Dominant‐negative pathogenic variant BRIP1 c.1045G>C is a high‐risk allele for non‐mucinous epithelial ovarian cancer: A case‐control study

BRIP1 is a moderate susceptibility epithelial ovarian cancer (EOC) gene. Having identified the BRIP1 c.1045G>C missense variant in a number of families with EOC, we aimed to investigate the frequency of this and BRIP1.2392C>T pathogenic variant in patients with breast cancer (BC) and/or EOC. A case‐control study of 3767 cases and 2043 controls was undertaken investigating the presence of these variants using Sanger sequencing and gene panel data. Individuals with BC and/or EOC were grouped by family history. BRIP1 c.1045G>C was associated with increased risk of BC/EOC (OR = 37.7; 95% CI 5.3–444.2; P = 0.0001). The risk was highest for women with EOC (OR = 140.8; 95% CI 23.5–1723.0; P < 0.0001) and lower for BC (OR = 11.1; 95% CI 1.2–106.5; P = 0.1588). BRIP1 c.2392C>T was associated with smaller risks for BC/EOC (OR = 5.4; 95%CI 2.4–12.7; P = 0.0003), EOC (OR = 5.9; 95% CI 1.3–23.0; p = 0.0550) and BC (OR = 5.3; 95%CI 2.3–12.9; P = 0.0009). Our study highlights the importance of BRIP1 as an EOC susceptibility gene, especially in familial EOC. The variant BRIP1 c.1045G>C, rs149364097, is of particular interest as its dominant‐negative effect may confer a higher risk of EOC than that of the previously reported BRIP1 c.2392C>T nonsense variant. Dominant‐negative missense variants may confer higher risks than their loss‐of‐function counterparts.


| INTRODUCTION
Epithelial ovarian cancer (EOC) is a highly heritable cancer, with a threefold increase in risk for first-degree relatives of affected women. 1 Approximately 10-15% of EOC is considered to be hereditary. This is somewhat higher in the most common high grade serous (HGSOC) subtype, although the precise figure is unknown. [1][2][3][4] For many years BC and EOC have been noted to affect multiple individuals in some families, 5 leading to the identification of hereditary breast and ovarian cancer syndrome (HBOC). Pathogenic variants (PVs) in BRCA1 and BRCA2 explain approximately 25% of HBOC, and PVs in other genes in homologous recombination, mismatch repair and cell cycle checkpoint pathways account for additional significant contributions to risk. 6,7 However, the cause of approximately 35% of familial EOC remains unexplained.
The gene BRCA1 interacting protein 1 (BRIP1) encodes the protein BRIP1, which interacts with BRCA1 through BRCT repeats at the c-terminal end of BRCA1 and is required for normal repair of doublestrand DNA breaks. 8 PVs have been found in the first two-thirds of the gene, between nucleotides 68-2508, predicted to truncate the protein before the BRCA1 binding domain. 9 The gene is part of the Fanconi anaemia complement group family of proteins and is also known as FANCJ and BACH1. 10 c.2392C>T; (p.Arg798Ter) and ER-negative and triple-negative disease.
However, overall there was no significant association between BRIP1 and BC risk. 15 Conflicting or no significant evidence of BRIP1 as a BC risk gene was confirmed in further studies, 16 p.(Ala349Pro), (rs149364097). This variant is described as pathogenic/ likely pathogenic in ClinVar. 24 We hypothesised that this variant may be enriched in our patient population as an EOC and/or BC PV. We aimed to investigate the frequency of this variant as well as the established BRIP1 c.2392C>T variant through a case-control study.

| Study design and participants
Women were recruited through the 'Investigation of genetic modifiers in BRCA1/2 breast cancer and non BRCA1/2 high risk families' study for whole exome sequencing and Sanger sequencing at MCGM, the Predicting Risk of Cancer at Screening (PROCAS) study and FH-Risk. 25 Five hundred and twenty-one women from our centre had been included in the study by Seal et al. 11 and as such had had DNA analysed for variant BRIP1 c.2392C>T; p.(Arg798Ter). The primary source of participants included in our study is summarised in Table 1.
The PROCAS study recruited women aged 46-73 years attending BC screening in Greater Manchester, who were not affected with either BC or EOC at study entry. Women who developed BC after entry were included as cases. 25 Their samples underwent panel testing as part of the Breast Cancer Risk after Diagnostic Gene Sequencing (BRIDGES) study. 18 Women who were recruited to the Predicting Risk of Cancer at Screening (PROCAS) study with no BC diagnosis were included as controls; this meant that the controls would be from the same geographical region as the cases and similarly known to MCGM.
Cases comprised all women with non-mucinous EOC or BC, aged 18 years or over known to MCGM. They were recruited 1990-2020, but the ovarian cancer cases predominantly from 2016. Controls were women with no history of EOC or BC and no prior known PVs. They were recruited through PROCAS 2009-2013, and cancer-free as of 2020. Controls (group 1) were compared with cases in five groups: women with BC and no family history of EOC (group 2), women with BC and a family history of EOC (group 3), women with EOC and no BC/EOC family history (group 4), women with EOC and family history of BC only (group 5), and women with familial EOC (≥2 family members) (group 6).

| Data extraction
Relevant clinical information (histology of EOC and BC diagnoses, age at diagnosis, family history of BC and/or EOC) was obtained from local clinical record systems. Data for PROCAS patients was obtained from a questionnaire completed at study entry. Prospective cancer diagnoses were updated through the cancer registry.

| DNA extraction
DNA was extracted from peripheral blood lymphocytes from women who attended MCGM with EOC and ≥2 affected relatives with EOC.
DNA from the women in PROCAS was extracted from saliva using an Oragene kit (DNA Genotek) according to the manufacturer's protocols.

| Sanger sequencing
Amplification of exon 8 of BRIP1 to genotype c.1045G>C was done by using the following forward and reverse primers:  3 | RESULTS

| Patient demographics
A total of 3767 individuals with BC and/or EOC and 2043 controls were tested for the two variants. These are described further in Table 2. 3.2 | BRIP1 c.1045G>C frequency Cases and controls were categorised by diagnosis and relevant family history, as shown in Table 3.  The two women with an EOC diagnosis had HGSOC and were diagnosed at 59 and 60 years.

| DISCUSSION
Our study highlights the importance of BRIP1 as an EOC susceptibility gene, in particular in familial EOC cases. The BRIP1 c.1045G>C PV was present in 3.23% of our cases with familial EOC. The finding of this variant in this population is of particular interest and it may be that this variant is more common in North-West of England as it has not been described as significantly in other studies. Although to the best of our knowledge our BRIP1 c.1045G>C PV carriers are unrelated, it is possible that they could link several generations before.
As the majority of the data were generated through Sanger sequencing for these specific variants, insufficient data were available for haplotype analysis. No samples were available from other affected family members to perform segregation analysis. The BRIP1 c.2392C>T PV showed a significant increase of risk for BC and a significant but smaller risk for the combined group in our study. We had a detection rate for this variant of 10/3700 (0.27%), five times higher assuming the same OR than the detection rate of 0.047% in the Easton et al study. 15 As we found an OR of 5 for BC and this was not found in the Easton study there may be another factor such as a potential founder effect contributing, or our ORs are incorrect and a more accurate OR for this variant would be a fifth of what we found. We did nonetheless not find this variant in over 2000 controls meaning a local population rate below 0.05%. It is possible therefore the truth lies in between.
There are some limitations to the present study. Both BRIP1 variants may be local founders and the true odds ratios of their effects may be lower. Nevertheless, we did not find either variant in over 2000 control samples and the larger gnomAD control dataset, restricted to the European non-Finnish population, was the best alternative. Even still the EOC conferred by the nonsense variant BRIP1 c.2392C>T is similar to previous estimates.
While this needs further validation, the dominant-negative effect of the BRIP1 c.1045G>C variant is likely to mean a more significant EOC risk than that of the well-known stop-gain BRIP1 c.2392C>T variant. Although rarer, due to the dominant-negative effect it has, it appears to confer a higher risk of EOC and should be considered in investigation of women with familial EOC. Finally, clinicians should not assume that a missense variant that is classified as pathogenic or likely pathogenic confers the same risk as loss of function variants. In particular, an assessment of whether the variant may be associated with a dominant-negative effect is important.