclusion, our comprehensive strategy combining germline and somatic mutational status of CRC-associated genes by means of a subexome panel allows the elucidation of up to 86% of MSH2-deficient suspected LS tumors.

Lynch syndrome associated with two MLH1 promoter variants and allelic imbalance of MLH1 expression

Lynch syndrome is a hereditary cancer syndrome caused by a constitutional mutation in one of the mismatch repair genes. The implementation of predictive testing and targeted preventative surveillance is hindered by the frequent finding of sequence variants of uncertain significance in these genes. We aimed to determine the pathogenicity of previously reported variants (c.-28A>G and c.-7C>T) within the MLH1 5′untranslated region (UTR) in two individuals from unrelated suspected Lynch syndrome families. We investigated whether these variants were associated with other pathogenic alterations using targeted high-throughput sequencing of the MLH1 locus. We also determined their relationship to gene expression and epigenetic alterations at the promoter. Sequencing revealed that the c.-28A>G and c.-7C>T variants were the only potentially pathogenic alterations within the MLH1 gene. In both individuals, the levels of transcription from the variant allele were reduced to 50% compared with the wild-type allele. Partial loss of expression occurred in the absence of constitutional epigenetic alterations within the MLH1 promoter. We propose that these variants may be pathogenic due to constitutional partial loss of MLH1 expression, and that this may be associated with intermediate penetrance of a Lynch syndrome phenotype. Our findings provide further evidence of the potential importance of noncoding variants in the MLH1 5′UTR in the pathogenesis of Lynch syndrome.

Allele Quantification Pyrosequencing® at Designated SNP Sites to Detect Allelic Expression Imbalance and Loss-of-Heterozygosity

Pyrosequencing(®) is able to quantitate the level of a nucleotide at a designated germ-line or somatic variant, including single nucleotide polymorphisms (SNPs). SNPs within a gene of interest may be used to distinguish between the two genetic alleles and study their behavior in heterozygous individuals. With regard to cancer etiology and development, identification of alleles and the detection of allelic imbalances, such as transcriptional loss from one allele or loss-of-heterozygosity (due to deletion of one allele), within a tumor are particularly useful. Lynch syndrome, the most common form of hereditary bowel and uterine cancer, is caused by heterozygous germ-line mutations within the DNA mismatch repair genes and tumors develop following inactivation of the remaining functional allele within somatic tissues, usually by acquired loss-of-heterozygosity. MLH1 is the most frequently mutated gene in Lynch syndrome; however, some cases whose tumors display immunohistochemical loss of the MLH1 protein have no apparent mutation within the coding region of MLH1. Allelic loss of expression or reduced function of MLH1 can also result in the propensity to develop Lynch syndrome associated cancers. In this chapter we describe allele quantification Pyrosequencing assays designed at a common benign SNP within the MLH1 coding region for application to either DNA or mRNA (cDNA) templates, which enabled us to detect pathological allelic imbalances in such cases with suspected Lynch syndrome. Our allele quantification Pyrosequencing assays at the MLH1 c.655A > G (rs1799977) exonic SNP were applied to clinical specimens and detected both constitutional allelic expression loss and tumor loss-of-heterozygosity in some cases, facilitating the identification of the mechanistic cause underlying their cancer development. We provide detailed protocols for implementing these Pyrosequencing assays and illustrative examples of their application in patients.

Epigenetic mechanisms in the pathogenesis of Lynch syndrome

Inherited defects in the DNA mismatch repair (MMR) system, MLH1, MSH2, MSH6, and PMS2 genes, underlie Lynch syndrome, one of the most prevalent cancer syndromes in man. The syndrome offers a model for cancers arising through MMR defects and microsatellite instability, which applies to ~ 15% of all colorectal, endometrial, and other cancers. Lynch syndrome also illustrates the significance of the epigenetic component in cancer development. Inactivation of tumor suppressor genes by epigenetic mechanisms is an acquired property of many tumors developing in Lynch syndrome. Furthermore, constitutional epimutations of MMR genes may explain a proportion of mutation-negative families lacking MLH1 or MSH2 protein expression in tumor tissue. This review provides an update of the molecular basis of Lynch syndrome by focusing on the role of epigenetic mechanisms in the pathogenesis of the disease.

The MLH1 c.-27C>A and c.85G>T variants are linked to dominantly inherited MLH1 epimutation and are borne on a European ancestral haplotype

Germline mutations of the DNA mismatch repair genes MLH1, MSH2, MSH6 or PMS2, and deletions affecting the EPCAM gene adjacent to MSH2, underlie Lynch syndrome by predisposing to early-onset colorectal, endometrial and other cancers. An alternative but rare cause of Lynch syndrome is constitutional epimutation of MLH1, whereby promoter methylation and transcriptional silencing of one allele occurs throughout normal tissues. A dominantly transmitted constitutional MLH1 epimutation has been linked to an MLH1 haplotype bearing two single-nucleotide variants, NM_000249.2: c.-27C>A and c.85G>T, in a Caucasian family with Lynch syndrome from Western Australia. Subsequently, a second seemingly unrelated Caucasian Australian case with the same MLH1 haplotype and concomitant epimutation was reported. We now describe three additional, ostensibly unrelated, cancer-affected families of European heritage with this MLH1 haplotype in association with constitutional epimutation, bringing the number of index cases reported to five. Array-based genotyping in four of these families revealed shared haplotypes between individual families that extended across ≤2.6-≤6.4 megabase regions of chromosome 3p, indicating common ancestry. A minimal ≤2.6 megabase founder haplotype common to all four families was identified, which encompassed MLH1 and additional flanking genes and segregated with the MLH1 epimutation in each family. Our findings indicate that the MLH1 c.-27C>A and c.85G>T variants are borne on a European ancestral haplotype and provide conclusive evidence for its pathogenicity via a mechanism of epigenetic silencing of MLH1 within normal tissues. Additional descendants bearing this founder haplotype may exist who are also at high risk of developing Lynch syndrome-related cancers.

Identification of constitutional MLH1 epimutations and promoter variants in colorectal cancer patients from the Colon Cancer Family Registry

PURPOSE

Constitutional MLH1 epimutations manifest as promoter methylation and silencing of the affected allele in normal tissues, predisposing to Lynch syndrome-associated cancers. This study investigated their frequency and inheritance.

METHODS

A total of 416 individuals with a colorectal cancer showing loss of MLH1 expression and without deleterious germline mutations in MLH1 were ascertained from the Colon Cancer Family Registry (C-CFR). Constitutive DNA samples were screened for MLH1 methylation in all 416 subjects and for promoter sequence changes in 357 individuals.

RESULTS

Constitutional MLH1 epimutations were identified in 16 subjects. Of these, seven (1.7%) had mono- or hemi-allelic methylation and eight had low-level methylation (2%). In one subject the epimutation was linked to the c.-27C>A promoter variant. Testing of 37 relatives from nine probands revealed paternal transmission of low-level methylation segregating with a c.+27G>A variant in one case. Five additional probands had a promoter variant without an MLH1 epimutation, with three showing diminished promoter activity in functional assays.

CONCLUSION

Although rare, sequence changes in the regulatory region of MLH1 and aberrant methylation may alone or together predispose to the development of cancer. Screening for these changes is warranted in individuals who have a negative germline sequence screen of MLH1 and loss of MLH1 expression in their tumor.Genet Med 2013:15(1):25-35.

Dominantly inherited constitutional epigenetic silencing of MLH1 in a cancer-affected family is linked to a single nucleotide variant within the 5'UTR

Constitutional epimutations of tumor suppressor genes manifest as promoter methylation and transcriptional silencing of a single allele in normal somatic tissues, thereby predisposing to cancer. Constitutional MLH1 epimutations occur in individuals with young-onset cancer and demonstrate non-Mendelian inheritance through their reversal in the germline. We report a cancer-affected family showing dominant transmission of soma-wide highly mosaic MLH1 methylation and transcriptional repression linked to a particular genetic haplotype. The epimutation was erased in spermatozoa but reinstated in the somatic cells of the next generation. The affected haplotype harbored two single nucleotide substitutions in tandem; c.-27C > A located near the transcription initiation site and c.85G > T. The c.-27C > A variant significantly reduced transcriptional activity in reporter assays and is the probable cause of this epimutation.

A novel aspect of GNAS imprinting: higher maternal expression of Gαs in human lymphoblasts, peripheral blood mononuclear cells, mammary adipose tissue, and heart

The human GNAS gene is imprinted in a tissue-specific manner, being expressed primarily from the maternal allele in pituitary, thyroid, renal proximal tubules, and gonads, but is supposed to be biallelically expressed with an equal allelic expression in most other tissues. We analysed allelic expression of Gαs using Pyrosequencing. By genotyping the GNAS T393C polymorphism we quantified mRNA transcripts in lymphoblasts (Ly, n=11), peripheral blood mononuclear cells (PBMC, n=18), mammary adipose tissue (MAT, n=23) and heart tissue (HT, n=44). Allelic expression analysis revealed an unequal allelic expression (ratio maternal/total×100±SEM: 55.7±1% (95% CI 53.4-58.1%) in Ly, 56.1±0.8 (95% CI 54.5-57.7%) in PBMC, 54.5±0.8% (95% CI 53-56.1%) in MAT and 54.1±0.6% (95% CI 53-55.3%) in HT). Maternal ratio differed significantly from the mean (p<0.0001). This phenomenon may be a general feature existing in all tissues.

De novo constitutional MLH1 epimutations confer early-onset colorectal cancer in two new sporadic Lynch syndrome cases, with derivation of the epimutation on the paternal allele in one

Lynch syndrome is an autosomal dominant cancer predisposition syndrome classically caused by germline mutations of the mismatch repair genes, MLH1, MSH2, MSH6 and PMS2. Constitutional epimutations of the MLH1 gene, characterized by soma-wide methylation of a single allele of the promoter and allelic transcriptional silencing, have been identified in a subset of Lynch syndrome cases lacking a sequence mutation in MLH1. We report two individuals with no family history of colorectal cancer who developed that disease at age 18 and 20 years. In both cases, cancer had arisen because of the de novo occurrence of a constitutional MLH1 epimutation and somatic loss-of-heterozygosity of the functional allele in the tumors. We show for the first time that the epimutation in one case arose on the paternally inherited allele. Analysis of 13 tumors from seven individuals with constitutional MLH1 epimutations showed eight tumors had lost the second MLH1 allele, two tumors had a novel pathogenic missense mutation and three had retained heterozygosity. Only 1 of 12 tumors demonstrated the BRAF V600E mutation and 3 of 11 tumors harbored a mutation in KRAS. The finding that epimutations can originate on the paternal allele provides important new insights into the mechanism of origin of epimutations. It is clear that the second hit in MLH1 epimutation-associated tumors typically has a genetic not epigenetic basis. Individuals with mismatch repair-deficient cancers without the BRAF V600E mutation are candidates for germline screening for sequence or methylation changes in MLH1.