Evaluation of copy-number variants as modifiers of breast and ovarian cancer risk for BRCA1 pathogenic variant carriers.

Genome-wide studies of patients carrying pathogenic variants (mutations) in BRCA1 or BRCA2 have reported strong associations between single-nucleotide polymorphisms (SNPs) and cancer risk. To conduct the first genome-wide association analysis of copy-number variants (CNVs) with breast or ovarian cancer risk in a cohort of 2500 BRCA1 pathogenic variant carriers, CNV discovery was performed using multiple calling algorithms and Illumina 610k SNP array data from a previously published genome-wide association study. Our analysis, which focused on functionally disruptive genomic deletions overlapping gene regions, identified a number of loci associated with risk of breast or ovarian cancer for BRCA1 pathogenic variant carriers. Despite only including putative deletions called by at least two or more algorithms, detection of selected CNVs by ancillary molecular technologies only confirmed 40% of predicted common (>1% allele frequency) variants. These include four loci that were associated (unadjusted P<0.05) with breast cancer (GTF2H2, ZNF385B, NAALADL2 and PSG5), and two loci associated with ovarian cancer (CYP2A7 and OR2A1). An interesting finding from this study was an association of a validated CNV deletion at the CYP2A7 locus (19q13.2) with decreased ovarian cancer risk (relative risk=0.50, P=0.007). Genomic analysis found this deletion coincides with a region displaying strong regulatory potential in ovarian tissue, but not in breast epithelial cells. This study highlighted the need to verify CNVs in vitro, but also provides evidence that experimentally validated CNVs (with plausible biological consequences) can modify risk of breast or ovarian cancer in BRCA1 pathogenic variant carriers.

A novel pathogenic MLH1 missense mutation, c.112A > C, p.Asn38His, in six families with Lynch syndrome.

BACKGROUND: An unclassified variant (UV) in exon 1 of the MLH1 gene, c.112A > C, p.Asn38His, was found in six families who meet diagnostic criteria for Lynch syndrome. The pathogenicity of this variant was unknown. We aim to elucidate the pathogenicity of this MLH1 variant in order to counsel these families adequately and to enable predictive testing in healthy at-risk relatives. METHODS: We studied clinical data, microsatellite instability and immunohistochemical staining of MMR proteins, and performed genealogy, haplotype analysis and DNA testing of control samples. RESULTS: The UV showed co-segregation with the disease in all families. All investigated tumors showed a microsatellite instable pattern. Immunohistochemical data were variable among tested tumors. Three families had a common ancestor and all families originated from the same geographical area in The Netherlands. Haplotype analysis showed a common haplotype in all six families. CONCLUSIONS: We conclude that the MLH1 variant is a pathogenic mutation and genealogy and haplotype analysis results strongly suggest that it is a Dutch founder mutation. Our findings imply that predictive testing can be offered to healthy family members. The immunohistochemical data of MMR protein expression show that interpreting these results in case of a missense mutation should be done with caution.