Hereditary breast and ovarian cancer: assessment of point mutations and copy number variations in Brazilian patients

One hundred twenty unrelated breast cancer patients fulfilling criteria for hereditary breast and ovarian cancer (HBOC) were recruited from 2007 to 2010 for this study. The inclusion criteria were: 1) Breast cancer diagnosed ≤ 45 years of age (no family history); 2) Breast cancer diagnosed ≤ 45 years of age with 1 or more close blood relative with breast/ovarian/fallopian tube/primary peritoneal cancer at any age; 3) Breast cancer diagnosed <45 ≤ 50 years of age with 1 or more blood relative with breast/ovarian/fallopian tube/primary peritoneal cancer ≤ 50 years of age; 4) Breast cancer diagnosed >50 of age with 1 or more blood relative with breast/ovarian/fallopian tube/primary peritoneal cancer at any age; 5) Two primary BC when the first occurrence was prior to age 50; 6) Breast cancer with a history of ovarian/ fallopian tube/primary peritoneal cancer at any age; 7) For an individual with an ethnicity that is associated with a higher mutation frequency (e.g., Ashkenazi Jewish); 8) Personal history of ovarian/fallopian tube/primary peritoneal cancer; 9) Personal history of male breast cancer. All enrolled individuals received genetic counseling and signed an informed consent. This study was performed in compliance with the Helsinki Declaration and was approved by the ethics committee of the A C Camargo Cancer Center (approval number: 870/06-B). The complete clinical and molecular information of the patients is given in the Additional file 1.

DNA from peripheral blood was purified using the Puregene Genomic DNA Isolation kit (Quiagen, Hilden, Germany) according to manufacturer’s instructions. The entire coding sequence and exon-intron boundaries of the BRCA1 (U14680 or NM_007294.3) and BRCA2 (U43746 or NM_000059.1) genes were evaluated. The CHEK2 gene (NM_007194.3) and the TP53 gene (NM_000546.5) were screened solely for the c.1100delC and p.R337H mutations, respectively. All PCR products were sequenced in both forward and reverse directions on an ABI Prism 3130xl genetic analyzer (Life Technologies, Foster City, USA). Mutations were recorded and referenced with respect to the cDNA sequence using the nomenclature proposed by the BIC database [14]. PCR conditions and primer sequences are available upon request.

Patients negative for BRCA1/2 mutations were investigated for CNVs in these genes. Exonic deletions and duplications affecting BRCA1 and BRCA2 genes were investigated on genomic DNA using the multiplex ligation-dependent probe amplification (MLPA) commercial kits P087-B1 and P045-B3 (MRC-Holland, Amsterdam, The Netherlands) according to the manufacturer’s recommendations.

The BIC database was searched for all BRCA1 and BRCA2 alterations [14]. Unreported mutations that generated a premature stop codon (nonsense and frameshift) were classified as pathogenic. Missense alterations classified as class 3 in the IARC_LOVD database [15] or unknown in the BIC database were considered to be variants of uncertain significance (VUSs). Variants classified as class 1 and 2 using the IARC_LOVD were considered to be wild type. In cases of inconsistency between the databases, the classification from IARC-LOVD prevailed. Additionally, the VUSs were characterized using three in silico protein prediction algorithms: SIFT [16], POLYPHEN-2 [17] and Align-GVGD [18].

For transcriptional analysis of the novel splice site variant, frozen tumor tissues were obtained from the carrier and from a sporadic breast tumor that was negative for mutations in the BRCA1 gene, which was used as a control sample. RNA samples were purified using the Precellys 24® equipment (Carlsbad, California, USA), followed by total RNA extraction using an RNeasy Mini kit (Qiagen, Venlo, The Netherlands). The first strand cDNA was synthesized from 2 μg of total RNA using a random hexamer primer with the Superscript first strand system for RT-PCR (Life Technologies, Foster City, USA). RT-PCR fragments of the index patient and the control sample were both obtained according to standard PCR protocols using primers adjacent to the exon involved in the splice site mutation. All RT-PCR products were inserted into the T/A plasmid vector pTZ57R/T using the InsT/Aclone PCR Product Cloning Kit (Thermo Fisher Scientific, USA), and the ligated plasmid was used for transformation in DH10B E. coli cells via electroporation (2.5 KV, 25 μ FD, 200 OHMS), followed by single-colony sequencing on the ABI 3130xl sequencer using M13 primers.

Genomic CNVs affecting 14 cancer susceptibility genes (PTEN, TP53, ATM, NBN, RAD50, BRIP1, PALB2, RAD51, MLH1, MSH2, MSH6, CDKN2A, CDH1 and CTNNB1) were investigated in oligo-based array-CGH data obtained from a previous study [19]. All hybridizations were gender-matched and processed in reverse labeling duplicates; experiments were carried out using the 180 K whole genome platform (design 22060, Agilent Technologies, Santa Clara, CA, USA). The Agilent Genomic Workbench software was used for the detection of CNVs (deletion and duplications) with the aberration detection method 2 (ADM-2) and a threshold of 6.7. The duplication or deletion of genomic segments were declared when one probe exhibited a log₂ ratio of Cy3/Cy5 > 0.70 or < -0.70, respectively. A careful visual inspection was performed to filter out poor quality hybridizations and noisy data. Additionally, alterations located more than 3 kb upstream or downstream of the coding exons were excluded. The Database of Genomic Variants (DGV-HG19, http://​projects.​tcag.​ca/​variation/​) was used to exclude common variants detected in the general population. Only alterations detected in both experiments of the same patient were considered.

To validate the genomic DNA CNVs, we used the quantitative duplex PCR method previously described [20]. GAPDH was used as a reference gene. For array-CGH probes located in introns, PCR primers were designed to encompass the closest exon.

The mean age of diagnosis of the first primary tumor in all 120 patients was 43 years-old (yo), with a range from 22 to 88. Thirty-one out of 120 patients (26%) were found to harbor pathogenic mutations, including 20 for BRCA1 (64.5%), seven for BRCA2 (22.5%), three for TP53 R337H (10%) and one for CHEK2 1100delC (3%). In the BRCA1 gene, 20 patients presented 18 different mutations, of which 16 were point mutations (89%) (Six nonsense, six frameshift, two splice site and two missense) and two were CNVs, one deletion encompassing exons 16 and 17 and a rare case of exon 24 amplification detected within BRCA1 (11%) (Table 1 and Additional file 2). The Ashkenazi Jewish mutation c.5382insC was the most recurrent and was found in three cases. The seven BRCA2 carriers presented six distinct mutations, of which three were nonsense and three were frameshift. The nonsense variant c.9709A > T was found in two cases. No CNVs were detected in BRCA2. Moreover, seven out of 28 (25%) BRCA1/2-carriers were found to harbor novel pathogenic mutations, five in BRCA1 and two in BRCA2.

Interestingly, patient MO-15 was found to harbor a germ line splice site mutation (c.560 + 2 T > A) in intron 7 of the BRCA1 gene (Figure 1A). RT-PCR products encompassing part of exon 6 to exon 8 revealed the presence of a 186-bp fragment in addition to the expected fragment of 258 bp (Figure 1B). Sequencing of the 186-bp fragment disclosed a frameshift deletion of the last 62 bp of exon 7 due to the activation of a novel cryptic splice site within exon 7 (Figure 1C). Figure 1D shows the schematic representation of the premature stop codon created in the mRNA after the deletion of 62 bp of exon 7 (r.[=, 499_560del); p.Ser127Thrfs*11) caused by the germ line splice site mutation c.560 + 2 T > A. The predicted protein from the aberrant mRNA apparently created an isoform of 137aa (Figure 1E).

Nineteen out of 120 HBOC patients (16%) harbored VUSs according to our criteria (based on BIC and/or IARC-LOVD databases - see the Methods section). Among the VUS carriers, 17 variants were distinct, and two of them were described for the first time (Table 2). The VUS evaluation using the three protein prediction algorithms (PolyPhen, SIFT and GVGD-Align) showed that six were classified as likely pathogenic by at least one algorithm (three in one algorithm, two in two algorithms and one in all three algorithms). The schematic representation of all VUSs in the BRCA1 and BRCA2 genes and their functional domains on the protein are shown in Figure 2.

Overall, the mutation detection rate was 26%; however, when selecting according to age of cancer onset, young women (≤35 yo) had the highest rate at 35% (Figure 3). Regarding specific HBOC criteria fulfilled by each family, the majority of mutation carriers (45%) met criterion 2 (Breast cancer diagnosed ≤ 45 yo and familial history positive for BC). Four out of five patients fulfilling the criterion for Ashkenazi Jewish ancestry (criterion 7) harbored pathogenic mutations, and patients fulfilling criterion 6 showed a detection rate of 44% (4/9). Patients younger than 45 yo without a family history of cancer (criterion 1) did not present pathogenic mutations in our cohort (Table 3).

Among the 120 cancer patients included in this study, 100 had array CGH data available from a previous study [19], which were used for detecting CNVs within the 14 breast cancer susceptibility genes. After a careful visual inspection, two of them were found to harbor CNVs in PTEN and ATM genes. The exon-4 heterozygous deletion in the ATM gene (Patient SM-46, carrier of a pathogenic BRCA2 mutation, c.K3161*) and exon-2 heterozygous deletion in the PTEN gene (Patient SM-62) were confirmed using the duplex qPCR gene dosage method (Figure 4). However, both the array CGH probe and gene dosage PCR primers for the PTEN gene were located in the deleted region that has recently been described as a polymorphic chromosomal deletion [21].