MLH1 Promoter Hypermethylation: Development and Preliminary Validation of a Methylation-Specific High Resolution Melt Curve Analysis (MS-HRM) Assay for use in a Lynch Syndrome Pre-Screen Pathway


Year:

Session type:

Theme:

Diana Pelka1,Sasha Hansel1,Ben Poskitt1,David Moore1,Phil Bennett1,Gareth Gerrard1
1Sarah Cannon Molecular Diagnostics, HCA Healthcare UK

Abstract

Background

Lynch Syndrome (LS) is associated with germline mutations in genes encoding the mismatch repair proteins, leading to mismatch repair deficiency (dMMR). dMMR can also arise somatically by methylation mediated silencing of the MLH1 gene promoter. In colorectal cancer, there is a strong association between the BRAFVal600Glu mutation and MLH1 promoter hypermethylation, thus analysis of these markers, along with microsatellite instability (MSI), constitute components of the NICE-mandated LS pre-screen pathway. In the absence of a CE-IVD solution, we sought to develop and validate an MS-HRM based, MLH1 promoter methylation assay for clinical use, using primers and control material from MethylDetect, and bisulphite conversion and HRM-PCR mastermix kits from Qiagen.

Method

Paired tumour/normal DNA was extracted from 20 FFPE samples: 12 colorectal, 4 prostate, 2 endometrial, 1 bladder, 1 unknown. Ten samples were MSH-H, 10 MSS; 5 were known dMMR, 2 MMR-normal; 4 were known BRAFVal600Glu, 10 BRAFWT; 1 from known LS, 3 from suspected LS. 20μL (5ng/μL) DNA was bisulphite converted (bsDNA) and 2-3μL used per HRM reaction. These were run on a Qiagen Rotor-Gene Q with 2-4 technical replicates for 45-50 cycles, with a melt start of 65-69oC and finish of 83-95oC. The paired-normal sample was used as the unmethylated baseline on a per-sample basis.

Results

Repeated optimisation runs showed that 3μL bsDNA input, with 4 technical replicates, x50 cycles, and 69-83oC melt were the optimal parameters. All 11 MSS samples scored as unmethylated, as did the 4 known and suspected LS samples. The remaining 6 MSI-H samples (including 3/4 BRAFVal600Glu) scored as hypermethylated. MethylDetect control samples showed a limit-of-detection of 1% methylation.

Conclusion

All validation samples scored as expected, except for one BRAFVal600Glu sample that scored unmethylated; this however came from a suspected LS patient and thus warrants further investigation. This assay proved to be rapid, cost-effective, sensitive, and specific.