Selected aspects of inherited susceptibility to prostate cancer and tumours of different site of origin

Epidemiologic research conducted over the last two decades has led us to believe that inherited factors play an important role in the aetiology of prostate cancer, but the genes which underlie the inherited susceptibility are elusive. The most compelling associations to date are with genes involved in DNA damage repair, including BRCA2. In Poland we have initiated a programme to identify DNA variants which confer an increased risk of prostate cancer and other cancers. Here we review our recent results. We found that germline mutations in BRCA1, CHEK2 and NBS1 confer an increased prostate cancer risk in Polish men. We provide evidence that CHEK2 is a multi-organ cancer susceptibility gene. We show that inherited variation in RNASEL and MSR1 genes do not contribute to prostate cancer development in Poland.

I In nt tr ro od du uc ct ti io on n Research conducted over the last two decades has led us to believe that inherited factors play an important role in the aetiology of cancer [1][2][3][4][5][6][7][8]. Prostate cancer is among the leading causes of morbidity and mortality in men. Relatively little is known about the genetic determinants of this disease, but epidemiologic data suggest that dominant susceptibility genes may be responsible for up to 5% of all of cases [9][10]. Through linkage analysis, numerous chromosomal loci have been identified, but no clear prostate susceptibility gene has emerged. Three candidate susceptibility genes have been positionally cloned -HPC1, HPC2/ELAC2 and MSR1 -but a clear role for any of these genes in hereditary prostate cancer has not been established [11,12]. There is evidence that rare mutations of genes in the DNA damage signalling pathway and cell cycle control pathway (BRCA2, CHEK2 and NBS1) predispose to prostate cancer, but the contribution of these two genes to prostate cancer aetiology is relatively small [13]. Common variants in the genes in these pathways (CDKN1B, CDKN1A, ATMATM, XRCC1, ERCC2) also have been associated with an increased risk of prostate cancer [14][15][16]. The DNA damage signalling and cell cycle control pathways play a crucial role in the maintenance of the integrity of the genome in response to DNA damage and has been implicated in the pathogenesis of prostate cancer and of cancers at other sites. This paper reviews a range of studies which have been performed in Polish population with the following objectives: 1) to investigate the association between inherited variation in RNASEL, MSR1, NBS1 and BRCA1 genes and prostate cancer risk in the Polish population; 2) to investigate the role of CHEK2 mutations in inherited susceptibility to prostate cancer and malignancies of other sites in the Polish population.  [44][45][46][47][48][49] RNASEL and MSR1 were identified through linkage studies of prostate cancer families. Two mutations in RNASEL (Met1Ile and Glu265X) were originally described in familial prostate cancer cases from the USA [17]. Other more common variants in RNASEL (R462Q and D541E) were found to influence the risk of prostate cancer in men [18]. Germline mutations in the MSR1 gene (six rare missense variants and R293X truncating mutation) were first shown to segregate with hereditary prostate cancer in the USA. Common polymorphisms in MSR1 have been associated with increased prostate cancer risk [19]. To date, the roles of RNASEL or MSR1 genes in prostate cancer aetiology have not been investigated in Slavic populations. We investigated if inherited variation in these genes influences prostate cancer risk in Poland.

M Ma at te er ri ia al ls s a an nd d m me et th ho od ds s
The case group consisted of 737 prostate cancer cases. Of the 737 cases, 506 were from Szczecin and 231 were from other countries (Opole, Bialystok, Olsztyn). Study subjects were unselected for age or family history. Family histories were obtained from each subject. 110 patients had one or more first-or seconddegree relatives with prostate cancer (familial cases). The control group consisted of 511 unselected healthy elderly men aged 50 and above, taken from three family doctors practicing in Szczecin. None of the controls had cancer.
The polymorphisms in MSR1 and RNASEL were selected after sequencing of the entire coding region of these genes in 52 and 94 Polish men with familial prostate cancer, respectively. Sequencing revealed only two variants in the RNASEL gene (R462Q and D541E) and two variants in the coding sequence of the MSR1 gene (P275A and R293X). These DNA variants were then genotyped by restriction fragment length polymorphism polymerase chain reaction (RFLP-PCR) using the restriction enzymes AvaI (R462Q variant), MboI (D541E variant), Hpy8I (P275A variant) and MvaI (R293X variant).

R Re es su ul lt ts s
The prevalence of the sequence variants in cases and controls is shown is Table 1. The R462Q and D541E Selected aspects of inherited susceptibility to prostate cancer and tumours of different site of origin variants of RNASEL were seen in similar frequency in cases and controls. We saw a trend towards a higher frequency of the P275A variant in unselected cases than in controls (10 vs. 7.2%; OR=1.4, p=0.1). A truncating MSR1 R293X mutation was present in 1.6% of controls, and 1.6% of unselected cases. 1 1. .2 2. . G Ge er rm ml li in ne e 6 65 57 7d de el l5 5 m mu ut ta at ti io on n i in n t th he e N NB BS S1 1 g ge en ne e a an nd d s su us sc ce ep pt ti ib bi il li it ty y t to o p pr ro os st ta at te e c ca an nc ce er r Individuals with inherited recessive clinical syndromes, such as Nijmegen breakage syndrome (NBS), which is characterized by spontaneous chromosomal instability, immunodeficiency and a predisposition to cancer, carry a mutation in one of the genes in the DNA damage signalling pathway [20]. The product of the NBS1 gene is responsible for DNA damage repair [21].
A 5-bp deletion in exon 6 of NBS1 (657del5) is present in the majority of NBS patients from Eastern Europe [22]. This variant is present in approximately 0.6% of individuals (heterozygous carriers) from the general population in Poland [23,24]. It has been suggested that heterozygous carriers of the founder mutation of NBS (657del5 allele) might be at increased risk of cancer [25,26], but prostate cancer specifically has not been studied to date.

M Ma at te er ri ia al ls s a an nd d m me et th ho od ds s
The case group consisted of 340 men diagnosed with prostate cancer at the University Hospital in Szczecin. Family histories of cancer were obtained from each subject. Thirty-five patients had one or more firstor second-degree relatives with prostate cancer (familial cases). We also included a second set of 21 familial cases of prostate cancer from men who were referred for evaluation at the Hereditary Cancer Centre by family doctors or urologists because of familial aggregation of prostate cancers. There were 1500 unaffected control subjects. One thousand control subjects were selected at random from the computerized patient lists of three family practices in Szczecin. A second control group comprised 500 newborns from Szczecin for whom a sample of umbilical cord blood was obtained.
Allele-specific PCR was used to detect the NBS1 founder mutation in DNA isolated from peripheral blood leukocytes of cases and controls. A separate DNA sample was sequenced to confirm the presence of the NBS1 mutation.
For the loss of heterozygosity (LOH) studies, for each of the nine prostate cancers in men with an NBS1 mutation, a single non-carrier control tumour was selected. The control subject was born within 2 years of the patient and had a tumour of the same Gleason score as the matched patient. DNA was obtained from eight of the nine paraffin-embedded, microdissected tumours from NBS1 mutation carriers and from all of the nine non-carrier control subjects. For the LOH analyses, two primer pairs were used, corresponding to the polymorphic microsatellite markers D8S88 and D8S1811. PCR was performed using fluorescent primers. PCR products were separated in an ABI PRISM 377 DNA Sequencer (Applied Biosystems). Data collection and analysis were performed using ABI PRISM 377 Collection Software and GenScan Analysis Software Version 3.0 (Applied Biosystems). A signal reduction in one allele of at least 70% was taken as the threshold of recognition for LOH. The NBS1 mutant allele is five nucleotides shorter than the wild-type NBS1 allele. For the LOH analysis of mutation-positive cases, additional primers were designed specifically to amplify exon 6 of NBS1, which contains the deleted sequence. PCR conditions using this primer set were as for allele-specific PCR. This primer set generates two distinct fragments from constitutional DNA from men with an NBS1 deletion.

R Re es su ul lt ts s
The NBS1 mutation was present in 9 of 340 unselected patients with prostate cancer (2.6%) compared with only 9 of 1500 (0.6%) control subjects from the general population (OR=4.5; p=0.002). The 657del5 germline mutation was present in 5 of the 56 (9%) familial cases (OR=16; p<0.0001). We investigated the segregation of the NBS1 mutant allele with prostate cancer in four families. We were able to establish the mutation status in two affected males from each family; in each family, the NBS1 mutation was present in both affected members ( Figure 1).
To analyze whether the wild-type allele of NBS1 is lost in prostate cancer, we performed LOH analysis of microdissected prostate tumours from eight patients who carried the NBS1 mutation and from nine patients who were found not to carry the NBS1 mutation. The wild-type NBS1 allele was lost in seven of eight prostate tumours from carriers of the 657del5 allele, but loss of heterozygosity was seen in only one of nine tumours from noncarriers (p=0.003). BRCA1 mutations confer high risk of breast and ovarian cancer [27]. Several studies suggested an increased risk of prostate cancer in Ashkenazi Jewish men with a BRCA1 mutation (185delAG or 5382insC) [28][29][30]. Other studies, in non-Jewish populations, have found little or no evidence of an increased risk for prostate cancer in BRCA1 carriers [31][32][33][34]. In Poland, there are three common founder alleles in BRCA1 (C61G, 4153delA and 5382insC), which, in total, account for 90% of all BRCA1 mutations [27]. Here we investigated if Polish men who carry one of these three alleles in BRCA1 are at an increased risk of prostate cancer. Poznan, Rzeszow, Sucha Beskidzka). Study subjects were unselected for age or family history. Family histories were obtained from each participant. Two hundred and twenty-nine patients had one or more first-or seconddegree relative with prostate cancer (familial cases).
The control group consisted of a mix of 2000 newborn children from 10 hospitals throughout Poland (Szczecin, Bialystok, Gorzow, Katowice, Wroclaw, Poznan, Opole, Lodz and Rzeszow), 1570 adults selected at random from the patient lists of three family doctors practicing in the Szczecin region and 1000 individuals from Szczecin who submitted blood for paternity testing. In total there were 4570 population controls.

M Me et th ho od ds s
The 4153delA and 5382insC mutations were detected using a multiplex-specific polymerase chain reaction (PCR) assay. The third mutation (C61G) generates a novel restriction enzyme site in exon 5. This mutation is detected after digesting amplified DNA with AvaII. All mutations were confirmed by sequencing.

R Re es su ul lt ts s
A BRCA1 mutation was seen in eight of 1793 (0.45%) cases and in 22 of 4570 (0.48%) controls (OR=0.9; p=1.0) - Table 2. 5382insC is the most frequent mutation of the three Polish founder mutations. The 5382insC mutation was detected only in 0.06% of cases, compared with 0.37% of controls (OR=0.15; p=0.06). In contrast, 4153delA was more common in cases than in controls (0.22 vs. 0.04%; OR=5.1; p=0.1). The C61G mutation was also more frequent in cases than in controls (0.17 vs. 0.07%; OR=2.6; p=0.5). A statistical test of homogeneity of the OR rejected the null hypothesis that the ORs associated with the three mutations were similar (p=0.008).
A BRCA1 mutation was found in three of 229 (1.3%) familial prostate cancer cases, compared with five of 4570 controls (OR=12; 95% confidence interval (CI) 2.9-51; p=0.0004). The 4153delA mutation was present in one familial case (OR=10.0; p=0.3) and C61G was responsible for two other prostate cancer families (OR=13.4; p=0.008). The family with the 4153delA mutation contained two men with prostate cancer and the families with the C61G mutation contained four and five men with prostate cancer. The C61G segregated with prostate cancer in the two families ( Figure 2). Germline mutations in the CHEK2 gene have been described in several populations. For example, in the United States, 18 different CHEK2 mutations were found [35]. Two founder variants in the CHEK2 gene (1100delC and I157T) are present in Finland [36]. In the Ashkenazi Jewish population, a single S428F mutation was detected [37]. In order to identify CHEK2 variants present in the  Recently, a large deletion in CHEK2 was identified in several unrelated patients with breast cancer of Czech or Slovak origin. Haplotype analysis confirmed that the mutation had a single source. The geographical and ethnic extent of this founder allele has not yet been determined [38]. We sought to establish if this deletion is present in the Polish population.

M Ma at te er ri ia al ls s a an nd d m me et th ho od ds s
Three samples of pooled DNA, each including pooled DNA from about 500 people from Poland, were amplified with primers described previously for the detection of the deletion of exons 9 and 10 [38]. Ninety unpooled DNA samples representing individual controls were also included.
Short extension times (2 min) were applied during polymerase chain reaction (PCR) to amplify only a short allele containing the large deletion. A single PCR product of about 1.3 kb was amplified from all samples with the pooled DNA. The short product was seen in 1 of 90 DNA samples from single patients. The PCR products from all positive cases were sequenced.

R Re es su ul lt ts s
A large deletion of exons 9 and 10 of the CHEK2 gene (del5395) is also a founder mutation in Poland. A single PCR product of about 1.3 kb was amplified from all samples with the pooled DNA. The short product was seen in 1 of 90 DNA samples from single patients. The PCR products from all positive cases were sequenced. The deletion breakpoints were characterized at a nucleotide level (Figure 3). By our estimate, the length of this deletion is 5395 bp, and not 5567 bp as described in the original report [38]. CHEK2 gene mutations have been associated with an increased prostate cancer risk in men from the United States [35] and Finland [39], but the Polish population has not been studied. In this study, we investigated whether CHEK2 plays an important role in the development of prostate cancer in Poland.

Cezary Cybulski
variants were analyzed by restriction fragment length polymorphism PCR. 1100delC was analyzed using an allele-specific PCR assay. Positive results were confirmed by sequencing.

R Re es su ul lt ts s
Protein truncating mutations in the CHEK2 gene (del5395, 1100delC, IVS2+1G>A) were more frequent in 1864 men with prostate cancer than in 5496 control individuals (2.4 vs. 1.1%; OR=2.3; p<0.0001). The I157T missense mutation was also more common in unselected cases than in controls (7.6 vs. 4.8%; OR=1.6; p<0.0001). The frequency of CHEK2 mutations was also found to be higher in a series of 249 familial prostate cancer cases than in controls, both for truncating variants (4.8 vs. 1  A founder allele in CHEK2, 1100delC, has been reported to be a low-penetrance breast cancer susceptibility allele in several studies, and in many ethnic groups [37,[40][41][42][43][44][45]. Other CHEK2 variants (IVS2+1G>A, I157T, and S428F) have also been suggested to confer increased breast cancer risks [44,45]. Recently, a large deletion of exons 9 and 10 of CHEK2 was identified in two USA families at high risk of breast cancer [38]. The aim of the study was to establish the relationship between CHEK2 mutations and the risk of breast cancer Poland.

M Ma at te er ri ia al ls s a an nd d m me et th ho od ds s
This study included prospectively ascertained cases of invasive breast cancer diagnosed throughout Poland from 1996 to 2003. The case group consisted of two groups of women with breast cancer. The first group had 3228 cases diagnosed at age 50 or below. The second group included a sample of 1978 patients, unselected for age. Of these, 752 were under the age of 51 and were also enrolled in the early-onset study described above. All cases were unselected for family history. The control group consisted of 5496 patients. The controls are described in detail in section 2.3.
Genotyping methods are described in detail in section 2.3.

R Re es su ul lt ts s
Protein truncating mutations (del5395, 1100delC, IVS2+1G>A) were observed more frequently in 1978 unselected breast cancer cases compared to controls (2.3 vs. 1.1%; p<0.0001). The frequency of missense variant I157T (p=0.0001) was also greater among unselected breast cancer cases than in the control group (6.8 vs. 4.8%; p=0.001). The odds ratio for breast cancer associated with truncating mutations (OR=2.4) was greater than that associated with the missense variant I157T (OR=1.4) - Table 4.
Protein-truncating mutations (del5395, 1100delC, IVS2+1G>A) were detected in 2.2% of 3228 patients with breast cancer diagnosed before the age of 51. This frequency was significantly higher (p<0.0001) than in the control group (1.1%). The missense variant I157T was also significantly more common (p=0.002) among patients with early onset breast cancer (6.4%) than in controls (4.8%). The odds ratio for early onset breast cancer associated with a truncating CHEK2 mutation (OR=2.2) was greater than that associated with the I157T mutation (OR=1.4) - Table 4. Germline mutations in CHEK2 have been associated with a range of cancer types, in particular of the breast and the prostate [35][36][37][38][39][40][41][42][43][44][45][46][47]. Protein-truncating mutations in CHEK2 have been reported to confer higher risk of cancer of the breast and the prostate than the missense I157T variant. Recent studies from Finland suggest an increased risk of colon cancer among carriers of missense variant I157T, but not in carriers of CHEK2 protein truncating mutations [36].
The aim of the study was to evaluate the association between specific CHEK2 alleles and colon cancer.

M Ma at te er ri ia al ls s a an nd d m me et th ho od ds s
We studied 1085 colorectal cancer cases diagnosed between 1998 and 2005 in three centres in North-Western Poland. Patients were unselected for age and family history. 964 colon cancer cases were diagnosed in Szczecin and 121 cases were diagnosed in Koszalin and Kolobrzeg. One hundred and ten cases had first-degree relatives diagnosed with colon cancer (familial cases). The control group consisted of 5496 patients. The controls are described in detail in section 2.3. Genotyping methods are described in detail in section 2.3.

R Re es su ul lt ts s
The missense mutation I157T was overrepresented both in unselected cases (6.6%) and familial cases (10%) with colorectal cancer compared to controls (4.8%). The OR for unselected colorectal cancer cases with the missense mutation I157T was 1.5 (p=0.002) - Table 5. The OR for familial cases with the I157T variant was 2.2 (p=0.01). We saw no association between colorectal cancer risk and CHEK2 truncating alleles. CHEK2 mutations originally were found in families with Li-Fraumeni syndrome [46,47]. The CHEK2 protein is expressed in a wide range of tissues and participates in the DNA damage response in many cell types; therefore CHEK2 is a good candidate for a multi-site cancer susceptibility gene.
The aim of the study was to assess the range of cancers associated with inactivating mutations in the CHEK2 gene.

M Ma at te er ri ia al ls s a an nd d m me et th ho od ds s
To establish the range of cancer types associated with CHEK2 mutations, we genotyped unselected cases of cancer with the most common types of cancer in Poland. In this part the results of 2001 cancer patients are described (excluding cases of the prostate, breast and colon cancers, which are described in detail in previous sections).
The control group consisted of: 2 000 newborn children from 10 hospitals throughout Poland (Szczecin, Bialystok, Gorzow, Katowice, Wroclaw, Poznan, Opole, Lodz and Rzeszow); 1000 adult patient lists of three family doctors practicing in the Szczecin region; 1000 adults unselected from family history from Szczecin who submitted blood for paternity testing.
Genotyping methods are described in detail in section 2.3. To confirm the chromosomal location of the observed CHEK2 mutations, we analyzed the polymorphic marker D22S275, which maps to intron 4 of CHEK2. We genotyped: 36 patients with the I157T variant; 24 patients with 1100delC; 52 patients with IVS2+1G>A and 50 individuals from the general population.

R Re es su ul lt ts s
The frequencies of the three CHEK2 variants in cases and controls are presented in Table 6. Because of their different effects on protein synthesis, the two truncating mutations (IVS2+1G>A or 1100delC) were considered separately from the missense mutation (I157T).
CHEK2 truncating alleles were associated with increased risk of breast and prostate cancer (described in the previous sections) and thyroid cancer (OR=4.9; p=0.0006). The missense variant I157T was associated with an increased risk of prostate, breast and colon cancer (described in the previous sections) and in addition with an increased risk of kidney cancer (OR=2.0; p=0.0006) and thyroid cancer (OR=1.9; p=0.04).
Although any individual finding might be due to chance, our study on the whole suggests that mutations in CHEK2 increase the risk of cancer in many different organs. A total of 52 comparisons were made. Of these, 13 were significant at the p=0.05 level (2.6 expected by chance) and 5 were significant at the p=0.01 level (0.5 expected). Furthermore, in all three sites for which a significant association was seen with the truncating mutation, a significant association was also seen with the missense mutation. This would be unlikely to be the case if the observations were due to chance.
To confirm the chromosomal location of the observed CHEK2 mutations we analyzed the polymorphic marker D22S275, which maps to intron 4 of CHEK2. All individuals with the 1100delC variant or with the I157T variant carried the 165 bp allele of D22S275, which we estimate has a frequency of 15% in the Polish population. All individuals with the IVS2+1G>A mutation carried the 171bp allele of the D22S275 marker, which we estimate has a frequency of 8% in the general population. These observations support the chromosome 22 assignment for the three variant alleles. A founder allele in CHEK2, 1100delC, has been reported to be a low-penetrance breast cancer susceptibility allele in several studies, and in many ethnic groups. Recent studies reported that patients with the CHEK2 1100delC variant had a 6-times higher risk of contralateral breast cancer and 3-times higher risk of distant metastasis [42,48,49] and more frequently developed ER-positive tumours than non-carriers (91 vs. 69%; p=0.03) [48]. Although the relationship between CHEK2 mutations and breast cancer is well documented, little is known about clinical characteristic of cancers in carriers of CHEK2 mutations.
To investigate the contribution of CHEK2 mutations to early-onset breast cancer in Poland, and to establish the characteristic features of these cancers, we compared clinical and pathological features of CHEK2-positive and CHEK2-negative breast cancers.

M Ma at te er ri ia al ls s a an nd d m me et th ho od ds s
The case group consisted of 3228 patients with breast cancer diagnosed before age 50, unselected from family history, who tested negative for BRCA1 mutation (4153delA, 5328insC, and C61G). The medical and pathology reports of the cases were reviewed locally by the physician associated with the study and relevant information was forwarded to the study centre in Szczecin. Information was recorded on age at diagnosis, stage, grade and lymph node status, oestrogen-receptor status, multicentricity and bilaterality. The data were collected from at least 70% of patients. Tumour blocks and/or paraffin-embedded slides were requested from the corresponding pathology centres. A central pathology review of was conducted in Szczecin by two pathologists associated with the study.

R Re es su ul lt ts s
A CHEK2 mutation was identified in 252 of 3,228 women with breast cancer (7.8%), including I157T (207 times), IVS2+1G>A (31 times), and 1100delC (16 times). The mean age of diagnosis in women with a CHEK2 mutation was similar to that of the noncarrier cases. However, the mean age of diagnosis of women with a truncating mutation was 1.8 years lower than women without a CHEK2 mutation (42.5 vs. 44.3 years; p=0.01).
The characteristics of the breast cancer cases in the 252 women with a CHEK2 mutation are presented in Table 7 and compared with non-carriers. Breast cancers in women with a CHEK2 mutation were more commonly of lobular histology (21.5 vs. 15.8%; p=0.05), of size greater than 2 cm (54.8 vs. 43.5%; p=0.01) or of multi-centric origin (28.7 vs. 19.5%; p=0.01) than were cancers from women without a CHEK2 mutation. Intraductal cancers (DCIS) with micro-invasion were also more common in women with a CHEK2 mutation than in non-carriers (11.3 vs. The great majority of women with a CHEK2 mutation did not have a strong family of cancer -13.8% of the women with breast cancer and a CHEK2 mutation were from a family with two or more firstdegree relatives with breast cancer. However, this was more frequent than reported by the non-carriers (8.9%) and the difference was statistically significant (OR=1.6; p=0.02).   Groups of individuals with an increased risk of prostate cancer in the Polish population can be identified by testing of specific variants in the NBS1, BRCA1 and CHEK2 genes. It seems that analysis of the RNASEL and MSR1 genes is not justified for this purpose. The list of known genetic markers of high risk of prostate cancer (in addition to strong family history of prostate cancer and germline mutations in the BRCA2 gene) may be extended by specific mutations in the NBS1, BRCA1 and CHEK2 genes in men with a positive family history of prostate cancer in at least one 1 st or 2 nd degree relative (the risk increased about 5-15 fold).
CHEK2 is multi-organ cancer susceptibility gene. It seems justified to consider surveillance of the prostate, breast, thyroid, kidney and colon as an option for carriers of CHEK2 mutations. It appears reasonable to consider magnetic resonance imaging of the breast beginning from the age of ~35 as an option for mutation carriers, as CHEK2 mutations may predispose to lobular and multifocal cancers in young women. However, the establishment of surveillance protocols for carriers of a CHEK2 mutation requires further studies.
Most carriers of NBS1 and CHEK2 mutations report negative family history of prostate, breast and colorectal cancers. The use of DNA tests is the only way to find carriers of these changes conferring increased risk of tumour development. Identification of low penetrance DNA variants may be very important if the simultaneous presence of such alternations and/or combinations with external risk factors in a carrier would add up to a clinically significant high risk of tumour development.