Somatic inactivation of breast cancer predisposition genes in tumors associated with pathogenic germline variants

Abstract

Background: Breast cancers (BCs) that arise in individuals heterozygous for a germline pathogenic variant in a susceptibility gene, such as BRCA1 and BRCA2, PALB2, and RAD51C, have been shown to exhibit biallelic loss in the respective genes and be associated with triple-negative breast cancer (TNBC) and distinctive somatic mutational signatures. Tumor sequencing thus presents an orthogonal approach to assess the role of candidate genes in BC development.

Methods: Exome sequencing was performed on paired normal-breast tumor DNA from 124 carriers of germline loss-of-function (LoF) or missense variant carriers in 15 known and candidate BC predisposition genes identified in the BEACCON case-control study. Biallelic inactivation and association with tumor genome features including mutational signatures and homologous recombination deficiency (HRD) score were investigated.

Results: BARD1-carrying TNBC (4 of 5) displayed biallelic loss and associated high HRD scores and mutational signature 3, as did a RAD51D-carrying TNBC and ovarian cancer. Biallelic loss was less frequent in BRIP1 BCs (4 of 13) and had low HRD scores. In contrast to other established BC genes, BCs from carriers of CHEK2 LoF (6 of 17) or missense (2 of 20) variant had low rates of biallelic loss. Exploratory analysis of BC from carriers of LoF variants in candidate genes such as BLM, FANCM, PARP2, and RAD50 found little evidence of biallelic inactivation.

Conclusions: BARD1 and RAD51D behave as classic BRCA-like predisposition genes with biallelic inactivation, but this was not observed for any of the candidate genes. However, as demonstrated for CHEK2, the absence of biallelic inactivation does not provide definitive evidence against the gene's involvement in BC predisposition.

Figure 1.

Case-control analysis of rare LoF variants (minor allele frequency [MAF] ≤ 0.005) and MS variants (MAF ≤ 0.001) in known or strongly proposed breast cancer genes, including subcategories of estrogen receptor–positive (ER+), ER-negative (ER-), and triple-negative (TN) breast tumor where diagnosis was available. ER+ and ER- groups were mutually exclusive, and the ER- groups include the TN samples. Participants without sufficient pathological information were only included in the overall LoF group and excluded from the subcategory analysis. CHEK2, BARD1, and BRIP1 were screened in 6689 cases and 14 381 controls; RAD51C and RAD51D were screened in 5726 cases and 13 428 controls. The sample sizes of ER+, ER-, and TN were 2146, 1246, and 862, respectively. CI = confidence interval; LoF = loss of function; MS = missense; OR = odds ratio (3,4,11,12,18,19).