Novel and recurrent BRCA1/BRCA2 germline mutations in patients with breast/ovarian cancer: a series from the south of Tunisia

The incidence of breast cancer (BC) and/or ovarian cancer (OC) is increasing in Tunisia especially in young women and mostly those with family history. However, the spectrum of BRCA mutations remains little explored in Tunisian patients in particular in the southern region. We sequenced the entire coding regions of BRCA1and BRCA2 genes using next generation sequencing (NGS) in 134 selected patients with BC and/or OC. Among the 134 patients, 19 (14.17%) carried pathogenic mutations (10 are BRCA1 mutation carriers and 9 are BRCA2 mutation carriers) that are mainly frameshift index (76.9%). Interestingly, 5 out of the 13 variants (38.46%) were found at least twice in unrelated patients, as the c.1310-1313 delAAGA in BRCA2 and the c.5030_5033 delCTAA that has been identified in 4/98 BC patients and in 3/15 OC patients from unrelated families with strong history of cancer. Besides recurrent mutations, 6 variant (4 in BRCA1 and 2 in BRCA2) were not reported previously. Furthermore, 3 unrelated patients carried the VUS c.9976A > T, (K3326*) in BRCA2 exon 27. BRCA carriers correlated significantly with tumor site (p = 0.029) and TNBC cases (p = 0.008). In the groups of patients aged between 31 and 40, and 41–50 years, BRCA1 mutations occurred more frequently in patients with OC than those with BC, and conversely BRCA2 carriers are mostly affected with BC (p = 0.001, and p = 0.044 respectively). The overall frequency of the BRCA germline mutations was 14.17% in patients with high risk of breast/ovarian cancer. We identified recurrent mutations as the c.1310_1313 delAAGA in BRCA2 gene and the c.5030_5033 delCTAA in BRCA1 gene that were found in 4% and 20% of familial BC and OC respectively. Our data will contribute in the implementation of genetic counseling and testing for families with high-risk of BC and/or OC.

Ben Ayed-Guerfali et al. J Transl Med (2021) 19:108 DNA repair process that increases the mutation rates and thus contributes to tumor development [9,10]. Indeed, women carrying pathogenic germline mutations in BRCA1/BRCA2 genes have an increased risk of BOC compared to those with wild type BRCA genes [11][12][13]. Screening of BRCA1/BRCA2 genes for the identification of pathogenic mutations is essential to provide genetic counseling to the members of affected families, for medical follow-up, and for the targeted therapy selection which is based on the inhibitor of poly-ADP-ribose polymerase inhibitor namely the Olaparib [14]. Despite the importance of the mutational spectrum of BRCA1/ BRCA2 genes, only few studies using Sanger DNA sequencing, investigated Tunisian patients with HBOC. In a recent study, Laitamn et al., reported all the pathogenic variants in both BRCA1 and BRCA2 genes in the Middle East, North Africa, and Southern Europe [15]. The authors identified 232 and 239 pathogenic sequence variants in BRCA1 and BRCA2, respectively that include only few variants that were found in Tunisian patients [15].
Besides, Riahi et al., reported twelve pathogenic mutations (25%); among them, three were found in BRCA 1 (c.211dupA, c.5266dupC and c.1504_1508delTTAAA) and two were novel mutations detected in BRCA 2 (c.1313dupT and c.7654dupT). The study was carried out on 92 families and was performed by direct sequencing [16]. Furthermore, two other studies reported deleterious mutations of the BRCA1/BRCA2 genes in 19% and 18% of familial patients from the north region of Tunisia respectively [17,18]. Moreover, Fourati et al. by targeting the exons 5, 11, and 20 of the BRCA1 gene and exons 10 and 11 of the BRCA2 gene, showed that among 66 patients, only 12 patients (18%) had deleterious mutations in the BRCA1 or BRCA2 genes [19].
The study of Mahfoudh et al., included 16 Tunisian high-risk BC families and they were screened for only BRCA1 gene. The authors showed that the prevalence of BRCA1 carriers was 37.5% and identified 3 truncating mutations (c.916 delTT, c.3450 delCAAG, c.5382 insC) and one splice site mutation c.212 + 2insG [20].
Technological advances in DNA sequencing have improved and facilitated the screening procedure of BRCA1/BRCA2 genes since the large size of these genes and the absence of hotspots regions make this procedure expensive and time-consuming. In fact, NGS has resolved these problems and allows, henceforth the detection of mutations with higher sensibility [21][22][23]. Our study is the first using NGS to screen the whole exons of BRCA1/ BRCA2 genes in 134 selected patients with high risk of BOC in the south region of Tunisia.

Patients
A total of 134 patients were selected (between 2016 and 2019) among them 113 patients with HBOC (98 cases) and/or OC (15 cases) that meet one of the following criteria: (1) Presence of at least three related first or second-degree BC cases; (2) BC in young patients aged less than 35 years, (3) Presence of male BC among the first or second-degree relatives, (4) Presence of at least two cases of BC or OC, regardless of age, and at least one case of prostate cancer in a related first or second degree patient.
In addition, 21 cases without evidence of family history were included in this study, 12 were diagnosed with BC at the age ≤ 40, and 9 patients diagnosed with OC at the age ≤ 60 years. All patients were recruited from the department of Medical Oncology of the CHU Habib Bourguiba of Sfax (Tunisia), and had provided written informed consent for participation in the research study and for the genetic testing.

Genetic testing
Genomic DNA was isolated from 0.2 mL of peripheral blood of the proband from each selected family using the "QIAamp DNA Blood Mini kit" (Qiagen), following the manufacturer's instructions. Isolated DNA was quantified by Qubit 3.0 Fluorometric quantitation (Thermo Fisher Scientific). The BRCA/BRCA2 germline mutations were detected using the next-generation sequencing approach (NGS). Briefly, 50 ng of genomic DNA from each sample was used to prepare library using the "AmpliSeq ™ BRCA Panel", "AmpliSeq ™ Library Plus" and "AmpliSeq ™ CD Indexes" according to the protocol provided by Illumina. The adaptors and sample unique DNA barcodes were incorporated into the amplified amplicons with a second PCR. Libraries were quantified with the Qubit ® dsDNA HS Assay Kit (Life Technologies).The DNA library was pooled and prepared for sequencing using the Illumina MiSeq sequencer with Miseq Reagent Kit v2 (300-cycles) (Illumina, San Diego, CA) according to the manufacturer's instructions to generate paired-end reads with a 151-bp read length. Reads were trimmed to remove low-quality sequences and then aligned to the human reference genome (GRCh37/hg19) using the Burrows-Wheeler alignment (BWA) package. The total PF Reads is 16,755,084 and the Q30 is 93.56%.
Sanger sequencing was performed to confirm the BRCA1/BRCA2 pathogenic mutation identified by NGS (see Additional file 1: Fig.1 in Supplementary data). Sequencing primers were designed using Primer 5.0 software and provided on request. PCR products were purified using the QIAquick PCR Purification Kit (Qiagen) and labeled using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems), and the sequence analysis was performed by BioEdit.

Variant annotation and classification
We annotated the mutations using in silico prediction tools to evaluate the functional effects of the candidate variants: PolyPhen-2 (http:// genet ics. bwh. harva rd. edu/ pph2/), which predicts the effects of an amino acid substitution on the structure and function of a protein. The PolyPhen score represents the probability that a substitution is damaging, so values between 0.9 to 1 are more confidently predicted to be deleterious. The program sorting intolerant from tolerant (SIFT) predicts whether an amino acid substitution is likely to affect protein function based on sequence homology and the physicochemical similarity between the alternate amino acid. The score is the normalized probability that the amino acid change is deleterious if the score is < 0.05 and tolerated if the score is > 0.05. Mutation Taster (http:// www. Mutat ionta ster. org/) was used to assess the impact of mutations on protein function.

Statistical analysis
Associations between patient characteristics and presence of pathogenic BRCA1/BRCA2 mutations were tested by Chi-square test. A significance level of p ≤ 0.05 was used in the analysis. All analyses were performed using SPSS (version 20.0).

Patients characteristics
A total of 134 breast and/or ovarian cancer subjects were included in this study. The characteristics of the study group are presented in Table 1 (Table 1).

Variants of uncertain significance in BRCA1/BRCA2 genes
In addition to deleterious BRCA1/BRCA2 mutations, 10 VUS (6 in BRCA1 and 4 in BRCA2) were predicted to be pathogenic by PolyPhen, SIFT or Mutation Taster (Table 4). Each VUS was identified in more than one patient and the c.9976A > T in exon 27 of the BRCA2 gene was detected in 3 unrelated patients, 2 with BC diagnosed at an age ≤ 40 years and 1 patient with OC ( Table 4). All VUS have been reported in the BIC database except the variant N550H in BRCA1 exon 11.    [9,10]. The prevalence of BRCA1/BRCA2 mutations varies in different populations due to founder mutation effect [15,[24][25][26][27][28]. Genetic testing of patients with family history for breast/ovarian cancer have become standard clinical management in Western countries, however, in Tunisia studies of BRCA -associated breast/ovarian cancer remain less investigated. Clearly, the majority of studies that reported the spectrum of BRCA mutations in Tunisia included a small number of patients and used Sanger method for DNA sequencing [16][17][18][19][20]. Riahi et al. reported that the rate of BRCA mutations in patients from the north of Tunisia was 25% (12/48) [16]. Furthermore, in another study, it was showed that deleterious BRCA1 mutations were detected in 37% (6/16) of patients selected from high-risk breast cancer families [20]. In this study, we used NGS to determine the BRCA mutation rates in high risk breast and/ or ovarian cancer in patients from the south of Tunisia. Thirteen truncating mutations and one splice acceptor mutation were identified in 19 among 134 patients (14.17%). The rate of BRCA1 mutations found in this study (7.46%, 10/134) was lower than those reported in previous studies. In fact, in a review of the literature, Cherbal et al., reported that in breast/ovarian cancer families from Algeria, Morocco and Tunisia, BRCA1   was mutated in 17.43% of cases (34/195) and that the c.798_799delTT was a recurrent mutation [29]. In our study none of the previously reported mutations in North African patients was identified except for the c.1310_1313delAAGA in BRCA2. Indeed, this mutation was detected in 4% (4/98) of patients from unrelated families with high risk of BC and among 9 BRCA2 carriers, 4 (44.4%) had the c.1310_1313delAAGA mutation suggesting that it might be recurrent in our population. Our finding is in line with a previous study showing that the c.1310_1313delAAGA mutation was detected in 6% (4/66) of patients from the north of Tunisia [19]. Furthermore, it was reported that the c.1310_1313delAAGA mutation was detected in 11.4% (14/122) of patients from the North-East region of Morocco, and in one Algerian family among 10 carrying BRCA mutation [30,31]. According to the Breast Cancer Information Core database (BIC; http:// resea rch. nhgri. nih. gov/ bic/), this mutation was found in different European patients and was recorded several times in the French UMD-BRCA2 database and classified as founder mutation [32]. Furthermore, we found that 2 BC patients are double heterozygous carrying both the c.1310_1313delAAGA frameshift mutation and the splice site acceptor variant c.-632G < A. This splicing mutation has been reported two times in the ClinVar database therefore, and in the best of our knowledge this is the first report describing the association of the splice site and a frameshift mutation in the BRCA2 gene in 2 unrelated patients with strong family history of BC. The pathogenic mutation c.17-20 delAAGA (Lys6Xfs) in BRCA2 exon 2 was detected in 2 unrelated young patients (39 year-old) with strong family history of BC. This mutation has not been reported previously in the ClinVar and BIC databases.
We also detected in the BRCA2 exon 27, the c.9976A > T (K3326X) in 3 unrelated patients, 2 were diagnosed with family BC at 28 and 36 years old and one patient with OC at 56 years old and without family history. The BRCA2 variant (K3326X), was firstly interpreted as pathogenic, but its identification in control populations led to its classification as a benign polymorphism [33]. However, recent studies showed the association of K3326X variant with the risk of developing melanoma, pancreatic, breast and ovarian cancers [34][35][36]. Moreover, the large study of Meeks et al., provided evidence that the K3326X variant is associated with the risk of developing BC and OC independently of other pathogenic variants in BRCA2 [37]. Altogether, it was suggested that the K3326X variant is not neutral and that it may be included in SNP panels for evaluating BC risk.
In BRCA1gene, we identified the c.5030_5033delCTAA mutation in 3 out of 15 (20%) unrelated patients with early onset HBOC. This mutation has not been reported in North Africa but was found in patients from Greece, Italy, Jordan, Lebanon, Kuwait, and Saudia Arabia [15]. In addition, the nonsense mutation c.2338C > T was shared by 2 young patients (< 50 years-old) diagnosed with OC. The c. 2338C > T has been reported in Caucasian patients according to the BIC and ClinVar databases. Altogether, these findings may suggest ethnic and genetic associations between unrelated populations, or that these mutations occur in mutational hotspots region.
It is well established that TNBC exhibit an aggressive behavior and has a worse prognosis [39]. In our cohort, among 110 patients with only BC, 26 (23.6%) were TN including 14 patients with family history of BC and 12 without evidence of family history of BOC. Five patients (5/26, 19.23%) carried BRCA mutations and 4 among them (80%) were BRCA1 carriers. Consistent with our findings, in a previous study, the BRCA1 mutation carriers had a higher incidence in TNBC than BRCA2 mutation carriers [40][41][42][43]. Moreover, in a recent work, Toss et al., showed that in TNBC Italian patients, BRCA mutation prevalence was 22.6% (21.4% BRCA1) and that 64.2% are ≤ 30 years old, which is in line with our findings [44]. Furthermore, Mahfoudh et al., reported that the 5382insC mutation in BRCA1 was detected in 25% (2/8) of Tunisian patients with TNBC [45]. In our study, none of the patients carried the BRCA1 5382insC mutation, however, in TNBC we detected 3 mutations in BRCA1 (c.3254delG, c.3364_3370dupACA GAT T, c.4067_4071delAAGAA), and no BRCA2 pathogenic mutations were detected in TNBC cases except the VUS c.9976A > T identified in 2 unrelated patients (Additional file 2: Table 1 in Supplementary data).
It is important to note that previous studies used Sanger sequencing to identify BRCA mutations in Tunisian families with a history of BOC unlike the present study where the whole BRCA genes were screened by the NSG which allowed us to identify new mutations in Tunisian patients. However, some recent studies have performed the whole exome sequencing to identify candidate genes in few BRCA negative families. Hamdi et al., performed whole exome sequencing on seven Tunisian families with HBOC and identified four novel BC candidate genes (MMS19, DNAH3, POLK and KATB6) [46]. Furthermore, other studies identified by exome sequencing, RCC1 and RAD50 as BC candidate susceptibility genes in Tunisian BRCA negative families [47,48].

Conclusions
This study assessed the prevalence of germline mutations and identified novel and recurrent mutations for BRCA1/ BRCA2 genes in patients with high risk of BOC from the south region of Tunisia. Giving the high prevalence of pathogenic mutations in BRCA1/BRCA2 genes (14.17%), our data will contribute to the establishment of a service dedicated to the genetic screening and to the counseling of the families with high risk of HBOC in Tunisia.