A Review of the Hereditary Component of Triple Negative Breast Cancer: High- and Moderate-Penetrance Breast Cancer Genes, Low-Penetrance Loci, and the Role of Nontraditional Genetic Elements

Triple negative breast cancer (TNBC), representing 10-15% of breast tumors diagnosed each year, is a clinically defined subtype of breast cancer associated with poor prognosis. The higher incidence of TNBC in certain populations such as young women and/or women of African ancestry and a unique pathological phenotype shared between TNBC and BRCA1-deficient tumors suggest that TNBC may be inherited through germline mutations. In this article, we describe genes and genetic elements, beyond BRCA1 and BRCA2, which have been associated with increased risk of TNBC. Multigene panel testing has identified high- and moderate-penetrance cancer predisposition genes associated with increased risk for TNBC. Development of large-scale genome-wide SNP assays coupled with genome-wide association studies (GWAS) has led to the discovery of low-penetrance TNBC-associated loci. Next-generation sequencing has identified variants in noncoding RNAs, viral integration sites, and genes in underexplored regions of the human genome that may contribute to the genetic underpinnings of TNBC. Advances in our understanding of the genetics of TNBC are driving improvements in risk assessment and patient management.


Introduction
Breast cancer is a complex disease characterized by clinical, pathological, and molecular heterogeneity, which may influence risk assessment, diagnosis, treatment, and clinical outcomes [1]. Pathological characterization of breast disease includes a number of variables such as histological architecture, degree of cellular differentiation, tumor size, presence of local or distant metastasis, and hormone receptor and human epidermal growth factor receptor 2 (HER2) status.
Triple negative breast cancers (TNBC), which do not express the estrogen (ER) or progesterone receptors (PR) and have little or no HER2 protein expression, account for 10-15% of breast cancers diagnosed each year [2]. TNBC represents an aggressive form of disease, often diagnosed at a later stage, characterized by high-tumor grade, larger size, poorly differentiated histology, more frequent lymph node metastases, and younger age at diagnosis [3]. TNBC is more likely to present as an interval cancer, appearing between screening mammograms, possibly due to higher proliferation rates than other tumor types [4]. Risk of distant metastasis and death are significantly higher in patients with TNBC within five years of diagnosis [3], and TNBC displays distinctive patterns of metastasis with a higher affinity for lung, brain, and distant lymph nodes compared to other subtypes.
Like all types of breast cancer, TNBC exhibits marked heterogeneity in terms of histology, patterns of metastatic dissemination, response to therapies, and patient outcomes. While the majority of TNBC are invasive ductal carcinomas, other histologies may be triple negative as well, with fiveyear survival outcomes ranging from 100% in patients with medullary tumors to 56% in those with metaplastic TNBC [5]. Although there is significant overlap between TNBC and basal-like tumors, as defined by immunohistochemistry (IHC) and patterns of gene expression, 28% of TNBC are classified as luminal A, luminal B, HER2-enriched, or normal-like [6]. Evaluation of TNBC at the gene expression level has shown variability in levels of estrogen related genes, genes involved in oxidation reduction, and proliferation genes, suggesting that additional subclassification of TNBC is warranted [7]. Cluster analysis of 587 TNBC identified six subtypes including basal-like 1, basal-like 2, immunomodulatory, mesenchymal, mesenchymal stem-like, and luminal androgen receptor, each of which may be responsive to different targeted or chemotherapeutic agents [8].
A number of risk factors have been associated with TNBC that have not been linked to increased risk for other cancer subtypes. In contrast to luminal A tumors, TNBC/basallike tends to be associated with younger age at diagnosis, African ancestry, younger age at menarche and at first fullterm pregnancy, higher parity, lack of breastfeeding, and higher BMI and waist-to-hip circumference ratio [2,[9][10][11]. The frequency of TNBC in African Americans (29.8%) is intermediate between that in West African women (53.2%) and White American women (15.5%), suggesting a genetic component to TNBC [12]. Strong associations between TNBC and BRCA mutation status have been reported: 70-90% and 16-23% of breast tumors in BRCA1 and BRCA2 mutation carriers are TNBC [11]; however, germline mutations in BRCA1 and BRCA2 only account for 15.4% of patients with TNBC [13], and the prevalence of BRCA1 and BRCA2 mutations is lower in African American women (20.4%) with TNBC compared to European American women (33.3%) [14]. These data suggest that TNBC has a genetic component and genes other than BRCA1 and BRCA2 may play a role in disease etiology. In this review we examine current data regarding the contribution of germline mutations in highand moderate-penetrance genes to TNBC. In addition, we evaluate the latest genome-wide association studies (GWAS) and candidate gene approaches to identify low-penetrance genes. Finally, we consider the role of nontraditional genetic variants including single nucleotide polymorphisms (SNPs) in microRNA (miRNA) binding sites, retroelements, and novel sequences not present in the current reference genome in the etiology of TNBC.

Methods
Relevant literature was identified by searching the PubMed database (https://www.ncbi.nlm.nih.gov/pubmed). Search terms included TRIPLE NEGATIVE BREAST CANCER, GENETICS, HEREDITARY CANCER, miRNA, and VIRUS INTEGRATION. Only articles written in English were included. To ensure data presented here were current, articles published within the last 12 months and/or metaanalyses are highlighted.

High-Penetrance Breast Cancer Genes
3.1.1. BRCA1 and BRCA2. BRCA1 and BRCA2 are known to be tumor suppressor genes that function in DNA repair pathways. Cells lacking functional BRCA1 or BRCA2 are deficient for double-stranded break repair, resulting in genomic instability that leads to cancer predisposition. Current clinical data suggest BRCA1-and BRCA2-deficient tumors may have heightened sensitivity to platinum agents or poly (ADPribose) polymerase I (PARP) inhibitors [25]. In a large collection of families with hereditary breast cancer (n=237), 52% of families had disease that was likely attributable to mutations in BRCA1 while 32% had disease linked to BRCA2 [26]. Rebbeck et al. investigated whether the location or type of BRCA1/BRCA2 mutations is associated with variation in breast and ovarian cancer risk. Patients carrying mutations in exon 11 of BRCA1 appeared to have different disease phenotypes than patients carrying other BRCA1 mutations. Similarly, mutations in exon 11 of BRCA2 were associated with variability in breast and ovarian cancer risk [27]. Mutations in both genes have been associated with increased risk of TNBC albeit at different frequencies and within different age groups.
The first breast cancer susceptibility gene, BRCA1, was identified in 1994 [28]. BRCA1 is located on chromosome 17q21 and is comprised of 24 exons, 22 of which encode an 1863 amino acid protein. The BRCA1 protein has multiple sequence motifs including RING, DNA-binding, and BRCA1 C-terminus (BCTR) domains that allow BRCA1 to interact with other proteins and assist in subcellular localization. BRCA1 is a tumor suppressor gene that contributes to repair of damaged replication forks and double-strand breaks, transcriptional regulation in response to DNA damage, chromatin remodeling, and regulation of cell division, apoptosis, and transcription [29]. BRCA2 is a tumor suppressor gene located on chromosome 13q12 that was identified in 1995 [30]. BRCA2 has 27 exons and the BRCA2 protein interacts with RAD51 through the BRC motif. BRCA2 is also a transcriptional coregulator involved in DNA repair through homologous recombination.
As early as 1998, histological characterization revealed that in comparison with sporadic tumors, tumors in BRCA1 mutation carriers exhibited a distinct phenotype that includes high mitotic counts, pushing margins, and lymphocytic infiltration [31]. Histologic characteristics of TNBC also include high-grade with high mitotic indices, regions of central necrosis, conspicuous lymphocytic infiltrate, and pushing borders [32]. In fact, TNBC represents the predominant tumor type in patients with BRCA1 mutations, accounting for 71% (range 42-100%) of tumors, while TNBC has been diagnosed in only 25% of patients with germline BRCA2 mutations [33]. In contrast, the frequency of BRCA1 or BRCA2 mutations in women with TNBC is generally lower, with an average of 35% (range 9-100%) and 8% (2-12%) of women with TNBC harboring germline BRCA1 and BRCA2 mutations, respectively. In addition to differences in mutation frequency, age distribution differs between BRCA1 and BRCA2 positive patients with TNBC, with an average age at diagnosis of 47.2 years and 58.8 years in those with BRCA1 and BRCA2 mutations, respectively [34]. In a recent study that included over 10,000 patients with TNBC, germline mutations were associated with odds ratios of 16.27-26.90 for BRCA1 and 5.42-6.33 for BRCA2 [35].

PALB2.
The BRCA2 binding protein known as partner and localizer of BRCA2 (PALB2) stabilizes and regulates BRCA2 through localization and stabilization within important nuclear structures such as chromatin and the nuclear matrix and by promotion of recombination repair and checkpoint functions [36]. PALB2 was recently classified as a high-risk breast cancer gene with an odds ratio (OR) of 7.46 [95% confidence interval (CI)=5.12-11.19] [37]. Mutations in PALB2 are associated with aggressive disease. Over half (54.5%) of the familial and sporadic breast cancer patients from Finland who carried the 1592delT PALB2 mutation presented with TNBC compared to other familial (12.2%) or sporadic (9.4%) breast cancer patients [38]. Similarly, in a predominantly European-Caucasian cohort of women, 46% of tumors in women with PALB2 mutations were TNBC [39]. When mutation profiles were determined in a cohort of 4,797 women diagnosed with TNBC, panel testing performed at Myriad Genetics, found that 1.3% of women had pathogenic variants in PALB2 ( Figure 1) [40]. In a similar cohort of 1,824 women with TNBC unselected for family history of cancer, deleterious mutations in PALB2 were detected in 1.2% of patients [41].

TP53.
Tumor protein p53 (TP53), which encodes the p53 phosphoprotein, is a tumor suppressor gene that plays a critical role in each of the 10 Hallmarks of Cancer as defined by Hanahan and Weinberg [43]. As a result of this functional diversity, the p53 signaling pathway is at least partially disrupted in most human cancers and TP53 mutations are the most frequent genetic changes seen in human cancers [44]. Although the frequency of somatic mutations in TP53 is higher in basal-like tumors than any other subtype [45], germline mutations in TP53 have not been associated with an increased risk of TNBC. The mutation rate for TP53 in a cohort of 2,134 BRCA1/BRCA2 mutation negative women with familial breast cancer was 0.52% and TP53 mutations carriers showed enrichment for HER2+ tumors [39]. In a large cohort of 35,409 women with a single diagnosis of breast cancer, mutations in TP53 were detected in 0.7% of women with TNBC compared to 2.1% of those with non-TNBC subtypes [40]. In 133 women from Taiwan with early-onset and/or family history of breast cancer, only two women carried a pathogenic mutation in TP53 and both had ER+/HER2+ tumors [46]. Similarly, only one of 1,824 women with TNBC evaluated by Couch et al. [41] carried a TP53 mutation. These results suggest that germline mutations in TP53 are not associated with increased risk of TNBC.

PTEN.
Phosphatase and tensin homolog (PTEN) is a tumor suppressor gene involved in the regulation of the phosphoinositol-3-kinase and AKT signaling pathways and control of cellular proliferation and survival. PTEN is the second most frequently mutated gene in human cancers (after TP53) and germline mutations in PTEN are frequently observed in cancer susceptibility syndromes [47] [57]. In a study of 105 women with TNBC from Spain, BARD1 mutations were detected in two patients (1.9%) [50]. Likewise, nine (0.5%) of the 1,824 women in the Triple Negative Breast Cancer Consortium (TNBCC), had BARD1 mutations [41]. Although not exclusive to TNBC, the mutation frequency in BARD1 was 3.3% in women with TNBC compared to 1.7% in women with non-TNBC [40]. In a study of 4,032 Caucasian women with TNBC, the mutation rate of BARD1 was 0.7% compared to 0.2% in women with non-TNBC and the OR for an association with TNBC compared to non-TNBC disease was 3.73 (95% CI=2.3-5.95) [35].

Low-Penetrance Breast Cancer Loci. Mutations in high-
and moderate-penetrance breast cancer genes account for ∼14% of all TNBC cases [40,41]. Genome-wide association studies (GWAS) over the last decade have identified SNPs that are associated with breast cancer risk in an additive fashion.

Journal of
In an early study to identify susceptibility loci for breast cancer, ∼266,000 SNPs across the genome were genotyped in 408 breast cancer cases with a strong family history and 400 controls from the United Kingdom. In the second phase of this study, ∼12,000 SNPs that showed an association with breast cancer in phase I were genotyped in an additional 3,990 cases and 3,916 controls [58]. To determine whether any SNPs were reliably associated with breast cancer risk, the 30 most significant SNPs from phase II were further validated in an additional 21,860 cases and 22,578 controls. Six SNPs were associated with increased risk (P<10 −5 ), including SNPs in or near the fibroblast growth factor receptor 2 (FGFR2; rs2981582) gene, lymphocyte-specific protein (LSP1; rs3817198), mitogen-activated protein kinase kinase kinase 1 (MAP3K1; rs889312), and tox high mobility group box family member 3 (TOX3; rs12443621 and rs8051542) and in the chromosome 8q24 region (rs13281615  [60]. In a second metaanalysis of three GWAS including 4,193 ER-breast cancer cases and 35,194 controls, combined with 40 follow-up studies, variants at rs4245739 located in the 3 region of the mouse double minute 4 homolog (MDM4) oncogene on chromosome 1q32.1 seemed to be specific to TNBC [16]. In addition to the loci summarized above, a GWAS approach identified the 19p13 chromosomal region as a modifier of breast cancer risk in BRCA1 mutation positive individuals [23]. Five SNPs from 19p13 were genotyped in 2,301 women with TNBC and 3,949 controls to evaluate the association between the 19p13 locus and TNBC in the general population. Minor alleles for SNPs rs8170 (OR per A allele =1.28, 95% CI =1. 16-1.41) and rs2363956 (OR per C allele=0.80, 95% CI 0.74-0.87) were associated with TNBC risk in women without BRCA1 mutations. In the TNBCC, 22 known breast cancer susceptibility loci were studied in 2,980 Caucasian women and 4,978 controls to assess relationships with TNBC. Two SNPs from the 19p13.1 locus [rs8170 (P=2.25x10 −8 ) and rs8100241 (P=8.66x10 −7 )] were associated with risk of TNBC, as were SNPs from the estrogen receptor (ESR1; rs2046210 and rs12662670), RAD51L1 (rs999737), and TOX3 (rs3803662) [21]. Subsequent studies in the BCAC using 48,869 breast cancer cases and 49,787 controls demonstrated that rs8170 was a TNBC-specific risk variant (OR=1.25; 95% CI=1.18-1.33) [24]. A haplotype analysis in this study that included both rs8170 and rs8100241 found that the C-G and T-G haplotypes were both associated with risk of TNBC (C-G OR=1.17; 95% CI=1.09-1.25 and T-G OR=1.35; 95% CI=1.25-1.46) compared with the C-A haplotype.
Pooled analysis of the Collaborative Oncological Gene-Environment Study (COGS) and TNBCC SNP data further refined the GWAS data [17]. Multiple data sets, consisting of 22 studies from 7 different countries were combined in a two-stage analysis. Evaluation of SNPs from 3,677 women with TNBC and 4,708 controls supported the association of 25 known breast cancer susceptibility loci, including 2q35, LGR6, MDM4, TERT, ESR1, TOX3, and 19p13.1 with TNBC. Newly identified associations with TNBC were observed for an additional 15 SNPs from 14 loci. Interestingly, SNPs in CASP8, MAP3K1, LSP1, and FTO were not found to be associated with risk of TNBC. More recently, Milne et al. performed GWAS in 21,468 patients with ER-disease and 18,908 BRCA1 mutation positive individuals combined with 100,594 controls [15]. When evaluating the subset of individuals with TNBC, associations with 10 previously reported loci were replicated and 10 new susceptibility loci were identified. To date, seven chromosomal loci have been associated with risk for TNBC in multiple studies (Table 1).

Other Genetic Elements
Genetic elements other than susceptibility genes within the germline may also contribute to risk of TNBC Recent studies suggest that retroviral sequence elements from ancient retroviral infections may contribute to heritable TNBC. Some members of the human endogenous retrovirus HERV-K family are related to the endogenous mouse mammary tumor virus (MMTV) which can function as a mammary carcinogen in mice. While HML-2 proviruses have not been found at significantly higher frequencies in the genomes of patients with breast cancer compared to healthy controls and have not been associated with breast cancer histology [65,66], the frequencies of detection for Journal of Oncology 7 HERV-K113 and HERV-K115 are significantly higher in individuals of African ancestry (21.8% and 34.1%, respectively) compared to individuals from the United Kingdom (4.2% and 1.0%) [67]. Given the enrichment of the TNBC subtype in women of African ancestry, the presence of these sequences should be evaluated in larger populations with available ER, PR, and HER2 status. More recently, Marchi et al. mined whole genome sequence data generated by next-generation sequencing and identified 17 sites of viral integration not present in the human reference sequence [68] that may contribute to breast cancer risk.
Sequences not present in the human reference genome may harbor additional genes and/or genetic elements that contribute to risk of TNBC. De novo assembly of whole genome sequences from Asian and African individuals revealed ∼5 Mb of novel sequences from both individuals and populations [69]. The authors estimate that a complete human pan-genome would contain 19-40 Mb of novel sequence not present in the current reference genome. Sixtynine genes, any of which may increase risk of TNBC, were located within unmapped regions of the African genome. Data from our own laboratory demonstrated that the insertion frequency of a 30 Kb region of chromosome 7p11 that harbors the promoter and first three of the four exons that compromise the phophoserine phosphatase-like (PSPHL) gene was 76% in African American women compared to only 21% in European American women [70]. While presence of an intact PSPHL gene was not associated with increased risk of breast cancer or TNBC, other uncharacterized genes from regions variably represented between populations may contribute to increased risk of TNBC

Conclusions
Identification of the BRCA1 gene 25 years ago revolutionized the field of cancer risk assessment. Associations between germline BRCA1 mutations and TNBC led the National Comprehensive Cancer Network to amend their BRCA1/2 testing criteria in 2011 to include individuals diagnosed with TNBC [71]. Current guidelines allow testing for patients diagnosed with TNBC at ≤60 years of age with or without a significant family history of breast cancer [72]. BRCA1 and BRCA2 testing in 439 women with TNBC from the Australian Breast Cancer Tissue Bank supports TNBC pathology as a sufficient criterion for testing as 59% of women with pathogenic mutations did not have a family history of breast or ovarian cancer [34]. In a recent study of 10,901 TNBC patients, pathogenic variants were detected in TNBC riskassociated genes in 4.3% of patients not meeting NCCN testing criteria (diagnosed at >60 years of age without a family history) [35]. Because many of the mutations detected in this group were clinically actionable, the authors suggest that all patients with TNBC may benefit from genetic testing.
The importance of identifying the genetic etiology of TNBC extends beyond risk assessment. Surgical options are not dictated by tumor subtype; however, BRCA1 and BRCA2 mutation carriers are at increased risk for contralateral breast and ovarian cancer; thus the option of mastectomy with or without prophylactic removal of the contralateral breast and salpigo-oophorectomy should be considered. For carriers of mutations in other TNBC genes, such as BARD1 and PALB2, evidence is not yet sufficient to recommend mastectomy or oophorectomy [72]. Because patients with TNBC do not respond to hormone-or HER2-targeted treatments, chemotherapy is the primary treatment option. Carcinomas from patients with TNBC from patients with pathogenic BRCA1/BRCA2 mutations have demonstrated unique sensitivity to platinum agents in both the neoadjuvant and adjuvant settings [73]. More recently, PARP inhibitors, which exploit DNA repair deficiencies in cells with dysfunctional BRCA1 or BRCA2 proteins, leading to synthetic lethality, have shown promise with Olaparib approved by the FDA for the treatment of TNBC in patients with germline BRCA1/BRCA2 mutations [74].
Despite recent achievements in identifying additional TNBC susceptibility genes and optimizing patient management for mutation carriers, future studies are needed. For example, what is the clinical utility of the low-penetrance genes/loci SNPs? In perhaps the largest GWAS to date, comprised of 94,075 cases and 75,017 controls of European ancestry that were derived from 69 different studies, 313 SNPs were assembled into a polygenic risk score (PRS) for both ER positive and ER negative tumors [75]. Preliminary studies evaluating the utility of the PRS are underway in Canada and Europe; however, given that the models were developed using cases and controls of European ancestry, the ability of this assay to accurately determine risk of TNBC in patients of other ancestries, especially African with its higher frequency of TNBC in young women, may be suboptimal. In conjunction, additional studies of nontraditional elements of the genome including retrotransposons and pseudogenes may reveal additional heritable risk factors for TNBC. Finally, for the ∼4% of TNBC patients with germline mutations in genes other than BRCA1 and BRCA2 [35,41], effective management strategies and novel therapeutics are urgently needed.

Disclosure
The contents of this publication are the sole responsibility of the author(s) and do not necessarily reflect the views, opinions, or policies of Uniformed Services University of the Health Sciences (USUHS), The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., the Department of Defense (DoD), the Departments of the Army, Navy, or Air Force. Mention of trade names, commercial products, or organizations does not imply endorsement by the U.S. Government.

Conflicts of Interest
The authors declare no conflicts of interest.