Polyamine metabolism in prostate cancer

Abstract

Purpose of review:

Normal and malignant prostate engage in high rates of de novo polyamine synthesis. This review considers how polyamine metabolism regulates prostate cancer initiation and progression.

Recent findings:

The androgen receptor (AR) establishes a metabolic program to drive robust polyamine synthesis in the normal prostate. Upon malignant transformation, this AR-driven metabolic program persists and is optimized for oncogenesis by the proto-oncogene MYC and/or alterations to PI3K signaling. A deeper understanding of the function of polyamines in prostate cancer may be obtained by considering their function in the normal prostate.

Summary:

Recent findings support ongoing research into the role of polyamines in driving prostate cancer initiation and progression and suggest targeting polyamine metabolism remains a promising therapeutic strategy for prevention and treatment of prostate cancer.

Introduction

An oncogenic metabolic program is essential for cancer cell fitness including the abilities to survive, proliferate, invade, metastasize, and evade the immune system. Increasing data suggest that a common aspect of oncogenic metabolism is the maintenance of intratumoral polyamine abundance. Polyamines are ubiquitous positively charged metabolites that serve important signaling functions in mammalian cells. Mechanisms to maintain intratumoral polyamines and other aspects of oncogenic metabolism are heavily influenced by the tissue of origin of the cancer. This is because the tissue of origin shapes both the initial microenvironmental metabolic demands on the cancer cell, as well as access to signaling pathways to meet these demands. Interestingly, the human prostate boasts both the highest abundance of polyamines and the highest rate of malignant transformation of all tissues in men. This review will consider whether these phenomena are linked and how polyamine metabolism and function in prostate cancer is shaped by pre-existing pathways of the normal prostate.

Polyamine metabolism

The primary polyamines in mammalian cells are putrescine, spermidine, and spermine. They are ubiquitous, with intracellular concentrations in the millimolar range, although free concentrations are low due to their binding to cellular anions. The reader should refer to several excellent prior reviews for general background on the metabolism and function of polyamines [1–3]. Briefly, polyamines can be synthesized de novo in the cell or, in some contexts, taken up from outside the cell. Pathways of mammalian synthesis and catabolism of polyamines are summarized in Figure 1. Two reactions to note are the decarboxylation of ornithine to generate putrescine by ornithine decarboxylase (ODC; encoded by ODC1) and the decarboxylation of S-adenosylmethionine (SAM) to generate decarboxylated SAM by S-adenosylmethionine decarboxylase 1 (AMD1, encoded by AMD1), as these are key nodes of regulation that connect polyamine synthesis to the urea cycle and one-carbon metabolism, respectively.

Figure 1. Pathways of mammalian polyamine synthesis and catabolism.

Ornithine decarboxylase (ODC) and S-adenosylmethionine decarboxylase 1 (AMD1) are highlighted as key regulatory nodes of polyamine synthesis that connect polyamine synthesis to the urea and methionine cycles. Ornithine, the substrate for ODC, can be generated from the amino acids arginine, proline, or glutamine. Arginine is converted to ornithine by arginase (ARG1/2). Proline and glutamine generate glutamate-1-semialdehyde (GSA), which can be converted to ornithine by ornithine aminotransferase (OAT). The polyamines putrescine, spermidine, and spermine are generated downstream of ODC. ODC decarboxylates ornithine to create putrescine. Spermidine is made from putrescine by spermidine synthase (SRM) and spermine is made from spermidine by spermine synthase (SMS). Both SRM and SMS utilize decarboxylated S-adenosylmethionine (dcSAM), which is generated by AMD1 from S-adenosylmethionine (SAM). Methylthioadenosine (MTA) is the product resulting from removal of an aminopropyl group from SAM by SRM or SMS. Spermine can be converted back to spermidine by spermine oxidase (SMOX). Polyamine excretion from the cell is enhanced by their acetylation by spermidine/spermine acetyltransferase (SSAT). Acetylated spermine and spermidine can be converted to spermidine and putrescine, respectively, by the peroxisomal enzyme polyamine oxidase (PAOX).

The molecular functions assigned to polyamines are prolific. These include binding and altering structure and function of nucleic acids, driving protein translation by activating the initiation and elongation factor eIF5A, regulating ion channel function, and altering redox potential [1]. Arruabarrena-Aristorena et al. describe polyamines as the “oil” that facilitates running of cellular machinery [3]. Polyamine synthesis seems to be essential for life as homozygous deletion of ODC1 or AMD1 is embryonic lethal in mice [4,5]. Disruption of this pathway in humans either by loss-of-function alterations to SMS (i.e., Snyder-Robinson Syndrome) or gain-of-function alterations to ODC1 (i.e., Bachmann-Bupp Syndrome) produces profound neurodevelopmental phenotypes [6–9]. General cellular processes supported by polyamine-mediated signaling include cellular growth, proliferation, maintenance of an undifferentiated epigenetic cell state, and immune suppression [1–3,10,11]. As such, polyamines support many of the hallmarks of cancer, and their biosynthesis is known to be maintained in varied cancer types, including prostate cancer. Inhibition of polyamine synthesis with the ODC inhibitor, difluoromethylornithine (DFMO, also known as eflornithine), was recently shown to prolong survival of children with high-risk neuroblastoma when used as maintenance therapy and received US Food and Drug Administration approval for this indication in 2023 [12], providing proof-of-principle that polyamine synthesis can drive cancer progression in humans. Given that oncogenic metabolic programs are heavily influenced by established pathways of the tissue of cancer origin, we will begin with a discussion of polyamine metabolism in normal prostate.

Polyamines in the normal prostate

The prostate is a small gland that sits at the base of the bladder and contributes secretions to seminal plasma during ejaculation. It exhibits the highest rate of malignant transformation of any tissue in men, excluding skin. It is a mysterious organ because the functions of its secretions are poorly defined, and it is unclear why it is so susceptible to cancer. The prostate is not essential for life and is also not essential for reproduction, as reproduction can be achieved by intrauterine insemination of epididymal spermatozoa without any exposure to seminal plasma. Notably, however, exposure to seminal plasma in women undergoing in vitro fertilization treatment increases the rate of pregnancy [13]. This suggests that seminal plasma increases reproductive success.

In humans, prostate gland secretions comprise approximately 25% of the ejaculate and are composed of the serine protease prostate specific antigen, prostatic acid phosphatase, zinc, citric acid, lipids, cholesterol, immunoglobulins, and polyamines [14]. Of the polyamines, spermine has the highest concentration in seminal plasma of 0.5–3.5 mg/ml, which is about 50,000-fold higher than serum. Notably, spermine was first discovered in seminal plasma and named after this association with spermatozoa. It was first described as a crystalline substance in semen by Antoni van Leewenhoek in 1677, then rediscovered by multiple scientists including Vanquelin in 1791, Charcot in 1853, Leyden in 1872, and Boettcher in 1865 [15]. In 1941, Hamalainen published an autopsy survey of spermine concentration in various human tissues obtained at autopsy of 69 people with trauma as the most common cause of death. This suggested that the prostate has the highest concentration of spermine of any measured tissue with a median concentration 115.5 mg/100g, with pancreas ranked as a distant second with a median concentration of 26.7 mg/100g [16,17].

The high concentration of spermine in prostate and prostatic secretions begs the question of what are the functions of prostatic polyamines? Given that seminal plasma at large is not required for reproduction, prostatic polyamines are also not a requirement for reproduction. Further, other animals including the dog, bull, rabbit, and mouse produce very little prostatic polyamines and have no known consequent reproductive disadvantage compared with humans or rats that produce high quantities of prostatic polyamines [18]. Nonetheless, prostatic polyamines probably enhance the reproductive success of humans. Prostatic polyamines in seminal fluid come in contact with many cell types during coitus. Some studies have suggested that spermine enhances spermatozoa function by increasing spermatozoa motility [19] and preventing premature acrosomal exocytosis [20]. Others have suggested that prostatic polyamines regulate transglutaminase activity and seminal plasma clotting to enhance transit to the female reproductive tract [21,22]. Notably, a key function of seminal fluid is to facilitate induction of maternal immune tolerance to paternal antigens to enable growth and development of a fetus that will express these paternal antigens [23]. Components of seminal vesicle secretions including TGF-β and prostaglandins are well known to participate in this skewing of the female immune response to semen [23,24], however whether prostatic secretions, and in particular prostatic polyamines, also contribute is unknown. Yet in other contexts of infection and cancer, polyamines can drive differentiation and survival of immune-suppressive macrophages and restrain Th1 differentiation of T cells [25–32]. Therefore, a distinct possibility is that prostatic polyamines facilitate maternal immune tolerance to paternal antigens expressed by spermatozoa. Lastly, a notable aspect of semen biochemistry is the high concentration of diamine oxidase [33], which catabolizes spermine to hydrogen peroxide and toxic aldehydes such as acrolein. Human spermatozoa can detoxify acrolein due to expression of the aldo-keto reductase AKR1B7 [34], however acrolein is toxic to most other cells including microbes. Thus, another possibility is that spermine, through its conversion to acrolein, may create a cytotoxic chemical barrier to cellular attack or competition within the female reproductive tract.

Though the function of prostatic polyamines is not well defined, their regulation is better understood. The major regulator of polyamine synthesis in prostate is the androgen receptor (AR). The AR is a nuclear hormone receptor that is activated by androgens and exerts downstream effects through regulation of transcription. Inhibition of the AR by androgen deprivation leads to apoptosis of prostate luminal epithelial cells and massive shrinkage of the prostate, illustrating its function as a survival factor for these cells [35]. At 6 hours, prior to initiation of cell death, androgen deprivation decreases ODC and AMD1 activity and polyamine abundance in rat ventral prostate, which is reversible by administration of androgens [36,37]. Androgen administration to rodents was subsequently shown to acutely increase ODC1 mRNA, and prolonged administration of androgens increased the half-life of ODC protein from 15 minutes to 100–150 minutes [38–40]. Our recent studies indicate that AR binds two enhancer sites approximately 4 and 10 kb upstream of the ODC1 promoter to directly regulate ODC1 transcription in normal prostate [38]. The induction of ODC by AR was shown to be required for normal prostate development [41]. Therefore, the AR is also described to be a differentiation factor, in addition to a survival factor, facilitating prostatic development and function, including biosynthesis of polyamines.

Polyamines in prostate cancer initiation

Given that the prostate is associated with both the highest rate of cancer formation and the highest abundance of polyamines of all tissues in men, one may question whether these phenomena are linked. Some data suggest that enhanced ODC activity or increased putrescine is required and sufficient for prostate cancer development. Shukla-Dave et al. observed that putrescine treatment of the immortalized epithelial prostate cell line RPWE-1 increased proliferation and invasion, and overexpression of ODC in these cells was sufficient to transform them, enabling tumor formation upon injection into immunocompromised mice [42]. Young et al. attempted the definitive experiment to assess sufficiency of ODC in driving prostate cancer initiation in the absence of prior immortalization through generation of mice with prostate-specific overexpression of ODC, however technical challenges to overexpressing ODC protein precluded a definitive answer [43]. Notably, humans with a polymorphism in the ODC1 gene that increases ODC1 transcription and fewer CAG repeats in exon 1 of the AR gene may have a higher risk of developing prostate cancer [44], supporting an idea that high ODC activity may drive prostate cancer initiation. ODC activity is robust in the normal prostate, and further increases in local prostate cancer [45]. In the TRAMP (TRansgenic Adenocarcinoma of Mouse Prostate) murine model, in which the SV40 early genes are under regulation by the AR-regulated probasin promoter, inhibition of ODC with DFMO markedly reduced the incidence and severity of prostate cancer development [46]. On the basis of these data, some studies have been designed to explore the potential of DFMO as a prostate cancer chemopreventive agent in humans [47,48]. These studies confirmed that DFMO at a dose of 500 mg/m² by mouth daily or at a flat dose of 500 mg daily reduced polyamine abundance in prostate tissue without significant toxicities but were not powered to assess its ability to reduce the incidence of prostate cancer. Nonetheless, preclinical data support an idea that the AR-driven program of polyamine biosynthesis in normal prostate may predispose this tissue to development of cancer and provide a rationale for further clinical investigation of polyamine synthesis inhibition for prostate cancer prevention.

When assessing alterations to the metabolome of prostate cancer versus adjacent normal prostate, a notable change is a decrease in spermine (Fig 2A) [49–51]. While Sanford et al. initially reported in 1975 that urine polyamines, including spermine, are higher in people with prostate cancer [52], studies with larger numbers and more sophisticated methods indicate that spermine in urine is also reduced in people with prostate cancer compared with healthy controls [53–55]. Tumor spermine abundance has a negative correlation with tumor grade and appears lower in metastases than local prostate tumors [49,50]. This decrease in spermine may be due to its diversion toward spermidine and activation of the translation factor eIF5A through SMOX. SMOX expression is markedly higher in prostatic intraepithelial neoplasia (PIN) and prostate cancer compared to normal prostate [56]. This increase in SMOX occurs concurrent with an increase in MYC activity [57], a well-known early driver of prostate cancer. Interestingly, MYC was recently shown to drive lymphoma initiation through regulation of spermidine flux to eIF5A activation [58]. eIF5A can function as an initiation and elongation factor, and the eIF5A regulon appears to be composed of diverse protein products that enable oncogenic behavior including proliferation, angiogenesis, invasion and metastasis [59]. Interestingly, a comparison of the transcriptome of prostate cancer versus adjacent normal prostate [60] shows that the increase in tumoral MYC directly correlates with an increase in SRM, SMOX, and DHPS (Fig 2B), which encode enzymes that drive spermidine flux to hypusination and activation of eIF5A [60]. Therefore, it is possible that spermine is reduced in prostate cancer due to activation of pathways leading to eIF5A activation, which may be in part mediated by MYC (Fig 2C). In addition to activation of MYC, activation of PI3K through loss of PTEN function is a common early event in prostate cancer. PI3K signaling increases stability and activity of AMD1 [61], which may ensure dcSAM substrate is not limiting for SRM and spermidine synthesis in prostate cancer initiation. Furthermore, Li et al recently suggested that spermine can actually be a tumor suppressive metabolite in prostate cancer through inhibition of AR signaling [62]. However, it was not clear in this study whether tumor suppression was due to increased spermine or rather perhaps alterations to linked metabolites such as spermidine. Nonetheless, given that the AR regulates spermine synthesis through regulation of ODC1 and SMS [62], the presence of feedback-inhibition on AR is an attractive idea to regulate physiologic polyamine synthesis of the normal prostate and may impinge on pathways of oncogenesis. In sum, these data suggest a model whereby the AR programs polyamine synthesis in the normal prostate that persists and is optimized for oncogenesis by activation of MYC and PI3K (Fig 2C).

Figure 2. Alterations to polyamine metabolic flux upon prostate cancer development.

A. Arginine and spermine abundance is lower in local prostate cancer (PCa) compared with adjacent normal prostate tissue, while ornithine, putrescine, and spermidine are unchanged. Data reanalyzed from Priolo et al. [51] with N=25 normal samples and N=60 prostate cancer samples and p values by unpaired two-tailed t test. B. SRM, SMOX, DHPS, and MYC transcript abundance increase in local PCa compared with adjacent normal prostate tissue. The change in SRM, SMOX, and DHPS directly correlate with the change in MYC. Data reanalyzed from Taylor et al. [60] with N=122 prostate cancer samples with matched normal adjacent tissue samples and r and p values determined by Pearson’s correlation calculation. C. Schematic of possible alterations to polyamine metabolic flux upon PCa development. The enzymes ARG2, ODC, AMD1, and SMS are positively regulated by the androgen receptor (AR) in both normal and malignant prostate. Early events in prostate cancer initiation are activation of MYC and PI3K. PI3K activation reinforces flux through AMD1, while MYC enhances flux from spermidine toward activation of eIF5A through deoxyhypusine synthase (DHPS). MYC positively regulates SRM and may regulate SMOX to enhance flux from putrescine and spermine, respectively, to spermidine. The net effect is reduction of arginine and spermine in PCa compared with adjacent normal. Spermine exerts negative feed-back on AMD1, such that its depletion can increase AMD1 activity. Some data suggest that spermine is a tumor suppressive metabolite such that its depletion enhances the process of prostate cancer initiation.

Polyamines in prostate cancer progression

Although prostate cancer is a common disease, it is generally only lethal after development of metastatic disease. Thus, it is also important to consider whether polyamine metabolism drives prostate cancer progression after initiation. Numerous studies have assigned oncogenic functions to polyamines and polyamine metabolism in prostate cancer and other cancer types. Broadly, polyamines can play direct and indirect roles in cancer progression by affecting cancer cell signaling or the tumor microenvironment (TME) [63]. Direct cancer cell-intrinsic oncogenic roles for polyamines include binding and stabilizing nucleic acids and hypusination of eIF5A to drive translation of oncogenic proteins such as MYC [64,65]. Direct TME effects include promotion of immune-suppressive tumor-promoting myeloid cells including M2 macrophages and myeloid-derived suppressor cells [25–31] and angiogenic endothelial cells [66]. Although polyamines are required for T cell proliferation, dampening polyamine synthesis and consequent eIF5A activity leads to Th1 differentiation and increased production of IFNγ [32]. Elevated polyamine metabolism in cancer can also have indirect effects due to substrate depletion. Most notable is the effect of polyamine metabolism on one-carbon metabolism through the use of SAM by AMD1, which can alter nucleotide metabolism and epigenetic regulation of gene transcription through DNA and histone methylation. In fact, in models of colon cancer, the mechanism by which DFMO inhibited cell growth was found to be through alterations of SAM- and folate-associated one-carbon metabolism with marked depletion of thymidine, such that DFMO-induced cell cycle arrest could be rescued by supplementation with thymidine [67]. Prostate cancer has been suggested to be highly dependent on methionine salvage from MTA by methylthioadenosine phosphorylase (MTAP) due to high flux through AMD1, creating a vulnerability to MTAP inhibition [68] that can be augmented by increasing SSAT activity by the polyamine analogue N¹-N¹¹-bisethynorspermine (BENSpm), essentially creating a sink for methyl donors [69]. Enhanced polyamine synthesis and catabolism not only can reduce SAM pools for methylation reactions, but also can reduce acetyl coenzyme A pools for acetylation reactions. This is illustrated nicely by SSAT overexpression in the TRAMP mouse model, which resulted in 18-fold increase in SSAT activity and associated 70% reduction of acetyl-co A abundance and 40% reduction of SAM, and overall leading to markedly decreased tumor growth [70]. Another indirect effect of polyamine synthesis may be arginine depletion. Tumors seem to be deplete of arginine relative to the bloodstream [71], with serum levels in patients with metastatic prostate cancer lower than healthy controls [71]. Arginine depletion has been attributed to high activity of ARG1 of tumor-infiltrating suppressive myeloid cells [72], and which may also be augmented by utilization of arginine by cancer cells for polyamine synthesis. This is notable because arginine depletion has been described to decrease T cell survival and antitumor cytotoxic function [73], implying that increasing arginine in the tumor microenvironment by polyamine synthesis inhibition may enhance tumor immunity.

The question at the crux of these considerations is whether targeting polyamine metabolism is a viable therapeutic strategy for patients with advanced, life-threatening, prostate cancer. Herr et al. considered this question in the early 1980s [74]. Treating established tumors in the Dunning R-3327 rat prostate adenocarcinoma model, they observed that inhibition of ODC with DFMO (2% in the drinking water) was well tolerated but did not slow tumor growth. In contrast, inhibition of AMD1 with the spermidine analogue methylglyoxyalbisguanylhydrazone (MGBG) (15 or 40 mg/kg IP daily for 21 days) slowed tumor growth but was toxic. Notably, the combination of DFMO and MGBG was actually less toxic than MGBG monotherapy and provided even more effective control of tumor growth. The mechanism governing both reduced toxicity and increased efficacy was not elucidated, but their experiment provided rationale for a phase I clinical trial testing safety of DFMO in combination with MGBG for patients with advanced solid tumors or lymphomas. A phase I trial of DFMO monotherapy had determined that DFMO up to a dose of 2.25 g/m² PO QID was safe for this patient population; no objective responses were observed but no patients with prostate cancer were included [75]. Unfortunately, the combination of DFMO 1 g/m2 PO QID with a dose escalation of MGBG 200–700 mg/m2 on day 4 and every 2 weeks thereafter proved to be quite toxic with significant myelosuppression, fatigue, and GI toxicity [76], halting further clinical investigation into this drug combination. While no objective responses were observed, detailed study of the 5 patients on this study with prostate cancer showed that 3 of these 5 patients experienced significant pain relief on trial [77]. Scher et al subsequently simplified the regimen and tested efficacy of MGBG monotherapy 500 or 600 mg/m² IV weekly for patients with metastatic CRPC. The treatment was better tolerated, albeit with toxicities of weakness and mild myelosuppression, and notably 6 of 25 evaluable patients (24%) experienced objective responses although they were short-lived [78]. Notably, to date, there has not been a clinical trial testing efficacy of DFMO monotherapy for patients with advanced prostate cancer. This is despite some pre-clinical studies that have shown, contrary to Herr et al.’s initial study discussed above, that DFMO monotherapy can reduce the rate of growth of established prostate cancer tumors, including castration-resistant tumors, in animal models [79]. Clinical studies that target polyamine metabolism relevant to patients with prostate cancer are summarized in Table 1.

Table 1.

Clinical studies that target polyamine metabolism relevant to patients with prostate cancer.

Patient Population (N)	Design and Treatment	Results	Reference
Advanced solid tumor or lymphoma (22)	Phase I: Single-arm dose-escalation DFMO 0.75–3.0 g/m2 PO QID	Recommended dose for phase II (MTD): 2.25 g/m2 PO QID. No objective responses. Dose-limiting toxicities thrombocytopenia and GI toxicity.	Abeloff et al. 1984 [75]
Advanced solid tumor or lymphoma (22 total; 5 with PCa)	Phase I: Single arm DFMO 1.0 g/m2 PO QID with dose-escalation MGBG 200–700 mg/m2 on day 4 and every 2 weeks thereafter	No recommended dose for phase II. 1 objective response. 3 of 5 patients with PCa had pain relief. Response evaluation limited by toxicity of myelosuppression, fatigue, nausea/vomiting, diarrhea.	Warrell et al. 1983; Herr et al. 1986 [76,77]
Treated colon, bladder or prostate cancer (38)	Phase I: Single arm dose-escalation DFMO 0.125–0.75 g/m2 PO QID	DFMO 0.5 g/m2 PO daily caused minimal toxicity and reduced TPA-induced ODC activity in skin by 57%.	Love et al. 1993 [88]
Local prostate or bladder cancer (25)	Pharmacology: Randomized 3:2 DFMO 0.5 g/m2 PO daily or no treatment for 14 days before surgery	DFMO 0.5 g/m2 PO daily for 14 days decreased prostate putrescine by 27%.	Messing et al. 1999 [89]
Advanced solid tumor or lymphoma (39)	Phase I: Single-arm dose-escalation SAM486A 3 mg/m2 −400 mg/m2 IV over 120h every 4 weeks	Recommended dose for phase II (MTD): 400 mg/m2 IV over 120h every 4 weeks. No objective responses. Dose-limiting toxicity of neutropenia.	Paridaens et al. 2000 [90]
Prostate atypia (9)	Pharmacology: Prostate biopsies before and after 28 days of DFMO 0.5 g/m2 PO daily	DFMO 0.5 g/m2 PO daily for 28 days decreased prostate polyamines (putrescine −97.6%, spermidine −73.6%, spermine −55%).	Simoneau et al. 2001 [47]
Advanced solid tumors (46)	Phase I: Single arm dose-escalation PG-11047 50–750 mg IV on day 1, 8, 15 of a 28-day cycle	Recommended dose for phase II (MTD): 610 mg IV on day 1, 8, 15 of a 28-day cycle. No objective responses. Dose-limiting toxicities mucositis and GI toxicity.	Stewart et al. 2020 [91]
Family history of PCa (66)	Phase II: Randomized 1:1 DFMO 0.5 g PO daily or placebo for 12 months	DFMO decreased prostate putrescine (−60.8%) and prostate volume. No change in spermidine or spermine. Underpowered to assess efficacy.	Simoneau et al. 2008 [48]
CRPC (35)	Phase II: Single arm MGBG 500 or 600 mg/m2 IV weekly without ADT	6/25 (24%) ORR. No responses in bone. Toxicities include weakness and mild myelosuppression.	Scher et al. 1985 [78]

Currently the most important therapeutic target for advanced prostate cancer is the AR. The backbone of systemic therapies for metastatic disease is suppression of androgen production by the testes and therapeutic effect can be enhanced by therapy intensification with concurrent suppression of adrenal androgen production or direct AR inhibition [80]. When prostate cancer acquires resistance to AR inhibition, it often maintains signaling through the AR by varied mechanisms including AR amplification, overexpression, and expression of ligand-independent splice variants and mutant proteins. We have demonstrated that the marked enhanced activity of AR that occurs in castration-resistant disease can constitute a vulnerability to exposure to high doses of androgens, clinically called Bipolar Androgen Therapy, which results in growth-inhibition in this context [81–83]. To better understand this nuanced AR function in advanced castration-resistant prostate cancer, we have been studying molecular consequences of AR activation in this advanced disease setting. Notably, we observed that polyamine abundance is markedly increased by treatment with androgens in models of castration-resistant prostate cancer and serves to facilitate resistance to androgen treatment [81,82]. This occurs by AR binding upstream and driving expression of ODC1, a mechanism that is conserved from normal prostate. This indicates that AR maintains regulatory control of polyamine synthesis throughout all stages of normal prostate to advanced prostate cancer. We also observed that MYC paradoxically can be a negative regulator of ODC and AMD1 in this context due to transcriptional interference with the AR. Importantly, the extent to which efficacy of AR inhibition is driven by loss of positive regulation of polyamine synthesis across the disease spectrum is not known. Along the same lines, the extent to which inhibition of polyamine synthesis with agents such as DFMO or MGBG is redundant with AR inhibition, limiting their efficacy in this context, is also not known, but should be considered as these and similar agents are clinically developed.

Although advanced prostate cancer seems to actively engage in de novo synthesis of polyamines, it may not be dependent on this pathway for progression. This is because polyamines derived from diet or the bacterial microbiome may be supplied to the tumor in the tumor microenvironment and taken up by the cancer cell. Foods such as wheat germ, corn, and grapefruit or orange juice contain high amounts of polyamines and low-polyamine diets have been designed excluding such foods [84,85]. A small single-arm study suggested that a low-polyamine diet may extend overall survival for patients with CRPC [86]. Notably, in vitro, the growth-suppressive effects of DFMO on prostate cancer cells can be completely reversed by supplementing the culture media with putrescine [86] and some studies have suggested that DFMO has increased anti-tumor efficacy in vivo when combined with a low polyamine diet [87]. Whether bacterial microbiome synthesis of polyamines can compensate when both de novo synthesis and dietary polyamines are limited is unknown.

Conclusions

Upon review of polyamine metabolism and function in normal and malignant prostate, more similarities than differences are noted in these two states. In the normal prostate, polyamines seem to enhance the viability, function, and immune evasion of spermatozoa. In prostate cancer, polyamines continue these functions, but now also enhance cancer cell, rather than spermatozoa, fitness. Moreover, the regulatory control of polyamine metabolism by the AR is a feature of normal prostate that seems to be persistent through most stages of prostate cancer even through late-stage castration-resistant disease. Thus, robust polyamine synthesis may be a defining feature of both the normal and malignant prostate that is poised as a target for therapeutic intervention.

Summary Points.

1. Polyamines are polycations that support several hallmarks of cancer including cellular growth, proliferation, maintenance of an undifferentiated epigenetic state, and immune evasion.

2. The androgen receptor (AR) drives a high rate of polyamine synthesis in normal prostate, which may enhance spermatozoa function and evasion from the maternal immune system but also increase susceptibility to cancer.

3. The AR continues to drive polyamine metabolism in prostate cancer, which is optimized for oncogenesis through the actions of MYC and PI3K.

4. Targeting polyamine metabolism is a promising strategy to prevent and treat prostate cancer that requires further research.