The Importance of Extended Analysis Using Current Molecular Genetic Methods Based on the Example of a Cohort of 228 Patients with Hereditary Breast and Ovarian Cancer Syndrome

In about 20–30% of all women with breast cancer, an increased number of cases of breast cancer can be observed in their family history. However, currently, only 5–10% of all breast cancer cases can be attributed to a pathogenic gene alteration. Molecular genetic diagnostics underwent enormous development within the last 10 years. Next-generation sequencing approaches allow increasingly extensive analyses resulting in the identification of additional candidate genes. In the present work, the germline molecular diagnostic analysis of a cohort of 228 patients with suspected hereditary breast and ovarian cancer syndrome (HBOC) was evaluated. The 27 pathogenic gene variants initially detected are listed, and their distribution in the high-risk BRCA1 and BRCA2 genes is presented in this study. In ten high-risk patients, in whom, to date, no pathogenic variant could be detected, an extended genetic analysis of previously not considered risk genes was performed. Three variants of uncertain significance and one pathogenic variant could be described. This proves the importance of extended analysis using current molecular genetic methods.


Introduction
Human genetic analyses are an integral part of everyday clinical practice due to their scope and importance for subsequent therapeutic decisions. The introduction of highthroughput analytical techniques allows increasingly comprehensive testing, which in turn contributes to a better medical understanding of the disease. Around 30% of all women with breast cancer in Germany have a family history of breast cancer. They thus fulfill the inclusion criteria for genetic testing regarding hereditary breast and ovarian cancer syndrome (HBOC) [1]. In hereditary breast and ovarian cancer syndrome, high-risk genes (BRCA1, BRCA2, TP53, and PALB2) are distinguished in addition to moderately penetrant risk genes (currently ATM, BARD1, BRIP1, CDH1, CHEK2, RAD51C, and RAD51D) [2]. In the scope of the German Consortium for Hereditary Breast and Ovarian Cancer, other candidate genes (e.g., NBN, FANCM, XRCC2, and RECQL) are currently co-analyzed for research purposes and their significance in contributing to breast and ovarian cancer is under investigation [3]. Women with a BRCA1 or BRCA2 pathogenic variant develop disease approximately 10 to 20 years earlier than women without familial risk [4]. Due to the increased lifetime risk of breast cancer of up to 70% [5], these women are eligible for an intensified screening and follow-up program [2].
We performed a comprehensive statistical evaluation in a cohort of 228 patients whose data were collected in 2012-2017 at the Institute of Human Genetics of the University Medical Center of Freiburg in the context of molecular genetic diagnostics regarding hereditary breast and ovarian cancer. The aim was to provide an overview of the overall spectrum and to identify relevant and previously undetected pathogenic variants in highrisk patients.

Data Collection and Risk Assessment
Initially, data were collected from 228 patients. These were fully considered in the statistical analysis. For 213 patients, detailed information of the family history was available, allowing an additional risk assessment using the checklist of the German Consortium for Hereditary Breast and Ovarian Cancer [6]. In addition to the age of onset of the disease, this checklist is based on other associated cases of the disease in the maternal and paternal line and takes these into account with a single, double, or triple weighting. For reasons of clarity and informative value, the score values obtained were grouped into specially determined subgroups (Table 1). This grouping is introduced and determined exclusively to differentiate the cohort under investigation.

Statistical Evaluation
The average age at analysis was 50.6 years for women and 57.4 years for men. At the time of analysis, 164 patients had breast cancer, 17 had ovarian cancer, and 7 had breast and ovarian cancer. A total of 27 people had predictive testing performed. For 13 patients, a reference to the respective background for the analysis performed was not possible, due to missing information. The classification of detected sequence variants was performed according to the standards and guidelines of the American College of Medical Genetics and Genomics (ACMG) [13].
In total, pathogenic variants (ACMG class 5) were detected in 37 (16.2%) patients. These were mainly found in the high-risk genes BRCA1 (54.1%, n = 20) and BRCA2 (37.8%, n = 14). In the CHEK2 gene, two pathogenic variants were detected (5.4%, n = 2) and, in the RAD51C gene, one pathogenic variant was identified (2.7%, n = 1). In total, eleven patients with detected BRCA1 variants had breast cancer, five patients had ovarian cancer, and two patients had breast and ovarian cancer. For two patients, no conclusion could be drawn about the cancer type due to missing information. Regarding BRCA2 variants, twelve patients developed breast cancer. One patient had breast and ovarian cancer. For one patient, no conclusion could be drawn about the cancer type due to missing information. BRCA1 mutation carriers are thus more likely to develop ovarian cancer than BRCA2 mutation carriers, as previously reported [14,15]. All patients carrying a pathogenic CHEK2 or RAD51C variant had breast cancer. All detected pathogenic variants are summarized with their respective reference in Table 2. The pathogenic variants c.5609_5610delTCinsA as well as c.7878G>C of the BRCA2 gene were initially described in the context of Fanconi anemia disease. Later, they were associated with hereditary breast cancer. Each of the detected pathogenic variants had been described elsewhere before.

Graphical Representation of the Pathogenic Variants in the BRCA1 and BRCA2 Genes
In the BRCA1 gene, 70% of the pathogenic variants are located in the functional RING domain at the N-terminus (6 of the 20 pathogenic variants) or in the BRCT domains at the C-terminus (8 of the 20 pathogenic variants) ( Figure 1). Furthermore, one pathogenic variant was detected in the nuclear localization sequence (c.1510del, p.(Arg504Valfs*28)) and another pathogenic variant (c.4183C>T, p.(Gln1395*)) in the coiled-coil domain in the region of the serine cluster ( Figure 1). For the BRCA2 gene, the pathogenic variants are mainly located in the BRC repeat domain (Figure 2), where 7 of the 14 detected pathogenic variants are found. Another four pathogenic variants are located in the functional domains at the C-terminus (Figure 2). Equal numbers of them are located in the helical domain and in the oligonucleotide binding fold (Figure 2). C-terminus (8 of the 20 pathogenic variants) (Figure 1). Furthermore, one pathogenic var-130 iant was detected in the nuclear localization sequence (c.1510del, p.(Arg504Valfs*28)) and 131 another pathogenic variant (c.4183C > T, p.(Gln1395*)) in the coiled-coil domain in the 132 region of the serine cluster (Figure 1). For the BRCA2 gene, the pathogenic variants are 133 mainly located in the BRC repeat domain (Figure 2), where 7 of the 14 detected pathogenic 134 variants are found. Another four pathogenic variants are located in the functional do-135 mains at the C-terminus (Figure 2). Equal numbers of them are located in the helical do-136 main and in the oligonucleotide binding fold (Figure 2  Seventeen of the twenty detected pathogenic variants in the BRCA1 gene (=85.0%) and eleven of the fourteen detected pathogenic variants in the BRCA2 gene (=78.6%) can be localized in the conserved regions of the genes. These are essential for correct protein structure and function. Thus, in the presence of such a pathogenic variant in the case of a missense variant, a loss of function of the protein can be assumed.

Evaluation of the Molecular Genetic Analyses of the High-Risk Patient Group
Because of the further development of sequencing technologies, a renewed and extended gene analysis with the TruRisk®panel of the German Consortium of Hereditary Breast and Ovarian Cancer was performed for ten patients who showed a particularly high-familial-risk constellation without the detection of a pathogenic variant. As a result, four variants could be detected (Table 3). Pathogenic ATM variants increase the risk for breast cancer by 2-3 times [42]. Evidence regarding an increased risk of prostate or pancreatic cancer cannot be confirmed currently due to insufficient data [43,44]. Pathogenic BARD1 variants also lead to increased risk for Ewing sarcoma, osteosarcoma, and neuroblastoma, in addition to an increased risk for breast cancer [45][46][47]. The MUTYH and SMARCA4 genes are among the candidate genes of the TruRisk®panel. In addition to MUTYH-associated polyposis (MAP) and the associated increased risk for colon cancer, pathogenic MUTYH variants have been described to confer increased risks for, among others, urinary bladder and gastric cancer [48]. Pathogenic variants in the SMARCA4 gene are increasingly associated with small-cell carcinoma of the ovary, hypercalcemic type, or adenocarcinomas of the lung and endometrium [49][50][51].
The detected ATM variant is a class 5 pathogenic variant generating a premature stop codon. The databases HGMD®Professional 2020.3 and ClinVar contain several entries describing the variant as pathogenic [52,53]. For the BARD1 variant c.212G>T, the sequence variant interpretation tools SIFT, Polyphen2, and MutationTaster predict a putative deleterious effect of the variant. In the ClinVar database, the variant was reported five times, most recently in November 2018, and interpreted with unclear significance in all cases. There is no entry in the HGMD®Professional 2020.

Discussion
The further development of analytical methods opens up a multitude of new possibilities. The analysis of the present cohort shows that within a five-year period, the standard analysis method has evolved from Sanger sequencing to next-generation sequencing. This high rate of technical change results not only in a large amount of data with different sequence variants but also in a steady stream of new candidate genes, the clinical significance of which is sometimes still unclear at the time of analysis. A large data resource is needed to achieve an assessment of candidate genes in terms of their clinical relevance and to incorporate them into routine diagnostics. At the same time, a more comprehensive panel diagnosis increases the probability to detect a variant of clinical relevance. However, the more genes are sequenced, the more variants of uncertain significance can be identified, which in turn cannot be clearly classified clinically. Thus, the challenge lies in the risk calculation and the statement of clinical evaluation. The diagnostic range may change over a few years, and a possible extended second or third analysis, and a re-evaluation of already described varieties, based on current scientific knowledge, may be useful. A possible recommendation would be a routine review as well as a supplement of the molecular genetic diagnostics according to the current state of research, for example in a 5-to 10-year rhythm. Currently, mostly only case-control studies are available for interpretation, which could promote misinterpretation. In order for the statements to gain a higher level of evidence, further investigations are needed, for example, in the context of prospective cohort studies. These are another important aspect to illustrate the existing possibility of multifactorial inheritance and pathogenicity, which may be causative for the disease. In this context, a single variant would moderately increase the risk of disease, whereas the presence of additional variants could potentiate the risk of tumor development. In addition, with many pathogenic variants detected, there is not only an increased risk for breast or ovarian cancer, but also other associated tumor diseases are possible. The statement of a genotype-phenotype relationship is also becoming increasingly important. Thus, different ovarian cancer cluster regions (OCCRs) and breast cancer cluster region (BCCRs) can be formulated for both the BRCA1 and BRCA2 genes. These are characterized in this specific region by a significantly higher rate of breast or ovarian cancer compared to the other entity [56,57]. In addition to mutation position analysis, there are other considerations for predicting specific cancer risks. For example, the prospective cohort studies from the International BRCA1/2 Carrier Cohort Study (IBCCS), Breast Cancer Family Registry (BCFR), and Kathleen Cuningham Foundation Consortium for Research Into Familial Breast Cancer (kConFab) suggest that both family history and variant location in the gene influence cancer risk [58][59][60]. Further research is needed in this regard. The present cohort analysis shows that familial clustering of breast and ovarian cancer can be accounted for by analyzing additional risk genes. It is worthwhile to revisit the variants found after a certain period of time (e.g., 5-10 years) as well as to extend the analysis in unexplained cases to verify the results with the current state of science, even if only one patient out of 228 benefits from it.