Genetic and epigenetic profiling of BRCA1/2 in ovarian tumors reveals additive diagnostic yield and evidence of a genomic BRCA1/2 DNA methylation signature

• Published in: Journal of human genetics Vol. 65; no. 10; pp. 865 - 873
• Main Authors: Aref-Eshghi, Erfan, McGee, Jacob D., Pedro, Victor P., Kerkhof, Jennifer, Stuart, Alan, Ainsworth, Peter J., Lin, Hanxin, Volodarsky, Michael, McLachlin, Catherine Meg, Sadikovic, Bekim
• Format: Journal Article
• Language: English
• Published: England Nature Publishing Group 01.10.2020; Springer Singapore
• Subjects: Adenocarcinoma; Adenocarcinoma - diagnosis; Adenocarcinoma - drug therapy; Adenocarcinoma - genetics; Adult; Aged; Antineoplastic Agents - therapeutic use; Area Under Curve; BRCA1 protein; BRCA1 Protein - genetics; BRCA2 Protein - genetics; Computer applications; Copy number; Deoxyribonucleic acid; DNA; DNA Copy Number Variations; DNA Methylation; DNA repair; DNA sequencing; DNA, Neoplasm - chemistry; DNA, Neoplasm - genetics; Epigenetics; Female; Genes, BRCA1; Genes, BRCA2; Genetic Predisposition to Disease; Genomes; Genomic analysis; High-Throughput Nucleotide Sequencing; Humans; INDEL Mutation; Middle Aged; Molecular Diagnostic Techniques; Mutation; Neoplastic Syndromes, Hereditary - diagnosis; Neoplastic Syndromes, Hereditary - drug therapy; Neoplastic Syndromes, Hereditary - genetics; Nucleotide sequence; Ovarian cancer; Ovarian Neoplasms - diagnosis; Ovarian Neoplasms - drug therapy; Ovarian Neoplasms - genetics; Ovaries; Phenotypes; Point Mutation; Poly(ADP-ribose) Polymerase Inhibitors - therapeutic use; Predictive Value of Tests; Promoter Regions, Genetic - genetics; Ribose; ROC Curve; Sensitivity and Specificity; Tumors
• ISSN: 1434-5161; 1435-232X; 1435-232X
• DOI: 10.1038/s10038-020-0780-4

Poly-ADP-ribose-polymerase inhibitor (PARPi) treatment is indicated for advanced-stage ovarian tumors with BRCA1/2 deficiency. The “BRCAness” status is thought to be attributed to a tumor phenotype associated with a specific epigenomic DNA methylation profile. Here, we examined the diagnostic impact of combined BRCA1 /2 sequence, copy number, and promoter DNA methylation analysis, and evaluated whether genomic DNA methylation patterns can predict the BRCAness in ovarian tumors. DNA sequencing of 172 human tissue samples of advanced-stage ovarian adenocarcinoma identified 36 samples with a clinically significant tier 1/2 sequence variants (point mutations and in/dels) and 9 samples with a CNV causing a loss of function in BRCA1 / 2 . DNA methylation analysis of the promoter of BRCA1 /2 identified promoter hypermethylation of BRCA1 in two mutation-negative samples. Computational modeling of genome-wide methylation markers, measured using Infinium EPIC arrays, resulted in a total accuracy of 0.75, sensitivity: 0.83, specificity: 0.64, positive predictive value: 0.76, negative predictive value: 0.74, and area under the receiver’s operating curve (AUC): 0.77, in classifying tumors harboring a BRCA1/2 defect from the rest. These findings indicate that the assessment of CNV and promoter DNA methylation in BRCA1 / 2 increases the cumulative diagnostic yield by 10%, compared with the 20% yield achieved by sequence variant analysis alone. Genomic DNA methylation data can partially predict BRCAness in ovarian tumors; however, further investigation in expanded BRCA1/2 cohorts is needed, and the effect of other double strand DNA repair gene defects in these tumors warrants further investigations.