Introduction

Germline mutations in the BRCA1 (MIM #113705) and BRCA2 (MIM #600185) genes account for a substantial portion of familial breast and ovarian cancer cases. BRCA1 and BRCA2-associated tumors are characteristic in their early age of onset with aggressive profile [1]. The cumulative lifetime risk of BRCA1/BRCA2 carriers for developing cancer is clearly elevated above the average risk population: up to 80% of developing breast cancer [2] and up to 54% of developing ovarian cancer [2]. In addition to incomplete penetrance, the age of onset of first malignancy varies even among carriers of identical mutations [3]. Thus, additional genetic or environmental factors may play a role in modifying breast and ovarian risk in BRCA1/BRCA2 mutation carriers’ tumorigenesis and account for the differences.

One of the goals of individualized medicine is to more accurately predict individual’s risk for being diagnosed with cancer and identify means to favorably alter that risk by tailoring early detection and prevention schemes. Defining the precise modifier genes that affect cancer risk in BRCA1/ BRCA2 carriers has obvious clinical implications To date, except for the ovarian cancer region within BRCA2 (OCCR), the penetrance of the different mutations within the BRCA1/2 genes has not been documented. It is conceivable that variations in additional genes will affect the penetrance of BRCA1 and BRCA2. Several examples have been reported: a polymorphism in the 5′ UTR of the RAD51 gene (135 GC, rs1801320) increases the risk for breast cancer, among BRCA1/2 carriers [4–6]. The length of the polyglutamine repeat in the AIB1 gene is associated with younger age at breast cancer diagnosis among Ashkenazi-Jewish and non Jewish BRCA1 carriers [7, 8] and recently conflicting results were reported on the additional risk conferred by the functional snp in the AURKA gene, F31I, to BRCA1/2 carriers. A small study found an increased breast cancer risk in BRCA1/2 carriers [9], while a larger study did not detect any association with a modified risk of breast cancer in BRCA1 and BRCA2 carriers [10].

The BRCA1 and BRCA2 genes encode for large multifunctional proteins involved in DNA damage repair, cell cycle regulation and overall maintenance of genome stability [11]. Phenotypic similarities as embryonic lethality and gross chromosomal rearrangements between Brca1- and Brca2-deficient murine models, imply that the two gene products function in a common pathway necessary for maintenance of genomic stability [12, 13].

The tumor suppressor p53 has a pivotal role in cells decision to live or die. p53 deficiency eliminates the cell cycle arrest and the ability of cells to undergo apoptosis in response to DNA damage. Thus, p53 deficiency renders cells hypersensitive to DNA damage, and allow for cell survival and accumulation of genomic instability [14]. Thus, cooperation between the BRCA1/2 and p53 pathways in tumor suppression has been suggested [13]. Mutations in the p53 gene are highly frequent in breast cancer as in many types of cancers. It has been noted that BRCA1-tumors have higher frequency of p53 mutations than sporadic breast cancer [15, 16]. Moreover, deficiencies in both pathways seem to enhance tumor development. Mutant p53 rescued the embryonic lethality of Brca1 as well as Brca2 knockout mice [12] and in a different model, p53 deficiency accelerated mammary tumor formation in Brca1 conditional mutant mice [17]. Thus, it is plausible that the p53 pathway modulate BRCA1-related tumorigenesis.

MDM2 is a negative regulator of p53 and attenuates the p53 tumor suppression pathway [18]. Recently, Bond et al. [19] reported that the G-allele of SNP309 at the promoter of the MDM2 gene, introduces an SP1 binding site, which results in increased MDM2 transcript and protein [19]. The original report as well as subsequent reports described that the presence of the GG genotype influences the age of diagnosis and tumor formation in Li-Fraumeni syndrome and in variety of additional cancers including soft tissue sarcoma, endometrial cancer, gastric cancer, renal cell carcinoma and hepatocellular cancer [20–23]. Conflicting results were published in regard to risk association of SNP309 with breast cancer. While some studies found no association between SNP309 and cancer risk, others reported earlier age of onset among the GG genotype carriers [24–27].

Taken together, the present study aims to investigate whether SNP309 modulates the age at diagnosis of breast and ovarian tumors in BRCA1 and BRCA2 mutation carriers as well as in non-carrier breast cancer patients among AJ women. The population in the study is unique in that the BRCA1/BRCA2 mutation spectrum is limited to three predominant mutations: 185delAG and 5382insC in the BRCA1 gene and 6174delT in the BRCA2 gene. Moreover, the haplotypes associated with these mutations are limited [28–30] and recently we reported that the MDM2 haplotype associated with GG genotype among AJ show little variability [31].

Materials and methods

Study population

All study participants were of Ashkenazi–Jewish (AJ) origin consulted and tested at the Oncogenetics unit or at the Genetic Institute of the Sheba Medical Center in Tel-Hashomer, Israel. Three populations were studied: (1) BRCA1/2 mutation carriers, (2) non-carriers breast cancer patients and (3) control group without cancer. The BRCA1/2 carriers group (n = 260) is further divided into two subgroups of women who developed breast and/or ovarian cancer (n = 140) and unaffected women (n = 120). Carriers who underwent preventive surgeries were removed from the study. The distribution of mutations within the carrier groups is as follow: 183 women are carriers of mutations in the BRCA1 gene (140 carriers of 185delAG, 42 carriers of 5382insC) and 74 women are carriers of the mutation 6174delT in the BRCA2 gene. Of the 140 BRCA1/2 carrier cases: 93 were affected with breast cancer, 33 with ovarian cancer and 14 had both cancers.

A group of consecutive non-carrier breast cancer patients (n = 187) were identified through the Oncology Department of the Sheba Medical Center between the years 2001–2004. Patients were genotyped for the presence of predominant BRCA1/2 mutations at the Shaare Tsedek Medical Center in Jerusalem, Israel and all were found to be non-mutation carriers.

The Control group (n = 138) consisted of cancer-free females who tested at the Genetic institute of the Sheba Medical Center for parental diagnostic tests. The median age of this group was significantly younger than the other two groups and cancer expectancy is considered as in the general Jewish–Ashkenazi population. All participants signed an informed consent approved by the Institutional Review Board and by the Israeli National Genetic Review Board. Characteristics of patients and controls in the different study groups are shown in Table 1.

Table 1 Characteristics of case and control subjects
Full size table

Genotyping

Genomic DNA was extracted using the PUREGENE DNA isolation kit (Gentra Systems) according to manufacturer instructions. SNP 309 (rs2279744) was described previously [19]. SNP genotyping was performed in 384-well microplates with a high-throughput system of chip-based mass spectrometry (MALDI-TOF) using the MassARRAY by SequenomTM Inc. (La-Jolla, CA) as described previously [31]. For determination of allele-specific primer products, PCR primers were tagged with 5′-ACGTTGGATG-3′ at the 5′ end to avoid interference with the mass spectra. The primer sequences were as follows: forward: 5′-AGTTCAGGGTAAAGGTCACG-3′ and reverse: 5′-TCACACTAGTGACCCGACAG-3′. Amplification of 2.5 ng of cDNA was performed in 2.75 mM MgCl2, 200 μM dNTP, 0.1 U HotStart Taq DNA polymerase (Qiagen) and 1 pmol of each forward and reverse PCR primers in 5 μl total volume. PCR conditions: 95°C 15 min, followed by 45 cycles of 95°C for 30 s, 56°C for 1 min, then 72°C for 1:30 min, with an extension at 72°C for 7 min. Products were then treated with 0.04 U shrimp alkaline phosphatase, SAP (Sequenom) followed by extension cycle, to which 1.2 μM final concentration of extension primer 5′-GGGCTGCGGGGCCGCT-3′ and 0.6 U of ThermoSequenase (Sequenom) were added to a total reaction of 9 μl with the termination mixture. The extension conditions include a 94°C 2 min with 75 cycles of the following: 94°C for 5 s, 52°C for 5 s and 72°C for 5 s. Quality control and quality assurance were provided by randomly including non-DNA containing wells in the chip as well as re-genotyping about 10% of samples on different chips. Reproducibly was determined as 97%.

Statistical analysis

Difference in allele frequencies between younger and older BRCA1/2 carrier’s cancer patients and non-carrier cancer patients were calculated by χ 2-square analysis. To evaluate specific genotype associated risk, age-adjusted odds ratio (OR) with corresponding 95% confidence intervals (95% CI) and P values were calculated. All statistical analyses were done using SPSS software.

Results

The genotype distribution of the rs2279744 in the entire population sample set did not deviate from the Hardy–Weinberg equilibrium (P = 0.34). However, the frequency of the G allele in the AJ population is significantly different from the frequency originally published for mixed North American population [19]. While the G allele is the minor allele in the North American population with frequency of 0.34, in the Jewish–Ashkenazi population, the G allele is the major allele with frequency of 0.54. The distribution of genotypes among the populations of healthy individuals (n = 138), breast cancer cases (n = 187) and BRCA1/2 mutation carriers (n = 260) is shown in Table 2. In all the groups the genotype distribution did not deviate from Hardy–Weinberg equilibrium (Table 2) and we did not find any significant differences in genotype or allele distribution among the different study groups.

Table 2 Distribution of MDM2 SNP309 in cases and controls
Full size table

Next, we determined the frequency of the MDM2 SNP309 within the BRCA1/2 carriers cohort, stratified by type of mutation and disease occurrence (Table 3). As expected, the frequency of the three common mutations in BRCA1 and BRCA2 genes: 185delAG, 5382insC and 6174delT is equal among cancer patients and healthy carriers (Table 3). The MDM2 SNP309 genotype distribution among the BRCA1 cancer patients and healthy carriers did not vary significantly. However, we noticed that the GG genotype is two times more prevalent in healthy BRCA2 carriers than in cancer-diagnosed BRCA2 carriers. (Table 3).

Table 3 MDM2 SNP309 genotype among BRCA1 and BRCA2 carriers
Full size table

To determine whether the GG genotype is associated with early onset breast/ovarian cancer in BRCA1/2 carriers, we compared the frequency of the MDM2 SNP309 genotype in symptomatic BRCA1/2 carriers diagnosed with cancer at or before age 51 old with the frequencies of the SNP309 genotype in women diagnosed with cancer older than 51 years. Age 51 years was selected as our cut-off because it is considered as the average age of menopause and was shown previously to be the age discriminator in B-cell lymphomas, sarcomas and invasive ductal breast carcinoma [19]. Seventy percent (98/140) of breast and ovarian cancer cases in the BRCA1/2 carriers group were diagnosed before age of 51 years, while only 33.7% (63/187) of sporadic breast cancer cases were diagnosed before that age. These expected results reflect the genetic predisposition and the predicted early age at diagnosis of cancer among BRCA1/2 carriers.

We found that the frequency of the GG genotype is as much as 2.9 times higher in the group of BRCA1/2 carriers diagnosed with cancer at or before 51 years (35.5%), than in the group of symptomatic BRCA1/2 carriers with late onset breast cancer, older than 51 years old (11.9%) (P = 0.019) (Fig. 1a). Among the early onset cancer patients 35 were GG (35.5%) 47 were GT (49%) and 16 were TT (16.3%), while among cancer patients diagnosed after the age of 51 years old 5 were GG (11.9%) 27 were GT (64.3%) and 10 were TT (23.8%). The overall OR for young BRCA1/2 patients with the GG genotype was 3.93 (95% CI 1.41–10.90, P = 0.009) .

Fig. 1
figure 1

MDM2 SNP309 genotype frequencies in early and late onset of BRCA1/2 and non-familial cancer cases. (a) MDM2 SNP309 genotype frequencies in early and late onset of BRCA1/2 cancer cases. Frequencies of SNP309 genotypes were calculated in BRCA1 and BRCA2 carriers according to age at diagnosis of breast and ovarian cancers before (dark bar) and after (light bar) age of 51 years old representing early and late onset. Chi-square analysis of genotypes indicate significant differences in allelic distribution in early and late onset cases (P = 0.0083). (b) MDM2 SNP309 genotype frequencies in early and late onset of non-familial breast cancer cases. Frequencies of SNP309 genotypes were calculated in consecutive breast cancer patients that do not carry the three predominant mutations in the Jewish–Ashkenazi population. Frequencies were calculated according to age at diagnosis of breast cancer before (dark bar) and after (light bar) age of 51 years old representing early and late onset. Chi-square analyses of genotypes indicate no significant differences in allelic distribution in early and late onset cases (P = 0.127)

Full size image

A similar analysis was carried out for breast cancer patients who are non-mutation carriers. No differences in genotype frequency and distribution by age at diagnosis using 51 years as a cut-off were noted in these cases: 63 of sporadic breast cancer patients who were diagnosed by age 51 years and 124 patients were diagnosed after that age. Among the early onset cancer patients 23 were GG (36%) 28 were GT (44%) and 12 were TT (19%). Similar distribution was observed in cancer patients diagnosed after the age of 51 years old where 38 were GG (31%) 49 were GT (40%) and 37 were TT (29%) (Fig. 1b). The overall OR for young breast cancer patients without BRCA1/2 mutations with the GG genotype was 1.30 (95% CI 0.69–2.47, P = 0.419).

To determine whether the effect of SNP309 on age at diagnosis is associated with breast cancer or ovarian cancer, we analyzed the SNP309 genotype in carriers according to their cancer type and age at diagnosis. Patients who were diagnosed with both breast and ovarian cancer were analyzed according to age at detection of the first malignancy. The numbers for ovarian cancer were too small (n = 35) and hence this analysis was omitted (data not shown). We found that the GG genotype is 2.8 times more frequent in early onset breast cancer patients among BRCA1/2 carriers than in older BRCA1/2 carriers’ breast cancer patients (P = 0.02). The overall OR for young BRCA1/2 breast cancer patients with the GG genotype was 3.83 (95% CI 1.05–13.94, P = 0.024) (Fig. 2a).

Fig. 2
figure 2

The GG genotype at MDM2 SNP309 associates with early onset breast cancer in BRCA1 carriers. (a) MDM2 SNP309 genotype frequencies in early and late onset of BRCA1 cancer cases. Frequencies of SNP309 genotypes were calculated in BRCA1 carriers according to age at diagnosis of breast and ovarian cancers before (dark bar) and after (light bar) age of 51 years old representing early and late onset. Chi-square analyses of genotypes indicate significant differences in allelic distribution in early and late onset cases (P = 0.026). (b) MDM2 SNP309 genotype frequencies in early and late onset of breast cancer cases in BRCA1/2 carriers. Frequencies of SNP309 genotypes were calculated in BRCA1/2 carriers according to age at diagnosis of breast and ovarian cancers before (dark bars) and after (light bars) age of 51 years old representing early and late onset. Chi-square analysis of genotypes indicates significant differences in allelic distribution in early and late onset cases (P = 0.019)

Full size image

Additional analysis for assessing the age-at diagnosis effect of the MDM2 SNP309 on breast cancer by gene (BRCA1 Vs. BRCA2) was carried out. This analysis showed that the GG genotype was more prevalent in early onset BRCA1 carriers compared with late-onset cases (P = 0.04). Among early onset cases BRCA1 carriers, 30 were GG (40.5%) 33 were GT (44.5%) 11 were TT (15%) while 5 were GG (18%) 16 were GT (57%) and 7 were TT (25%) among late onset Fig. 2b). The overall OR for young BRCA1 cancer patients with the GG genotype was 3.07 (95% CI 1.05–8.96, P = 0.04).

The number of BRCA2–related cancer cases in our cohort is small (n = 37) thus, a larger cohort of BRCA2 tumors is required to determine whether MDM2 SNP309 modifies the risk already conferred by BRCA2.

Discussion

The GG genotype of the functional MDM2 SNP309 was recently found as associated with accelerated tumorigenesis in individuals diagnosed with the Li-Fraumeni syndrome [19]. This genotype was also associated with earlier age at diagnosis in sporadic soft-tissue sarcoma and B-cell lymphoma patients [19]. In the present study, a statistically significant difference in the frequency of the GG genotype between BRCA1/2 mutation carriers diagnosed with cancer at or before age 51 years compared with mutation carriers diagnosed with cancer after that age. The GG genotype is three-times more frequent among younger BRCA1/2 cancer patients than in older BRCA1/2 cancer patients and is associated with a significant risk of developing early cancer (OR 3.93 CI 1.41–10.9 P = 0.009). No such differences in genotype distribution were noted among sporadic breast cancer patients. These results suggest that the GG genotype contributes to breast and ovarian cancer risk in genetically predisposed BRCA1/2 mutation carriers, and accelerates breast and ovarian carcinogenesis. Taken together, the MDM2 SNP309 may be viewed as a modifier of penetrance of mutant BRCA1/2 alleles.

Mutations in the BRCA1/2 genes increase cancer risk for both breast and ovarian cancers, but incomplete penetrace suggests that additional genetic modifiers may alter the site-specific cancer risk [2]. Indeed, site-specific effects have been observed for several modifier genes such as the rare alleles of the oncogene Ha-RAS1, which increase the risk for ovarian cancer, but not breast cancer, among BRCA1 carriers [32]. A polymorphism in the 5′ UTR of the RAD51 gene (135 GC, rs1801320) increases the risk for breast cancer, but not for ovarian cancer, among BRCA2 carriers [6, 33]. Our study suggests that the GG genotype at SNP309 of MDM2 increases the risk for breast cancer, predominantly in BRCA1 mutation carriers.

As the GG genotype of MDM2 SNP309 clearly affects age at diagnosis in breast cancer, one may speculate that hormonal changes in estrogen-signaling pathway due to menopause, account for this association [22]. Given the estrogen receptor status (ER negative) in majority of BRCA1 associated breast cancer [34], this explanation seems unlikely. Alternatively, genomic instability accumulation may account for cancer risk conferred by the different MDM2 alleles (GG, TT or GT) in BRCA1/2 mutation carrying women. The SNP309 G- allele generates an SP1 binding site that potentially alters the expression of the p53 pathway by up-regulating the MDM2 oncogene and down-regulating the p53 tumor suppressor gene. Loss of p53 creates a permissive environment for inappropriate proliferation and survival, impaired cell cycle checkpoint control signaling and attenuated apoptosis. Thus, loss of p53 or attenuation of the pathway leads to accumulation of genomic instability [14]. As BRCA1 and BRCA2 participate in the DNA damage response and specifically in DNA repair of double strand breaks (DSB) [35, 36], the combined effect of defects in both BRCA1/2 and p53 pathway could explain the acceleration in carcinogenesis that clinically manifests as an earlier age at diagnosis. Indeed, in murine models, p53 deficiency is highly cooperative with both Brca1 and Brca2 in promoting tumorigenesis [13, 17].

Subsequent to the original finding by Bond et al. [19], several conflicting reports were published on the association between SNP309 and breast cancer acceleration and outcome. Reports by Wilkening et al., Petenkaya et al. and Campbell et al. [24, 37, 38] found no association between SNP309 polymorphism and breast cancer risk and outcome regardless of family history. Both Wilkening et al. [37] and Campbell et al. [38] reported no association between familial breast cancer or patients at high risk without BRCA1/2 mutations and SNP309 while, Petenkaya et al. [24] reported no association between MDM2 SNP309 and sporadic breast cancer. Our results also indicate that there is no association between the G allele and risk of breast cancer among sporadic breast cancer patients. However, in contrast to a study published by Copson et al. [25] that reported no association between MDM2 SNP309 and risk in BRCA1 carriers, we find a three time increased risk of early onset breast cancer among BRCA1/2 carriers with a GG genotype. The discrepancy between of the two studies could stem from the fact that our population is relatively homogeneous and the mutation spectrum is limited to three predominant mutations with a strong founder effect of highly penetrant mutations (185delAG and 5382insC in BRCA1 and 6174delT in BRCA2). Moreover, even the haplotypes associated with BRCA1 mutations and MDM2 GG genotype are highly common [28–31]. Whereas, in the other studies, the population analyzed consists of male and female BRCA1 carriers with wide spectrum of mutations and varied penetrance. Therefore, it is conceivable that the added risk of MDM2 SNP309 is masked in that population, and analysis of risk should be carried in conjunction with haplotype analysis. Second, the G allele frequency in the Jewish Ashkenazi is higher than in mixed Caucasian population and thus, even the small added risk could have a high impact.

The fact that asymptomatic carriers are about 8 years younger than cancer-diagnosed carriers could potentially bias our results, especially as our patient’s follow-up time is limited. However, the current median age of the asympthomatic group is already older than the median age at diagnosis among the cancer-diagnosed carriers. Given the similar frequencies of BRCA1/2 mutations and MDM2 polymorphisms between the two groups, we suggest that yet an additional modifier or modifiers play a role in altering cancer risk among carriers. The small number of BRCA2 carriers in our cohort limited our ability to test the association between SNP309 and BRCA2 6174delT mutation. As the frequency of SNP309 was two times higher in asymptomatic carriers, analyzing a larger group of BRCA2 carriers is of particular interest as it may potentially clarify whether SNP309 have an opposing effect on disease course in BRCA2 carriers. Nevertheless, the additive risk of SNP309 to cancer-diagnosed BRCA1/2 carriers was stronger when the two groups of patients were combined rather than assessed individually.

In conclusion, our results show that Jewish–Ashkenazi BRCA1/2 carriers that harbor the GG genotype of SNP309 in the MDM2 intronic promoter have an increased risk of being diagnosed with early onset cancer relative to heterozygotes and TT genotype. Thus, SNP309 influences age at diagnosis among the high-risk BRCA1 and BRCA2 mutation carriers of Ashkenazi descent.