Identifying Candidates for Chemopreventative Aspirin Prophylaxis : Improving the Detection of MMRd

Abstract

Lynch syndrome (LS) is an inherited cancer syndrome arising from germline mutations in genes of the DNA mismatch repair (MMR) system. To optimise clinical management, the efficient diagnosis of LS individuals is essential, with known carriers of LS-associated gene defects benefiting from various practices including daily aspirin prophylaxis. Current guidelines in the UK advocate the use of two techniques in the screening for LS: MMR protein immunohistochemistry and DNA microsatellite instability (MSI) testing. However, several limitations compromise the efficacy of this clinical guidance, including the restricted tumour spectrum recommended for analysis, and the ambiguous genetic diagnoses returned by existing assays. This project aimed to enhance the existing techniques for detecting mismatch repair deficiency (MMRd) and LS individuals by addressing both these issues. In considering the potential cancer spectrum for LS screening, I initially analysed 122 extracolonic cancer samples from LS gene carriers using a new MSI assay (the Newcastle Assay) to review MSI as a biomarker of MMR deficiency in these tumour types. An MSI-H classification was returned for the majority of tumours, including 80% (41/51) of those considered of the LS spectrum, but a comparatively low proportion of endometrial cancer (EC) samples (26/35 - 72%) were found with this phenotype. Further investigation of MSI in EC specifically involved the analysis of 363 well-characterised samples, from two external clinical trial cohorts, using the Newcastle Assay. In this study, the frequency of instability between the cohorts varied for samples with confirmed MMR deficiency, highlighting caveats when using MSI as a biomarker for MMRd in EC. With this approach also often failing to detect MMRd for which MSH6-deficiency is responsible, these findings support the prioritisation of IHC for LS screening of this tumour type as recently recommended by NICE, but also suggest that additional MSI testing could have clinical benefit. For the diagnosis of LS, confirmation of a germline pathogenic MMR variant is required, but this is complicated for the MMR gene PMS2 by the presence of multiple pseudogenes. To improve the analysis of this gene, I attempted to develop a sequencing-based assay using a combination of long-range PCR sequencing and Molecular Inversion Probe (MIP) technology. I established a MIP pool consisting of 42 ii exonic and 100 intronic probes to assess for sequence variants and CNVs/loss of heterozygosity respectively. In the review of samples for which PMS2 mutation was confirmed, detection of variants by the exon-tiling component of this assay was demonstrated. Validated point mutations were accurately identified, with all but one of the 22 pathogenic variant calls returned from the 66/138 samples with verified PMS2 changes. However, the proficiency of this assay for detecting copy-number variation and loss of heterozygosity remains to be analysed, curtailed by time constraints. Information derived from intronic SNPs has thus far been inconclusive, and in no samples has homozygosity or substantial deletions been indicated. Results from the assay did however suggest that LS patients may have mutations in both PMS2 alleles, and these will require further investigation. Both these studies endeavoured to further the guidance and techniques for LS screening, and subsequently the application of treatment practices such as aspirin prophylaxis. The results of MSI analysis in extracolonic tumour samples demonstrated the facility of this biomarker in various Lynch-spectrum cancers, with EC analysis highlighting conditions for its use. In contrast, development of a PMS2 assay was less successful, and alternative long-read sequencing approaches are likely to supplant it. This suggests that continued development of the assay is likely not viable, although the complexity of the results highlight the need for accurate mutation screening of this MMR gene.