A novel germline PALB2 deletion in Polish breast and ovarian cancer patients

PALB2 protein [OMIM #610355, a partner and localizer of BRCA2] was recently identified in a complex with BRCA2 (Breast Cancer 2) protein [1]. PALB2 supports BRCA2 stability and determines its localization in the nucleus after DNA damage [1]. Relocation of PALB2 and BRCA2 to damaged chromatin is regulated by BRCA1 (Breast Cancer 1) protein [2]. These three proteins form a complex in which PALB2 acts as a bridge between BRCA1 and BRCA2 [2]. In cells depleted of PALB2, DNA repair pathway dependent on the BRCA1/2 is disrupted [1, 2]. This suggests that inactivation of the PALB2 gene may cause similar biological and phenotypic changes as inactivation of the BRCA1 or BRCA2 genes; the latter is responsible for a fraction of breast and ovarian cancers, and a dysfunction of BRCA2 alone - for the Fanconi anemia (FA) syndrome.

FA syndrome is a genetic disorder caused by germline biallelic mutations in one of 13 genes (including FANCD1/BRCA2 and FANCN/PALB2) of the so-called FA/BRCA pathway; this pathway controls the repair of double strand-breaks and the response to DNA crosslinking agents [3, 4]. Studies on FA patients provide evidence that biallelic mutations of the PALB2 gene are responsible for one of the syndrome's subtypes, which is very similar to the one caused by inactivation of the BRCA2 gene [5, 6].

Mutations of the BRCA1 and BRCA2 genes are the best known alterations of the FA/BRCA pathway. Heterozygous germline mutations of these genes predispose to breast and ovarian cancers. The cumulative risk of ovarian cancer to age 70 years for women with a BRCA1 or BRCA2 mutation is estimated at about 40% and 10%, respectively [7]. The cooperation between BRCA2 and PALB2 during the DNA repair process and similarity in clinical features of the FA syndrome caused by biallelic mutations of BRCA2 and PALB2 genes suggest the role of PALB2 alterations in carcinogenesis. To date, PALB2 alterations have been associated with familial breast [8‐13] and pancreatic [14, 15] cancers. To our knowledge, PALB2 gene has not been extensively studied in ovarian cancer patients.

The initial screening of 70 ovarian carcinomas revealed nine substitutions and one deletion of the PALB2 gene. These alterations, and the final number of the samples studied, are shown in Table 2.

We discovered a novel PALB2 germline deletion, c.509_510delGA, located in exon 4. It may be expected that due to this deletion PALB2 protein is shortened to 169 amino acids from 1186 amino acids of the wild type and it is devoid of C-terminal domain (with WD40 repeats). WD40 motifs are thought to be responsible for protein-protein interactions and they seem to be necessary for BRCA2/PALB2 complex formation [11]. It was shown that PALB2 mutants (229delT, 2521delA and c.1592delT) that lacked C-terminal domain had highly reduced BRCA2 binding capacity and were defective in the repair of double-strand breaks and mitomycin C-induced damages [10, 11]. To date, all PALB2 gene alterations detected in families with breast cancer or FA disease were frameshift or nonsense changes, leading to the expression of a truncated protein. These data strongly support the assumption that the deletion detected in this study results in the production of an inactive protein. Silent and missense sequence variants of the PALB2 gene have been detected in previous studies in both cancer samples and in control tissues, but none of those variants have been strongly associated with cancer susceptibility [9, 10, 12].

Our study proves that PALB2 alterations contribute to the familial but not to the sporadic breast cancer in Poland. This is in agreement with previous reports, in which alterations of the PALB2 gene were predominantly associated with familial breast cancer [9‐13]. Some authors reported PALB2 mutations also in pancreatic cancer [14, 15], male breast cancer [9] and in a single prostate cancer family [10]. Pancreatic cancer was mentioned at least twice in the family histories of our PALB2 mutation carriers.

The low frequency of PALB2 mutations (0.6%) observed in our study is similar to that noticed for breast cancer by other authors [8, 9, 11‐13]. Only Erkko et al. [10] found a higher rate of mutations in the familial breast cancer: 2.7%. In their study, however, the rate of the mutations in the control population was also relatively high (0.2%). No other study except Erkko's et al. [10] and our study found PALB2 truncating mutations in control populations.

We are the first to present a PALB2 truncating mutation in a patient with medullary breast carcinoma. To date, such alterations have been detected predominantly in ductal breast carcinomas. In agreement with data presented by Heikkinen et al. [23], our study suggests that breast cancers with PALB2 mutations are predominantly triple-negative ones; however, their phenotype does not completely overlap with the basal type.

In our study, PALB2 alterations did not associate with ovarian cancer risk. In addition, the PALB2 deletion was accompanied by a germline BRCA2 nonsense mutation in one ovarian cancer patient (Moes et al., unpublished data). A similar observation was published by Heikkinen et al. [23] who found two breast cancer patients with a PALB2 deletion among 104 BRCA2 mutation carriers. Interestingly, the presence of these two alterations in carcinomas appears to be more frequent than just one occurring by chance, considering the low frequency of PALB2 and BRCA2 mutations in control populations.

To date, PALB2 gene alterations have not been extensively studied in ovarian cancer patients. Erkko et al. [24] found three (0.5%) ovarian cancer patients with PALB2 gene mutations in 593 unselected ovarian cancer cases screened for the PALB2 c.1592delT alteration, identified earlier in breast cancers [10]. Another dysfunction of PALB2 detected in ovarian cancer was its expression silencing by promoter hypermethylation (in 4 of 53 sporadic cases) [25]. These studies, together with our data suggest that PALB2 protein may be defective or insufficient in some rare ovarian carcinomas; however, a possible contribution of this insufficiency to ovarian cancer development needs elucidation.

It is unclear how the PALB2 deficiency contributes to cancer development. It has been suggested that PALB2 participates in breast carcinogenesis through haploinsufficiency and/or the dominant-negative effect. In the majority of breast cancers with PALB2 gene mutations, the loss of heterozygosity (LOH) was not observed [10, 11]. The only evidence of deletion of the wild-type allele was presented by Garcia et al. [12]. In our study, there was a suspicion of LOH at PALB2 locus in one ovarian carcinoma with PALB2 deletion. The regular allele of PALB2 gene may also be eliminated by promoter hypermethylation (as above). Such dysfunction was revealed not only in ovarian but also in sporadic breast cancers [25]. Since wild-type PALB2 proteins oligomerize at damaged DNA [26], truncated PALB2 mutants may possibly complex with wild type protein, thus, disturbing its proper function. In addition, other molecular changes may cooperate with PALB2 deficiency in cancer development.

Some of PALB2 gene alterations in breast cancer were suggested to be founder mutations for other ethnic groups [8, 10]. The presence of the same deletion in seven unrelated women in our study might suggest that this was a founder mutation for the examined population from central Poland. Although the genotype analysis of the mutation carriers showed differences in the haplotypes, one cannot exclude an ancient founder mutation. More detailed analysis is necessary to determine the origin of this alteration.

Our study had some limitations. The entire coding sequence of the PALB2 gene was screened in 70 non-consecutive ovarian cancers only; the sensitivity of the SSCP ranges from 70% to 95%, according to different publications [27, 28], and it is 90% in our laboratory [20]. Thus, some PALB2 alterations (particularly of the missense type) could have been missed. Nevertheless, a practical value of this study is that the c.509_510delGA should be searched for in Polish families with breast cancer.

The c.509_510delGA is a novel PALB2 mutation that increases the risk of familial breast cancer. Occurrence of the same PALB2 alteration in seven unrelated women suggests that c.509_510delGA (p.R170fs) is a recurrent mutation for population from Poland. It may be useful to include this PALB2 mutation to a list of alterations that should be searched for in Polish families with breast cancer.