Women who carry pathogenic variants in BRCA1 and BRCA2 have a greatly increased risk of developing breast cancer. There is a significant variation of age at breast cancer diagnosis between pathogenic variant carriers, suggesting additional factors influence disease penetrance. Extensive genome-wide association studies, facilitated by the Consortium of Investigators of Modifiers of BRCA1/BRCA2 (CIMBA), have demonstrated that single-nucleotide polymorphisms (SNPs) and small insertions/deletions (indels) that are associated with breast cancer risk in the general population are also associated with risk for BRCA1 and BRCA2 pathogenic variant carriers. However, these small variants account for less than 10% of heritable risk variation. Copy number variants (CNVs) are another important source of genetic variation that also influences cancer risk. Although previous CNV studies of BRCA1 and BRCA2 pathogenic variant carriers did not detect significant risk associations they suggested that CNVs are an important modifier of hereditary breast cancer risk and highlighted the need for larger and more comprehensive studies.
What did we do?
We conducted genome-wide CNV analyses of 15,342 BRCA1 and 10,740 BRCA2 pathogenic variant carriers. Data generated by the OncoArray Network were used to identify deletions and duplications and these variants were used to identify candidate modifier gene regions using retrospective likelihood analyses. The accuracy of variant calling was evaluated using single-gene tests and whole-genome sequencing and assessed using a database of known CNVs. These results were used to identify and prioritise candidate breast cancer risk-modifier genes for further in silico and in vitro analysis.
What did we find?
Thirty-one gene regions were identified from our retrospective likelihood analysis as potential risk modifier genes and were prioritised for further analysis. Deletions overlapping BRCA1 suggested an elevated breast cancer risk estimate compared with non-CNV pathogenic variants. By contrast, deletions overlapping SULT1A1 showed a decreased breast cancer risk in BRCA1 pathogenic variant carriers. This discovery is intriguing as SULT1A1 is an important SULT isoform that is expressed widely in human tissues and plays an important role in the metabolism, bioactivation, and detoxification of carcinogens, medications, and steroid hormones (e.g., Estrogen).
Investigating genes impacted by DNA copy number variants
The findings that women with a deletion overlapping BRCA1 were at the higher end of breast cancer risk, and women carrying a deletion overlapping SULT1A1 had decreased breast cancer risk, were investigated further.
As our genome-wide analysis did not directly compare the effect of BRCA1 deletions and BRCA1 non-deletion pathogenic variants we categorised the clinically diagnosed variants for BRCA1 and BRCA2 carriers by type (deletions, duplications, and small variants) and assessed the risk for CNV versus non-CNV pathogenic variants for BRCA1 and BRCA2, separately. Our additional analyses suggested elevated breast cancer risk for BRCA1 deletions but not BRCA2 deletions. Risk estimates were also elevated for duplications versus non-duplication pathogenic variants for both BRCA1 and BRCA2 duplication carriers; however, results for BRCA2 were less definitive. These data provide evidence that CNVs in BRCA1 are associated with a higher risk than smaller pathogenic variants.
Deletions overlapping SULT1A1 suggested a decreased breast cancer risk in BRCA1 pathogenic variant carriers. SULT1A1 was selected as a novel candidate modifier for functional analysis based on potential biological mechanisms of action and because overlapping CNVs had a population frequency above 1%. The effect of SULT1A1 deletions in BRCA1 pathogenic variant carriers was modeled in breast cell lines with and without a pathogenic BRCA1 variant. Consistent with the epidemiological data, laboratory experiments showed that reduced expression of SULT1A1 in cells containing a BRCA1 pathogenic variant was associated with reduced cellular proliferation and reduced DNA damage after treatment with DNA-damaging agents. These data provided important evidence that deletion variants in BRCA1 and SULT1A1 contribute to variable breast cancer risk in BRCA1 carriers.
Evaluating uncommon genetic variation, such as CNVs, that overlap SULT1A1 and other potential modifier genes in BRCA1 and BRCA2 pathogenic variant carriers remains a challenge. Newer and larger studies, such as the new Confluence project, may lead to improved statistical power to validate our findings as well as detect further potential CNV modifiers. Furthermore, our findings suggest that the balance between the generation of catecholestrogens and catecholestrogen sulfation may be an important mechanism for modulating breast cancer risk and is worthy of future investigation.
Our study has shown that CNVs are an important source of genetic variation that impacts cancer risk for BRCA1/2 pathogenic variant carriers. Characterising pathogenic variant type in BRCA1, and future screening for deletions overlapping SULT1A1 may be useful to develop a more comprehensive model of breast cancer risk. For example, integrating these genetic data into the CanRisk Web Tool along with family history, lifestyle/hormonal risk factors, common genetic susceptibility variants, and mammographic density may further improve breast cancer risk predictions and better inform patient decisions regarding breast cancer risk management. Additionally, if verified, future therapeutic intervention studies targeting SULT1A1 in BRCA1 pathogenic variant carriers may lead to new medical options for reducing breast cancer risk.