Prostate specific antigen reduction following statin therapy: Mechanism of action and review of the literature
Prostate specific antigen (PSA) is a serine protease that is exclusively produced in the prostate, and its detection is the only laboratory test available for screening men for prostate cancer (PC). The interpretation of the assay is difficult since it is specific for prostate tissue and cellular growth, but not for PC. Pharmacologic therapy for hyperlipidemia, such as statins, may influence prostate cellular growth and subsequently PSA levels in patients. Dysregulated cellular growth in the prostate is mediated by inhibiting the rate-limiting pathway step in cholesterol synthesis, thereby decreasing isoprenylate intermediates, decreasing cholesterol rich cellular membrane domains, and down-regulating androgen and estrogen receptors. Statins, with variable efficacy, have been previously shown to inhibit cellular inflammation, angiogenesis, proliferation, migration/adhesion, and invasion, while promoting apoptosis in prostate cells by inhibiting the conversion of HMG-CoA to mevalonate. An individual statin's molecular structure, need for enzymatic conversion, bioavailability, and peripheral tissue concentration may partially account for differing properties. By inhibiting prostatic cellular growth and promoting apoptosis, statins may subsequently decrease PSA levels, an effect recently observed in cohorts. There is scientific and clinical evidence supporting the observations that statins are associated with an overall reduction in serum PSA in men, when used for greater than 6 months, and especially when used for greater than 2 years. © 2010 IUBMB IUBMB Life, 62(8): 584–590, 2010.
Prostate specific antigen (PSA) is a serine protease that is exclusively produced in the prostate and can be quantified in human serum by assays using several monoclonal antibodies. PSA elevation is commonly observed in men in a variety of clinical situations, including benign prostatic hypertrophy (BPH), prostatitis, prostate infarction, prostate intraepithelial neoplasia, prostate cancer (PC), oxidative stress (1, 2), recent ejaculation (3), and manipulation of the gland through instrumentation or digital examination. Furthermore, serum PSA is highly correlated with increasing patient age and prostate volume, and can be expected to rise by ∼ 3.2% (0.04 ng/mL) per year for a healthy 60-year-old man (4) or rise 0.2 ng/mL for every 1 g of benign hyperplastic tissue (5). PSA measurements are stable in men over a defined short period of time; however, it has been observed that one third of patients may have considerable variability of greater than 1.0 ng/mL between PSA measurements within 90 days (6).
Currently, the American Cancer Society and the American Urologic Association recommend that PSA levels be assessed in men older than 50 years of age in an attempt to screen individuals at risk for developing PC (7, 8). Unfortunately, the interpretation of the assay is difficult since it is specific for prostate tissue, but not for PC, and varies by race (9). Interestingly, ∼40% of men with organ-confined PC have normal levels (10) and 25% of men with BPH have elevated levels (3). Thus, the test has poor sensitivity and specificity in regards to assessing overall risk of PC in men.
While PSA is not an ideal tumor marker, it represents one of the only laboratory screening tests for PC. Typical management dictates that men with PSA levels higher than 4.0 ng/mL should be ruled out for PC with biopsy. Pharmacologic therapy, such as statins, may influence PSA levels in patients. It is important for clinicians to be aware of such effects in evaluating the risk of PC in patients.
EFFECTS OF STATIN THERAPY ON THE PROSTATE
HMG-CoA reductase inhibitors, otherwise known as statins, inhibit the rate-limiting pathway step in the conversion of HMG-CoA to mevalonate. Inhibition of HMG-CoA reductase is the only known mechanism of bioactivity of these agents (11). Thus, statins have been demonstrated to decrease total and low density lipoprotein cholesterol, while raising high density lipoprotein in patients. Following the use of statins, observations were noted that this class of medications may have unintended, yet beneficial effects, that extend beyond treatment for hyperlipidemia.
Statins have been previously shown to inhibit cellular inflammation (12), angiogensis (13), proliferation (14), migration/adhesion (15), and invasion, while promoting apoptosis (16). Inhibition of HMG-CoA reductase reduces cellular concentration of mevalonate, thereby influencing cellular signaling pathways by reducing downstream isoprenylate intermediates (17). Two of the more important intermediate products of the mevalonate pathway are farnesylpyrophosphate and geranylgeranylpyrophosphate. Farnesylpyrophosphate is a precursor for the formation of cholesterol, heme A, dolichols, and ubiquionones. The coupling of farnesylpyrophophate and geranylgeranylpyrophosphate to intracellular proteins, reactions otherwise known as isoprenylation, is essential to the membrane binding and subsequent functionality of Ras/Rho G-proteins (18–20). Inhibition of HMG-CoA reductase by statins results in the inability of the Ras protein to be farnesylated (18), which may be responsible for cellular growth arrest, apoptosis, and reduced angiogenesis in prostate tissue.
Reduced prostate cellular growth through inhibition of isoprenoid synthesis alone may be questionable given the poor penetration of statins in peripheral tissues (11). Poor tissue penetration of statins may be the result of extensive hepatic extraction (11) resulting in limited bioavailabity and short duration in the circulation. For unknown reasons, the prostate may accumulate statins at higher concentrations and for longer duration than has currently been demonstrated, thereby allowing isoprenylation-mediated effects of statins to be realized in vivo. While this postulated mechanism is plausible, there is no evidence to date to support this occurs in humans. It has also been suggested that the antitumor effects of statins within peripheral tissues may result from a decrease in circulatory low density lipoprotein levels (11). In response to circulating androgens, the prostate normally synthesizes cholesterol and lipids by increasing transcription of genes responsible for fatty acid synthase and HMG-CoA reductase (21–23). Androgen mediated lipogenesis is influenced by sterol response element binding proteins, namely SREBP-1a, SREBP-1c, and SREBP-2 (24). Dysregulation of SREBP-2 may be responsible for elevated prostate tissue cholesterol levels. Interestingly, increased intraprostatic cholesterol deposits are observed in patients with aggressive PC (25).
Statins may regulate prostate cellular growth and induce cytostasis by decreasing cholesterol rich cellular membrane domains (26). These membrane domains, otherwise known as lipid rafts, consist heavily of cholesterol and sphingolipids which serve as platforms for signaling protein assembly (27). These lipid rafts have been postulated to influence intracellular signaling (28), including androgen (29), epidermal growth factor (30), and luteinizing hormone receptor activity (18, 31). Lowering circulating cholesterol levels may modulate hyperplastic prostate growth in patients by inhibiting serine/threonine kinase Akt1 activation (32) and translocation (33). Akt1 activation normally recruits SREBP transcription factors, thereby inducing expression of lipogenic enzymes (32).
Increased prostate cellular concentration of cholesterol is thought to promote unregulated cellular proliferation (34). Disruption of membrane cholesterol inhibits the cellular proliferation effects of epidermal growth factor receptor activation (30). Epidermal growth factor receptor disruption leads to inactivation of Akt, normally a crucial cell survival signal, solid tumor growth promoter in prostate cells (35), and regulator of angiogenesis (36). Interestingly, statins have biphasic capabilities whereby they are able to predominately affect cholesterol synthesis (37) and promote angiogenesis through Akt activation at low concentrations. At higher concentrations, statins inhibit angiogenesis and biosynthesis of non-sterol products through decrease protein prenylation (33, 37). High doses (20 μM) of simvastatin have been shown to decrease Akt activation and induce cellular apoptosis (38), in contrast to doses of less than 0.1 μM which consequently leads to rapid Akt activation (39).
Cholesterol is also an essential precursor in sex steroid synthesis; however, studies have failed to demonstrate that statin use decreases circulating hormones in patients, such as free testosterone, dehydroepiandrosterone sulfate, or luteinizing hormone (40). Androgen independent prostate cellular growth has been associated with cellular lipid membranes and thus cellular cholesterol concentration (41). Furthermore, sex hormone binding globulin was found to be 10.6% lower in statin users compared to non-users. However, this may be more easily explained by baseline higher insulin levels acting as a negative regulator of sex hormone binding globulin production in the liver (40). Interestingly, statins may lower intraprostatic androgen levels, despite not lowering serum androgen levels in patients (42). The androgen receptor has been shown to recruit transcription factors, such as sterol response element binding proteins, that regulate lipogenesis (24). Furthermore, alteration of sex steroid synthesis may change networks of epithelial, stromal, and luminal factors necessary for prostate growth (43), and subsequently decreasing PSA levels in men. Dysregulated cellular growth in the prostate is mediated by inhibiting the rate-limiting pathway step in cholesterol synthesis, thereby decreasing isoprenylate intermediates, decreasing cholesterol rich cellular membrane domains, and decreasing androgen and estrogen receptors.
PHARMACOLOGY OF THE STATIN MEDICATIONS
HMG-CoA reductase inhibitors, otherwise known as statins, received United States Food and Drug Administration approval in 1987 for the management of hyperlipidemia (44). The statin family of medications are comprised of lovastatin, simvastatin, atorvastatin, fluvastatin, rosuvastatin, and pravastatin. In 2006 alone, atorvastatin was the top selling medication, with greater than 74 million prescriptions filled by patients (45). Pharmacologically, not all statin medications are equivalent; subtle differences in structural properties of statins affect bioavailability, peripheral circulation concentration, and effectiveness in lowering low density lipoprotein levels.
Statins have a side chain that either contains an open ring (acid) or a closed ring (lactone) structure, which is an inactive prodrug that requires conversion to the active β-hydroxyacid form by carboxyesterases found predominately in the liver and in the plasma (17). Simvastatin and lovastatin are administered as prodrugs and require esterase-dependent conversion from their lactone form to the β-hydroxyacid open form. However, pravastatin does not require enzymatic conversion and is structurally similar to the open form of lovastatin, with the difference being a methyl group in lovastatin versus a hydroxyl group in provastatin on the C6 position of the common naphthalene ring (11). Unlike simvastatin and lovastatin, pravastatin is already in an active form, resulting in higher serum concentrations and greater bioavailability (11). Similar to pravastatin, atorvastatin does not require esterase mediated conversion since it is not a prodrug but does undergo first pass liver extraction which results in low bioavailability in peripheral tissues. The clinical relevance of this enzymatic conversion and activation is that HMG-CoA reductase binds ∼1000 times more effectively by statins with open ring structures than by HMG-CoA itself (46), possibly accounting for some of the pharmacologic differences among competing statins.
All statins, except pravastatin, are lipophilic and thus concentrate primarily in the liver and poorly in the peripheral tissues resulting in lower circulation concentration (11). This property plays a large role in absorption and bioavailability of individual statins. For example, the bioavailability of lovastatin is less than 5% (Merck) compared to 40% for atorvastatin (17). Lipophilic statins may be the most effective in lowering PSA levels due to greater intracellular access (47). Thus, atorvastatin has shown to provide the greatest mean reduction in low density lipoprotein cholesterol (53%), than simvastatin (38%), lovastatin (28%), pravastatin (15%), or fluvastatin (15%) in patients (48). Most clinically relevant, simvastatin, lovastatin, pravastatin, and fluvastatin have been shown to have about 85%, 60%, 50%, and 33% of the efficacy of atorvastatin, respectively (49).
Individual statin medications are not equipotent in treating hyperlipidemia, inhibiting cellular growth, promoting apoptosis (50), and decreasing cell viability of prostate cells. The inhibitory potency for cellular growth of simvastatin, lovastatin, fluvastatin, and atorvastatin is similar, but pravastatin has been shown to be less potent (51). Differences in inhibitory potency may be partially due to variations in the equilibrium dissociation constants (Ki) of statins: Simvastatin, 1.2 × 10−10 M; Atorvastatin, not determined; Lovastatin, 3–6 × 10−10 M; Pravastatin, 23 × 10−10 M (11). Simvastatin, lovastatin, and mevastatin induce dose-dependent apoptosis in the TR-PCT1 pericyte cell line, which is thought to be cholesterol, caspase-3, and caspase-7 mediated. At a 5 μM concentration, mevastatin was observed to induce apoptosis in 20% of pericytes, compared to 30% with lovastatin and 45% with simvastatin (52). Additionally, lovastatin and simvastatin have been shown to decrease cell viability in PC cell lines (PC3, DU145, LNCaP) by inducing apoptosis and cell growth arrest, mainly by over-expression of Cdk inhibitors such as p27 and p21 (53, 54). Lastly, lovastatin and simvastatin suppress Rb, phosphorylated Rb, cyclin D1, cyclin D3, CDK4, CDK6, and c-jun expression, thereby inhibiting growth of PC cells (20). While generally successful in treating hyperlipidemia in patients, statins have additional biologic effects that vary by individual statin medication, depending on molecular structure, need for enzymatic conversion, bioavailability, Ki, and concentration in the peripheral tissues. By inhibiting prostatic cellular growth and promoting apoptosis, statins may subsequently decrease PSA levels, which is highly correlated with prostate cellular growth.
IMPACT OF STATIN THERAPY ON PSA
There are only a handful of studies that directly address the relationship of the effect of statins on PSA levels. Observational evidence and recent studies to date in human populations appear to support biologic evidence that statins reduce PSA levels. Hypercholesterolemia is one of the most common medical conditions treated in the United States. It is not uncommon for primary care providers to simultaneously or near simultaneously test lipid and PSA levels. Thus, the association of these two variables provides valuable insight into a possible causal relationship.
In 2005, Cyrus-David et al. published the first retrospective study evaluating the effect of statins on PSA levels in a cohort of airline pilots (55). A convenience sample of 15 airline pilots who were diagnosed with hypercholesterolemia were compared with 85 patients with normal cholesterol levels from 1992 to 2001. All men included were 40 years or older who did not have a history of PC, prostatitis, or BPH. After fifteen months of follow-up (six follow-up visits), there was a 41.6% decrease in total serum PSA between the baseline and sixth visits in the treatment group. Furthermore, during this same period of time, the serum PSA level increased by 38% in the comparison control group. The investigators concluded that the decrease in serum PSA in the treatment group may be the result of apoptosis induction and the inhibition of undetectable premalignant prostate cells. The reasoning for a hypothesized effect on PSA levels may be based on statins' growth-inhibitory effects on epithelial cells.
Moyad et al. (2005) studied a total of 512 men at Schiffler Cancer Center from 1995 through December 2001, who were treated with permanent brachytherapy for PC, assessing biochemical progression free survival which included serial PSA measurements and use of statin medications (56). Among statin users, which consisted mostly of atorvastatin, the investigators observed a statistically significant difference in pretreatment PSA level at 5.7 ng/mL compared to 6.9 ng/mL in men who did not use statin lowering medications (56). Soto et al. (2009) observed as well that patients on preoperative statin therapy presenting for radiation therapy had a 1.5 ng/mL decrease in PSA compared to men who did not use cholesterol lowering agents over a 26 month follow-up period (57). It has been suggested that statins and lower cholesterol may act as a radiation sensitizer for these men (58).
In 2007, Mills et al. published a phase 2, double-blind, randomized placebo controlled clinical trial evaluating atorvastatin treatment for men with lower urinary tract symptoms and benign prostatic enlargement (59). Included among many of the endpoints were differences in PSA levels among subjects after 6 months of treatment with 80 mg daily of atorvastatin or placebo. Patients were excluded, who had a history of PC and history of urologic surgery or procedures that may alter prostate anatomy. A total of 350 subjects were treated and randomized from over 35 centers in nine countries to treatment and placebo arms. Baseline PSA level was 0.08 ng/mL lower in the treatment arm than the placebo control arm. After 26 weeks of treatment with atorvastatin, there was a mean decrease of 0.1 ng/mL compared to a mean decrease of 0.0 ng/mL in the placebo control group (59). However, these results were not statistically significant but provide supporting evidence of the role that statins may play in decreasing PSA levels.
A longitudinal study consisting of 1214 men at the Durham Veterans Affairs Medical Center who were prescribed a statin between 1990 and 2006 was published by Hamilton et al. (2008) evaluating the influence of statin medications on PSA levels (60). Men were included who had underwent PSA testing within 2 years before starting a statin and 1 year after starting a statin. Subjects were excluded who received androgen or finasteride therapy, men with PC, men who received any non-statin lipid lowering agent within 3 years, men with a pre-statin PSA level greater than 10 ng/mL, or men who recently underwent any urologic surgical procedure. After starting a statin, the median low density lipid decline was 27.5% with a mean PSA level decline of 4.1%. The median time from starting a statin to the final PSA level taken was 214 days. Interestingly, the PSA decline after statin initiation was positively associated with low density lipid cholesterol decline; for every 10% low density lipid cholesterol decline, the PSA level responded by a decline of 1.64% (60). It is important to note that this study found that men whose low density lipid cholesterol levels did not change after statin initiation experienced a rise in median PSA levels. Statin doses equal to and greater than 20 mg of simvastatin were associated with a greater than 8.5% greater decline in PSA levels and among men with a PSA level higher than 4.0 ng/mL, a decline of 12.5% was observed (60).
Stamatiou et al. (2008) investigated the effects of lovastatin on conventional medical treatment of lower urinary tract symptoms (61). Men older than 50 years of age with a total prostate volume greater than 40 mL and a serum PSA higher than 1.5 ng/mL were recruited. Patients were divided into 2 groups based on a diagnosis of lipidemia. Eighteen eligible patients with BPH with lipidemia were prescribed lovastatin 80 mg daily and finasteride 5 mg daily, compared to 15 patients without lipidemia who were prescribed 5 mg of finasteride only. After four months of follow-up, men with lipidemia experience a statistically significant 0.98 ng/mL decrease in mean PSA from baseline. Men without lipidemia were also observed to have a decrease in PSA of 0.72 ng/mL, but this difference was not statistically significant (61). The investigators concluded that the latter result may have been observed due to the low power and duration of the study.
In 2009, Mener et al. used the computerized medical records at the University of Rochester Medical Center to identify 962 men, who were prescribed a statin medication and underwent PSA testing within 2 years before and 1 year after starting statin therapy (62). Patients were excluded who were received androgen or finasteride therapy, men with PC, BPH, and men who underwent any recent urologic surgical procedure. Statin medication dosages were translated into dose equivalents using 20 mg of simvastatin as a reference value equal to 1. The mean time between initiation of statin therapy and post-statin PSA assay measurement was 174 days, with a mean time between pre and post statin PSA measurement of 1.91 years. The mean PSA decline was 8.04%. Patients prescribed atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, and simvastatin were observed to have PSA declines of 0.22 ng/mL, 0.27 ng/mL, 0.383 ng/mL, 0.316 ng/mL, 0.286 ng/mL, and 0.316 ng/mL, respectively. However, fluvastatin and lovastatin did not reach statistical significance in regards to PSA level decline in men (62). Supporting these observations that lovastatin produced the greatest decline in PSA levels, cellular growth arrest and apoptosis has been demonstrated in vitro at concentrations ranging from 0.1 to 100 μM (63). Most surprising, a statin dose equivalent of 1.0 yielded the maximal decline of PSA levels in men (62). In contrast to these clinical observations, low doses of atorvastatin in the range of 0.01–0.1 μM that correlate well with the serum concentration of patients on chronic therapy promote angiogenesis and cellular growth through Akt dependent activation in contrast to doses greater than 0.1 μM (64). The reason for the discrepancy from the basic science laboratory findings to the observations in a clinical population are unknown, but it is clear that evaluating the differing pharmacologic nature of individual statins is crucial, since the lipophilic properties and cellular growth effects of each medication can vary widely.
Preoperative statin use on total preoperative PSA level and risk of biochemical recurrence in patients with PC presenting for radical prostatectomy was recently studied (65). A retrospective review of 1,031 men using statins were compared to 2,797 men not using statins, who were undergoing radical prostatectomy from 2001 to 2008. In men who used statin medications, the median serum PSA was 4.7% lower compared to men, who did not use cholesterol lowering agents. This observation was supported by a study examining 1,351 men undergoing radical prostatectomy from 2003 to 2009, evaluating preoperative serum PSA levels, statin use, tumor volume, and PC aggressiveness. Men who used statins had a statistically significant decrease of 0.2 ng/mL in PSA level (44).
Most recently, Murtola et al. (2010) assessed the relative risk of overall PC in a study cohort of 23,320 men participating in the Finnish Prostate Cancer Screening Trial from 1996 to 2004 (66). Men aged 55–67, who did not have PC were randomized to the PC screening arm of the trial, which included serial PSA measurements; men who consistently used statins were compared to those not using statins. While not a statistically significant difference, men aged 60–72, who used statin medications were observed to have a slightly lower overall PSA level compared to men who did not use cholesterol lowering agents. Interestingly, in a subgroup of men who were later diagnosed with PC, the pre-diagnostic PSA was lower at 5.64 ng/mL among current statin users compared with 5.76 ng/mL in men who did not use cholesterol lowering agents. This study attempted to show that the inverse association between statin use and PC incidence is not solely dependent on the observation that statins lower PSA in men. The authors suggested that lowering serum cholesterol in men may in fact explain the observation. It is important to note that current statin users who had a history of 6 months without medication purchases were reclassified to non-users until their next purchase, which may bias PSA level comparisons between the two groups towards the null. Furthermore, 19.8% of men in the cohort used statin medications for less than 1 year, which may have been an insufficient amount of time to observe a substantial decrease in PSA (66).
CONCLUSIONS AND FUTURE PROSPECTS
There is scientific and clinical evidence supporting the hypothesis that statins are associated with an overall reduction in serum PSA in men, when used for greater than 6 months, and especially for greater than 2 years in men. The basic science evidence discussed provides plausible mechanisms of actions elucidating the clinical observations that HMG-CoA reductase inhibitors decrease PSA levels in men, most likely from inhibiting prostatic cellular growth. The public health implications are significant given the prevalence of PSA testing in the United States. Not all of the variance in PSA levels in men is accounted for by prostatic volume and age, which only accounts for 30% and 5% of the variance, respectively (4). Use of statins in men may in fact account for part of the observed variance, but this hypothesis needs to be further investigated. Taking into the account, the change in follow-up and plan for biopsy in men who exceed the threshold for PSA levels, it is important to determine the overall effect of statins on PSA levels. Further investigation is needed to determine if the threshold level of PSA necessitating prostate biopsy should be lowered for men taking statins.