Li-Fraumeni-like syndrome associated with a large BRCA1 intragenic deletion

Germline mutations of the tumor suppressor gene TP53 account for more than half of the families with classic Li-Fraumeni syndrome (LFS) [1], which is an inherited condition characterized by the development of sarcomas and other early-onset tumors, including breast cancer [2, 3]. Families presenting incomplete features of LFS are referred as having Li-Fraumeni-like syndrome (LFL). Depending on the criteria adopted to classify the cancer phenotype in a given family, up to 22% of LFL pedigrees have detectable TP53 mutations [4‐6]. Several cancer predisposition syndromes that involve breast cancer have been described to date, and include, in addition to LFS/LFL, the hereditary breast and ovarian cancer (HBOC), hereditary diffuse gastric cancer, and the Cowden and Peutz-Jeghers syndromes [7]. Due to the high frequency of breast and other cancers in LFS/LFL individuals, there may be an overlap of phenotypes, and often some families fulfill genetic testing criteria for more than one hereditary breast cancer syndrome [1, 8, 9].

Several studies have investigated the frequency of BRCA1/BRCA2 and TP53 germline mutations in families with multiple early-onset breast cancers [6, 8, 10, 11]. Approximately 5-10% of breast cancer is estimated to result from dominant mutations in known single genes [12‐14], particularly in the BRCA1 or BRCA2 genes. Germline TP53 mutations have been considered to be responsible for only a small fraction of the hereditary breast cancer cases overall [15], and have mostly been described in families with the other core-cancers of LFS/LFL [1, 8, 9]. Germline mutations of the BRCA2 gene have been described in families presenting both breast cancer and sarcomas, suggesting that BRCA2 mutations account for a proportion of LFS/LFL families negative for TP53 mutations [16, 17]. As far as we are aware, germline BRCA1 mutations have not been detected in LFS/LFL kindreds, not even among families presenting a complex cancer history consistent both with LFL and other syndromes that constitute the HBC phenotype [6, 8, 11, 18].

All known breast cancer susceptibility genes present germline point mutations in only approximately 20-25% of the cases fulfilling the criteria for genetic testing [12]. Gene rearrangements can contribute to disease through different mechanisms, resulting in either imbalance of gene dosage or gene disruption, and they are not usually detected by routine molecular diagnostic methods such as gene sequencing. In particular, large rearrangements, most often deletions, have been reported as a cause of cancer susceptibility, occurring in at least 30% of highly penetrant Mendelian cancer-predisposing genes [19].

BRCA1 germline rearrangements have been implicated in up to 30% of HBC families in certain populations [19‐23]. The aim of the present study was to determine the frequency of germline copy number changes of TP53 BRCA1, and BRCA2 genes in breast cancer patients with clinical diagnosis of Li-Fraumeni or Li-Fraumeni-like syndrome, and without detectable germline TP53 point mutations.

All studied patients were females affected by breast cancer, two of them with bilateral disease, and 11 (45.8%) with more than one primary tumor. The average age at breast cancer diagnosis was 41 years (SD: 11.5; range: 26–61 years). Nineteen of the 23 families met genetic testing criteria for both LFL and another hereditary breast cancer syndrome (Table 1); two families met criteria for both classic LFS and another hereditary breast cancer syndrome, and two fulfill only the criteria for LFL.

In the MLPA analysis none of the patients showed TP53, or BRCA2 deletions or duplications. We identified a single patient carrying a heterozygous intragenic BRCA1 microdeletion (Y54). Analysis using two different sets of MLPA probes (kits P087 and P002) and array-CGH allowed confirming a deletion that spanned from exon 9 to 19 (Figure 1 depicts the chromosome 17q21.31 array-CGH profile of the patient, indicating the position of the BRCA1 microdeletion). We tested two non-affected relatives of patient Y54 (III.13 and III.16) and found that one of them carries the BRCA1 deletion (III.16). Unfortunately, affected relatives of the patient Y54 could not be investigated for the presence of the BRCA1 deletion either because they were deceased or were not available.

The DNA fragment containing the rearrangement breakpoints was sequenced and the results showed that the deletion starts at intron 8 and ends at intron 19 of the BRCA1 gene, resulting in a deletion-block identified as: g.29197_65577del36381 (Figure 2). Detailed in silico assessment of the genomic sequences surrounding the breakpoints showed that consensus Alu sequences flanked them.

Clinically, this family fulfilled genetic testing criteria for both hereditary breast and ovarian cancer (HBOC) and LFL (Eeles 1 criteria) syndromes; the cancer family history was significant for the presence of two individuals with multiple primary tumors, including the proband (Figure 3).

In families with a breast cancer history that suggests the involvement of high risk genes such as TP53, BRCA1 and BRCA2, a more extensive analysis of these genes should be considered. In this study we have screened three major breast cancer predisposition genes for copy number changes in a group of 23 breast cancer patients with the clinical diagnosis of LFS/LFL who had no germline TP53 point mutations.

We did not identify large rearrangements encompassing TP53, which is in line with previous reports of low prevalence of such alterations, encountered in less than 5% of LFS/LFL families [24, 25].

Similarly, large rearrangements in other breast cancer predisposition genes seem to be infrequent. A few BRCA2 deletions have been previously reported in families with male breast cancer [26], and contribute to inactivate this gene in breast cancer families [21, 27]. Rearrangements affecting the BRCA2 gene have also been reported in breast/sarcoma families, causing a Li–Fraumeni type of cancer pattern [16]. Although none of the families included in this study had male breast cancer cases, nine of them had a breast cancer/sarcoma phenotype; however, no BRCA2 rearrangements were identified, which may be related to the relatively small sample size.

BRCA1 rearrangements, on the other hand, are more prevalent mostly due to the high density of Alu elements throughout the BRCA1 locus [28]. A large study by Walsh et al (2006) [11] suggested that the mutation spectra of BRCA1/BRCA2 includes several genomic rearrangements, and those alterations seem to be particularly frequent in certain populations (due to founder effect), and in families presenting individuals with multiple primary tumors [20, 21, 29, 30]. Indeed, the “multiple primary tumors” phenotype was observed in the BRCA1 rearrangement-positive family identified in our series. Interestingly, the BRCA1 microdeletion identified here appears to be the same as the one identified in a breast cancer Italian patient [20]. Our patient is originally from southern Brazil, and since Italians have strongly contributed to the ethnic make-up of southern Brazilian population [31] it is possible that the Brazilian and the Italian patients have a common ancestry. Considering that we could not establish the parental origin of the rearrangement, this large genomic deletion may represent a breast cancer susceptibility allele rather than a more general cancer predisposition factor.

This study contributes to the understanding of the etiology of cancer susceptibility in Li-Fraumeni (LFS) and Li-Fraumeni-like (LFL) families, and their possible relation to large genomic rearrangements in high risk breast cancer susceptibility genes.