The evolving role of familial history for prostate cancer

Abstract

Background. Family history of prostate cancer is a risk factor for prostate cancer occurrence. Differently from other neoplasms no major predisposing gene has been identified. Material and methods. This review article presents the controversial results of studies about the prognostic and predictive role of family history in prostate cancer, reports the discovered predisposing genes, and biologic and pathologic findings. Results. Mortality from PC remains a significant health care problem, but no trial investigated if it changed in presence of positive family history. The largest family study yet published concluded that men with family history are diagnosed and die at earlier ages than men without it. However, it failed to stress the prognostic value of family history. Genome-wide association studies of prostate cancer have identified a number of genetic variants at different loci in different populations. Prostate neoplasms of patients with positive family history exhibit a different pattern of expression of genes related with estrogen and androgen metabolism within the tumor. High-penetrance and low-penetrance genes in diagnosis and prognosis of prostate cancer, difficulties to define a classification and to quantify relative risks of single genes, documented gene-environment interactions are discussed. Conclusion. Family history stands for both shared genetic and environmental factors and their interaction. The availability of prostate-specific antigen test could explain partly the high familial risk, among brothers or shortly after the diagnosis of prostate cancer. Polymorphisms in genes associated with prostate cancer probably represent the most part of familial prostate cancer burden. An increasing knowledge of disregulated cellular pathways of lethal prostate cancer could define which of all genetic alterations have a role in defining new preventive and therapeutic strategies.

Prostate cancer (PC) is the most common male malignancy in Western countries. In 2008 the estimated new cases and deaths from PC in the European Union were 382 300 and 89 300, respectively [1]. Disease-specific mortality was confined to metastatic PC hormonal-sensitive and hormonal-resistant states.

Familial prostate cancer refers to a clustering of this disease within families. Hereditary prostate cancer (HPC) refers to a specific subtype of familial PC marked by a pattern consistent with passage of a susceptibility gene via Mendelian inheritance. In 1993 Carter et al. [2] provided a restrictive definition for HPC based on three criteria, resumed in Table I.

Table I. Criteria for Hereditary Prostate Cancer diagnosis [2].

HPC Diagnostic Criteria
1 Family including prostate cancer in >3 first-degree relatives
2 Family with prostate cancer in three successive generations of the maternal or paternal lineages
3 Family with two first-degree relatives affected at age £55 years

To be diagnosed with HPC the family should meet at least one of the three criteria.

Family history of prostate cancer

Family history (FH) of PC is a risk factor for PC occurrence. FH reflects not only shared genes but also shared environments and common behaviors.

Epidemiology of familial prostate cancer

Epidemiologic studies have determined that having a first-degree relative, brother or father, with PC increases the risk of PC for an individual by approximately two- to three-fold [3]. The risk is further increased by early age at onset among relatives and multiple relatives with the disease. However, the main population-based cohort studies were conflicting about the role of age and kinship [4–6]. The analysis of the Swedish Family-Cancer Database (SFCD) concluded that the age to reach the same cumulative incidence of PC as the general population at 55 years decreased with decreasing age at diagnosis of the relative. As well as incidence, PC specific mortality was related to the number of affected relatives, and it was higher for brothers than for sons of PC patients [7]. Additionally, a meta-analysis of 13 case-control and cohort studies reported pooled relative risk (RR) in first-degree relative of 2.5. RR was higher in relatives of patients with diagnosis of PC before age 60, and it declined with age. The risk for men with two relatives with PC increased 3.5 fold, and RR for sons appeared to be lower than for brothers [8]. Another meta-analysis found a more than two-fold increased PC risk among individuals with a FH of PC among first-degree relatives, and doubled risk for an age at diagnosis of less than 65 [9].

Approximately 10–15% of men with PC have at least one relative who is also affected by the disease, and it has been found that 5–10% of PC is accounted for by genetic susceptibility [5]. Carter and colleagues have suggested that the cumulative proportion of prostate cancer cases within the U.S. population that is attributable to high-risk susceptibility alleles is 43% for men diagnosed ≤55 years, 34% for men ≤70 years, and 9% for men ≤85 years [10].

In a study among 44 788 pairs of twins in Scandinavia, 42% of cases of PC were attributed to inheritance [11]. As twins studies show, the hereditary risk of PC is higher than for any other cancer [12–14]. Registry-based estimates of RR of PC incidence among first-degree relatives is 1.65 in Sweden [15] and 1.77 in the Netherlands [16]. By using the SFCD, Hemminki et al. concluded that PC showed the highest familial proportion (20.15%), followed by breast (13.58%) and colo-rectal (12.80%) cancers [17].

Genes involved in prostate cancer predisposition

At the genetic level there is some evidence about differences between sporadic and familial PC. A large body of evidence confirms that PC has a strong genetic basis. HPC is likely to be caused by multiple genes, interacting among themselves and with environmental factors, and this could explain the difficulty to identify susceptibility genes of HPC.

Distinct models of Mendelian inheritance and varying levels of penetrance are associated with different loci. An autosomal dominant inheritance is strongly suspected in familial early-onset PC [10,18], while an eterosomal allele could explain the presence of late-age PC in some families [19]. However, other models of inheritance have been postulated [20]. A recent segregation analysis of two early- and late-onset cohorts in Finnish families with clinically diagnosed PC cases indicates that Mendelian recessive inheritance could be consistent with the sex-limited X-linked region of HPC-X [21].

Susceptibility loci have been reported in more than a decade, but polymorphisms in genes associated with PC probably represent the most part of familial PC burden [22]. Genome-wide association studies of PC have identified a number of loci and genetic variants that individually contribute to a small increase in PC risk, but still do not explain the familial risk of this cancer. Recently, a secondary analysis of 1 233 PC pedigrees from the International Consortium for Prostate Cancer Genetics (ICPCG) using two novel statistics, sumLINK and sumLOD, confirmed significant linkage evidence at chromosome 22q12 and at other loci, concluding that over 70% of results are likely true positive findings [23].

High penetrance genes

The genetics of HPC remains unresolved. Discrepancies of studies could be attributable to ethnic differences, limited sample size, and other modifying factors. With many genome scans completed to date, suggestive evidence for loci has been described on nearly every chromosome [24]. This has allowed making several predictions regarding the genetic nature of PC in high-risk families. PC susceptibility is caused by mutations in multiple genes; different genes are more likely associated with distinct population frequencies in various ethnic groups.

The search for PC susceptibility genes by linkage studies has been inconclusive, mainly due to the difficulty of replicating promising regions of linkage in different populations. The ICPCG has emphasized that one of the major difficulties in studying PC is genetic heterogeneity, possibly due to multiple, incompletely penetrant PC susceptibility genes. In the case of HPC1, it was shown that only very large families with no evidence for linkage to the X chromosome can more likely attribute their disease to mutations at HPC1. In the case of HPC20 on chromosome 20, the Consortium found little evidence for replication [25]. In the last few years, a series of sequence variants located along chromosome 8q24 have been described. These variants are significantly associated with an increased risk of PC [26–29], as reported by more recent genome-wide association studies [28,30,31]. This discovery has been considered a major breakthrough in PC genetics. Although the high number of the identified 8q24 genetic variants, they do not explain the entire incidence of PC, and some authors concluded that germ-line mutations of 8q and 17q are able to explain around 16% of all HPC [32]. BRCA2 germ-line mutations explain less than 2% of early onset PC, its mutations confer a RR of PC diagnosis as higher as 20 [33]. Many other regions have also been implicated as sites of high penetrant genes conferring PC susceptibility in different populations, and are reported in Table II [31,34,35].

Table II. High penetrance genes predisposing to PC in patients with positive FH.

Locus	Gene or candidate gene	AR	Age	SNP/Mutation	Odds Ratio
1p36 (CABP)	Unknown	–	Early	–	–
1q23–q25 (HPC1)	RNASEL	4–13%	Early	Glu265X	1.15–3.07*
				Met1Ile	–
				D1S413	2.38
				471delAAA	–
1q42.2–43 (PCAP)	PCTA-1	40–50%°	Early	D1S2785	–
2p15–21	THADA	–	–	rs1465618	1.15–1.89
	EHBP1	–	–	rs2710646	–
	EHBP1	–	–	rs721048	0.98–1.22
2q31	ITGA6	–	–	rs12621278	0.35–0.45
3p12.1	CHMP2, POU1F1	–	–	rs2660753	1.13–1.54
3q21.3	Unknown	–	–	rs10934853	1.12
4q22	PDLIM5	–	–	rs7679673	0.69–0.83
4q24–25	TET2	–	Late	rs12500426	1.17–1.35
		–	–	rs17021918	0.67–0.85
6q26–q27	SLC22A3	–	–	rs9364554	1.02–1.22
7p15	JAZF1	–	–	rs10486567	1.11–1.42
	BHLHB8	–	–	rs12155172	1.01–1.24
7q21.3	PON1	–	–	Ileu102Val	6.3*
	LMTK2	–	–	rs6465657	1.06–1.26
8p21	Unknown	–	–	rs1512268	1.29–1.49
	NKX3.1	–	–	rs2928679	1.11–1.53
8p22–23	MSR-1	–	Late	Arg239X	–
				Glu265X	–
8q24	Unknown	–	–	rs10086908	0.65–0.76
	Unknown	–	–	rs10090154	1.32–3.72
	Unknown	–	–	rs1016343	1.42–1.88
	Unknown	–	–	rs10505476	0.70–1.32
	Unknown	–	–	rs12543663	1.07–1.64
	Unknown	–	–	rs13254738	1.00–1.37
	Unknown	–	–	rs13281615	0.69–1.69
	Unknown	9%	Early	rs1447295	1.40–3.22
	Unknown	–	–	rs1456310	0.58–0.92
	Unknown	–	–	rs16901966	0.48–12.57
	HapC	11–31%	–	rs16901979	1.34–2.07
	Unknown	–	–	rs16902094	1.21
	Unknown	–	–	rs3891248	0.98
	Unknown	–	–	rs4242382	–
	Unknown	–	–	rs4242384	1.66–2.10
	Unknown	–	Early	rs445114	1.14
	Unknown	–	–	rs620861	0.61–1.42
	Unknown	–	–	rs6470494	1.37–2.45
	POU5F1P1	21%	–	rs6983267	1.12–1.33
	Unknown	–	Early	rs6983561	0.82–2.26
	Unknown	–	–	rs7000448	1.19–1.26
	Unknown	–	–	rs7008482	1.20–2.60
	Unknown	–	–	rs7017300	1.45–2.28
	Unknown	–	–	rs7837688	1.49–1.90
	MYC (IVS1-355)	–	–	rs979200	0.68–1.14
	Unknown	–	–	broad11934905	2.24
	Unknown	8%	–	DG8S737-A8	1.23–1.62
9q33	Unknown	–	–	rs1571801	1.05–1.26
10q11.2	MSMB	–	–	rs10993994	1.05–1.25
	MSMB	–	–	rs7920517	1.03–1.22
10q26	Unknown	–	–	rs1276910	0.99–1.25
	CTBP2	–	–	rs4962416	0.97–1.18
11p15	IGF2/AS, INS, TH	–	–	rs7127900	1.27–1.52
11q13	MYEOV	–	–	rs10896469	1.05–1.25
	Unknown	–	–	rs11228565	1.23
	Unknown	–	–	rs7931342	1.04–1.24
12	CpG	–	–	rs902774	1.16–229
13q12	BRCA2	5%	Early	999del5	4.6*
				6174delT	1.9*
16p13	Unknown	–	Late	rs4782780	0.69–0.95
17p11(HPC2)	ELAC2	0.5–4%	Early	Ser217Leu	1.04–3.11
				Ala541Thr	1.18–2.40
				Glu633Val
17q12	Unknown	–	–	rs11649743	–
	Unknown	–	–	rs3760511	–
	HNF1B	–	–	rs4430796	1.14–1.35
	HNF1B	–	–	rs7501939	0.71
17q21	BRCA1	–	Early	Codon-356	1.8–3.3*
17q24.3	Unknown	–	–	rs1859962	1.08–1.28
	Unknown	–	–	rs6501455	–
	Unknown	–	–	rs7214479	–
	Unknown	–	–	rs983085	–
19q13	KLK2	–	–	rs198977	1.08
	KLK2/3	–	–	rs2735839	0.91–1.18
	KLK15/3	–	–	rs2659056	0.83–2.01
	Unknown	–	–	rs266849	0.42–1.08
	Unknown	–	–	rs8102476	1.12
20q13 (HPC20)	Unknown	–	Late	–	–
22q13	APOL3	–	–	rs5759167	0.70–0.77
	Unknown	–	–	rs9623117	–
Xp11.22	NUDT10/11	–	–	rs5945572	1.05–1.35
	NUDT11	–	–	rs5945619	1.06–1.36
Xq27-28 (HPCX)	Unknown	16%	–	DXS1113	–

Legend AR: attributable risk (accounting for a percentage of PC cases in familial clusters).

SNP: single nucleotide polimorphysm.

°Result not confirmed by other studies.

*Relative risk.

Low penetrance genes

Differently from initial segregation analyses supporting the hypothesis of rare highly penetrant loci, the more recent experimental evidences better support the hypothesis that some of the familial risks may be due to inheritance of multiple moderate-risk genetic variants, particularly proteins involved in androgen biosynthesis [36]. Molecular epidemiologic studies have yet to provide consistent inferences about the role of low penetrance genes in PC aetiology [37]. Furthermore, it is certainly possible that the common polymorphisms involved in androgen metabolism might modulate the effects of susceptibility genes for HPC [24]. Table III reports some examples of low penetrance genes in PC.

Table III. Low penetrance genes related to PC.

Cellular Pathway	Gene and protein	PC Incidence
Androgen metabolism	Androgen receptor. Rare mutations, frequent polymorphisms and short CAG/CGN-repeats.	Increased
	SRD-5A2. Role of A49T polymorphism and TA-repeats.	Increased
	CYP-17. Important for African Americans risk of PC, not for general population.	Increased
	CYP-3A4. Involved in testosterone metabolism; indirectly can increase PC aggressiveness.	Aggressive
	Other. CYP-19, PSA-AREI, HSD3B1, UGT, etc.	Variable
Estrogen metabolism	ER-beta.	–
	CYP-1B1. Polymorphism-432 favours PC incidence.	Increased
Vitamin-D receptor	VDR. Two polymorphisms increase PC incidence with a RR 2.0–2.5, many studies inconclusive.	Variable
DNA repair	XRCC1. Polymorphism-399 predisposes to PC.	Increased
	ERCC3. Polymorphism-241 favours or reduce PC depending on diet.	Variable
	ERCC5. Polymorphism-72 reduce PC incidence.	Reduced
	MGMT. Polymorphism-84 increases incidence of PC.	Increased
	XPD. Polymorphism-312 favours PC.	Increased
	CPA4.	–
DNA methylation	Many genes are substrates of HDAC and methylation system, are inactivated by methylation and PC incidence is consequently increased (GSTP1, FLNC, EFS, ECRG4, PITX2, PDLIM4, KCNMA, EDNRB, PTGS2, COX2, RASSF1A, RUNX3, APC, HOXD3, BMP7, CDH1, CDKN2A, etc). In particular APC methylation is associated with aggressive PC.	Increased
	NKX3.1	–
Carcinogen	GST. GST-M1 polymorphism increases incidence of PC.	Increased
metabolism	NAT1/2.	–
	CYP-2D6.	–
	CYP-2C19.	–
	UGT	–
Inflammation	RANTES	–
	TNF	–
	RNASEL	–
	MSR1	–
	TGFB1. Its polymorphism reduce PC incidence	Reduced
	Other. MIC-1, IL-8, IL-10, VEGF, ICAM, TLRs.	–
Hypoxia	HIF-2betae
	NOS
Other	KLK-3. Promoter variants are associate with more aggressive PC.	Aggressive
	CYR2. A polymorphism of this circadian gene favours incidence of PC.	Increased
	PTEN. Its reduction/absence is related with high Gleason scores.	Aggressive
	TGFB1. Polymorphism-509 reduce the risk of aggressive PC	Reduced

Genetic variation in key genes in the androgen pathway is important for development of PC and may account for a considerable proportion of all cases. Carriers of five high-risk alleles in the AR, CYP17, and SRD5A2 genes are at two-fold excess risk to develop PC [38]. Some variants of genes involved in androgen action have been reported conferring higher PC risk [39,40].

Other biological alterations, like aberrant DNA-hypermethylation, play a role in the interaction between genes and environment [41,42]. A recent literature review reported conflicting results about the role of vitamin-D receptor gene polymorphisms and PC development or progression, whereas polymorphisms of genes related to cell adhesion or angiogenesis appear more promising [43].

Environmental and dietary risk factors and common genetic polymorphisms of genes are likely to play a major role in common forms of PC [38,44–48]. Gene-environment interactions are basic in cancer development, especially for low penetrance genes and their polymorphisms. In a surgery cohort, among carriers of a mutated PC susceptibility locus on chromosome 17q12 there were few men with diabetes mellitus compared with U.S. population (7% vs 21%); it is unclear if this reflects selection bias, genetic protection from PC among patients with diabetes mellitus, or both [49]. Complexity of interactions between genes and environment emerges from a Chinese study [50]; it suggested that genetic variants of DNA repair pathways are involved in PC aetiology, and that PC risk of patients carrying some gene variants may be modulated by preserved foods and insulin resistance. The elevated risk of aggressive PC in men with a high energy intake could be attributable to certain metabolic profiles that favor enhanced growth factor production over an increase in adiposity [51]. The risk associated with FH is somewhat pronounced among taller and overweight men [4,52]. In many trials, obesity increases the risk of more aggressive PC, by a modulating effect of adiponectin, insulin or IGF-I, and may decrease either the occurrence or the likelihood of diagnosis of less-aggressive tumors [53].

Cellular pathways disregulations of prostate cancer

While men with HPC are at increased risk for PC development at an early age, biologic behavior of these cancers remains unclear.

Except an earlier age of onset, no anatomoclinical and tumor progression peculiarities between hereditary and sporadic PC have been previously identified [54]. However, regarding the expression of both androgen and estrogen receptor-related genes in sporadic and hereditary PC, the immunohistochemistry findings showed that the percentage of AR-positive cancer cells is higher in hereditary PC than in sporadic forms, whereas the mean number of oestrogen-alpha-receptor-positive stromal cells is higher in sporadic PC rather than in the hereditary one [55].

Inherited mutations in candidate genes may be associated with disease prognosis if they are involved in metabolic events that lead to tumor progression. These events include regulation of somatic DNA damage or repair directly and metabolism of steroid hormones that induce the growth of PC. Therefore, some of the genes that may be considered candidates for PC initiation may also be candidates for PC progression and prognosis.

Furthermore, these associated variants at 8q24 could affect regulation or transcription of a causal gene outside the region. A strong candidate is the proto-oncogene c-myc. Whereas the individual associations of the 8q24 variants with PC are relatively modest, the risk alleles are fairly common and a combination of multiple variants showed considerably larger associations.

Role of family history in the clinical management of prostate cancer

The role played by FH in prediction of PC is largely recognized, but the role of FH on prognosis of PC remains controversial. A poor prognosis in PC patients with a positive FH was reported from two large studies, both in the pre-prostate specific antigen (PSA) era [56,57]. Reports of PSA era have failed to confirm such a negative impact of FH on prognosis of PC [58–61]. In a more recent series of PC patients of Cleveland Clinic, it was demonstrated that familial PC has the same clinical behavior of sporadic PC, except for early age at onset [62]. Patients with familial PC had a slightly higher rate of surgical positive margins, and an earlier age of onset. Otherwise, in more recent studies, sporadic cases appeared similar to familial cases in terms of clinical and pathological features [63–65]. Radical prostatectomy has the same efficacy in sporadic and familial PC [64], as well as brachytherapy [66], biochemical progression-free survival at ten years is similar [65].

The largest family study yet published used the nation-wide SFCD, with a total of 26 651 PC cases, of which 5 623 were familial, concluded that men with a first-degree relative affected by PC are diagnosed and die at earlier ages than men without FH of PC. However, it failed to stress the prognostic value of FH due to the overall number of deaths among men with a FH was small and the maximum age at death in the study was 74, probably lower than the median age at death from PC in Sweden [7].

To date, although no major susceptibility gene has been identified, an assessment of the risk of PC according to the number and age of affected first-degree relatives is warranted for clinical counselling and screening recommendations [67].

Role of family history on prognosis and treatment of prostate cancer

Development of PC could differ among men with and without a hereditary predisposition to PC.

Within an Alpha-Tocopherol, Beta-Carotene Cancer Prevention prospective cohort study, on 1 111 incident PC cases of 19 652 men along a 12.3 years of follow-up, a first-degree FH of PC was associated with a RR of 1.91 for PC incidence, and a RR of 4.16 for advanced disease, after adjustment for age and trial intervention. Age at diagnosis was earlier among men with a FH of PC, stage at diagnosis was more advanced (58% vs. 31%; p= 0.0005) [4].

Early diagnosis fails to predict accurately the outcome of individual patients. A major challenge is to distinguish clinically indolent PC from aggressive PC with the potential to kill the patient. Early localized PC, as detected by screening programs, could harbor tumors with aggressive genetic characteristics [68], and the available diagnostic tools fail to provide consistent predictive information for clinical decision-making in individual cases.

On the other hand, FH allowed selection of families with aggressive PC, then identification of segregation of aggressiveness loci.

Every effort to replicate results about the identified predisposing genes using apparently similar data sets has been challenging. Therefore, the PC genetics community has focused on identifying loci associated with aggressive disease. Some analyses of genetic expression profiles have identified clinically relevant PC [69,70]. In identifying a profile of 86 genes that distinguish high- from low-grade PC, a set of potential targets for modulating the development and progression of the lethal PC phenotype has been generated [71]. In clinical practice, Gleason score (GS) has been identified as surrogate of aggressiveness of PC. Recent findings reported that GS is linked to several genomic regions. A genome-wide scan for PC aggressiveness loci has been attempted in a U.S. population [72], considering that genes linked to GS and, thus, tissue architecture could serve as molecular markers of PC aggressiveness. Previously, genomic regions containing tumor-suppressor genes for more aggressive PC on chromosome 5, 7, 10 and 16 were identified. Many genes and loci have been detected in association with PC aggressiveness, but the analysis found evidence for linkage with GS in three regions on chromosomes 5q, 7q and 19q, a slight linkage to GS was evident also for 1p and 9q. Genetic changes affecting PC prognosis have been reported in Table IV.

Table IV. Genes predisposing to aggressive PC in patients with positive FH for PC.

Locus	Gene or candidate gene	SNP/Mutations	Age	Odds Ratio (Gleason > 7)
1p13–q21	Unknown	–	–	–
1q23–q25 (HPC1)	RNASEL	D1S202	–	–
		Arg462Gln	–	14.8
		Asp174Tyr	–	2.3
1q42.2–43	Unknown	–	–	–
4q	Unknown	D4S403	–	–
5p13–q21	Unknown	D5S407	–	–
5q31–33	Unknown	D5S1480-820	–	–
6p22.3	Unknown	–	–	–
6q23.3	Unknown	D6S292	–	–
		D6S1040	<58
7q32–33	CPA4	D7S3056	–	–
		D7S1824	>65	–
		D7S3061-180	–	–
		D7S2531	–	1.86
8q24	Unknown	rs1447295	–	1.4
		rs6983267	Early	0.66
		rs4242382	–	2.22
		DG8S737	–	1.9
9q34	Unknown	–	–	–
10q11.2	MSMB (PSP94 protein)	rs10993994	–	1.57–1.64
10q26	CTBP2	rs4962416	–	1.46
11q13	Unknown	rs10896449	–	1.11
11q14.1–14.3	Unknown	–	<65	–
13q12	BRCA2	999del5	Early	3.42
		5174delT	–	3.20
17	BRCA1	185delAG	–	3.50
17p12	Unknown	rs4054823	–	2.11
17q12	HNF1B	rs4430796	Early	0.72
18	Unknown	–	–	–
19q12–13.11	Unknown	D19S433		–
		D19S245-902
20p11.21–q11.21	Unknown	ATTC013	–	–
20 p13–12.4	Unknown	–	–	–
22q11–13	Unknown	–	–	–
22q11.1	Unknown	–	–	–
22q12.3–13.1	Unknown	–	–	–
Xq12–13	Unknown	–	–	–
Xq27–28	HPC-X	–	–	–

Aggressiveness genes could provide important information about the most appropriate treatments. However, little information exists on whether inherited genotypes can be used to improve the ability to predict PC outcomes [73]. Two studies have shown that germ line DNA polymorphisms of androgen pathway can influence the response to androgen deprivation therapy (ADT), like HSD3B1, HSD17B4 and CYP19A1 [74], or like SLCO1B3 [75].

Limited data exist that address the role of inherited genotypes in determining optimal PC treatment or chemoprevention strategies. Response to hormonal therapy improved by 25% for each increase in CAG triplet repeat length in AR, and there was a four-fold better prognosis comparing men with 25 versus 19 AR-CAG repeats [76]. Recent studies describe germ line polymorphisms that determine the response to ADT. Coding and non-coding germ line polymorphisms in genes involved in androgen pathway affect the response to ADT, and have the potential to be useful for prognosticating the response to ADT, designing clinical trials for patients with early castration-resistant PC [77].

Role of family history in prevention of prostate cancer

FH is an important public health screening tool for identification of high-risk groups, particularly in the post-genomic era with the discovery of inherited causes of many diseases. Since PC is a result of complex genetic and environmental interactions [78], FH can be a personalized genomic tool that captures many of these interactions and it is essential to individualized cancer prevention strategies.

In recent years, five-year survival rates for PC have been ranked third highest of all cancers [79]. Much of this improvement could be attributable to an increase in the number of men being diagnosed with early-stage PC as a result of widespread use of PSA testing. PC screening with serum PSA has been accompanied by a dramatic stage migration, and today less than 20% of men show radiological evidence of metastasis at initial diagnosis.

Screening for PC in the general population remains controversial, the reduction of PC mortality and the possible harms have not been established [80]. Most guidelines recommend offering screening, PSA test and digital rectal examination, earlier for men with FH for early-onset PC or multiple affected individuals [67].

FH was a significant risk factor for PC at PSA levels within 0–4 ng/mL [81]. A recent prospective trial including 87 men demonstrated a high frequency of PC in men with a positive FH, normal digital rectal examination findings and a PSA level of 4.0 ng/mL or less [82]. This result raises the possibility of risk stratification for PC screening programs. Many efforts to ameliorate the positive predictive value of PSA test have been attempted, and many of them included FH, as the Prostate Cancer Prevention Trial calculator [83], or the inclusion of FH and determination of single nucleotide polymorphysms (SNPs) for individual risk assessment [84]. As recently showed by Baltimore Longitudinal Study of Aging including 505 men, SNPs on chromosome 10 and 19 do influence the risk of PC per unit increase in PSA [85].

Familial risk estimates by proband status and age are useful for clinical counselling, due to the different age-specific risks. A case-control study revealed that FH is a major risk factor of PC for Italian men younger than 60 years, with odds ratio (OR) 6.9 vs. 4.0 [86].

It should be remembered that the accuracy of defining familial cases is one of the problems that is not easy to resolve. A recommendation for mutation testing for a disease gene is usually based on the FH, and the results may turn out negative because the FH has been wrong. Some trials reported a higher prevalence of PSA testing among men with a FH of PC compared to those without [87]. Several studies and a meta-analysis have found that a sibling history of PC confers greater risk than a paternal history [4,7,8]. This would be consistent with the additional contribution to risk of a shared environment during childhood and adolescence [88]. However, a PSA effect could account for a consistent part of this difference. The diagnosis of cancer in a family member raises concerns about the risks among other members. In these families, application of opportunistic PSA testing may bias familial estimates, mainly for brothers than for sons of affected fathers, masking a true pattern of inheritance and conducting to an early detection [89]. This PSA effect could also explain the overestimated familial RR shortly after the first diagnosis. An analysis of data from PC Database Sweden concluded that an increased diagnostic activity among men with a FH of PC contributed to an increased risk of PC and produced detection bias in studies of familial PC. In particular, brothers of index patients with PC were at increased risk of PC, especially brothers with a T1c disease and with a diagnosis in the year after the diagnosis of the index patient [90].

Finally, perceived cancer risk is negatively associated with age, and may vary as a function of the salience of the disease in one's larger racial/ethnic community and peer groups [91]. Psychosocial factors such as FH, worry or concern about PC and marital status play a role in white men's decisions about PC screening [92].

Future directions

Mortality from PC remains a significant health care problem. The risk of death from PC is surely a less biased indicator of the risk for invasive PC. Moreover, no trial reported it. This is not a secondary matter: as in other neoplasms, in the PSA era it is necessary to know the role of FH on PC prognosis, if any. However, a possible lack of prognostic relevance could mirror the existence of different predisposing genes, some of them predisposing to aggressive PC and other ones predisposing to indolent PC.

The genetic profile of tumors of patients with aggressive metastatic castration-resistant prostate cancer (mCRPC) may help to define which predisposing gene could have a relation with the diagnosis of lethal PC. Several recent articles reported on a frequent occurrence of a TMPRSS2-ERG fusion gene in PC, a molecular event that has been associated with a more aggressive clinical phenotype, implying the existence of a distinct molecular subclass of PC. In particular, the fusion gene expression has been shown to be regulated by an oestrogen-receptor dependent mechanism [93]. It is interesting that the level of oestrogen-receptor-alpha expression in stromal cells of familial PC is different from sporadic cases [55]. Then, it is mandatory to elucidate the different activation of hormonal pathways in mCRPC with and without TMPRSS2-ERG fusion gene, and which genetic alteration among those reported has a role in the activation of this cellular pathway. New insights on this topic could change the treatment of mCRPC moving towards a tailored treatment.

PSA doubling time, TMPRSS2-ERG fusion gene occurrence, as well as GS and other prognostic parameters, define a more aggressive PC, a disease that has a marked ability to escape hormonal manipulations. A retrospective assessment of FH of patients with a rapid development of castration-resistance appears as a shortcut to better understand the clinical and prognostic significance of the high number of genetic predisposing alterations to PC, as a means to clearly define the role of c-myc, PI3K/Akt and src pathways. It is interesting to depict a map of the complex network of gene inactivations and their relative importance at diagnosis of PC or at emergence of mCRPC. To date, CpG island hypermethylation status at a defined panel of genes may be useful biomarker in men with mCRPC [94] as well as an ‘apoptotic methylation signature’ could have a diagnostic and prognostic role [95].

In summary, FH could address preclinical research in defining PC populations with different prognosis and activation of different molecular pathways. Even for every single gene the phenotypic effects vary depending on mutation or environmental influences on the mutated protein. Considering the wide prognostic and diagnostic of every mutation of PC predisposing high penetrance genes, beside environmental effects, the consistent results of studies about prognostic effects of genetic mutations should be translated immediately in clinical oncology and preventive medicine. It could be possible if every result, either deletion or methylation, will be classified by the effect it produces on prognosis and by populations in which the effect is proven. A molecular classification of PC will be reliable as a consequence of the future studies in the domain of genetic epidemiology and prognostic effect of more frequent mutations predisposing to PC or to mCRPC. If such a classification would be available, a role of environmental risk factors could be more easily assessed in prospective studies.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.