Prostate cancer is the most frequently diagnosed cancer, excluding basal and squamous cell skin cancers, and the second leading cause of cancer death in men in the United States. Significant familial clustering of prostate cancer cases has been observed [1
] suggesting that highly penetrant prostate cancer predisposition genes exist. Although a number of genes have been implicated in prostate cancer (see Review articles [2
]), most require validation and functional studies to determine their precise role in the etiology of prostate cancer.
One gene that has consistently been associated with prostate cancer is BRCA2
(see Review article[5
]). Early studies in Iceland identified a BRCA2
founder mutation, 999del5, that segregates with prostate cancer particularly in cases diagnosed ≤ 65 years. [6
] The Breast Cancer Linkage Consortium has reported an increased risk of prostate cancer in families with any known deleterious BRCA2
] First-degree male relatives of women with breast and/or ovarian cancer and a known BRCA2
mutation were found to have an increased risk of prostate cancer (RR = 4.65), and rates were found to be even higher in male relatives less than 65 years of age (RR = 7.33)[8
] However, a subsequent study of 22 families with a minimum of three men affected with prostate cancer that also included at least two cases of breast and/or ovarian cancer observed no BRCA2
] Another common BRCA2
founder mutation, 6174delT, has been found to increase prostate cancer risk in some individuals with Ashkenazi Jewish ancestry,[10
] but these results have not been confirmed in other studies [11
It has been estimated that germline mutations in BRCA2
account for about 5% of prostate cancer in familial clusters[15
] However, other than the founder mutation in Iceland and possibly another mutation in those with Ashkenazi Jewish ancestry, the role of segregating BRCA2
mutations in familial prostate cancer remains unclear. To further assess involvement of BRCA2
in familial prostate cancer, we screened for BRCA2
variants in high-risk prostate cancer pedigrees that showed at least nominal linkage to the BRCA2
region on chromosome 13q.
Utah Prostate Cancer Resource
The vast majority of studied Utah high-risk prostate cancer families have been identified using the Utah Population Database (UPDB)[17
] The UPDB is a computerized database that links cancer information from the Utah Cancer Registry (UCR) to genealogy information for almost four million individuals who are, for the most part, descendants of the nineteenth century pioneers to Utah. The UCR began collecting statewide cancer records in 1966, and in 1973 the UCR became part of the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) Program. As a SEER site, all cancers in the state of Utah, except basal and squamous cell skin cancers, are required by law to be reported. Approximately 60% of cancer records in the UCR link to a UPDB genealogy record. The Utah population represented in the UPDB have been shown to be similar to other Caucasian populations of Northern and Western European descent [18
] with normal levels of inbreeding[19
] On rare occasions, families with an excess of prostate cancer who reside outside of Utah are contributed to the Utah Prostate Cancer High-Risk Pedigree Resource by collaborators.
From the Utah High-Risk Prostate Cancer Pedigree Resource of over 300 sampled pedigrees, we selected a subset of the most informative prostate cancer pedigrees for a genome-wide linkage search. [20
] These pedigrees satisfied the following two criteria: all cases were no more distantly related than two meioses from another case and the resulting cluster contained at least four cases. There were no other exclusion criteria for the linkage analysis including age at diagnosis, or other cancers or disease states (e.g., breast cancer) within the family. We did not remove any prostate cancer pedigrees from our High-Risk Prostate Cancer Resource because of prior knowledge of a BRCA2
mutation. We note that our high-risk prostate cancer pedigree definition is more strict than the usual definition of a high-risk prostate cancer pedigree[5
] We identified 59 pedigrees, from 3 to 6 generations in size, containing 464 cases and an additional 1,261 unaffected relatives. We genotyped 246 of the 464 cases, as well as 640 of the 1,261 unaffected relatives to infer genotypes of deceased relatives or phase for genotyped cases. All enrolled subjects gave informed consent, and this study was approved by the University of Utah Institutional Review Board.
Genotyping was performed by the Center for Inherited Disease Research (CIDR) using a panel of 404 fluorescent microsatellite (STR) markers spaced ~10 cM apart across the genome. The laboratory methods used by CIDR are described in detail at http://www.cidr.jhmi.edu
. Mendelian inheritance was verified for all markers, and samples with an inheritance incompatibility were set to missing.
Linkage analysis was performed using MCLINK, [21
] a Markov chain Monte-Carlo method that allows for fully informative multilocus linkage analysis on large extended pedigrees, using the Smith et al
. prostate cancer inheritance model. [22
] MCLINK computes the robust multipoint linkage statistic proposed by Göring and Terwilliger, [23
] referred to as theta LOD or TLOD[24
] The results of this genome-wide linkage scan are reported elsewhere. [20
We identified 5 pedigrees with at least nominally significant linkage evidence (TLOD score ≥ 0.588) within approximately 10 cM of the BRCA2 gene on chromosome 13q12.3 at 24.4–24.6 cM, using the Marshfield map. A TLOD score of 0.588 corresponds to a p-value of 0.05, not accounting for multiple testing. Table provides details of the linkage evidence for each pedigree. For two of the five pedigrees (9445402 and 9990205), the highest genome-wide TLOD score was at the chromosome 13q12 region. For the remaining three pedigrees, the chromosome 13q12 peak was the second (9775803), third (9434307) and fourth (9435901) highest of all peaks. For kindred 9775803, the 13q12 peak was preceded by a peak at 2q36-q37. For kindred 9434307, the 13q12 peak was preceded by peaks at 12q24 and 2p25-p24. Finally, for kindred 9435901, the 13q12 peak was preceded by peaks at 18p11-q22, 19q13, and 1p35-p34.
Characteristics of five high-risk prostate cancer pedigrees showing linkage evidence to chromosome 13q, BRCA2 region1
In each linked pedigree, we selected the youngest prostate cancer case that carried a segregating haplotype of interest as well as the most distantly related individual in the pedigree who also carried the same segregating haplotype for BRCA2 mutation screening. One pedigree had two 13q segregating haplotypes; both were screened for mutations. Hence, twelve individuals were selected for BRCA2 mutation screening. Ten of the subjects for mutation screening were prostate cancer cases and two of the subjects were unaffected relatives (1 male and 1 female) who carried the 13q segregating haplotype.
BRCA2 mutation screening
All coding regions and intron/exon boundaries of the BRCA2
gene were screened for mutations using 47 primer pairs that have been described previously. [25
] Shorter exons were amplified in a single fragment and longer exons were amplified as sub fragments. There were five fragments for exon 10, 16 fragments for exon 11, and two fragments for each of the following exons: 14, 18, and 27. PCR reactions were performed in a reaction volume of 10 μl containing 5 ng of genomic DNA, 1 mM dNTPs, 1.5 mM or 2.5 mM MgCl2
, 1 μM of both the forward and reverse primer, and 0.25 U Amplitaq DNA polymerase (Applied Biosystems, Foster City CA).
PCR products were cleaned using the ExoSAP-IT method (USB Corporation, Cleveland, OH). DNA sequencing reactions were conducted using BigDye®
Terminator v3.1 Cycle Sequencing Kits (Applied Biosystems, Foster City CA) and sequence data were obtained on both forward and reverse PCR primers on ABI 3730xl Genetic Analyzer (Applied Biosystems, Foster City CA). Contigs were assembled using PhredPhrap/Consed[26
] Polymorphisms were detected using Polyphred [27
] and MutationSurveyor [28
] by two independent researchers. All variants were unanimously confirmed.
Summary characteristics of the five screened pedigrees are shown in Table . In four of the five pedigrees, the linked 13q segregating haplotype was carried by the majority of prostate cancer cases. The age at diagnosis varied from what would be considered an early-onset case (i.e., ≤ 55 years – two prostate cancer cases), to cases with a diagnosis over the age of 80 years (six cases).
Next, for each pedigree we identified the common ancestor hypothesized to carry the 13q segregating haplotype. For all of the descendants of this ancestor, we identified all individuals with diagnosis of a known BRCA2-associated cancer (breast cancer, prostate cancer, ovarian cancer, pancreatic cancer, and melanoma) using linked UCR data. These results are shown in Table . In each pedigree, the number of prostate cancer cases among descendants of this most distant ancestor significantly exceeded that expected in a population matched by birth year (within five years), sex, and place of birth (Utah versus outside of Utah): p-value range: 0.007–0.000006. For pedigree 9445402, the number of melanoma cases also exceeded the matched population rate (p-value: 0.031). Although four of the five families had at least one known breast cancer case, there were no pedigrees with a significant excess of breast cancer cases. Furthermore, we observed no male breast cancer cases in any pedigree. The Utah Prostate High-Risk Cancer Resource is continually updated from the UPDB with new descendant information and newly diagnosed prostate cancer cases. Differences in the number of prostate cancer cases exist between Tables and . Table contains information on the pedigree structure and sampling design used in the original linkage analysis. Table reflects current cancer information for each pedigree.
Risk of BRCA2 associated cancers in descendants of most distant ancestor carrying the 13q haplotype1
No known deleterious mutations in BRCA2
were observed among the 12 carriers of the chromosome 13q haplotypes. We observed one BRCA2
variant of unknown significance as defined by the Breast Cancer Information Core (BIC) [29
] database and one variant not listed in the BIC database (see Table ). The variant of unknown significance (5972C→T) was detected in an unaffected male 1st
cousin (last known age to be cancer free: 72 years) but was not carried by the sampled prostate cancer case in the pedigree (diagnosis age: 54 years). The intronic variant not previously reported in the BIC database (IVS14 -62A→G) was observed in a single prostate cancer case, but it was not detected in the other haplotype carrier tested in the pedigree.
BRCA2 variants of unknown significance observed in Utah high-risk prostate cancer pedigrees
We also observed a number of BRCA2 polymorphisms that are likely of no biological consequence (see Table ). Many of these polymorphisms have been reported in other individuals with prostate cancer, and the majority segregated within pedigrees.
BRCA2 polymorphisms observed in Utah high-risk prostate cancer pedigrees
Our results are consistent with a growing pool of evidence that BRCA2
plays at most a minor role in causing prostate cancer in high-risk prostate cancer families. As with our results, Sinclair et al
] using high-risk prostate cancer pedigrees found no BRCA2
truncating mutations, but found one previously reported missense mutation and two previously unreported intronic polymorphisms. Gayther et al
] found two germ-line BRCA2
truncating mutations in 38 high-risk, familial prostate cancer pedigree in two individuals diagnosed ≤ 56 years, neither of which was found to segregate. A recent study by Agalliu et al
. using 266 subjects in high-risk prostate cancer families also reported no disease-associated protein truncating BRCA2
] Our study is unique in that we focused on pedigrees with strong linkage evidence near the BRCA2
gene region on chromosome 13q, and that all pedigrees met a very strict definition of high-risk.
While the five linked pedigrees studied here appear to have minimal involvement of the most likely candidate gene in the region (i.e., BRCA2), there is still unexplained linkage evidence on chromosome 13q that cannot be attributable to BRCA2 coding mutations. There may be yet unidentified regulatory BRCA2 mutations, deletions, and/or rearrangements that were not screened as part of this study. Alternatively, the nominal linkage evidence in these prostate cancer pedigrees may represent other prostate cancer predisposition genes located near BRCA2 including HMG1, LGR8, CCNA1, TRPC4, and FOXO1. Finally, the results may be spurious as only a limited number of pedigrees (i.e., five of the original 59 pedigrees) were studied, and segregation analysis with chromosome 13q may not be the most reliable way of identifying BRCA2 mutation families.