Identification of Eleven Novel BRCA Mutations in Tunisia: Impact on the Clinical Management of BRCA Related Cancers

Background Breast cancer is the world’s most common cancer among women. It is becoming an increasingly urgent problem in low- and middle-income countries (LMICs) where a large fraction of women is diagnosed with advanced-stage disease and have no access to treatment or basic palliative care. About 5-10% of all breast cancers can be attributed to hereditary genetic components and up to 25% of familial cases are due to mutations in BRCA1/2 genes. Since their discovery in 1994 and 1995, as few as 18 mutations have been identified in BRCA genes in the Tunisian population. The aim of this study is to identify additional BRCA mutations, to estimate their contribution to the hereditary breast and ovarian cancers in Tunisia and to investigate the clinicopathological signatures associated with BRCA mutations. Methods A total of 354 patients diagnosed with breast and ovarian cancers, including 5 male breast cancer cases, have been investigated for BRCA1/2 mutations using traditional and/or next generation sequencing technologies. Clinicopathological signatures associated with BRCA mutations have also been investigated. Results In the current study, 16 distinct mutations were detected: 10 in BRCA1 and 6 in BRCA2, of which 11 are described for the first time in Tunisia including 3 variations that have not been reported previously in public databases namely BRCA1_c.915T>A; BRCA2_c.-227-?_7805+? and BRCA2_c.249delG. Early age at onset, family history of ovarian cancer and high tumor grade were significantly associated with BRCA status. BRCA1 carriers were more likely to be triple negative breast cancer compared to BRCA2 carriers. A relatively high frequency of contralateral breast cancer and ovarian cancer occurrence was observed among BRCA carriers and was more frequent in patients carrying BRCA1 mutations. Conclusion Our study provides new insights into breast and ovarian cancer genetic landscape in the under-represented North African populations. The prevalence assessment of novel and recurrent BRCA1/2 pathogenic mutations will enhance the use of personalized treatment and precise screening strategies by both affected and unaffected North African cancer cases.

Background: Breast cancer is the world's most common cancer among women. It is becoming an increasingly urgent problem in low-and middle-income countries (LMICs) where a large fraction of women is diagnosed with advanced-stage disease and have no access to treatment or basic palliative care. About 5-10% of all breast cancers can be attributed to hereditary genetic components and up to 25% of familial cases are due to mutations in BRCA1/2 genes. Since their discovery in 1994 and 1995, as few as 18 mutations have been identified in BRCA genes in the Tunisian population. The aim of this study is to identify additional BRCA mutations, to estimate their contribution to the hereditary breast and ovarian cancers in Tunisia and to investigate the clinicopathological signatures associated with BRCA mutations.
Methods: A total of 354 patients diagnosed with breast and ovarian cancers, including 5 male breast cancer cases, have been investigated for BRCA1/2 mutations using traditional and/or next generation sequencing technologies. Clinicopathological signatures associated with BRCA mutations have also been investigated.
Results: In the current study, 16 distinct mutations were detected: 10 in BRCA1 and 6 in BRCA2, of which 11 are described for the first time in Tunisia including 3 variations that have not been reported previously in public databases namely BRCA1_c.915T>A; BRCA2_c.-227-?_7805+? and BRCA2_c.249delG. Early age at onset, family history of ovarian cancer and high tumor grade were significantly associated with BRCA status. BRCA1 carriers were more likely to be triple negative breast cancer compared to BRCA2 carriers. A relatively high frequency of contralateral breast cancer and ovarian cancer

INTRODUCTION
Breast cancer is the most common malignancy among women worldwide (1). Incidence and mortality rates of breast cancer differ between populations (1). In Tunisia, it remains the most common cancer among females and represents the first leading cause of cancer mortality among women. The mean age at diagnosis of Tunisian breast cancer cases is around 50 years old, a decade younger than Western countries (2,3).
BRCA1 and BRCA2 are the most prominent breast cancer susceptibility genes that convey high risk of breast and ovarian cancers (4). Since their discovery, a wide range of mutational spectrum have been described for both genes. So far, more than 1800 distinct BRCA1 and 2000 BRCA2 mutations have been reported in the Breast Cancer Information Core (BIC) database. These mutations explain around 20-30% of breast cancer genetic component and seem to be associated with different other cancers such as prostate, pancreatic, endometrial and melanoma (5). The identification of novel BRCA1/2 mutations has important clinical implications. Indeed, unaffected BRCA mutation carriers have various preventive options including extensive and regular surveillance, chemoprevention, and risk-reducing surgery (6)(7)(8), while, affected cases carrying BRCA mutations could benefit from personalized therapeutic options such as platinum-based chemotherapy and poly (ADP-ribose) polymerase (PARP) inhibitors (9,10). However, a full BRCA1 and BRCA2 gene screening remains a labor and time-consuming challenge due to the large gene size, diverse mutations, or variants of unknown significance (VUS) and complexity of large genomic rearrangements (LRs) and copy number variations (CNVs) requiring special technical approaches. Recent advances in high throughput sequencing technologies including Target panels and whole exome sequencing (WES) allowed rapid, sensitive, and costeffective screening of the large BRCA genes. In addition, the decreased cost of genotyping and sequencing offered affordable targeted testing options.
Sequencing thousands of cancer samples showed that the frequency of germline mutations in BRCA genes varies widely among populations. Some mutations are shared between different populations and others are ethnic specific (11,12). Indeed, in certain countries and ethnic communities, the BRCA mutation spectrum is limited to a few founder mutations (13,14). This is mainly observed in geographically, culturally, or religiously isolated populations and in countries with high rates of consanguinity and endogamy that undergo rapid expansion from a limited number of ancestors. Consequently, some alleles become more frequent which explain the high frequency of some founder mutations in these populations. The founder effect may, therefore, influence mutation prevalence and gene penetrance. Since cancer risk is a function of mutation prevalence and penetrance that seems to vary by ethnicity, investigating the prevalence, the frequency, and the penetrance of novel BRCA mutation in different populations will bring new insights on cancer risk and etiology.
In Tunisia, previous studies on BRCA genes have focused only on breast cancer patients. In some studies, the genetic investigation concerned all the coding regions of BRCA1 and/or BRCA2 genes and in others only hotspot exons have been investigated. These reports have revealed a total of only 18 distinct mutations of which 12 are localized within BRCA1 gene including 2 large rearrangements encompassing exons 5 and 20. Among the identified mutations c.211dupA, c.5266dupC in BRCA1 and c.1310_1313delAAGA in BRCA2 were the most recurrent mutations encountered among the hereditary breast cancer cases (11,(15)(16)(17)(18)(19)(20)(21)(22)(23). Despite these efforts, the mutational spectrum of BRCA1/2 genes is still not well established. The main goal of the present study was to identify additional novel BRCA mutations and to investigate the contribution of these mutations to the missing heredity of breast and ovarian cancers. We also aimed to compare breast cancer clinicopathological characteristics in BRCA+ vs BRCA-Tunisian breast cancer cases. Written informed consents were obtained from all participants. The study has been conducted according to the Declaration of Helsinki Principles and ethical approval was obtained from the biomedical ethics committee of Institut Pasteur de Tunis (2017/ 16/E/Hopital A-M). Clinico-pathological characteristics and follow-up data were collected from patients' medical records. Probands were selected based on the following selection criteria (1): Presence of at least 3 related first or second-degree breast cancer cases at any age (2), Young cancer patients aged less than 35 years (3), Presence of at least two cases of breast or ovarian cancer, regardless of age, and at least one case of pancreatic cancer or prostate cancer in a related firstor second-degree patient (4), one case with triple negative breast cancer (TNBC) at an age ≤40 years (5), one breast cancer and one ovarian cancer cases diagnosed at first or second degree relatives at any age. A study flowchart is illustrated in Figure 1.

DNA Isolation
Total genomic DNA was isolated from peripheral blood using DNeasy blood DNA extraction Kit (Qiagen) according to the manufacturer's instructions. DNA purity and concentration were measured using a NanoDrop ™ spectrophotometer.
PCR reactions were performed on genomic DNA (gDNA), following standard protocols. Sanger sequencing has been performed using an automated sequencer (ABI 3500; Applied Biosystems, Foster City, CA) and a cycle sequencing reaction kit (Bigdye Terminator v3.1 kit, Applied Biosystems). The data were analyzed using BioEdit software version 7.2.5.
Sanger sequencing technique was then used to validate the identified mutations resulting from NGS.
NGS was performed on 75 breast and ovarian cancer cases. Targeted BRCA1/2 sequencing and whole exome sequencing were performed on 60 and 15 patients respectively.

Targeted Gene Sequencing
Targeted gene sequencing was performed on BRCA1/2 for 60 breast and ovarian cancer patients with strong family history. All targeted coding exons and exon-intron boundaries of BRCA1/2 genes were amplified with 253 pooled primer pairs. After the targeted amplification and construction of a library through QIAGEN Library Kit v2.0, the libraries were pooled prior to emulsion PCR and bead enrichment steps that were carried out

Genetic Analysis
Genetic analysis results showed that 36 out of 354 tested breast and ovarian cancer patients were BRCA1/2 mutation carriers (31 breast cancer cases and 5 ovarian cancer patients), including 21 patients with BRCA1 mutation and 15 patients carrying BRCA2 mutation. A total of 16 mutations have been identified including 11 short indels, 4 single nucleotide variations (3 nonsense & 1 splicing) and 1 large rearrangement.

Identified BRCA1/2 Pathogenic Mutations in Breast Cancer Cases
Within the studied breast cancer cohort, 13 pathogenic mutations have been identified: 8 in BRCA1 and 5 in BRCA2 genes. Among the identified mutations, 9 are described for the first time in Tunisian population (6 in BRCA1 and 3 in BRCA2) ( Table 1).
Considering the BRCA1 gene, 6 patients belonging to 5 unrelated families were carriers of the recurrent c.211dupA mutation. Three patients belonging to 2 unrelated families were positive for c.5266dupC mutation. The missense c.1612C>T mutation has been identified in 2 related patients. c.19_47del, c.668dupA, c.2418dupA and c.5030_5033delCTAA mutations have been identified each in one patient. c.2433delC has been identified among 2 related patients. Except c.211dupA and c.5266dupC mutations, all remaining BRCA1 mutations are reported for the first time in the Tunisian population.
In the BRCA2 gene, 3 frameshift mutations as well as 1 splicing and 1 large rearrangement mutation were detected. Our results revealed 6 patients belonging to 5 unrelated families that are double heterozygous for BRCA2 gene. Indeed, these families were carrying two mutations classified as pathogenic in the ClinVar database namely c.632-1G>A and c.1310_1313delAAGA. Four additional patients carrying only the c.1310_1313delAAGA mutation have been identified including one male breast cancer. The c.7654dupA mutation was identified in 2 related patients with a strong family history of hereditary breast and ovarian cancer. The c.1389_1390delAG mutation has been identified in 1 additional male breast cancer case. Similarly, a large rearrangement mutation of BRCA2 gene (Del exons 1-16) has been identified in one patient. Among the identified BRCA2 mutations, BRCA2-Del exons 1-16 mutation is novel and was not described in public databases. c.632-1G>A and c.1389_1390delAG are described for the first time in the Tunisian population.

BRCA1/2 Pathogenic Mutations Identified in Ovarian Cancer Cases
A total of 19 ovarian cancer patients were screened for BRCA pathogenic mutations using Sanger and/or NGS. Four distinct deleterious mutations were identified: 3 mutations in BRCA1 gene (c.915T>A, c.1612C>T and c.3049G>T) and one BRCA2 mutation (c.249delG).
BRCA1-c.915T>A and BRCA2-c.249delG mutations are novel and not described in public databases. The other identified mutations are described for the first time in the Tunisian population. c.1612C>T mutation was identified among 2 patients. This same mutation was also identified in 2 related breast cancer patients. Screening for additional carriers of the identified mutations, based on their geographic origin, was performed using Sanger sequencing. Consequently, the geographic origin of the identified BRCA1/2 mutations has been clearly established ( Figure 2). We have also illustrated the distribution of BRCA mutations identified in hereditary breast and ovarian cases on BRCA1 and BRCA2 genes ( Figure 3). Breast cancer cluster regions (BCCRs) and ovarian cancer cluster regions (OCCRs) were assigned in Figure 3 according to the study of Rebbeck et al., 2015 (26

Polymorphisms and Variant of Unknown Significance Identified in BRCA1 and BRCA2 Genes
In addition to the pathogenic mutations that have been identified in BRCA genes, several SNPs and variants of unknown significance (VUS) have been observed (Supplementary Table 1). Among the 101 identified variants, 45.54% were coding, 50.49% were intronic and 3.96% were localized within regulatory regions. The majority of variations were classified as benign or likely benign in the ClinVar database (91.08%) and five intronic variations were not reported. One patient carried a VUS rs397507308 in BRCA2 and 3 other patients carried 2 intronic variations (rs276174878 and rs276174816) that have conflicting interpretations of pathogenicity. Another patient diagnosed with early onset bilateral breast cancer had an in-frame variant with conflicting interpretations of pathogenicity rs80358343 (c.5017_5019delCAC) in the BRCA1 gene.

Clinico-Pathological Features of BRCA Carriers Among Breast Cancer Cohort
Clinico-pathological characteristics of breast cancer cases carrying BRCA1 and BRCA2 mutations are described in Tables 2, 3 respectively. We investigated these clinicopathological features in BRCA1/2 mutation carriers vs BRCA negative patients ( Table 4) and between BRCA1 and BRCA2 mutation carriers (Table 5), as well.

BRCA+ vs BRCA-
Family history of ovarian cancer was significantly associated with BRCA positive status (p=0.004). Regarding the mean age at

BRCA1 vs BRCA2
Association between clinico-pathological features and BRCA status (BRCA1+, BRCA2+ and BRCAx) was shown in Figures 4,5. Our results showed that there were no significant differences between BRCA1 and BRCA2 mutated groups regarding the mean age at diagnosis, the family history of personal cancers, of breast cancer and ovarian cancer ( Table 5).  Pathology showed that the infiltrating ductal carcinoma was the most common histological type in both groups (100% and 92.86%). HER2 status, lymph node involvement, SBR grade, tumor size, Ki-67 index and metastatic status showed no statistically significant difference between both studied groups. However, BRCA1 carriers were more likely to have triple negative breast cancer (p=0.002) and BRCA2 carriers were more likely to have luminal B breast cancer tumors (p=0.000078). In addition, positive estrogen receptor (ER) status and positive progesterone receptor (PR) status studied separately were both associated with BRCA2 mutated tumors (p=0.000056 and p=0.000084), respectively.

Follow Up of BRCA1 and BRCA2 Carriers
Among BRCA carriers, contralateral breast cancer and ovarian cancer co-occurrence were observed respectively in 22.58% and 16.12% of cases. One patient diagnosed with early onset breast cancer has undergone a contralateral prophylactic mastectomy and is currently under regular surveillance. Both contralateral breast cancer and ovarian cancer occurrence were more frequent in BRCA1 than BRCA2 carriers. Also, 22.58% of the carriers have developed distant metastases and 5 cases died due to disease progression.

DISCUSSION
Detection of mutations in hereditary breast and ovarian cancer related BRCA1 and BRCA2 genes is an effective method of cancer prevention, early detection, and treatment. Mutations in the highly penetrant BRCA genes explain around a quarter of these cases (27). The frequency of germline mutations identified on both genes varies depending on the geographic and ethnic distributions. In some populations, a wide spectrum of different mutations is present, whereas in other groups specific recurrent BRCA mutations have been reported that may be due to the founder mutation effect (28)(29)(30)(31)(32).
Our previous studies, investigating breast cancer loci and Nucleotide Excision Repair pathway, have shown that the Tunisian population is an admixed and intermediate population between Sub-Saharan Africans and Europeans (33,34). This genetic diversity reflects the inter-ethnic variability in the frequency distribution of the studied polymorphisms. Indeed, allele frequencies of several variants were found to be statistically different between Tunisian and other populations including rs2046210 and rs941764 that site in breast cancer susceptibility loci (33). These findings are in favor of the    (38). In fact, at a young age, the human organism usually functions as well as it ever will. However, interactions between some genetic and environmental factors (GxE) may cause a physiological decline of some organism systems leading to early disease presentation. Therefore, the influence of specific genetic background, differences in variant penetrance and frequency between populations along with environmental factors may explain this early onset of the disease. Large cohorts of young breast cancer patients should be studied to elucidate these GxE factors. For ovarian cancer cases, the mean age at diagnosis was 52.62 years and the majority presented with serous ovarian carcinoma. Previous studies have shown that among all patients diagnosed with serous ovarian carcinoma, which is the most common subtype, over 15% will have germline BRCA mutations (39).
Among the 16 distinct deleterious mutations that have been observed c.19_47del, c.668dupA, c.915T>A, c.1612C>T, c.2418dupA, c.2433delC, c.3049G>T and c.5030_5033delCTAA in BRCA1 and c.-227-?_7805+? (Del exons 1-16), c.249delG, c.632-1G>A, c.1389_1390delAG in BRCA2, are reported for the first time in the Tunisian population. We have also identified an inframe deletion reported to have a conflicting interpretation of pathogenicity effect in early onset bilateral breast cancer patient BRCA1_c.5017_5019delCAC. This variation has been described in multiple breast and ovarian cancer cases, with some families showing incomplete co-segregation of the variation (40)(41)(42).
Among BRCA mutations identified in breast cancer patients BRCA1_c.211dupA, BRCA1_c.5266dupC,BRCA2_c.-227-?_7805 +?,BRCA2_1310_1313delAAGA, BRCA2_c.1389_1390delAG and BRCA2_c.7654dupA occurred in BCCRs that are considered to be associated with an increased likelihood of breast cancer compared to ovarian cancer. Considering BRCA mutations identified in ovarian cancer patients BRCA1_c.1612C>T and BRCA1_3049G>T arose in OCCRs. Other mutations, namely c.19_47del, c.668dupA, c.915T>A, c.2418dupA, c.2433delC, c.5030_5033delCTAA in BRCA1 and c.249delG, c.632-1G>A in BRCA2 do not overlap with previously reported breast or ovarian cancer cluster regions. This could be explained by ethnic differences in BRCA mutation spectrum or it may indicate shared cluster regions for both breast and ovarian cancer.
In the BRCA1 gene, the c.19_47del mutation was identified in one breast cancer patient. This mutation was previously described only in the Algerian population (43). The c.2433delC mutation was described in Korean breast and ovarian patients (44,45), and in Mexican patients (46,47). The pathogenic c.1612C>T mutation was identified in 4 breast and ovarian cancer patients. This mutation has been identified in Brazilian population (48), in ovarian cancer patients from Israeli population (49) and in Macedonian population (50). We also detected the c.668dupA mutation in one patient. This latter has not been reported in previous studies neither in Tunisia nor in other populations. Nevertheless, it is already listed and classified as pathogenic in ClinVar and predicted to result in the substitution of Alanine to Glycine (p.Ala224Glyfs) which leads to BRCA1 protein truncation. Another new mutation was identified in BRCA1 gene, c.2418dupA, that was reported by our group for the first time in the Tunisian population and was not reported previously in other populations (51). c.3049G>T has been identified in one ovarian cancer patient. This mutation has been reported in Thai patients with non-mucinous epithelial ovarian cancer (52). The c.5030_5033delCTAA mutation was identified among one patient with breast and ovarian cancers and it is reported in Brazilian population (48). The c.915T>A mutation is novel and not described in public databases.
In addition to the identification of rare and novel BRCA1 mutations, other mutations seem to be recurrent and/or were described in previous Tunisian reports. The c.211dupA mutation was shared by 6 patients belonging to the same geographical origin. This mutation has so far been reported only in hereditary breast/ovarian cancer families of Tunisian origin, particularly in the North-East region, suggesting a founder effect. In order to unravel the genetic specificities of this mutation and to trace its origin a haplotype analysis has been conducted by our group on the North Eastern region (51). Results have determined the founder haplotype segregating with this mutation and have revealed that it arose in the period of colonization approximately 130 years ago.
The c.5266dupC mutation has been identified among two families. This mutation was previously described in 8 Tunisian breast cancer families (11,16,17,20). It was originally described as an Ashkenazi founder mutation. Haplotype analysis has shown that this mutation arose approximately 1800 years ago in Northern Europe (53). Then, it has been reported in several other populations such as, Italian, Russian Slovenian and Greek (54).
Interestingly for BRCA2 gene, 6 breast cancer patients were double heterozygous carrying the two deleterious mutations c.632-1G>A and c.1310_1313delAAGA, and 4 other unrelated patients carried only the c.1310_1313delAAGA mutation including one male breast cancer (MBC). c.632-1G>A mutation appears to be rare in other populations since it was only reported in one patient with prostate cancer in the UK (55). However, c.1310_1313delAAGA seems to be a founder mutation in Maghrebin countries (16,17,56,57). It has been also identified in patients with Lebanese (58), European (59)(60)(61)(62), African (63), Asian (64) and Latino ancestry (65) as well as in Caribbean cohorts (66,67). These results show the genetic heterogeneity of breast and ovarian cancers in Tunisian patients and the admixed origins of BRCA mutations in Tunisia.
In addition, 5 male breast cancer cases were investigated among which 2 carried BRCA2 mutations (c.1310_1313delAAGA and c.1389_1390delAG). Male breast cancer is a rare disease accounting for less than 1% of all breast cancer cases and it was previously shown that nearly 90% of MBC arising in BRCA mutation carriers are found to harbor a BRCA2 mutation (68). Unfortunately, being a man with "a women's disease" makes MBC a disease surrounded by social taboo and lack of awareness especially in underdeveloped countries. Indeed, the treatment of MBC has been extrapolated from the knowledge of female breast cancer, despite the multiple differences in the pathogenesis, biology and genetics of these two disease entities. These evidence make MBC a gender issue that requires more attention from the scientific community.
The introduction of the c.1310_1313delAAGA mutation, that have been encountered in diverse populations, in the Tunisian population could be explained by the immigration of Andalusians in Tunisia which has been intensified after the fall of Granada in 1492 and lasted for two centuries before the total expulsion of all Andalusian Moriscos from the Iberian Peninsula in 1610. The diverse geographical distribution of this mutation may further suggest independent origins as shown for the 4184del4 BRCA1 mutation reported to have at least three independent origins in the study of Neuhausen et al. (69). The c.7654dupA BRCA2 gene mutation which was identified in a unique family with a strong family history of breast and ovarian cancer is reported previously    and exclusively in Algerian population (70) and could be therefore specific to North African countries. Through this report and despite the identification of novel mutations in Tunisian population, it is clear that the genetic susceptibility to breast cancer is explained in a vast majority of cases by recurrent mutations. Indeed, more than 44.44% of carriers harbor BRCA1-c.211dupA or BRCA2-1310_1313deAAGA mutations which highlights the importance of screening these mutations in the treatment workflow of cases with early onset or strong family history of breast cancer. In fact, identifying germline BRCA1 and BRCA2 pathogenic mutations is a crucial component in the medical management of affected patients. Regular surveillance and/or prophylactic mastectomy of the second breast or prophylactic salpingo oophorectomies, which have been shown to reduce the risk of developing cancer, are recommended to these carriers. Moreover, relatives who test positive for a germline BRCA pathogenic mutation may take appropriate action to prevent cancer or have cancer diagnosed as early as possible for better treatment options (59).
In addition, mutations in the BRCA genes and their associations with clinico-pathological features were reported in several studies (71)(72)(73)(74). However, in Tunisia this aspect was not previously investigated. This point was raised in the present study and our results showed that patients with BRCA1 and BRCA2 mutations were similar with regard to several epidemiological and clinico-pathological parameters. Nevertheless, BRCA1 carriers were more likely to be triple negative breast cancer compared to BRCA2 carriers (p=0.002) and BRCA2 carriers were more likely to be luminal B breast cancer tumors (p=0.000078). Consistent with our findings, various previous studies reported that there is a much higher rate of TNBC among BRCA1 mutation carriers (75,76) and BRCA2-related breast cancer is often luminal (77). Additionally, positive ER was significantly associated with BRCA2+ tumors (p=0.000056). PR status was significantly different between BRCA1 and BRCA2 mutation carriers; BRCA2 carriers are more likely to develop progesterone receptor (PR) positive tumors and PR-negative breast cancer are associated with BRCA1 mutation carriers (p=0.000084). It was reported that the ER positivity was predominantly seen in BRCA2 mutation carriers, which is consistent with our findings (71,78). Furthermore, a previous report has found that BRCA2-associated cancers are mainly PR positive (79). Other studies have raised some pathological differences between BRCA1/2 mutation carriers and BRCAx patients. In our study, BRCA carriers seem to be younger than BRCA-negative patients (p=0.049). Furthermore, patients with a positive family history of ovarian cancer are more likely to be BRCA positive (p=0.004). We also observed a significant predominance of SBR grade III tumors among BRCA1/2 mutations carriers (p= 0.027). These findings are in line with previous literature (35,80,81).
Furthermore, we have assessed disease outcomes in BRCA carriers, and we have observed a relatively high proportion of contralateral breast cancer and ovarian cancer occurrence that were more frequently observed in BRCA1 carriers. Previous reports have demonstrated that women carrying a pathogenic mutation in the BRCA1 or BRCA2 genes have an increased risk of developing a second primary cancer in the contralateral breast. The cumulative risk 20 years after breast cancer diagnosis was estimated to be 40% for BRCA1 carriers and about 26% for BRCA2 carriers (82). In accordance with our findings, it was shown also that the occurrence of both breast and ovarian cancer in a woman is associated with a high likelihood of a germline BRCA1 mutation (83).
Besides the BRCA genetic mutations that have been identified in our study, mutations on other high to moderate breast cancer genes such as TP53, ATM, BLM and CHEK2 have been also identified for the first time in North African populations (data not shown). All these findings reflect the genetic heterogeneity of cancer predisposition in Tunisia and highlights the importance of the use of NGS to identify clinically actionable genetic variants that have a crucial role in disease management. Therefore, technological advances in terms of array and DNA sequencing technologies made the route towards the examination of genetic risk largely clear. However, practical challenges related to marked population-specific differences still exist. In this context, Manolio and colleagues conveniently classified LRRK2 as a high penetrant gene associated with Parkinson disease (84) with G2019S mutation being the main cause of Parkinson familial cases. Recently the international LRRK2 consortium reported a worldwide frequency of 1% of LRRK2 G2019S, 30-40% in Arab patients from North Africa and 10-30% in Ashkenazi Jews, but is very rare in Asians (85,86). As a variant´s frequency has a direct impact on its penetrance, this example shows the ethnic-dependent penetrance of some important variants involved in complex diseases and the role of consanguinity and endogamy in shaping the genetic susceptibility to these diseases. Therefore, the same reflection can be applied on high and low penetrant breast cancer variants in order to review their penetrance in underrepresented populations such as North Africans. A disproportionate distribution of the identified mutations is observed between the Northern and Southern regions of Tunisia (Figure 2), with the vast majority found in the North. This can be explained by a selection bias because most of the recruited participants come from Northern governorates, but it can also be explained by the very high consanguinity rates in the South that reaches 98% in some cities and that may have an impact on BRCA mutations frequency and prevalence.
Additional limitations of our study have been observed. Indeed, to our best knowledge, this work represents the largest BRCA1/2 study in Africa. However, we believe that the sample size is still small and larger cohorts are needed to trace a clear and complete BRCA1/2 mutational spectrum in Tunisia. In addition, because of the limited resources dedicated to this work, we were not able to perform a complete sequencing of both genes for the whole cohort. Therefore, the frequency and prevalence of the identified mutations need to be assessed in larger studies. Clearly, the prevalence assessment of BRCA1 and BRCA2 mutations also rely on the quality of both cohort selection criteria and mutation ascertainment methods. The identification of novel BRCA mutations and the assessment of their penetrance in a specific population will help to implement more affordable and cost effective targeted genetic testing strategies.
Finally, up until now, most data on BRCA1/2 mutations associated with high risk for hereditary breast and ovarian cancer do not cover the North African populations. Accordingly, the novel mutations identified in this study will help to improve knowledge on the genetic component of hereditary breast and ovarian cancer in the North African region and will lead to a better clinical management of cancer patients. In addition, we are aiming to share genetic and phenotypic data with larger multi-ethnic Consortia of BRCA1/2 mutation carriers such as the Consortium of Investigators of Modifiers of BRCA1/2 (CIMBA) (87). This will make our findings more broadly useful and will give us a global overview of the similarities and differences that the Tunisian population has compared to other ethnicities.

CONCLUSION
In conclusion we have identified 16 distinct BRCA mutations in breast and ovarian cancer patients including 11 novel mutations in the Tunisian population. The recognition of the BRCA mutational spectrum and its geographical distribution in Tunisia is of keen interest for the scientific and medical communities as it helps to develop precise risk assessment tools, accurate genetic testing, cost-effective approaches for prevention and early detection of the disease as well as personalized treatments of BRCA related cancers for both affected and unaffected cancer cases.

DATA AVAILABILITY STATEMENT
The minimal dataset that would be necessary to interpret, replicate and build upon the findings reported in this study are included in this article and in its supplementary files. All identified mutations with their related details have been shared in the public database ClinVar under the following link "https://www.ncbi.nlm.nih.gov/clinvar/submitters/507986/". Any additional datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by The Biomedical Ethics Committee of Institut Pasteur de Tunis (2017/16/E/Hopital A-M). The patients/ participants provided their written informed consent to participate in this study. Written informed consent was obtained from participants for the publication of any potentially identifiable images or data included in this article.

AUTHOR CONTRIBUTIONS
YH prepared the study concept and design, supervised the study, did data analysis, data interpretation, drafted, and critically revised the manuscript. NMi and MB did the experiments, participated in participant recruitment, and participated in drafting and reviewing the manuscript. NMe contributed to clinical data analysis and reviewed the manuscript. SN contributed to participants recruitment and reviewed the manuscript. MR and OM contributed to data collection and reviewed the manuscript. HanB, YB, HR, OJ, ND, AZ, JA, HEB, SL and JBH contributed to the clinical investigation and recruitment of patients. AH, KR, FB, RM, SBA and HamB critically revised the clinicopathological section and the whole manuscript. SB and SA contributed to the study concept, design and supervision and critically revised the manuscript. All authors contributed to the article and approved the submitted version.