Precision intervention for prostate cancer: re-evaluating who is at risk

Abstract

The vast majority of new prostate cancer diagnoses are low-grade tumors that are monitored by active surveillance rather than undergoing immediate treatment. However, a subset of men will progress to advanced prostate cancer which may result in lethality, and these men are likely to benefit from early intervention to prevent or delay such progression. For this high-risk group, which includes aged men, men of African descent, and those with a hereditary predisposition to prostate cancer, informed risk stratification can be the cornerstone of clinical decision making and treatment intervention. In this review, we discuss the importance of a precision intervention approach that considers the cumulative risk for any given patient or population to develop prostate cancer or to progress to lethal disease, with particular focus on the interplay of major determinants of high-risk disease.

1. Introduction

Prostate cancer is the most frequently diagnosed cancer in more than half of the world’s nations, accounting for more than 7% of the global cancer burden, exceeded only by lung and colorectal cancers [1–3]. In the United States, the incidence of prostate cancer has risen sharply over the last two decades, due in part to the increase in the ageing population, since ageing is a major risk factor for prostate cancer (see below). Given the increasing life-expectancy, particularly in Western countries but also worldwide, the incidence of prostate cancer is expected to further increase in the future.

Another major factor contributing to the increased incidence of prostate cancer is the wide-spread screening for prostate-specific antigen (PSA) [4], which has resulted in a substantial increase in prostate cancer detection, albeit a significant level of overdiagnoses [5–7]. Indeed, high PSA can occur in benign as well as malignant contexts, and is not necessarily indicative of lethal disease. Many prostate cancers are relatively indolent and unlikely to progress to lethal disease within the patient’s lifetime. Furthermore, overdiagnosis can lead to overtreatment of non-lethal tumors via surgery or radiation therapy, which may result in significant morbidities, including persistent urinary, sexual and/or bowel problems. It has therefore become common practice for men who are diagnosed with low-grade, localized prostate cancer to be monitored by active surveillance, which involves close observation without immediate intervention unless there are signs of disease progression [8, 9]. Notably, the implementation of targeted biopsies via imaging-guided modalities, such as multi-parametric MRI, has improved detection of prostate cancer and thereby increased confidence in accurate monitoring of disease progression [10–12].

For patients that exhibit signs of progression during active surveillance, radical prostatectomy or radiotherapy can often be curative [13]. However, a significant percentage will fail local treatment and further progress to aggressive prostate cancer, which is likely to metastasize quickly and can be lethal over time. The majority of recent investigations define aggressive prostate cancer as any tumor that would require treatment according to National Comprehensive Cancer Network (NCCN) guidelines and would therefore not be eligible for active surveillance [14]. Currently, it is difficult to accurately identify men at greatest risk of progression to more aggressive and potentially lethal disease when it is early enough to implement appropriate interventive measures that would improve their survival chances. Thus, the real challenge of active surveillance is achieving the appropriate balance between minimizing the adverse consequences of overtreatment, while maximizing the opportunity to identify potentially lethal tumors at the earliest possible point to ensure their most effective treatment.

Considering the wide range of clinical outcomes from indolence to lethality, a “one size fits all” approach is unlikely to be effective for preventive management of prostate cancer. Rather, it would be beneficial to implement a “precision intervention” approach wherein for any given patient, a disease management plan would be adopted that is tailored to their individual risk, with the goal of following a conservative monitoring plan for those at lower risk, and more active and earlier intervention for those at moderate or higher risk for aggressive prostate cancer [15, 16]. While this may seem to be an obvious approach, in practice it has been more difficult to achieve since it requires an accurate understanding of potential risk factors and how their cumulative interactions impact the chances of disease onset or progression for any given individual or population. In this review, we discuss the current status of our understanding of the major factors associated with prostate cancer risk (Figure 1), and consider how these can be used in a precision intervention paradigm to more accurately anticipate patient outcomes, and help stratify men for more effective management of prostate cancer.

Figure 1. Prostate cancer progression and risk factors.

Progression pathway for prostate cancer. The stages of prostate cancer development and progression are shown, together with the risk factors likely to be significant at each stage and their suggested contribution to increasing prostate cancer risk (from low to high risk). Specific major genetic alterations are depicted for the risk of genetic predisposition.

2.0. Factors associating with prostate cancer risk

We use the term “risk factors” to refer to anything that may increase the probability of prostate cancer development or progression to aggressive disease. Throughout this review we distinguish risk that is associated with the development or onset of prostate cancer, or its progression to aggressive disease, and we consider how a knowledge of key risk factors can be used for precision interventions. Many studies have identified factors associated with prostate cancer risk, of which race, age, and genetic determinants are among the most significant [2, 17, 18]. However, despite an abundance of basic, translational, and clinical studies that have expanded our understanding of the processes underlying prostate cancer initiation and progression and have improved its early detection and treatment, significant gaps remain in meaningful risk assessment that will allow for more accurate stratification for given individuals or populations.

2.1. Age

Age constitutes the most significant risk factor associated with prostate cancer. Indeed, the incidence of prostate cancer displays a steep association with age, much more so than other epithelial cancers. Less than 25% of prostate cancers are detected before the age of 65 while more than 60% occur in men 70 years or older, and more than 80% in men 80 years or older [2, 17, 18], suggesting that the risk of developing prostate cancer increases exponentially with age. Older men are also more likely to be diagnosed with high-grade prostate cancer and tend to have worse outcome [19, 20]. It is conceivable that such outcome differences reflect their health at the time of diagnosis; in particular, men are more likely to undergo treatment if they are otherwise healthy, whereas those with co-morbidities that are common in aged men may necessitate more conservative management. These observations suggest that although age is a crucial risk factor for prostate cancer, it should not be the sole indicator for guiding treatment decisions and highlights the importance of taking into account associated risk factors, such as those discussed below.

2.2. Oxidative Stress and Inflammation

Aberrant oxidative stress is another key risk factor for prostate cancer, and particularly associated with aging [21, 22]. For reasons that are largely unknown, the prostate is highly susceptible to oxidative stress, which is a hallmark of aggressive, lethal prostate cancer and manifested as an excess of reactive oxygen species (ROS) [23]. The effects of oxidative stress are potentiated by aging in several respects. On the one hand, accumulation of excess ROS in the form of hydroxyl radicals, superoxides or peroxides can accelerate aging of prostatic epithelium [24]. On the other hand, antioxidant defense systems are often compromised in elderly individuals making them more susceptible to oxidative stress conditions that produce excess ROS, which in turn can trigger and accelerate tumorigenesis [22, 25]. Increased ROS production and oxidative stress can be caused by various factors, including metabolic alterations, androgen receptor activation and mutation-induced mitochondrial dysfunctions [22].

Another major source of age-associated oxidative stress is inflammation. The presence of chronic inflammation known as prostatitis, is prevalent in benign prostatic intraepithelial neoplasia in elderly men [26–28]. This condition is often accompanied by immune cell infiltrate into the prostate, the presence of which has been correlated with increased cancer risk and poor prognosis [29–33]. Notably, an increase in activated inflammatory cells has been associated with increased ROS levels in the prostate, which may in turn promote somatic mutations and mitochondrial dysfunction [22, 34]. These changes can promote tissue damage resulting in impaired differentiation and compensatory epithelial cell proliferation, therefore potentiating prostate carcinogenesis [35, 36]. Overall, these observations suggest that the aging prostate acquires an inflammatory microenvironment that promotes aberrant oxidative stress, which in turn increases the risk of developing clinically-relevant prostate cancer.

Although oxidative stress and inflammation have been associated with prostate cancer progression, in practice, chemopreventive approaches using agents that target these processes have not proven to be promising [37–39]. Nonetheless, while the mass implementation of such agents may not be effective or may even have opposing effects [40], their individualized use in certain contexts may be beneficial. For example, individuals with elevated levels of oxidative stress or reduced expression of the prostate-specific homeobox gene, NKX3.1, may benefit from anti-oxidant interventions in a precision prevention paradigm (see for example [41]).

2.3. Race

Epidemiological data suggest that racial differences contribute significantly to increased risk for prostate cancer onset and aggressiveness. In the United States, for example, men of African descent experience an approximately 60% higher incidence of prostate cancer and a 144% higher death rate compared to caucasian Americans [18, 42], while men of African descent from Caribbean and sub-Saharan African countries exhibit high rates of prostate cancer mortality and have higher tumor stage at diagnosis, respectively, compared to other racial groups [43]. In the United States, African American men are often diagnosed with more advanced and lethal prostate cancer compared to other racial groups [44, 45]. Although most prostate cancer patients are diagnosed at 55 years of age or older, the average age of diagnosis for African Americans is 40 years, and they are also about two times more likely to die from prostate cancer than Caucasian men [46, 47].

Exogenous factors that have been implicated as contributing to these observed racial disparities include differences in access to health care that are associated with demographic or socioeconomic factors [48–50]. In particular, epidemiologic evidence shows that African Americans residing in racially segregated areas with persistent poverty, stagnant social mobility and poor healthcare access, have worse outcomes than their Caucasian counterparts [51]. However, it is unlikely that disparities in prostate cancer outcome are solely due to exogenous factors [52]. Recent studies have identified transcriptomic and genomic differences associated with differential susceptibility men of African ancestry to prostate cancer [47, 52–54]. In particular, genome-wide association studies (GWAS) have identified susceptibility loci that are independently linked with prostate cancer risk (see below and [55]). In our view, the potential interplay between socioeconomic and genetic factors that may contribute to racial differences in prostate cancer risk is an important area for future investigation.

3.0. Genetic risk in prostate cancer onset

Another significant risk factor is genetic predisposition; notably, men with a familial history of prostate cancer tend to have earlier onset and more aggressive disease [56–58]. To clarify, familial prostate cancer refers to disease that occurs in with a family history of the disease but without an evident genetic basis, whereas hereditary prostate cancer refers to disease that occurs in individuals having germline pathogenic variants that are associated with higher risk for disease onset or progression [59–61]. Hereditary prostate cancer accounts for 10–15% of all cases, and is generally defined as occurring in three successive generations, or at least two first-degree relatives, each diagnosed with early-onset disease [17, 62]. Although individual germline variants may influence onset or mortality, prostate cancer is more likely polygenic, shaped by the cumulative effect of multiple genetic variants each with a relatively small effect. In particular, GWAS have identified numerous single nucleotide polymorphisms (SNPs) that cumulatively associate with increased incidence or lethality of prostate cancer [55, 63–65], while linkage analyses have identified rare variants of known genes that associate with a greater lifetime risk of developing prostate cancer [66–69].

3.1. Single nuclear polymorphisms associated with prostate cancer risk

Supporting the notion that genetic predisposition for prostate cancer is driven by the cumulative interactions of multiple alleles, each associating with low to modest risk, several SNPs have been identified that co-occur in hereditary prostate cancer [70–74]. Initial GWAS studies identified 16 SNPs that strongly associate with a family history of prostate cancer [74], which paved the way for subsequent, more comprehensive GWAS that have led to the identification of more than 170 loci linked to prostate cancer susceptibility [64, 75]. Some of these SNPs occur in known genes, including BRCA1, BRCA2, ELAC2 (locus HPC2), RNase L, and NKX3.1, which have been further associated with increased incidence of aggressive prostate cancer (see below and [76–81]). However, the majority of SNPs are located in non-coding regions of the genome and do not have an easily discernable mechanism by which they might contribute to prostate cancer incidence or progression. For example, SNPs on chromosomes 15q13 and 17p12, which are not associated with any known genes, occur at significantly higher frequency in men with high grade invasive prostate cancer compared to those with lower grade or indolent disease [82, 83]. Another SNP (rs11672691) of as yet unknown function, has been shown to significantly associate with increased risk for prostate cancer mortality [84]. In-depth functional characterization of these genetic variants is warranted to better understand whether or how they contribute to increased susceptibility for aggressive prostate cancer.

Notably, certain SNPs are present at a higher frequency in African American men as compared to men of other races. For example, several germline variants of BRCA1 and BRCA2 have been observed in African American men more frequently than in Caucasian American men [85], associating with disease aggressiveness (see below). Additionally, a novel SNP (rs7918885), localized on chromosomal region 10p14 within an intron of a long non-coding region of the GATA3 gene, has been detected exclusively in African American men [86, 87], which is notable since loss of function of GATA3 has been functionally implicated in prostate cancer progression [88]. Therefore, analyses of SNPs may inform on differences in the incidence of aggressive prostate cancer observed in African American men (Figure 1).

A promising application of these SNP analyses has been the introduction of a polygenic risk score (PRS) or a polygenic hazard score (PHS), which take into account the total number of genetic variants in a given individual to assess his risk of prostate cancer development or progression [89–92]. The PRS of prostate cancer patients at the time of diagnosis has been associated with their respective survival outcomes [93], and in combination with PSA levels, may help to identify men who are at high risk for developing lethal prostate cancer, thereby further informing PSA screening [70, 94]. Of note, PRS assessment has proven to be effective for predicting the risk for developing aggressive prostate cancer in men who carry a germline variant of BRCA1 [95]. One recent study also identified an association between PRS and metastatic prostate cancer at the time of diagnosis [90], likely due to the inclusion of SNPs known to associated with metastatic prostate cancer risk [96, 97]. Despite progress in establishing PRS models to predict disease progression, the performance of PRS for predicting metastatic disease remains limited compared to overall prostate cancer incidence or mortality.

The PHS, another iteration of an SNP-based polygenic risk calculator, has proven to be a significant predictor of aggressive prostate cancer at the age of diagnosis [92]. More specifically, in the Prostate testing for Cancer and Treatment (ProtecT) clinical trial, men with a high PHS exhibited a 3-fold greater risk of developing higher grade prostate cancer compared to those with average PHS. Thus, in combination with assessment of other risk factors, models based on analyses of SNPs may facilitate earlier and more accurate detection of patients at high-risk for developing prostate cancer and for identifying those at greater risk of progression to lethal disease.

With that in mind, most polygenic risk assessment models were constructed using data from Caucasian men raising the concern that they may be under-representative of genetic variants that are more relevant for men of other races [93]. Comprehensive multi-ethnic studies revealed that modified PRS or PHS can be applied to accurately assess the risk of prostate cancer onset in African American men [91, 93, 98, 99]. In the course of these studies, three new SNPs on chromosome 8 were identified that associate with increased risk of prostate cancer onset almost exclusively in men of African ancestry [99]. Interestingly, one of them (SNP rs76229939) is located in the protein-coding region of the PCAT2 gene and the other two (rs74421890 and rs5013678) are in the non-coding region of the PRNCR1 gene, neither of which have as yet well-known functions in prostate cancer. However, no polygenic risk or hazard score has been shown to accurately discriminate men at risk of aggressive prostate cancer from those at risk of indolent prostate cancer within African American populations, warranting further investigations to improve SNP-based risk stratification for aggressive prostate cancer.

3.2. Germline variants associated with prostate cancer risk

Linkage studies have identified several germline variants that cluster in relatively predictable heritance patterns in kindreds at high-risk for prostate cancer [100–102]. Following these early studies, several other genes have subsequently been shown to be associated with prostate cancer risk. In particular, variants of BRCA1 and BRCA2, which are tumor suppressor genes that have distinct roles in DNA repair [103], have been shown to associate with high risk of prostate cancer, especially in men under the age of 65 [104]. BRCA1 and BRCA2 were originally identified on the basis of their hereditary susceptibility to breast and ovarian cancer [105, 106]. In fact, the first evidence of the association of their germline mutations with prostate cancer emerged from the Breast Cancer Linkage Consortium, which reported an increased incidence of prostate cancer in kindreds with a high incidence of breast and ovarian cancers [107].

Men with pathogenic variants of BRCA1 or BRCA2 tend to present with prostate cancer having higher grade and stage at diagnosis, including local metastases, and tend to have worse outcome after prostatectomy or radiation therapy compared to non-carriers [108, 109]; however, men with BRCA2 variants appear to be at a higher risk for lethal prostate cancer than those with BRCA1 variants [110, 111]. Recent interim data from the IMPACT study, a clinical trial in which annual PSA screening is used to determine the prostate cancer incidence in men who harbor these germline variants, revealed that men with alterations of BRCA2 but not BRCA1 are more likely be diagnosed with clinically-significant prostate cancer [112]. Interestingly, a rare protein-truncating variant in BRCA2 was recently identified in African American men diagnosed with early-onset prostate cancer, providing further support that BRCA2 variants and, consequently, DNA damage repair gene defects, may be are associated with the more aggressive prostate cancer phenotypes observed in men of African descent [113].

In addition, a substantial proportion of men undergoing treatment for locally advanced prostate cancer harbor potentially pathogenic variants of TP53 [114]. While somatic inactivation of TP53 is a late event during prostate cancer progression [115, 116], emerging evidence indicates that TP53 alterations occur at a relatively high frequency in primary and, especially, in castration-naïve metastatic prostate cancer [114, 117, 118]. Moreover, a recent study revealed that germline variants and somatic mutations of TP53 were the most prevalent alterations in approximately 25% of the primary tumors of patients who later developed aggressive metastatic disease [119]. The relative risk of carrying germline TP53 pathogenic variants is comparable to that of BRCA2 mutations, predisposing men to lethal prostate cancer: five germline missense TP53 variants, 158R>H, 181R>H, 252R>Q, 283R>C and 337R>H, have been reported to be significantly enriched in prostate cancer cohorts compared to men with benign prostate [120].

Another key gene associated with prostate cancer susceptibility is the HOXB13 homeobox gene that harbors the germline mutation 84G<E, or G84E as reported in literature, which has been associated with a 3 to 4-fold increase risk of prostate cancer onset [67, 121]. This germline variant was first identified in four prostate cancer families after sequencing over 200 genes on chromosome 17 [67], and subsequent studies have shown its strong association with hereditary prostate cancer [67, 122]. Men harboring this variant have an almost 20-fold greater probability of developing prostate cancer compared to the general population [67, 123], with subsequent meta-analyses establishing a strong association between HOXB13 G84E and early onset prostate cancer [124]. Follow-up studies have identified additional HOXB13 variants that have more modest association with increased risk of advanced prostate cancer [123]. Importantly, men carrying the G84E variant have significantly higher PSA levels at diagnosis and higher Gleason scores compared to non-carriers, suggesting a possible association between the HOXB13 G84E variant and aggressive prostate cancer [125]. On the other hand, one study showed that the G48E variant was associated with a 2.9-fold increase for developing prostate cancer, but not correlated with clinical stage, Gleason score or PSA levels [126], suggesting that HOXB13 G84E may identify men at high risk of developing prostate cancer but not those at greater risk of progression to aggressive disease. Two additional HOXB13 variants (R229G and G216C), harboring missense mutations, were exclusively detected in African American, and linked to higher stage and more aggressive prostate cancer [67]. Notably, HOXB13 is known to function in prostate development and contributes to androgen-independent prostate cancer [127, 128]; however, the mechanisms by which its germline variants, and specifically G84E, may promote prostate carcinogenesis remains unknown.

Another homeobox gene associated with prostate cancer risk is NKX3.1, a prostate-specific homeobox gene that regulates prostatic differentiation and whose reduced expression promotes prostate cancer initiation [129]. Linkage and GWAS analyses have identified several germline variants of NKX3.1 [78, 80, 130]. Of these, the rs11781886 variant, located 15 bases upstream of the start codon, has been shown to associate with increased prostate cancer susceptibility potentially by reducing NKX3.1 expression. Retrospective analyses in the Selenium and Vitamin E Cancer Prevention Trial (SELECT) also revealed a role for the rs11781886 variant in modifying prostate cancer risk [77]. Another germline NKX3.1 variant (454A>G), encodes a variant protein, NKX3.1 T164A, which has a mutation in the homeodomain that results in impaired DNA binding activity [78]. A second NKX3.1 variant, 154C>T encoding for the NKX3.1 R52C variant protein, has been associated with increased risk of higher grade prostate cancer [80]. Although none of these variants appear to be associated with racial disparities [80], downregulation of NKX3.1 expression levels has been linked to increased aggressiveness and mortality in African American men [47, 131]. Recently we reported a novel function for NKX3.1 in mitochondria in response to oxidative stress [132] and found that, unlike wild-type NKX3.1, neither the NKX3.1 T164A nor the NKX3.1 R52C variant proteins were capable of protecting against oxidative stress or suppression of prostate cancer initiation.

4.0. Genetic risk in metastatic prostate cancer

In general, prostate cancer has a low mortality rate since most patients are diagnosed at a localized or locoregional stage. Men diagnosed with localized disease that progresses to more aggressive disease undergo surgery or radiation treatment, which is often curative. However, treatment failure is common and is manifested as rising PSA levels, termed biochemical recurrence (BCR) [133]. Additionally, some men will present at diagnosis with metastatic disease, which is called metastatic hormone-sensitive (mHSPC). For the past 65 years, androgen deprivation therapy (ADT) has been on the forefront of treatment for men that fail primary interventions or those with mHSPC [14]. While most men initially respond to ADT, nearly all will eventually develop castration resistance. New treatment options for men with metastatic castration-resistant prostate cancer (mCRPC) have improved clinical outcome [134]; however, there have been no significant advancements in the treatment of mHSPC. This necessitates a clear need to improve on current risk-stratification guidelines to better define treatment paradigms for patients with mCRPC and mHSPC.

Several studies have attempted to address this issue by analyzing the genomic profiles of primary tumor biopsies from lethal prostate cancers, either presenting as mCRPC or mHSPC, to identify somatic alterations that are associated with disease progression [119, 135–138]. With regards to mCRPC, somatic alterations of MYC, CCND1, and PRKDC copy number gains as well as TP53 and ZNRF3 loss have been shown to have prognostic value for predicting metastatic relapse of localized disease [138]. Interestingly, somatic loss of NKX3.1 was the most frequent genomic alternation event in localized prostate cancer and also present in around 70% of mCRPC tumors. Another study revealed that alterations in DNA damage repair pathway genes were present from initial diagnosis to metastatic relapse, with 11 truncating mutations in BRCA2, CDK12, ATM, MSH6, and PALB2 present in both primary mCRPC biopsies [119]. Interestingly, in the same cohort, two patients harbored pathogenic germline variants of BRCA2 that manifested as biallelic BRCA2 loss, while the remaining deleterious DNA damage repair gene mutations were only detected in somatic DNA. These data support the use of genomic profiling for diagnostic purposes in prostate cancer biopsies, focusing on DNA repair gene defects [139], as the prevalence BRCA2 alterations (both somatic and germline) in primary prostate tumors was similar to that reported for mCRPC.

Given that mHSPC tumors share common genomic alterations with mCRPC [139–141], it is reasonable to assume that patients with mHSPC could also benefit from molecular-guided patient stratification. In that respect, germline or/and somatic genomic alterations in DNA damage repair genes occur in 18% of mHSPC tumors, with BRCA2 variants occupying the top spot as the most altered gene at 7% [135]. In addition, a small number of mHSPC patients were shown to harbor somatic TP53, PTEN, RB1 and CTNNB1 mutations associated with a higher risk of progression to lethal prostate cancer [118]. A recent study demonstrated that men with high-volume mHSPC and a higher frequency of NOTCH and cell-cycle pathway genes exhibited shorter times to mCRPC and reduced survival [137]. Therefore, the established association between certain molecular signatures, specific clinical phenotypes and outcomes in mCRPC may also open new avenues for similar risk stratification approaches in mHSPC patients.

5.0. Mitochondrial genetic variation associated with cancer risk

Although less well studied than variants of nuclear DNA, alterations of mitochondrial DNA (mtDNA) have been associated with prostate cancer risk [142, 143], particularly in African American men [143]. The absence of genomic recombination of mtDNA and its strict maternal inheritance leads to accumulation of heritable mtDNA mutations that can define population-specific mtDNA genomic profiles, which are known as haplogroups. Based on correlation analyses between population haplogroups and prostate cancer risk, African American men have been found to be at highest risk [144]. Furthermore, reduced mtDNA content as well as increased mtDNA mutational load have been associated with higher grade prostate cancer and poorer prognosis in men of African descent [145, 146].

An association between the total number of somatic mtDNA variants and prostate cancer stage at diagnosis is evidenced by the correlation between higher mtDNA mutational load and risk of biochemical relapse after surgery [142, 146, 147]. Meta-analyses of these studies revealed that increased mtDNA mutational load is correlated with higher Gleason score, elevated PSA levels and age at diagnosis [148], suggesting that total mtDNA mutational load, rather than individual mutations may be a stronger indicator of prostate cancer risk. However, a recent study based on sequencing mtDNA from 384 individuals with localized prostate cancer found that each individual had an average of at least one mitochondrial SNP (mtSNP) [149]. Some of these individual mtSNPs, found in recurring mutational foci, were predominantly present in men diagnosed with locally invasive or metastatic disease at older age. It is therefore possible that analyses of specific mtDNA alterations can inform prostate cancer risk predictions, particularly when combined with other genetic indicators of high-risk prostate cancer. For example, a SNP located in an mtDNA transcriptional control region was found to co-occur with MYC oncogenic gain or NKX3.1 loss, and such correlation was associated with adverse outcome [149], suggesting the interplay of variants of both the mitochondrial and nuclear genomes in promoting risk of disease progression.

With respect to mechanism, knowledge of the aging process and accompanying cellular dysfunction suggests an important role for disruption of mitochondrial function in age-related decline and predisposition to cancer [150]. In particular, persistent oxidative stress may ultimately promote mutations in the mitochondrial genome that give rise to mitochondrial dysfunction, which can induce or aggravate prostatic tumorigenesis [132, 149, 151]. As one example, an inherited mtDNA variant in the CO1 gene (6124T>C) has been shown to alter ROS production and proliferation in patient-derived prostate cancer cells, implicating a functional role for mtDNA variants in prostate tumorigenesis [152]. Additional analyses of mutational signatures in patient-derived prostate cancer cells revealed two additional inherited mtDNA mutations (8993T>G and 6142T>C) that lead to increased oxidative stress via ROS accumulation [143]. Although a recent study identified more than 80 conserved mtDNA mutations among patients with metastatic prostate cancer, only a small percentage are predicted to impact the corresponding protein function [148]. Further investigation of the functional significance of mtDNA alterations and their association with oxidative stress and prostate tumorigenesis may identify novel pathways of aggressive prostate cancer, leading to improved patient risk assessment.

6.0. Precision prostate cancer intervention in practice

The overarching goal of precision intervention is to capitalize on the knowledge regarding risk factors associated with early onset, increased incidence, or increased tendency toward more aggressive prostate cancer to enhance patient stratification with the ultimate goal of identifying, as early as possible, those men who are at risk of developing clinically relevant prostate cancer, while minimizing unnecessary treatment for those at lower risk (Figure 2).

Figure 2. Cumulative prostate cancer risk assessment model for improved precision intervention.

A comprehensive risk assessment is employed for the population at risk for prostate cancer, which incorporates genetic risk profiling, to determine the cumulative risk for prostate cancer (low to high) for each individual or population. Of note is that genetic risk profiling includes the assessment of both nuclear (e.g., PRS, HOXB13 G84E, BRCA2, NKX3.1 154C<T) as well as mitochondrial (e.g., mtDNA mutations, mtSNPs) alterations. Patient stratification to the most effective mode of intervention based on their cumulative prostate cancer risk is depicted that highlights a precision intervention paradigm for prostate cancer.

6.1. Risk stratification for prevention of prostate cancer onset and progression

Moving towards reliance on risk-based strategies particularly for men at high risk of developing prostate cancer, including those of advanced age, of African descent or with a family history of prostate cancer, could help to inform PSA testing. For example, the Rotterdam prostate cancer risk calculator incorporates available pre-biopsy information, such as ultrasound volume and digital rectal examination results that complement PSA testing, resulting in a significant reduction of unnecessary biopsies while improving detection of clinically important prostate cancer [153]. However, this approach does not consider individual risk and may benefit from the incorporation of genetic screening to assess nuclear and mitochondrial variants to identify individuals at a higher risk for developing prostate cancer.

Indeed, as emerging evidence supports the importance of the germline genomic landscape for development or progression of prostate cancer, significant advancements have been made to establish guidelines for the genetic risk assessment of prostate cancer patients [154]. The most notable update has been the implementation of genetic testing for all men with a strong family history of prostate cancer and high-risk germline alterations, such as pathogenic variants BRCA1 or BRCA2 [155]. Moreover, certain histopathological features (i.e. intraductal carcinoma) have been associated with pathogenic germline DNA repair gene alterations and therefore used as an indicator for genetic testing [156]. Recent studies indicate that non-hereditary genetic testing can be successfully applied to prostate biopsies from prostate cancer patients to predict absence of adverse pathologic features [157–159]. In that regard, histopathological changes associated with increased oxidative stress, which is also associated with higher risk, have been linked to reduced expression levels of NKX3.1 [132]. Therefore, a key opportunity is to use these approaches to better define the characteristics that make patients eligible for active surveillance [160], and in conjugation with existing criteria such as low grade tumor on biopsy and low PSA levels, guide appropriate prevention interventions.

Chemopreventive strategies have gained focus in an effort to decrease the incidence morbidity and mortality of prostate cancer, as well as treatment-related complications [161–163]. However, this approach has not been promising, as exemplified by a large-scale study of finasteride, an inhibitor of 5-α-reductase type 2 (5-ARI) that halts the conversion of testosterone to dihydrotestosterone. The Prostate Cancer Prevention Trial (PCPT) revealed that finasteride intake reduced prostate cancer prevalence by almost 30% over 7 years [162]. However, follow-up analyses did not show any clear indication that finasteride could improve the long-term (15-year) survival of prostate cancer patients, while its administration was associated with an increased occurrence of side effects, such as urinary and bowel incontinence, erectile dysfunction and gynecomastia [163]. Interestingly, our co-clinical studies identified NKX3.1 expression status as a predictor of response to finasteride [41], suggesting that genetic features might help identify men most likely to benefit from such chemoprevention interventions. This also begs the question as to whether men harboring pathogenic germline NKX3.1 alterations, indicative of a potentially more aggressive prostate cancer (see above), could also have better responses to finasteride during active surveillance.

Increasingly, a plethora of agents have shown promising anti-cancer activities but their efficacy in men on active surveillance remains unclear. One such example is metformin, an anti-diabetic drug that has been shown to have anti-cancer properties and good tolerance [164, 165]. A randomized study aimed at evaluating the effect of metformin in delaying prostate cancer progression in men undergoing active surveillance is on-going Metformin Active Surveillance Trial (MAST) (NCT01864096), expected to be finalized in 2024. Given the antioxidant activity of metformin [166], one would also assume that it would be a suitable preventative intervention for prostate cancer, where oxidative stress is a prominent risk factor. Of note is that men with specific genetic alterations, such as increased mtDNA mutation load or NKX3.1 aberrations, have been shown to foster increased levels of prostatic oxidative stress and higher risk for aggressive disease (see above). Therefore, adopting a precision prevention approach of tailoring intervention approaches, like metformin, for each individual patient or population based on a cumulative prostate cancer risk that incorporates genetic risk profiling, could significantly improve prostate cancer prevention (Figure 2).

6.2. Risk stratification for diagnosis and treatment prostate cancer

Several nomograms have been created for estimation of BCR risk, such as the CAPRA-S and the Walz nomograms [167]. The most recent update of the latter provides refined estimates for BCR risk at one and two years post-radical prostatectomy based on PSA, Gleason score, disease stage and metastasis status [168], albeit with no clear correlation to prostate cancer mortality rates. In that respect, the application of non-hereditary genetic testing based on the differential expression of genes involved in distinct biological pathways of prostate tumorigenesis have been beneficial to predict disease outcome. The most widely used genetic classifier in clinical practice is the Decipher test, which aims at assessing the expression of a 22-gene panel in patients newly diagnosed with localized prostate cancer at the time of biopsy, as well as men who have undergone radical prostatectomy [159]. Decipher has been successfully applied to identify subsets of patients with histologically low risk prostate cancer at diagnosis that qualify for active surveillance [158], suggesting that molecular profiling of newly diagnosed prostate cancer can improve risk stratification and help guide treatment decisions. Additionally, OncotypeDx, a 17-gene Genomic Prostate Score (GPS) signature based on prostate cancer biopsies, was shown to be predictive of adverse prostate cancer pathology beyond conventional clinical factors [169]. Interestingly, a recent study found that the accuracy of prostate cancer outcome prediction by the presence of rare genetic variants, like HOXB13 G84E, could significantly be enhanced by the incorporation of an individual’s PRS profile [89]. We further suggest that incorporation of mitochondrial genome screening in conjugation with established nuclear genetic tests, may further improve risk assessment. Although currently there is less known about the frequency and functional relevance of mtDNA alterations, compelling evidence suggest collaboration between mitochondrial and nuclear mutational profiles in defining prostate cancer aggressiveness and outcome (see above).

To date, testosterone suppression using novel ADT therapies is the mainstay of initial treatment for patients with mHSPC [14]. The trend has been to follow clinical rather than molecular factors to aid in decision making when it comes to incorporating additional agents targeting the androgen receptor axis, such as abiraterone or enzalutamide. However, treatment intensification may be preferentially used for individuals with the more aggressive mHSPC phenotypes harboring distinct genetic alterations, such as TP53 or RB1 mutations, sparing overtreatment and toxicity for those who are less likely to benefit and eventually develop mCRPC. Although the choice for further treatment following the development of castration resistance is unclear, recent evidence suggests the need for routine genetic testing to inform sequential treatment in mCRPC patients [170]. For example, testing for germline DNA damage repair defects can help identify patients who could benefit from targeted treatments, while determining prostate cancer risk in family members. If DNA damage repair genetic alterations, such as pathogenic BRCA2 variants, are already detectable in the primary tumor, there is a strong rationale for testing synthetic lethal strategies with PARP inhibitors or platinum chemotherapy in metastatic prostate cancer, where the magnitude of benefit for patients could be larger.

7.0. Concluding remarks

Precision intervention is intended to incorporate an individuals’ risk profile, which can be defined as the culmination of several risk factors contributing to increased prostate cancer incidence and mortality (Figure 2). In particular, improvement of precision screening and intervention at different stages of prostate cancer may establish a successful connection between specific genetic factors, such high-risk germline genetic alterations, and other non-genetic factors, such as age or race. This could allow for a more accurate assessment of individual or population risk for lethal prostate cancer and fulfil the need for better stratification strategies in chemoprevention or therapeutic interventions, where men with high risk benefit the most. We believe that more comprehensive genetic screening approaches could improve the impact of current prostate cancer risk assessment approaches and counterbalance the risk of overdiagnosis and overtreatment morbidity.

Highlights:

• Age, race, oxidative stress and inflammation influence prostate cancer risk

• Genetic factors play a pivotal role in predicting prostate cancer onset and outcome

• Nuclear and mitochondrial genetic alterations point to prostate cancer susceptibility

• Integrative risk assessment may improve precision prostate cancer intervention