Introduction

[Note: Many of the medical and scientific terms used in this summary are found in the NCI Dictionary of Genetics Terms. When a linked term is clicked, the definition will appear in a separate window.]

[Note: Many of the genes described in this summary are found in the Online Mendelian Inheritance in Man (OMIM) database. When OMIM appears after a gene name or the name of a condition, click on OMIM for a link to more information.]

The public health burden of prostate cancer is substantial. A total of 180,890 new cases of prostate cancer and 26,120 deaths from the disease are anticipated in the United States in 2016, making it the most frequent nondermatologic cancer among U.S. males.[1] A mans lifetime risk of prostate cancer is one in seven. Prostate cancer is the second leading cause of cancer death in men, exceeded only by lung cancer.[1]

Some men with prostate cancer remain asymptomatic and die from unrelated causes rather than as a result of the cancer itself. This may be due to the advanced age of many men at the time of diagnosis, slow tumor growth, or response to therapy.[2] The estimated number of men with latent prostate carcinoma (i.e., prostate cancer that is present in the prostate gland but never detected or diagnosed during a patients life) is greater than the number of men with clinically detected disease. A better understanding is needed of the genetic and biologic mechanisms that determine why some prostate carcinomas remain clinically silent, while others cause serious, even life-threatening illness.[2]

Prostate cancer exhibits tremendous differences in incidence among populations worldwide; the ratio of countries with high and low rates of prostate cancer ranges from 60-fold to 100-fold.[3] Asian men typically have a very low incidence of prostate cancer, with age-adjusted incidence rates ranging from 2 to 10 cases per 100,000 men. Higher incidence rates are generally observed in northern European countries. African American men, however, have the highest incidence of prostate cancer in the world; within the United States, African American men have a 60% higher incidence rate than white men.[4] African American men have been reported to have more than twice the rate of prostate cancerspecific death compared with non-Hispanic white men.[1] Differences in race-specific prostate cancer survival estimates may be narrowing over time.[5]

These differences may be due to the interplay of genetic, environmental, and social influences (such as access to health care), which may affect the development and progression of the disease.[6] Differences in screening practices have also had a substantial influence on prostate cancer incidence, by permitting prostate cancer to be diagnosed in some patients before symptoms develop or before abnormalities on physical examination are detectable. An analysis of population-based data from Sweden suggested that a diagnosis of prostate cancer in one brother leads to an early diagnosis in a second brother using prostate-specific antigen (PSA) screening.[7] This may account for an increase in prostate cancer diagnosed in younger men that was evident in nationwide incidence data. A genetic contribution to prostate cancer risk has been documented, and there is increasing knowledge of the molecular genetics of the disease, although much of what is known is not yet clinically actionable. Malignant transformation of prostate epithelial cells and progression of prostate carcinoma are likely to result from a complex series of initiation and promotional events under both genetic and environmental influences.[8]

The three most important recognized risk factors for prostate cancer in the United States are:

Age is an important risk factor for prostate cancer. Prostate cancer is rarely seen in men younger than 40 years; the incidence rises rapidly with each decade thereafter. For example, the probability of being diagnosed with prostate cancer is 1 in 325 for men 49 years or younger, 1 in 48 for men aged 50 through 59 years, 1 in 17 for men aged 60 through 69 years, and 1 in 10 for men aged 70 years and older, with an overall lifetime risk of developing prostate cancer of 1 in 7.[1]

Approximately 10% of prostate cancer cases are diagnosed in men younger than 56 years and represent early-onset prostate cancer. Data from the Surveillance, Epidemiology, and End Results (SEER) Program show that early-onset prostate cancer is increasing, and there is evidence that some cases may be more aggressive.[9] Because early-onset cancers may result from germline mutations, young men with prostate cancer are being extensively studied with the goal of identifying prostate cancer susceptibility genes.

The risk of developing and dying from prostate cancer is dramatically higher among blacks, is of intermediate levels among whites, and is lowest among native Japanese.[10,11] Conflicting data have been published regarding the etiology of these outcomes, but some evidence is available that access to health care may play a role in disease outcomes.[12]

Prostate cancer is highly heritable; the inherited risk of prostate cancer has been estimated to be as high as 60%.[13] As with breast and colon cancer, familial clustering of prostate cancer has been reported frequently.[14-18] From 5% to 10% of prostate cancer cases are believed to be primarily caused by high-risk inherited genetic factors or prostate cancer susceptibility genes. Results from several large case-control studies and cohort studies representing various populations suggest that family history is a major risk factor in prostate cancer.[15,19,20] A family history of a brother or father with prostate cancer increases the risk of prostate cancer, and the risk is inversely related to the age of the affected relative.[16-20] However, at least some familial aggregation is due to increased prostate cancer screening in families thought to be at high risk.[21]

Although some of the prostate cancer studies examining risks associated with family history have used hospital-based series, several studies described population-based series.[22-24] The latter are thought to provide information that is more generalizable. A meta-analysis of 33 epidemiologic case-control and cohort-based studies has provided more detailed information regarding risk ratios related to family history of prostate cancer. Risk appeared to be greater for men with affected brothers than for men with affected fathers in this meta-analysis. Although the reason for this difference in risk is unknown, possible hypotheses have included X-linked or recessive inheritance. In addition, risk increased with increasing numbers of affected close relatives. Risk also increased when a first-degree relative (FDR) was diagnosed with prostate cancer before age 65 years. (See Table 1 for a summary of the relative risks [RRs] related to a family history of prostate cancer.)[25]

Among the many data sources included in this meta-analysis, those from the Swedish population-based Family-Cancer Database warrant special comment. These data were derived from a resource that contained more than 11.8 million individuals, among whom there were 26,651 men with medically verified prostate cancer, of which 5,623 were familial cases.[26] The size of this data set, with its nearly complete ascertainment of the entire Swedish population and objective verification of cancer diagnoses, should yield risk estimates that are both accurate and free of bias. When the familial age-specific hazard ratios (HRs) for prostate cancer diagnosis and mortality were computed, as expected, the HR for prostate cancer diagnosis increased with more family history. Specifically, HRs for prostate cancer were 2.12 (95% CI, 2.052.20) with an affected father only, 2.96 (95% CI, 2.803.13) with an affected brother only, and 8.51 (95% CI, 6.1311.80) with a father and two brothers affected. The highest HR, 17.74 (95% CI, 12.2625.67), was seen in men with three brothers diagnosed with prostate cancer. The HRs were even higher when the affected relative was diagnosed with prostate cancer before age 55 years.

A separate analysis of this Swedish database reported that the cumulative (absolute) risks of prostate cancer among men in families with two or more affected cases were 5% by age 60 years, 15% by age 70 years, and 30% by age 80 years, compared with 0.45%, 3%, and 10%, respectively, by the same ages in the general population. The risks were even higher when the affected father was diagnosed before age 70 years.[27] The corresponding familial population attributable fractions (PAFs) were 8.9%, 1.8%, and 1.0% for the same three age groups, respectively, yielding a total PAF of 11.6% (i.e., approximately 11.6% of all prostate cancers in Sweden can be accounted for on the basis of familial history of the disease).

The risk of prostate cancer may also increase in men who have a family history of breast cancer. Approximately 9.6% of the Iowa cohort had a family history of breast and/or ovarian cancer in a mother or sister at baseline, and this was positively associated with prostate cancer risk (age-adjusted RR, 1.7; 95% CI, 1.03.0; multivariate RR, 1.7; 95% CI, 0.93.2). Men with a family history of both prostate and breast/ovarian cancer were also at increased risk of prostate cancer (RR, 5.8; 95% CI, 2.414.0).[22] Analysis of data from the Women's Health Initiative also showed that a family history of prostate cancer was associated with an increase in the risk of postmenopausal breast cancer (adjusted HR, 1.14; 95% CI, 1.021.26).[28] Further analyses showed that breast cancer risk was associated with a family history of both breast and prostate cancers; the risk was higher in black women than in white women. Other studies, however, did not find an association between family history of female breast cancer and risk of prostate cancer.[22,29] A family history of prostate cancer also increases the risk of breast cancer among female relatives.[30] The association between prostate cancer and breast cancer in the same family may be explained, in part, by the increased risk of prostate cancer among men with BRCA1/BRCA2 mutations in the setting of hereditary breast/ovarian cancer or early-onset prostate cancer.[31-34] (Refer to the BRCA1 and BRCA2 section of this summary for more information.)

Prostate cancer clusters with particular intensity in some families. Highly penetrant genetic variants are thought to be associated with prostate cancer risk in these families. (Refer to the Linkage Analyses section of this summary for more information.) Members of such families may benefit from genetic counseling. Emerging recommendations and guidelines for genetic counseling referrals are based on prostate cancer age at diagnosis and specific family cancer history patterns.[35,36] Individuals meeting the following criteria may warrant referral for genetic consultation:[35-38]

Family history has been shown to be a risk factor for men of different races and ethnicities. In a population-based case-control study of prostate cancer among African Americans, whites, and Asian Americans in the United States (Los Angeles, San Francisco, and Hawaii) and Canada (Vancouver and Toronto),[39] 5% of controls and 13% of all cases reported a father, brother, or son with prostate cancer. These prevalence estimates were somewhat lower among Asian Americans than among African Americans or whites. A positive family history was associated with a twofold to threefold increase in RR in each of the three ethnic groups. The overall odds ratio associated with a family history of prostate cancer was 2.5 (95% CI, 1.93.3) with adjustment for age and ethnicity.[39]

Endogenous hormones, including both androgens and estrogens, likely influence prostate carcinogenesis. It has been widely reported that eunuchs and other individuals with castrate levels of testosterone before puberty do not develop prostate cancer.[40] Some investigators have considered the potential role of genetic variation in androgen biosynthesis and metabolism in prostate cancer risk,[41] including the potential role of the androgen receptor (AR) CAG repeat length in exon 1. This modulates AR activity, which may influence prostate cancer risk.[42] For example, a meta-analysis reported that AR CAG repeat length greater than or equal to 20 repeats conferred a protective effect for prostate cancer in subsets of men.[43]

(Refer to the PDQ summary on Prostate Cancer Prevention for more information about nongenetic modifiers of prostate cancer risk in the general population.)

The SEER Cancer Registries assessed the risk of developing a second primary cancer in 292,029 men diagnosed with prostate cancer between 1973 and 2000. Excluding subsequent prostate cancer and adjusting for the risk of death from other causes, the cumulative incidence of a second primary cancer among all patients was 15.2% at 25 years (95% CI, 5.015.4). There was a significant risk of new malignancies (all cancers combined) among men diagnosed before age 50 years, no excess or deficit in cancer risk in men aged 50 to 59 years, and a deficit in cancer risk in all older age groups. The authors suggested that this deficit may be attributable to decreased cancer surveillance in an elderly population. Excess risks of second primary cancers included cancers of the small intestine, soft tissue, bladder, thyroid, and thymus; and melanoma. Prostate cancer diagnosed in patients aged 50 years or younger was associated with an excess risk of pancreatic cancer.[44]

A review of more than 441,000 men diagnosed with prostate cancer between 1992 and 2010 demonstrated similar findings, with an overall reduction in the risk of being diagnosed with a second primary cancer. This study also examined the risk of second primary cancers in 44,310 men (10%) by treatment modality for localized cancer. The study suggested that men who received radiation therapy had increases in bladder (standardized incidence ratio [SIR], 1.42) and rectal cancer risk (SIR, 1.70) compared with those who did not receive radiation therapy (SIRbladder, 0.76; SIRrectal, 0.74).[45]

The underlying etiology of developing a second primary cancer after prostate cancer may be related to various factors, including treatment modality. More than 50% of the small intestine tumors were carcinoid malignancies, suggesting possible hormonal influences. The excess of pancreatic cancer may be due to mutations in BRCA2, which predisposes to both. The risk of melanoma was most pronounced in the first year of follow-up after diagnosis, raising the possibility that this is the result of increased screening and surveillance.[44]

One Swedish study using the nationwide Swedish Family Cancer Database assessed the role of family history in the risk of a second primary cancer after prostate cancer. Of 18,207 men with prostate cancer, 560 developed a second primary malignancy. Of those, the RR was increased for colorectal, kidney, bladder, and squamous cell skin cancers. Having a paternal family history of prostate cancer was associated with an increased risk of bladder cancer, myeloma, and squamous cell skin cancer. Among prostate cancer probands, those with a family history of colorectal cancer, bladder cancer, or chronic lymphoid leukemia were at increased risk of that specific cancer as a second primary cancer.[46]

Several reports have suggested an elevated risk of various other cancers among relatives within multiple-case prostate cancer families, but none of these associations have been established definitively.[47-49]

In a population-based Finnish study of 202 multiple-case prostate cancer families, no excess risk of all cancers combined (other than prostate cancer) was detected in 5,523 family members. Female family members had a marginal excess of gastric cancer (SIR, 1.9; 95% CI, 1.03.2). No difference in familial cancer risk was observed when families affected by clinically aggressive prostate cancers were compared with those having nonaggressive prostate cancer. These data suggest that familial prostate cancer is a cancer sitespecific disorder.[50]

Many types of epidemiologic studies (case-control, cohort, twin, family) strongly suggest that prostate cancer susceptibility genes exist in the population. Analysis of longer follow-up of the monozygotic (MZ) and dizygotic (DZ) twin pairs in Scandinavia concluded that 58% (95% CI, 5263) of prostate cancer risk may be accounted for by heritable factors.[13] Additionally, among affected MZ and DZ pairs, the time to diagnosis in the second twin was shortest in MZ twins (mean, 3.8 years in MZ twins vs. 6.5 years in DZ twins). This is in agreement with a previous U.S. study that showed a concordance of 7.1% between DZ twin pairs and a 27% concordance between MZ twin pairs.[51] The first segregation analysis was performed in 1992 using families from 740 consecutive probands who had radical prostatectomies between 1982 and 1989. The study results suggested that familial clustering of disease among men with early-onset prostate cancer was best explained by the presence of a rare (frequency of 0.003) autosomal dominant, highly penetrant allele(s).[15] Hereditary prostate cancer susceptibility genes were predicted to account for almost half of early-onset disease (age 55 years or younger). In addition, early-onset disease has been further supported to have a strong genetic component from the study of common variants associated with disease onset before age 55 years.[52]

Subsequent segregation analyses generally agreed with the conclusions but differed in the details regarding frequency, penetrance, and mode of inheritance.[53-55] A study of 4,288 men who underwent radical prostatectomy between 1966 and 1995 found that the best fitting genetic model of inheritance was the presence of a rare, autosomal dominant susceptibility gene (frequency of 0.06). In this study, the lifetime risk in carriers was estimated to be 89% by age 85 years and 3.9% for noncarriers.[51] This study also suggested the presence of genetic heterogeneity, as the model did not reliably predict prostate cancer risk in FDRs of probands who were diagnosed at age 70 years or older. More recent segregation analyses have concluded that there are multiple genes associated with prostate cancer [56-59] in a pattern similar to other adult-onset hereditary cancer syndromes, such as those involving the breast, ovary, colorectum, kidney, and melanoma. In addition, a segregation analysis of 1,546 families from Finland found evidence for Mendelian recessive inheritance. Results showed that individuals carrying the risk allele were diagnosed with prostate cancer at younger ages (<66 years) than noncarriers. This is the first segregation analysis to show a recessive mode of inheritance.[60]

Various research methods have been employed to uncover the landscape of genetic variation associated with prostate cancer. Specific methodologies inform of unique phenotypes or inheritance patterns. The sections below describe prostate cancer research utilizing various methods to highlight their role in uncovering the genetic basis of prostate cancer. In an effort to identify disease susceptibility genes, linkage studies are typically performed on high-risk extended families in which multiple cases of a particular disease have occurred. Typically, gene mutations identified through linkage analyses are rare in the population, are moderately to highly penetrant in families, and have large (e.g., relative risk >2.0) effect sizes. The clinical role of mutations that are identified in linkage studies is a clearer one, establishing precedent for genetic testing for cancer with genes such as BRCA1 and BRCA2. (Refer to the BRCA1 and BRCA2 section in the Genes With Potential Clinical Relevance in Prostate Cancer Risk section of this summary for more information about these genes.) Genome-wide association studies (GWAS) are another methodology used to identify candidate loci associated with prostate cancer. Genetic variants identified from GWAS typically are common in the population and have low to modest effect sizes for prostate cancer risk. The clinical role of markers identified from GWAS is an active area of investigation. Case-control studies are useful in validating the findings of linkage studies and GWAS as well as for studying candidate gene alterations for association with prostate cancer risk, although the clinical role of findings from case-control studies needs to be further defined.

The recognition that prostate cancer clusters within families has led many investigators to collect multiple-case families with the goal of localizing prostate cancer susceptibility genes through linkage studies.

Linkage studies are typically performed on high-risk kindreds in whom multiple cases of a particular disease have occurred in an effort to identify disease susceptibility genes. Linkage analysis statistically compares the genotypes between affected and unaffected individuals and looks for evidence that known genetic markers are inherited along with the disease trait. If such evidence is found (linkage), it provides statistical data that the chromosomal region near the marker also harbors a disease susceptibility gene. Once a genomic region of interest has been identified through linkage analysis, additional studies are required to prove that there truly is a susceptibility gene at that position. Linkage analysis is affected by the following:

Furthermore, because a standard definition of hereditary prostate cancer has not been accepted, prostate cancer linkage studies have not used consistent criteria for enrollment.[1] One criterion that has been proposed is the Hopkins Criteria, which provides a working definition of hereditary prostate cancer families.[2] Using the Hopkins Criteria, kindreds with prostate cancer need to fulfill only one of following criteria to be considered to have hereditary prostate cancer:

Using these criteria, surgical series have reported that approximately 3% to 5% of men will be from a family with hereditary prostate cancer.[2,3]

An additional issue in linkage studies is the high background rate of sporadic prostate cancer in the context of family studies. Because a mans lifetime risk of prostate cancer is one in seven,[4] it is possible that families under study have men with both inherited and sporadic prostate cancer. Thus, men who do not inherit the prostate cancer susceptibility gene that is segregating in their family may still develop prostate cancer. There are no clinical or pathological features of prostate cancer that will allow differentiation between inherited and sporadic forms of the disease, although current advances in the understanding of molecular phenotypes of prostate cancer may be informative in identifying inherited prostate cancer. Similarly, there are limited data regarding the clinical phenotype or natural history of prostate cancer associated with specific candidate loci. Measurement of the serum prostate-specific antigen (PSA) has been used inconsistently in evaluating families used in linkage analysis studies of prostate cancer. In linkage studies, the definition of an affected man can be biased by the use of serum PSA screening as the rates of prostate cancer in families will differ between screened and unscreened families.

One way to address inconsistencies between linkage studies is to require inclusion criteria that define clinically significant disease (e.g., Gleason score 7, PSA 20 ng/mL) in an affected man.[5-7] This approach attempts to define a homogeneous set of cases/families to increase the likelihood of identifying a linkage signal. It also prevents the inclusion of cases that may be considered clinically insignificant that were identified by screening in families.

Investigators have also incorporated clinical parameters into linkage analyses with the goal of identifying genes that may influence disease severity.[8,9] This type of approach, however, has not yet led to the identification of consistent linkage signals across datasets.[10,11]

Table 2 summarizes the proposed prostate cancer susceptibility loci identified in families with multiple prostate canceraffected individuals. Conflicting evidence exists regarding the linkage to some of the loci described above. Data on the proposed phenotype associated with each locus are also limited, and the strength of repeated studies is needed to firmly establish these associations. Evidence suggests that many of these prostate cancer loci account for disease in a small subset of families, which is consistent with the concept that prostate cancer exhibits locus heterogeneity.

Genome-wide linkage studies of families with prostate cancer have identified several other loci that may harbor prostate cancer susceptibility genes, emphasizing the underlying complexity and genetic heterogeneity of this cancer. The following chromosomal regions have been found to be associated with prostate cancer in more than one study or clinical cohort with a statistically significant (2) logarithm of the odds (LOD) score, heterogeneity LOD (HLOD) score, or summary LOD score:

The chromosomal region 19q has also been found to be associated with prostate cancer, although specific LOD scores have not been described.[8,11,95]

Linkage studies have also been performed in specific populations or with specific clinical parameters to identify population-specific susceptibility genes or genes influencing disease phenotypes.

The African American Hereditary Prostate Cancer study conducted a genome-wide linkage study of 77 families with four or more affected men. Multipoint HLOD scores of 1.3 to less than 2.0 were observed using markers that map to 11q22, 17p11, and Xq21. Analysis of the 16 families with more than six men with prostate cancer provided evidence for two additional loci: 2p21 (multipoint HLOD score = 1.08) and 22q12 (multipoint HLOD score = 0.91).[92,99] A smaller linkage study that included 15 African American hereditary prostate cancer families from the southeastern and southcentral Louisiana region identified suggestive linkage for prostate cancer at 2p16 (HLOD = 1.97) and 12q24 (HLOD = 2.21) using a 6,000 single nucleotide polymorphism (SNP) platform.[111] Further study including a larger number of African American families is needed to confirm these findings.

In an effort to identify loci contributing to prostate cancer aggressiveness, linkage analysis was performed in families with one or more of the following: Gleason grade 7 or higher, PSA of 20 ng/mL or higher, regional or distant cancer stage at diagnosis, or death from metastatic prostate cancer before age 65 years. One hundred twenty-three families with two or more affected family members with aggressive prostate cancer were studied. Suggestive linkage was found at chromosome 22q11 (HLOD score = 2.18) and 22q12.3-q13.1 (HLOD score = 1.90).[5] These findings suggest that using a clinically defined phenotype may facilitate finding prostate cancer susceptibility genes. A fine-mapping study of 14 extended high-risk prostate cancer families has subsequently narrowed the genomic region of interest to an 880-kb region at 22q12.3.[107] An analysis of high-risk pedigrees from Utah provides an overview of this strategy.[112] A linkage analysis utilizing a higher resolution marker set of 6,000 SNPs was performed among 348 families from the International Consortium for Prostate Cancer Genetics with aggressive prostate cancer.[44] Aggressive disease was defined as Gleason score 7 or higher, invasion into seminal vesicles or extracapsular extension, pretreatment PSA level of 20 ng/mL or higher, or death from prostate cancer. The region with strongest evidence of linkage among aggressive prostate cancer families was 8q24 with LOD scores of 3.093.17. Additional regions of linkage included with LOD scores of 2 or higher included 1q43, 2q35, and 12q24.31. No candidate genes have been identified.

In light of the multiple prostate cancer susceptibility loci and disease heterogeneity, another approach has been to stratify families based on other cancers, given that many cancer susceptibility genes are pleiotropic.[113] A genome-wide linkage study was conducted to identify a susceptibility locus that may account for both prostate cancer and kidney cancer in families. Analysis of 15 families with evidence of hereditary prostate cancer and one or more cases of kidney cancer (pathologically confirmed) in a man with prostate cancer or in a first-degree relative of a man with prostate cancer revealed suggestive linkage with markers that mapped to an 8 cM region of chromosome 11p11.2-q12.2.[114] This observation awaits confirmation. Another genome-wide linkage study was conducted in 96 hereditary prostate cancer families with one or more first-degree relatives with colon cancer. Evidence for linkage in all families was found in several regions, including 11q25, 15q14, and 18q21. In families with two or more cases of colon cancer, linkage was also observed at 1q31, 11q14, and 15q11-14.[113]

Linkage to chromosome 17q21-22 and subsequent fine-mapping and targeted sequencing have identified recurrent mutations in the HOXB13 gene that account for a fraction of hereditary prostate cancer, particularly early-onset prostate cancer. Multiple studies have confirmed the association between the G84E mutation in HOXB13 and prostate cancer risk. (Refer to the HOXB13 section of this summary for more information.) The clinical utility of testing for HOXB13 mutations has not yet been defined, but studies are ongoing to define the clinical role. For example, a study evaluated 948 unselected men scheduled for prostate biopsy. The G84E mutation was found in three men (0.3%) who had prostate cancer detected on biopsy, although none of the 301 men who had a family history of prostate cancer carried the mutation.[115] Furthermore, many linkage studies have mapped several prostate cancer susceptibility loci (Table 2), although the genetic alterations contributing to hereditary prostate cancer from these loci have not been consistently reproduced. With the evolution of high-throughput sequencing technologies, there will likely be additional moderately to highly penetrant genetic mutations identified to account for subsets of hereditary prostate cancer families.[116]

A case-control study involves evaluating factors of interest for association to a condition. The design involves investigation of cases with a condition of interest, such as a specific disease or gene mutation, compared with a control sample without that condition, but often with other similar characteristics (i.e., age, gender, and ethnicity). Limitations of case-control design with regard to identifying genetic factors include the following:[117,118]

Additionally, identified associations may not always be valid, but they could represent a random association and, therefore, warrant validation studies.[117,118]

Androgen receptor (AR) gene variants have been examined in relation to both prostate cancer risk and disease progression. The AR is expressed during all stages of prostate carcinogenesis.[120] One study demonstrated that men with hereditary prostate cancer who underwent radical prostatectomy had a higher percentage of prostate cancer cells exhibiting expression of the AR and a lower percentage of cancer cells expressing estrogen receptor alpha than did men with sporadic prostate cancer. The authors suggest that a specific pattern of hormone receptor expression may be associated with hereditary predisposition to prostate cancer.[121]

Altered activity of the AR caused by inherited variants of the AR gene may influence risk of prostate cancer. The length of the polymorphic trinucleotide CAG and GGN microsatellite repeats in exon 1 of the AR gene (located on the X chromosome) have been associated with the risk of prostate cancer.[122,123] Some studies have suggested an inverse association between CAG repeat length and prostate cancer risk, and a direct association between GGN repeat length and risk of prostate cancer; however, the evidence is inconsistent.[120,122-132] A meta-analysis of 19 case-control studies demonstrated a statistically significant association between both short CAG length (odds ratio [OR], 1.2; 95% confidence interval [CI], 1.11.3) and short GGN length (OR, 1.3; 95% CI, 1.11.6) and prostate cancer; however, the absolute difference in number of repeats between cases and controls is less than one, leading the investigators to question whether these small, statistically significant differences are biologically meaningful.[133] Subsequently, the large multiethnic cohort study of 2,036 incident prostate cancer cases and 2,160 ethnically matched controls failed to confirm a statistically significant association (OR, 1.02; P = .11) between CAG repeat size and prostate cancer.[134] A study of 1,461 Swedish men with prostate cancer and 796 control men reported an association between AR alleles, with more than 22 CAG repeats and prostate cancer (OR, 1.35; 95% CI, 1.081.69; P = .03).[135]

An analysis of AR gene CAG and CGN repeat length polymorphisms targeted African American men from the Flint Mens Health Study in an effort to identify a genetic modifier that might help explain the increased risk of prostate cancer in black versus white males in the United States.[136] This population-based study of 131 African American prostate cancer patients and 340 screened-negative African American controls showed no evidence of an association between shorter AR repeat length and prostate cancer risk. These results, together with data from three prior, smaller studies,[134,137,138] indicate that short AR repeat variants do not contribute significantly to the risk of prostate cancer in African American men.

Germline mutations in the AR gene (located on the X chromosome) have been rarely reported. The R726L mutation has been identified as a possible contributor to about 2% of both sporadic and familial prostate cancer in Finland.[139] This mutation, which alters the transactivational specificity of the AR protein, was found in 8 of 418 (1.91%) consecutive sporadic prostate cancer cases, 2 of 106 (1.89%) familial cases, and 3 of 900 (0.33%) normal blood donors, yielding a significantly increased prostate cancer OR of 5.8 for both case groups. A subsequent Finnish study of 38 early-onset prostate cancer cases and 36 multiple-case prostate cancer families with no evidence of male-to-male transmission revealed one additional R726L mutation in one of the familial cases and no new germline mutations in the AR gene.[140] These investigators concluded that germline AR mutations explain only a small fraction of familial and early-onset cases in Finland.

A study of genomic DNA from 60 multiple-case African American (n = 30) and white (n = 30) families identified a novel missense germline AR mutation, T559S, in three affected members of a black sibship and none in the white families. No functional data were presented to indicate that this mutation was clearly deleterious. This was reported as a suggestive finding, in need of additional data.[141]

Molecular epidemiology studies have also examined genetic polymorphisms of the steroid 5-alpha-reductase 2 gene, which is also involved in the androgen metabolism cascade. Two isozymes of 5-alpha-reductase exist. The gene that codes for 5-alpha-reductase type II (SRD5A2) is located on chromosome 2. It is expressed in the prostate, where testosterone is converted irreversibly to dihydrotestosterone (DHT) by 5-alpha-reductase type II.[142] Evidence suggests that 5-alpha-reductase type II activity is reduced in populations at lower risk of prostate cancer, including Chinese and Japanese men.[143,144]

A polymorphism in the untranslated region of the SRD5A2 gene may also be associated with prostate cancer risk.[145] Ten alleles fall into three families that differ in the number of TA dinucleotide repeats.[142,146] Although no clinical significance for these polymorphisms has yet been determined, some TA repeat alleles may promote an elevation of enzyme activity, which may in turn increase the level of DHT in the prostate.[120,142] A subsequent meta-analysis failed to detect a statistically significant association between prostate cancer risk and the TA repeat polymorphism, although a relationship could not be definitively excluded.[147] This meta-analysis also examined the potential roles of two coding variants: A49T and V89L. An association with V89L was excluded, and the role for A49T was found to have at most a modest effect on prostate cancer susceptibility. Bias or chance could account for the latter observation. A study of 1,461 Swedish men with prostate cancer and 796 control men reported an association between two variants in SRD5A2 and prostate cancer risk (OR, 1.45; 95% CI, 1.012.08; OR, 1.49; 95% CI, 1.032.15).[135] Another meta-analysis of 25 case-control studies, including 8,615 cases and 9,089 controls, found no overall association between the V89L polymorphism and prostate cancer risk. In a subgroup analysis, men younger than 65 years (323 cases and 677 controls) who carried the LL genotype had a modest association with prostate cancer (LL vs. VV, OR, 1.70; 95% CI, 1.092.66 and LL vs. VV+VL, OR, 1.75; 95% CI, 1.142.68).[148] A subsequent systematic review and meta-analysis including 27 nonfamilial case-control studies found no statistically significant association between either the V89L or A49T polymorphisms and prostate cancer risk.[149]

Polymorphisms in several genes involved in the biosynthesis, activation, metabolism, and degradation of androgens (CYP17, CYP3A4, CYP19A1, and SRD5A2) and the stimulation of mitogenic and antiapoptotic activities (IGF-1 and IGFBP-3) of normal prostate cells were examined for association with prostate cancer in 131 African American cases and 342 controls. While allele frequencies did not differ between cases and controls regarding three SNPs in the CYP17 gene (rs6163, rs6162, and rs743572), heterozygous genotypes of these SNPs were found to be associated with a reduced risk (OR, 0.56; 95% CI, 0.350.88; OR, 0.57; 95% CI, 0.360.90; OR, 0.55; 95% CI, 0.350.88, respectively). Evidence suggestive of an association between SNP rs5742657 in intron 2 of IGF-1 was also found (OR, 1.57; 95% CI, 0.942.63).[150] Additional studies are needed to confirm these findings.

Other investigators have explored the potential contribution of the variation in genes involved in the estrogen pathway. A Swedish population study of 1,415 prostate cancer cases and 801 age-matched controls examined the association of SNPs in the estrogen receptor-beta (ER-beta) gene and prostate cancer. One SNP in the promoter region of ER-beta, rs2987983, was associated with an overall prostate cancer risk of 1.23 and 1.35 for localized disease.[151] This study awaits replication.

Germline mutations in the tumor suppressor gene E-cadherin (also called CDH1) cause a hereditary form of gastric carcinoma. A SNP designated -160A, located in the promoter region of E-cadherin, has been found to alter the transcriptional activity of this gene.[152] Because somatic mutations in E-cadherin have been implicated in the development of invasive malignancies in a number of different cancers,[153] investigators have searched for evidence that this functionally significant promoter might be a modifier of cancer risk. A meta-analysis of 47 case-control studies in 16 cancer types included ten prostate cancer cohorts (3,570 cases and 3,304 controls). The OR of developing prostate cancer among risk allele carriers was 1.33 (95% CI, 1.111.60). However, the authors of the study noted that there are sources of bias in the dataset, stemming mostly from the small sample sizes of individual cohorts.[154] Additional studies are required to determine whether this finding is reproducible and biologically and clinically important.

There is a great deal of interest in the possibility that chronic inflammation may represent an important risk factor in prostate carcinogenesis.[155] The family of toll-like receptors has been recognized as a critical component of the intrinsic immune system,[156] one which recognizes ligands from exogenous microbes and a variety of endogenous substrates. This family of genes has been studied most extensively in the context of autoimmune disease, but there also have been a series of studies that have analyzed genetic variants in various members of this pathway as potential prostate cancer risk modifiers.[157-161] The results have been inconsistent, ranging from decreased risk, to null association, to increased risk.

One study was based upon 1,414 incident prostate cancer cases and 1,414 age-matched controls from the American Cancer Society Cancer Prevention Study II Nutrition Cohort.[162] These investigators genotyped 28 SNPs in a region on chromosome 4p14 that includes TLR-10, TLR-1, and TLR-6, three members of the toll-like receptor gene cluster. Two TLR-10 SNPs and four TLR-1 SNPs were associated with significant reductions in prostate cancer risk, ranging from 29% to 38% for the homozygous variant genotype. A more detailed analysis demonstrated these six SNPs were not independent in their effect, but rather represented a single strong association with reduced risk (OR, 0.55; 95% CI, 0.330.90). There were no significant differences in this association when covariates such as Gleason score, history of benign prostatic hypertrophy, use of nonsteroidal anti-inflammatory drugs, and body mass index were taken into account. This is the largest study undertaken to date and included the most comprehensive panel of SNPs evaluated in the 4p14 region. While these observations provide a basis for further investigation of the toll-like receptor genes in prostate cancer etiology, inconsistencies with the prior studies and lack of information regarding what the biological basis of these associations might be warrant caution in interpreting the findings.

SNPs in genes involved in the steroid hormone pathway have previously been studied in sporadic and familial prostate cancer using a sample of individuals with primarily Caucasian ancestry.[163] Another study evaluated 116 tagging SNPs located in 12 genes in the steroid hormone pathway for risk of prostate cancer in 886 cases and 1,566 controls encompassing non-Hispanic white men, Hispanic white men, and African American men.[164] The genes included CYP17, HSD17B3, ESR1, SRD5A2, HSD3B1, HSD3B2, CYP19, CYP1A1, CYP1B1, CYP3A4, CYP27B1, and CYP24A1. Several SNPs in CYP19 were associated with prostate cancer risk in all three populations. Analysis of SNP-SNP interactions involving SNPs in multiple genes revealed a seven-SNP interaction involving HSD17B3, CYP19, and CYP24A1 in Hispanic whites (P = .001). In non-Hispanic whites, an interaction of four SNPs in HSD3B2, HSD17B3, and CYP19 was found (P < .001). In African Americans, SNPs within SRD5A2, HSD17B3, CYP17, CYP27B1, CYP19, and CYP24A1 showed a significant interaction (P = .014). In non-Hispanic whites, a cumulative risk of prostate cancer was observed for men carrying risk alleles at three SNPs in HSD3B2 and CYP19 (OR, 2.20; 95% CI, 1.443.38; P = .0003). In Hispanic whites, a cumulative risk of prostate cancer was observed for men carrying risk alleles at two SNPs in CYP19 and CYP24A1 (OR, 4.29; 95% CI, 2.118.72; P = .00006). While this study did not evaluate all potentially important SNPs in genes in the steroid hormone pathway, it demonstrates how studies can be performed to evaluate multigenic effects in multiple populations to assess the contribution to prostate cancer risk.

A meta-analysis of the relationship between eight polymorphisms in six genes (MTHFR, MTR, MTHFD1, SLC19A1, SHMT1, and FOLH1) from the folate pathway was conducted by pooling data from eight case-control studies, four GWAS, and a nested case-control study named Prostate Testing for Cancer and Treatment in the United Kingdom. Numbers of tested subjects varied among these polymorphisms, with up to 10,743 cases and 35,821 controls analyzed. The report concluded that known common folate-pathway SNPs do not have significant effects on prostate cancer susceptibility in white men.[165]

Four SNPs in the p53 pathway (three in genes regulating p53 function including MDM2, MDM4, and HAUSP and one in p53) were evaluated for association with aggressive prostate cancer in a hospital-based prostate cancer cohort of men with Caucasian ethnicity (N = 4,073).[166] However, a subsequent meta-analysis of case-control studies that focused on MDM2 (T309G) and prostate cancer risk revealed no association.[167] Therefore, the biologic basis of the various associations identified requires further study.

Table 3 summarizes additional case-control studies that have assessed genes that are potentially associated with prostate cancer susceptibility.

Case-control studies assessed site-specific prostate cancer susceptibility in the following genes: EMSY, KLF6, AMACR, NBS1, CHEK2, AR, SRD5A2, ER-beta, E-cadherin, and the toll-like receptor genes. These studies have been complicated by the later-onset nature of the disease and the high background rate of prostate cancer in the general population. In addition, there is likely to be real, extensive locus heterogeneity for hereditary prostate cancer, as suggested by both segregation and linkage studies. In this respect, hereditary prostate cancer resembles a number of the other major adult-onset hereditary cancer syndromes, in which more than one gene can produce the same or very similar clinical phenotype (e.g., hereditary breast/ovarian cancer, Lynch syndrome, hereditary melanoma, and hereditary renal cancer). The clinical validity and utility of genetic testing for any of these genes based solely on evidence for hereditary prostate cancer susceptibility has not been established.

Admixture mapping is a method used to identify genetic variants associated with traits and/or diseases in individuals with mixed ancestry.[178] This approach is most effective when applied to individuals whose admixture was recent and consists of two populations who had previously been separated for thousands of years. The genomes of such individuals are a mosaic, comprised of large blocks from each ancestral locale. The technique takes advantage of a difference in disease incidence in one ancestral group compared with another. Genetic risk loci are presumed to reside in regions enriched for the ancestral group with higher incidence. Successful mapping depends on the availability of population-specific genetic markers associated with ancestry, and on the number of generations since admixture.[179,180]

Admixture mapping is a particularly attractive method for identifying genetic loci associated with increased prostate cancer risk among African Americans. African American men are at higher risk of developing prostate cancer than are men of European ancestry, and the genomes of African American men are mosaics of regions from Africa and regions from Europe. It is therefore hypothesized that inherited variants accounting for the difference in incidence between the two groups must reside in regions enriched for African ancestry. In prostate cancer admixture studies, genetic markers for ancestry were genotyped genome-wide in African American cases and controls in a search for areas enriched for African ancestry in the men with prostate cancer. Admixture studies have identified the following chromosomal regions associated with prostate cancer:

An advantage of this approach is that recent admixtures result in long stretches of linkage disequilibrium (up to hundreds of thousands of base pairs) of one particular ancestry.[182] As a result, fewer markers are needed to search for genetic variants associated with specific diseases, such as prostate cancer, than the number of markers needed for successful GWAS.[179] (Refer to the GWAS section of this summary for more information.)

Genome-wide searches have successfully identified susceptibility alleles for many complex diseases,[183] including prostate cancer. This approach can be contrasted with linkage analysis, which searches for genetic risk variants co-segregating within families that have a high prevalence of disease. Linkage analyses are designed to uncover rare, highly penetrant variants that segregate in predictable heritance patterns (e.g., autosomal dominant, autosomal recessive, X-linked, and mitochondrial). GWAS, on the other hand, are best suited to identify multiple, common, low-penetrance genetic polymorphisms. GWAS are conducted under the assumption that the genetic underpinnings of complex phenotypes, such as prostate cancer, are governed by many alleles, each conferring modest risk. Most genetic polymorphisms genotyped in GWAS are common, with minor allele frequencies greater than 1% to 5% within a given ancestral population (e.g., men of European ancestry). GWAS survey all common inherited variants across the genome, searching for alleles that are associated with incidence of a given disease or phenotype.[184,185] The strong correlation between many alleles located close to one another on a given chromosome (called linkage disequilibrium) allows one to scan the genome without having to test all tens of millions of known SNPs. GWAS can test approximately 1 million to 5 million SNPs and ascertain almost all common inherited variants in the genome.

In a GWAS, allele frequency is compared for each SNP between cases and controls. Promising signalsin which allele frequencies deviate significantly in case compared to control populationsare validated in replication cohorts. In order to have adequate statistical power to identify variants associated with a phenotype, large numbers of cases and controls, typically thousands of each, are studied. Because 1 million SNPs are typically evaluated in a GWAS, false-positive findings are expected to occur frequently when standard statistical thresholds are used. Therefore, stringent statistical rules are used to declare a positive finding, usually using a threshold of P < 1 10-7.[186-188]

To date, approximately 100 variants associated with prostate cancer have been identified by well-powered GWAS and validated in independent cohorts (see Table 4).[189] These studies have revealed convincing associations between specific inherited variants and prostate cancer risk. However, the findings should be qualified with a few important considerations:

The implications of these points are discussed in greater detail below. Additional detail can be found elsewhere.[192]

In 2006, two genome-wide studies seeking associations with prostate cancer risk converged on the same chromosomal locus, 8q24. Using a technique called admixture mapping, a 3.8 megabase (Mb) region emerged as significantly involved with risk in African American men.[69] In another study, linkage analysis of 323 Icelandic prostate cancer cases also revealed an 8q24 risk locus.[68] Detailed genotyping of this region and an association study for prostate cancer risk in three case-control populations in Sweden, Iceland, and the United States revealed specific 8q24 risk markers: a SNP, rs1447295, and a microsatellite polymorphism, allele-8 at marker DG8S737.[68] The population-attributable risk of prostate cancer from these alleles was 8%. The results were replicated in an African American case-control population, and the population attributable risk was 16%.[68] These results were confirmed in several large, independent cohorts.[70-73,80-83,193] Subsequent GWAS independently converged on another risk variant at 8q24, rs6983267.[73-75] Fine mapping, genotyping a large number of variants densely packed within a region of interest in many cases and controls, was performed across 8q24 targeting the variants most significantly associated with prostate cancer risk. Across multiple ethnic populations, three distinct 8q24 risk loci were described: region 1 (containing rs1447295) at 128.54128.62 Mb, region 2 at 128.14128.28 Mb, and region 3 (containing rs6983267) at 128.47128.54 Mb.[75] Variants within each of these three regions independently confer disease risk with ORs ranging from 1.11 to 1.66. In 2009, two separate GWAS uncovered two additional risk regions at 8q24. In all, approximately nine genetic polymorphisms, all independently associated with disease, reside within five distinct 8q24 risk regions.[86,87]

Since the discovery of prostate cancer risk loci at 8q24, other chromosomal risk loci similarly have been identified by multistage GWAS comprised of thousands of cases and controls and validated in independent cohorts. The most convincing associations reported to date for men of European ancestry are included in Table 4. The association between risk and allele status for each variant listed in Table 4 reached genome-wide statistical significance in more than one independent cohort.

Most prostate cancer GWAS data generated to date have been derived from populations of European descent. This shortcoming is profound, considering that linkage disequilibrium structure, SNP frequencies, and incidence of disease differ across ancestral groups. To provide meaningful genetic data to all patients, well-designed, adequately powered GWAS must be aimed at specific ethnic groups.[206] Most work in this regard has focused on African American, Chinese, and Japanese men. The most convincing associations reported to date for men of non-European ancestry are included in Table 5. The association between risk and allele status for each variant listed in Table 5 reached genome-wide statistical significance in more than one independent cohort.

The African American population is of particular interest because American men with African ancestry are at higher risk of prostate cancer than any other group. In addition, inherited variation at the 8q24 risk locus appears to contribute to differences in African American and European American incidence of disease.[69] A handful of studies have sought to determine whether GWAS findings in men of European ancestry are applicable to men of African ancestry. One study interrogated 28 known prostate cancer risk loci via fine mapping in 3,425 African American cases and 3,290 African American controls.[208] On average, risk allele frequencies were 0.05 greater in African Americans than in European Americans. Of the 37 known risk SNPs analyzed, 18 replicated in the African American population were significantly associated with prostate cancer at P .05 (the study was underpowered to properly assess nine of the remaining 19 SNPs). For seven risk regions (2p24, 2p15, 3q21, 6q22, 8q21, 11q13, and 19q13), fine mapping identified SNPs in the African American population more strongly associated with risk than the index SNPs reported in the original European-based GWAS. Fine mapping of the 8q24 region revealed four SNPs associated with disease that are substantially more common in African Americans. The SNP most strongly correlated with disease among African Americans (rs6987409) is not strongly correlated with a European risk allele and may account for a portion of increased risk in the African American population. In all, the risk SNPs identified in this study are estimated to represent 11% of total inherited risk.

Some of the risk variants identified in Table 5 have also been found to confer risk in men of European ancestry. These include rs16901979, rs6983267, and rs1447295 at 8q24 in African Americans and rs13254738 in Japanese populations. Additionally, the Japanese rs4430796 at 17q12 and rs2660753 at 3p12 have also been observed in men of European ancestry. However, the vast majority of the variants identified in these studies reveal novel variants that are unique to that specific ethnic population. These results confirm the importance of evaluating SNP associations in different ethnic populations. Considerable effort is still needed to fully annotate genetic risk in these and other populations.

Because the variants discovered by GWAS are markers of risk, there has been great interest in using genotype as a screening tool to predict the development of prostate cancer. In an attempt to determine the potential clinical value of risk SNP genotype, cases of prostate cancer (n = 2,893) were identified from four cancer registries in Sweden. Controls (n = 1,781) were randomly selected from the Swedish Population Registry and were matched to cases by age and geographic region.[78] Known risk SNPs from 8q24, 17q12, and 17q24.3 were analyzed (rs4430796 at 17q12, rs1859962 at 17q24.3, rs16901979 at 8q24 [region 2], rs6983267 at 8q24 [region 3], and rs1447295 at 8q24 [region 1]). ORs for prostate cancer for men carrying any combination of one, two, three, or four or more genotypes associated with prostate cancer were estimated by comparing them with men carrying none of the associated genotypes using logistic regression analysis. Men who carried one to five risk alleles had an increasing likelihood of having prostate cancer compared with men carrying none of the alleles (P = 6.75 10-27). After controlling for age, geographic location, and family history of prostate cancer, men carrying four or more of these alleles had a significant elevation in risk of prostate cancer (OR, 4.47; 95% CI, 2.936.80; P = 1.20 10-13). When family history was added as a risk factor, men with five or more factors (five SNPs plus family history) had an even stronger risk of prostate cancer (OR, 9.46; 95% CI, 3.6224.72; P = 1.29 10-8). The population-attributable risks (PARs) for these five SNPs were estimated to account for 4% to 21% of prostate cancer cases in Sweden, and the joint PAR for prostate cancer of the five SNPs plus family history was 46%.

A second study assessed prostate cancer risk associated with a family history of prostate cancer in combination with various numbers of 27 risk alleles identified through four prior GWAS. Two case-control populations were studied, the Prostate, Lung, Colon, and Ovarian Cancer Screening Trial (PLCO) in the United States (1,172 cases and 1,157 controls) and the Cancer of the Prostate in Sweden (CAPS) study (2,899 cases and 1,722 controls). The highest risk of prostate cancer from the CAPS population was observed in men with a positive family history and greater than 14 risk alleles (OR, 4.92; 95% CI, 3.646.64). Repeating this analysis in the PLCO population revealed similar findings (OR, 3.88; 95% CI, 2.835.33).[214]

However, the proportion of men carrying large numbers of the risk alleles was low. While ORs were impressively high for this subset, they do not reflect the utility of genotyping the overall population. Receiver operating characteristic curves were constructed in these studies to measure the sensitivity and specificity of certain risk profiles. The area under the curve (AUC) was 0.61 when age, geographic region, and family history were used to assess risk. When genotype of the five risk SNPs at chromosomes 8 and 17 were introduced, a very modest AUC improvement to 0.63 was detected.[78] The addition of more recently discovered SNPs to the model has not appreciably improved these results.[215] While genotype may inform risk status for the small minority of men carrying multiple risk alleles, testing of the known panel of prostate cancer SNPs is currently of questionable clinical utility.[216]

Another study incorporated 10,501 prostate cancer cases and 10,831 controls from multiple cohorts (including PLCO) and genotyped each individual for 25 prostate cancer risk SNPs. Age and family history data were available for all subjects. Genotype data helped discriminate those who developed prostate cancer from those who did not. However, similar to the series above, discriminative ability was modest and only compelling at the extremes of risk allele distribution in a relatively small subset population; younger subjects (men aged 50 to 59 years) with a family history of disease who were in 90th percentile for risk allele status had an absolute 10-year risk of 6.7% compared with an absolute 10-year risk of 1.6% in men in the 10th percentile for risk allele status.[217]

In another study, 49 risk SNPs were genotyped in 2,696 Swedish men, and a polygenic risk score was calculated. On the basis of their genetic risk scores, 172 men aged 50 to 69 years with PSA levels of 1 to 3 ng/mL underwent biopsy. Prostate cancer was diagnosed in 27% of these individuals, and 6% had Gleason 7 or higher disease.[218] The utility of this strategy for identifying who should undergo prostate biopsy is yet to be determined.

In July 2012, the Agency for Healthcare Research and Quality (AHRQ) published a report that sought to address the clinical utility of germline genotyping of prostate cancer risk markers discovered by GWAS.[216] Largely on the basis of the evidence from the studies described above, AHRQ concluded that established prostate cancer risk SNPs have poor discriminative ability to identify individuals at risk of developing the disease. Similarly, the authors of another study estimated that the contribution of GWAS polymorphisms in determining the risk of developing prostate cancer will be modest, even as meta-analyses or larger studies uncover additional common risk alleles (alleles carried by >1%5% of individuals within the population).[219]

GWAS findings to date account for only a fraction of heritable risk of disease. Research is ongoing to uncover the remaining portion of genetic risk. This includes the discovery of rarer alleles with higher ORs for risk. For example, a consortium led by deCODE genetics in Iceland performed whole-genome sequencing of 2,500 Icelanders and identified approximately 32.5 million variants, including millions of rare variants (carried by <1% of the population). These variants were analyzed in 5,141 prostate cancer cases and 54,444 controls (genotypes were imputed in cases in which they had not been genotyped in previous analyses). In addition to previously reported risk alleles at 8q24 and 17q12, significant associations with prostate cancer were observed for two rare 8q24 SNPsthe minor allele (the G allele) of rs183373024 (OR, 2.69; P = 1.5 1023) and the minor allele (the A allele) of rs188140481 (OR, 2.88; P = 1.5 1022).[220] These results were validated in independent cohorts of European cases and controls. The frequencies of the risk alleles of these two variants in controls ranged from 0.1% to 1.1% and were lowest in southern Europe and highest in northern Europe. These data, in which risk alleles had high ORs compared with previous GWAS, demonstrate that the bulk of inherited risk may reside in rare alleles.

In addition, other genetic polymorphisms, such as copy number variants, are becoming increasingly amenable to testing. As the full picture of inherited prostate cancer risk becomes more complete, it is hoped that germline information will become clinically useful.

Notably, almost all reported prostate cancer risk alleles reside in nonprotein coding regions of the genome, and the underlying biological mechanism of disease susceptibility remains unclear. Hypotheses explaining the mechanism of inherited risk include the following:

The 8q24 risk locus, which contains multiple prostate cancer risk alleles and risk alleles for other cancers, has been the focus of intense study. c-MYC, a known oncogene, is the closest known gene to the 8q24 risk regions, although it is located hundreds of kb away. Given this significant distance, SNPs within c-MYC are not in linkage disequilibrium with the 8q24 prostate cancer risk variants. One study examined whether 8q24 prostate cancer risk SNPs are in fact located in areas of previously unannotated transcription, and no transcriptional activity was uncovered at the risk loci.[222] Attention turned to the idea of distal gene regulation. Interrogation of the epigenetic landscape at the 8q24 risk loci revealed that the risk variants are located in areas that bear the marks of genetic enhancers, elements that influence gene activity from a distance.[223-225] To identify a prostate cancer risk gene, germline DNA from 280 men undergoing prostatectomy for prostate cancer was genotyped for all known 8q24 risk SNPs. Genotypes were tested for association with the normal prostate and prostate tumor RNA expression levels of genes located within one Mb of the risk SNPs. No association was detected between expression of any of the genes, including c-MYC, and risk allele status in either normal epithelium or tumor tissue. Another study, using normal prostate tissue from 59 patients, detected an association between an 8q24 risk allele and the gene PVT1, downstream from c-MYC.[226] Nonetheless, c-MYC, with its substantial involvement in many cancers, remains a prime candidate. A series of experiments in prostate cancer cell lines demonstrated that chromatin is configured in such a way that the 8q24 risk variants lie in close proximity to c-MYC, even though they are quite distant in linear space. These data implicate c-MYC despite the absence of expression data.[224,226] Further work at 8q24 and similar analyses at other prostate cancer risk loci are ongoing.

Additional insights are emerging regarding the potential interaction between SNPs identified from GWAS and prostate cancer susceptibility gene regulation. One study found that a SNP at 6q22 lies within a binding region for HOXB13. Through multiple functional approaches, the T allele of this SNP (rs339331) was found to enhance binding of HOXB13, leading to allele-specific upregulation of RFX6, which correlates with prostate cancer progression and severity.[227] Thus, this study supports the hypothesis that risk alleles identified from GWAS may play a role in regulating or modifying gene expression and therefore impact prostate cancer risk.

A 2012 study used a novel approach to identify polymorphisms associated with risk.[228] On the basis of the well-established principle that the AR plays a prominent role in prostate tumorigenesis, the investigators targeted SNPs that reside at sites where the AR binds to DNA. They leveraged data from previous studies that mapped thousands of AR binding sites genome-wide in prostate cancer cell lines to select SNPs to genotype in the Johns Hopkins Hospital cohort of 1,964 cases and 3,172 controls and the Cancer Genetic Markers of Susceptibility cohort of 1,172 cases and 1,157 controls. This modified GWAS revealed a SNP (rs4919743) located at the KRT8 locus at 12q13.13a locus previously implicated in cancer developmentassociated with prostate cancer risk, with an OR of 1.22 (95% CI, 1.131.32). The study is notable for its use of a reasonable hypothesis and prior data to guide a genome-wide search for risk variants.

Although the statistical evidence for an association between genetic variation at these loci and prostate cancer risk is overwhelming, the clinical relevance of the variants and the mechanism(s) by which they lead to increased risk are unclear and will require further characterization. Additionally, these loci are associated with very modest risk estimates and explain only a fraction of overall inherited risk. Further work will include genome-wide analysis of rarer alleles catalogued via sequencing efforts, such as the 1000 Genomes Project.[229] Disease-associated alleles with frequencies of less than 1% in the population may prove to be more highly penetrant and clinically useful. In addition, further work is needed to describe the landscape of genetic risk in non-European populations. Finally, until the individual and collective influences of genetic risk alleles are evaluated prospectively, their clinical utility will remain difficult to fully assess.

Prostate cancer is clinically heterogeneous. Many cases are indolent and are successfully managed with observation alone. Other cases are quite aggressive and prove deadly. Several variables are used to determine prostate cancer aggressiveness at the time of diagnosis, such as Gleason score and PSA, but these are imperfect. Additional markers are needed, as sound treatment decisions depend on accurate prognostic information. Germline genetic variants are attractive markers since they are present, easily detectable, and static throughout life. Several studies have interrogated inherited variants that may distinguish indolent and aggressive prostate cancer. Several of these studies identified polymorphisms associated with aggressiveness, after adjusting for commonly used clinical variables, and are reviewed in the Table 6.

Findings to date regarding inherited risk of aggressive disease are considered preliminary. Further work is needed to validate findings and assess prospectively.

Like studies of the genetics of prostate cancer risk, initial studies of inherited risk of aggressive prostate cancer focused on polymorphisms in candidate genes. Next, as GWAS revealed prostate cancer risk SNPs, several research teams sought to determine whether certain risk SNPs were also associated with aggressiveness (see table below). There has been great interest in launching more unbiased, genome-wide searches for inherited variants associated with indolent versus aggressive prostate cancer. While GWAS designed explicitly for disease aggressiveness have been initiated, most genome-wide analyses to date have relied on datasets previously generated to evaluate prostate cancer risk. The cases from these case-control cohorts were divided into aggressive and nonaggressive subgroups then compared with each other and/or with the control (nonprostate cancer) subjects. Several associations between germline markers and prostate cancer aggressiveness have been reported. However, there remains no accepted set of germline markers that clearly provides prognostic information beyond that provided by more traditional variables at the time of diagnosis.

In independent retrospective series (see Table 6) the prostate cancer risk allele at rs2735839 (G) was associated with less aggressive disease. This risk allele has also been associated with higher PSA levels.[198,238] A hypothesis explaining the association between the nonrisk allele (A) and more aggressive disease is that those carrying the A allele generally have lower PSA levels and are sent for prostate biopsy less often. They subsequently may be diagnosed later in the natural history of the disease, resulting in poorer outcomes.

To definitively identify the inherited variants associated with prostate cancer aggressiveness, GWAS focusing on prostate cancer subjects with poor disease-related outcomes are needed. Notably, in a genome-wide analysis in which two of the largest international prostate cancer genotyped cohorts were combined for analysis (24,023 prostate cancer cases, including 3,513 disease-specific deaths), no SNP was associated with prostate cancerspecific survival.[239] The authors concluded that any SNP associated with prostate cancer outcome must be fairly rare in the general population (minor allele frequency below 1%). As more data regarding rarer variants are generated and validated, the value of inherited variants for therapeutic decision making may be determined.

While genetic testing for prostate cancer is not yet standard clinical practice, research from selected cohorts has reported that prostate cancer risk is elevated in men with mutations in BRCA1, BRCA2, and on a smaller scale, in mismatch repair (MMR) genes. Since clinical genetic testing is available for these genes, information about risk of prostate cancer based on alterations in these genes is included in this section. In addition, mutations in HOXB13 were reported to account for a proportion of hereditary prostate cancer. Although clinical testing is not yet available for HOXB13 alterations, it is expected that this gene will have clinical relevance in the future and therefore it is also included in this section. The genetic alterations described in this section require further study and are not to be used in routine clinical practice at this time.

Studies of male BRCA1 [1] and BRCA2 mutation carriers demonstrate that these individuals have a higher risk of prostate cancer and other cancers.[2] Prostate cancer in particular has been observed at higher rates in male BRCA2 mutations carriers than in the general population.[3]

The risk of prostate cancer in BRCA mutation carriers has been studied in various settings.

In an effort to clarify the relationship between BRCA mutations and prostate cancer risk, findings from several case series are summarized in Table 7.

Estimates derived from the Breast Cancer Linkage Consortium may be overestimated because these data are generated from a highly select population of families ascertained for significant evidence of risk of breast cancer and ovarian cancer and suitability for linkage analysis. However, a review of the relationship between germline mutations in BRCA2 and prostate cancer risk supports the view that this gene confers a significant increase in risk among male members of hereditary breast and ovarian cancer families but that it likely plays only a small role, if any, in site-specific, multiple-case prostate cancer families.[6] In addition, the clinical validity and utility of BRCA testing solely on the basis of evidence for hereditary prostate cancer susceptibility has not been established.

Several studies in Israel and in North America have analyzed the frequency of BRCA founder mutations among Ashkenazi Jewish (AJ) men with prostate cancer.[7-9] Two specific BRCA1 mutations (185delAG and 5382insC) and one BRCA2 mutation (6174delT) are common in individuals of AJ ancestry. Carrier frequencies for these mutations in the general Jewish population are 0.9% (95% CI, 0.71.1) for the 185delAG mutation, 0.3% (95% confidence interval [CI], 0.20.4) for the 5382insC mutation, and 1.3% (95% CI, 1.01.5) for the BRCA2 6174delT mutation.[10-13] (Refer to the High-Penetrance Breast and/or Gynecologic Cancer Susceptibility Genes section in the PDQ summary on Genetics of Breast and Gynecologic Cancers for more information about BRCA1 and BRCA2 genes.) In these studies, the relative risks (RRs) were commonly greater than 1, but only a few have been statistically significant. Many of these studies were not sufficiently powered to rule out a lower, but clinically significant, risk of prostate cancer in carriers of Ashkenazi BRCA founder mutations.

In the Washington Ashkenazi Study (WAS), a kin-cohort analytic approach was used to estimate the cumulative risk of prostate cancer among more than 5,000 American AJ male volunteers from the Washington, District of Columbia, area who carried one of the BRCA Ashkenazi founder mutations. The cumulative risk to age 70 years was estimated to be 16% (95% CI, 430) among carriers and 3.8% among noncarriers (95% CI, 3.34.4).[13] This fourfold increase in prostate cancer risk was equal (in absolute terms) to the cumulative risk of ovarian cancer among female mutation carriers at the same age (16% by age 70 years; 95% CI, 628). The risk of prostate cancer in male mutation carriers in the WAS cohort was elevated by age 50 years, was statistically significantly elevated by age 67 years, and increased thereafter with age, suggesting both an overall excess in prostate cancer risk and an earlier age at diagnosis among carriers of Ashkenazi founder mutations. Prostate cancer risk differed depending on the gene, with BRCA1 mutations associated with increasing risk after age 55 to 60 years, reaching 25% by age 70 years and 41% by age 80 years. In contrast, prostate cancer risk associated with the BRCA2 mutation began to rise at later ages, reaching 5% by age 70 years and 36% by age 80 years (numeric values were provided by the author [written communication, April 2005]).

The rest is here:
Genetics of Prostate Cancer (PDQ)Health Professional ...