Molecular Basis of Prostate Cancer and Natural Products as Potential Chemotherapeutic and Chemopreventive Agents

Prostate cancer is the second most common malignant cancer in males. It involves a complex process driven by diverse molecular pathways that closely related to the survival, apoptosis, metabolic and metastatic characteristics of aggressive cancer. Prostate cancer can be categorized into androgen dependent prostate cancer and castration-resistant prostate cancer and cure remains elusive due to the developed resistance of the disease. Natural compounds represent an extraordinary resource of structural scaffolds with high diversity that can offer promising chemical agents for making prostate cancer less devastating and curable. Herein, those natural compounds of different origins and structures with potential cytotoxicity and/or in vivo anti-tumor activities against prostate cancer are critically reviewed and summarized according to the cellular signaling pathways they interfere. Moreover, the anti-prostate cancer efficacy of many nutrients, medicinal plant extracts and Chinese medical formulations were presented, and the future prospects for the application of these compounds and extracts were discussed. Although the failure of conventional chemotherapy as well as involved serious side effects makes natural products ideal candidates for the treatment of prostate cancer, more investigations of preclinical and even clinical studies are necessary to make use of these medical substances reasonably. Therefore, the elucidation of structure-activity relationship and precise mechanism of action, identification of novel potential molecular targets, and optimization of drug combination are essential in natural medicine research and development.


INTRODUCTION
It is nowadays evident that prostate cancer (PCa) is recognized as the second most common cancer and the fifth cause of cancer death in males (Torre et al., 2015)Initially, the usual therapy for prostate cancer is prostatectomy or radiation, which aims to remove or kill the malignant cells that have not spread or metastasized (Feldman and Feldman, 2001). However, numerous patients cannot be cured thoroughly by this treatment, and then followed by cancer recurrence and/or metastasis. The majority of prostate cancer growth is androgen dependent. Androgen deprivation therapy (ADT) such as surgery or gonadotropin-releasing hormone (GnRH) analog treatment, is the main therapeutic and dramatically effective intervention for the treatment of androgen dependent prostate cancer (ADPC) in putting patients with tumors in remission, as documented by the work on castration of Huggins, who was awarded the Nobel prize in 1966 (Huggins, 1978). Nevertheless, after this therapy, most of these prostate cancer patients gradually become androgen independent, go on to progress, metastasize and resist to ADT within 13-24 months accompanied by increased levels of prostate-specific antigen (PSA). Siegel et al. reported that failure of ADT is responsible for the ∼27,000 metastatic prostate cancer deaths in the United States annually (Siegel et al., 2017). This stage of prostate cancer is called castrationresistant prostate cancer (CRPC), which has poor prognosis (Small et al., 2004). At present, there is no effective therapy for CRPC besides docetaxel, which has been demonstrated to prolong overall survival in this patient population. However, the efficacy of docetaxel is not satisfactory and there are many severe adverse effects such as anemia, neutropenia, diarrhea, and sensory neuropathy. Although, therapeutic options have expanded rapidly since 2011, including AR inhibitors (enzalutamide, abiraterone), immunotherapy (sipuleucel-T), bone seeking radionuclides (radium-223), and second-line chemotherapy (cabazitaxel), all of these agents or interventions only have shown a median survival benefit of 2-5 months (Ritch and Cookson, 2016). So searching for more effective anti-prostate cancer drugs, especially with high efficacy and low toxicity, remains an urgent problem that needs to be resolved. Natural compounds represent an irreplaceable resource of structural scaffolds that can offer chemical agents for making prostate cancer less devastating and curable. In recent years, many natural products and extracts have been scientifically investigated in vitro and/or in vivo and proved as potential anti-prostate cancer agents, which are currently scattered across various publications. So a systematic summary and knowledge of future prospects are necessary to facilitate further chemical and pharmacological studies for anti-prostate cancer agents.
Herein, we reviewed the detailed molecular causes of prostate cancer and critically summarized the natural compounds (or extracts and Chinese herb preparations) that have been reported to inhibit prostate cancer cells proliferation/tumor growth, induce prostate cancer cells apoptosis or exhibit effects on specific signaling pathways involved in prostate cancer in vivo and in vitro. In addition, we also provided possible novel targets for screening natural compounds (or extracts or Chinese herb preparations) with anti-prostate cancer activity and discuss the future prospects for the application of these compounds and extracts and the novel available approaches and technological improvements that should be explored to treat prostate cancer.

OVERVIEW OF THE MOLECULAR BASIS OF PROSTATE CANCER Molecular Basis of Androgen Dependent Prostate Cancer
Androgens, principally testosterone (T) and dihydrotestosterone (DHT), are synthesized from cholesterol as the initial 27-carbon substrate via multiple enzymatic steps (Wadosky and Koochekpour, 2016). As a member of the ligand-activated nuclear hormone receptors superfamily, androgen receptor (AR) is a modular protein with four functional domains: an N-terminal regulatory domain (NTD), a DNA-binding domain (DBD), a small hinge region (H) and a ligand-binding domain (LBD) (Ho and Dehm, 2017). Upon binding to androgens, AR undergoes a conformational change, leading to nuclear translocation, phosphorylation, homodimer, and interaction with DNA (Li and Al-Azzawi, 2009). Subsequently, AR dimer binds to androgen-response elements (AREs), recruits essential cofactors and regulates the expression of androgen-regulated genes (Ho and Dehm, 2017).
The development and maintenance of the prostate is inseparable from androgen acting through the AR. Since Huggins and Hodges first demonstrated the responsiveness of prostate cancer to androgen deprivation, it has been clear that prostate cancer is dependent on androgen and AR activation for growth and survival (Huggins and Hodges, 1941). From then, hundreds of studies have demonstrated that androgen withdrawal results in initial regression of essentially all prostate cancers, albeit for a finite period, with the ultimate development of castration-resistant disease. Androgen deprivation therapy, via either orchidectomy or use of a gonadotropin-releasing hormone (GnRH) agonist has become the cornerstone of therapy in the treatment of prostate cancers. Newer agents, such as abiraterone, which block androgen synthetic pathways, have added clinical benefit in disseminated disease, demonstrating that even in "castration resistant disease" androgens may still be supporting prostate cancer growth  (Morgentaler, 2009). These data support the notion that prostate cancer, in most cases, is a hormone (androgen) sensitive disease.

Overview of the Mechanisms of Castration-Resistant Prostate Cancer
Historically, there are much debate about the mechanisms of castration resistance, which are mainly summarized as the following by recent studies: canonical AR signaling, relying on AR nuclear translocation and AR-DNA binding, and non-nuclear AR signaling which requires neither AR nuclear translocation nor AR-DNA binding (Qin and Bin, 2019;Pisano et al., 2021).

Canonical AR Signaling
The potential mechanisms of canonical AR signaling that lead to CRPC can be categorized into three parts. 1) common alterations in AR, which can lead to AR increase its sensitivity to very low levels of androgens or constitutive activation of AR signaling; 2) AR activation by androgens converted from adrenalandrogens or synthesized intratumorally via the de novo route; 3) alterations in cofactors of the AR pathway ( Figure 1).

Common Alterations in AR
One possible mechanism by which the prostate cancer becomes resistant to androgen deprivation therapy is alterations in AR, including AR amplification or overexpression, AR mutations and truncated AR lacking ligand-binding domains (ARΔLBD). Thus, these changes in AR increase its sensitivity to very low levels of androgens or lead to constitutive activation of AR signaling. Strictly speaking, this mechanism of prostate cancers is not androgen-independent and the responses still depend on AR and androgen.
AR Amplification or Overexpression. Despite low circulating androgens in the CRPC patients, one potential mechanism that would allow tumor cell proliferation is by promoting the expression of the AR itself, which increases ligand-occupied receptor content (Feldman and Feldman, 2001). Plenty of studies have shown that approximately 50% of tumors that become castration resistant after ADT have amplified the AR gene, the most frequent genetic alteration reported for CRPC tumors, whereas none of the untreated primary tumors before androgen ablation had an AR gene amplification (Robinson et al., 2015;Djusberg et al., 2017).
Numerous studies provide the simplest explanation of how increased androgen receptor expression leads to resistance to anti-androgen therapy. According to one study, a three-to-five-fold increase in receptor levels could compensate for low ligand levels and restore androgen receptor signaling in xenotransplantation models. (Chen et al., 2004). Although tumors with AR amplification have increased levels of AR, the signal for cell proliferation presumably continues to require androgen (Visakorpi et al., 1995). Maybe this can explain why tumors with castration resistance have increased sensitivity to androgens in a low androgen environment.
AR Mutations. In CRPC, the frequency of AR mutations in pretreated tumors is 5-30% (Grasso et al., 2012;Robinson et al., 2015;Kumar et al., 2016). Most mutations identified in CRPC were located in the AR-LBD. These alterations could facilitate AR signaling in CRPC by offering: 1) ligand facilitation, thereby inducing AR activation even in the presence of low or absent levels of androgens and 2) agonist properties to AR antagonists (Coutinho et al., 2016). In addition, mutations can also occur in the AR-NTD that account for about a third of all mutations described in AR. And mutations can usually cause alterations that contribute to AR transactivation, such as facilitated recruitment of co-factors and other components of the transcriptional machinery, promoted N/C interaction, increased response to DHT activation and enhanced protein stability and nuclear retention (Network C. G. A., 2015;Coutinho et al., 2016).
Truncated AR Lacking Ligand-Binding Domains (ARΔLBD). Latest RNA sequencing data from big data sets, strongly suggests that constitutively active ARΔLBD may play a role in 40-50% of patients with CRPC (Robinson et al., 2015).
Compared with hormone naïve PCa, ARΔLBDs are frequently upregulated in CRPC, and may serve as an adaptive response to therapies targeting the androgen/AR-signaling axis (Guo et al., 2009;. The recent genomic data on unique exon junctions reveals that at least 12 distinct AR-V mRNA species are detectable in primary PCa and 23 in CRPC (Abeshouse et al., 2015). However, among these variants, AR-V3/AR-V7 appears to be one of the most abundantly and ubiquitously expressed isoforms in our screening of a panel of human prostate cancer cell lines and tissues (Guo et al., 2009;Schweizer and Plymate, 2016). In addition, nonsense mutations leading to premature chain termination (Q641X, formerly Q640X) as well as enzymatic cleavage (tr-AR) were also shown to induce AR△LBDs (Haile and Sadar, 2011). F. Zengerling et al. reported that inhibition of IGF-1R resulted in a down-regulation of AR, Q641X and AR-V7 signaling in PCa cells (Zengerling et al., 2016), which suggests that IGF-1/IGF-1R axis is a modulator of the AR△LBD signaling, providing a rationale by targeting growth factor receptor for CRPC treatment.
AR Activation by DHT Synthesized Intratumorally via the de novo Route or Converted From Adrenal Androgens There are two possible ways for the initial substrates to convert to intratumoral DHT in CRPC. The first way is that androgens are synthesized intratumorally via the de novo route and the second is that androgens are converted from adrenal androgens.   DHT Intratumorally via the de novo Route. The use of cholesterol for de novo steroidogenesis requires the components of the steroidogenic machinery present in the adrenals and gonads, including steroidogenic acute regulatory proteins, CYP11A1 and CYP17A1 (Miller and Auchus, 2011), which may play important roles in prostate cancer (Locke et al., 2008). Comparisons between primary prostate cancer and CRPC demonstrate that the transcription levels of these proteins are upregulated in CRPC (Montgomery et al., 2008) and CYP17A1 protein is detectable in a subset of metastatic CRPC cases (Efstathiou et al., 2012). In contrast to steroidogenesis in the adrenals and gonads, CRPC expresses steroid-5α-reductase (SRD5A1) with obvious 5α-reductase activity (Chang et al., 2011). One of the functions of robust SRD5A enzyme activity is that any de novo steroidogenesis would likely occur through the back door pathway that bypasses the requirement for T and involves 5-reduction of a 21-carbon steroid (progesterone or 17α-hydroxyprogesterone) instead of a 19-carbon androgen (Shaw et al., 2000;Auchus, 2004). Although this biochemical pathway may be engaged in CRPC, the relatively lengthy eight enzymatic steps required for conversion from cholesterol to DHT, the abundance of adrenal precursors present in serum, and the much closer pathway proximity of adrenal precursors to DHT, together suggest that adrenal precursors serve as the major substrate pool.
DHT Converted From Adrenal Androgens. There are two possible pathways from adrenal precursor steroids to DHT (Luu-The et al., 2008;Chang and Sharifi, 2012). The canonical adrenal pathway is the route that results in T synthesis as the penultimate metabolite, which undergoes 5reduction to DHT (Scher and Sawyers, 2005;Ryan and Tindall, 2011;Hofland et al., 2012;Stein et al., 2012). This pathway is probably favored in the field because of the general notion that T must be the precursor to DHT and T is frequently detectable at concentrations greater than DHT in CRPC, as occurs with gonadal androgen physiology (Titus et al., 2005;Montgomery et al., 2008). In this pathway, DHEA is converted by 3βhydroxysteroid dehydrogenase (3βHSD) to androstenedione (AD), which is then 17-keto reduced by aldo-keto reductase 1C3 (AKR1C3) or 17βHSD3 to T, the immediate precursor to DHT. The second possible pathway is that AD, like T, a 3-keto, △4-steroid, is also a potential substrate for SRD5A (Tomkins, 1957). AD is reduced to 5α-androstanedione (5α-dione), which then becomes the immediate precursor to DHT. The 5α-dione pathway is the major pathway for the synthesis of DHT in CRPC (Chang et al., 2011).

Alterations in Cofactors of the AR Pathway
Resistance to AR-targeted therapies in CRPC was also associated with genomic and transcriptomic alterations in coregulators of the AR pathway. Expression of 50 of the ∼200 AR-associated coregulators is aberrant in clinical CRPC specimens (Liu et al., 2017). For example, a higher frequency of mutations in FOXA1, the gene encoding a pioneer factor that facilitates AR chromatin binding and transcriptional activation, was found in CRPC (12%) than in primary prostate cancer (4%) (Watson et al., 2013;Abeshouse et al., 2015). In addition, as one class of coregulators, steroid receptor coactivators (SRC-1, SRC-2 and SRC-3) play a key role in facilitating aberrant AR signaling in CRPC e. There have been studies reported that all 3 SRCs is elevated in CRPC (Taylor et al., 2010;Grasso et al., 2012;Abeshouse et al., 2015;Beltran et al., 2016;Bernemann et al., 2016). GATA2, another AR pioneer factor in the AR signaling axis, is aberrant expressed in CRPC and associated with poor outcome (He et al., 2014;Yan et al., 2014). Gupta et al. detected genomic copy number changes of circulating tumor cells from 16 patients with CRPC resistant to abiraterone or enzalutamide and revealed that multiple genes encoding AR coregulators had copy number alterations, including copy number gains BRD4 (43.75%) (Gupta et al., 2017). Moreover, changes of AR corepressors also play a key role in CRPC. For example, loss of activity of the key nuclear receptor corepressors NCOR1 and NCOR2, is prevalent

In vitro Antiproliferation Antimetasis
Inhibition of the androgen receptor element, resulting in decreased PSA velocity; inhibition signaling of insulin-like growth factor-I (IGF-I); decrease of the expression of αvβ3 and αvβ5 integrin. (Bureyko et al., 2009;Wertz, 2009;Tang et al., 2011;Holzapfel et al., 2013) In vivo Induction of apoptosis

In vivo
Induction of apoptosis 22Rv1 Ursolic Acid (62) Distributed in various plants

Non-nuclear AR Signaling
Trafficking from the nucleus into the cytoplasm, AR may have unexpected consequences because AR has known functions in the cytoplasm, which is called non-genomic signaling (Foradori et al., 2008). One of the main characteristics of non-nuclear signaling is its rapidity with which it occurs. When steroid receptors stay in the cytosol, they can undergo several protein-protein interactions within seconds to minutes after steroid stimulation, which activates a variety of signaling pathways to promote the development of CRPC ( Figure 2).

PI3K/AKT Signaling Pathway
The PI3K/AKT pathway is one of the most frequently activated signal transduction pathways in human cancer, including prostate cancer (Hoxhaj and Manning, 2020;Braglia et al., 2020). Alterations of the PI3K/AKT pathway, including altered expression, mutation, and copy number alterations, have been reported in 42% of primary prostate tumors and 100% of metastatic tumors (Taylor et al., 2010).
Androgens induce the accumulation of TORC2 complex, rapamycin insensitive chaperone of mTOR and stress activated protein kinase interacting protein 1 in the nucleus, thus stimulating TORC2 to activate Akt (Fang et al., 2012). Activated AKT can stimulate many downstream functions via Curcumin (77) Curcuma longa L.

DU145
In vitro Antiproliferation Inhibition of the expression of MT1-MMP and MMP2 proteins; inhibition of the DNA-binding ability of NICD.

DU145 PC3 22Rv1
In vitro Inhibition of migration Induction of apoptosis Activation of Chk1 and Chk2 and inhibition of Cdc25C and Cdc2 activities; blocking of the p38, JNK, PKC and PI3K/AKT signaling pathways and downregulation of NF-κB protein level; inhibition of MMP2 and MMP 9 gene expression. (Hastak et al., 2003;Kaur et al., 2009) In vivo Gossypol (80) Gossypium hirsutum L.

Src Signaling Pathway
Preclinical studies have confirmed that non receptor tyrosine kinase c-Src and Src family kinase (SFK) regulate a complex signal network, driving the development of castration-resistance prostate cancer and bone metastasis. (Cai et al., 2011). After the establishment of bone metastasis, prostate cancer cells destroy the balance of osteoclasts and osteoblasts by secreting a variety of molecules, such as growth factors and cytokines that disrupt the normal process of bone maintenance and reconstruction (Yang et al., 2001;Mundy, 2002). The balance is in favor of   (90) Silybum marianum (L.) Gaertn.

LNCaP 22Rv1
In  In vitro Inhibition of cell growth

PC-3
In vitro Antiproliferation Not investigated Li, (2020) osteoblastogenesis, which explains the usual condensing aspect of PCa-derived bone metastases. Src activity specifically affects ruffled borders of osteoclasts (essential for bone resorption), through dynamic regulating the interactions of actin cytoskeleton and formation of podosomes (Horne et al., 2005;Destaing et al., 2011). Src and other SFKs also play important roles in the antiapoptotic signal transduction of RANKL and other tumor necrosis factor family members in osteoclasts (Xing et al., 2001). One essential role for Src in osteoblasts has also been demonstrated that reduction of Src expression decreases osteoblast(responsible for bone formation) proliferation and increases differentiation (Marilena et al., 2000).
Recently, a large number of studies have shown that the activation of SRC is an important mediator of AR signaling. . AR can form a tertiary complex with the scaffold protein modulator of non-genomic actions of the estrogen receptor (MNAR/PELP1) and Src (Unni et al., 2004). Initially, Src is inactive within this complex. However, when AR binds to Src, this results in the activation of Src in this complex (AR/MNAR/Src) and the subsequent activation of a downstream effector, MEK (Unni et al., 2004). Subsequent studies have shown that AR-induced Src activation can promote cell proliferation through cell cycle progression from G1 phase to S phase (Migliaccio et al., 2002).

MAPKs Signaling Pathway
The MAPKs signaling cascade play important roles in regulating diverse biological functions including cell proliferation, motility and survival, which are essential to prostate carcinogenesis (Rossomando et al., 1989;Armenia et al., 2018;Abida et al., 2019). Studies of DHTresponsiveness in prostate cancer cells show that DHT treatment induces phosphorylation of ERK-1/2 within 1-2 min and peak levels of phosphorylation within 5-10 min (Liao et al., 2013). Activated ERK-1/2 then translocate to the nucleus and directly interact with and phosphorylate transcription factors (TFs), such as nuclear ETS domain-containing Elk1 (Marais et al., 1993;Gille et al., 1995;Yang et al., 1998). Elk1 transcriptionally regulates immediate early genes (IEGs) such as c-fos (Gille et al., 1995;Unni et al., 2004), and regulates the expression of several genes related to cell proliferation (Marais et al., 1993;Unni et al., 2004). In addition, recent studies showed that other molecules induce prostate cancer via MAPK signaling. Jason et al. reported that ADP-ribosylation factor 1 (ARF1), a crucial regulator in vesicle-mediated membrane trafficking and involved in the activation of signaling molecules, promotes the occurrence of prostate cancer via targeting oncogenic MAPK signaling (Davis et al., 2016). Gonzalo et al. reported that epidermal growth factor (EGF) could stimulate G0/G1-S transition via p38 MAPK to overcome the growth restriction of androgen deprivation in prostate cancer cells (Rodriguez-Berriguete et al., 2016).

JAK-STAT3 Signaling Pathway
Janus kinases (JAK) signal transducers and activator of transcription (STAT) pathway play an important role in differentiation, hematopoiesis, immune function and cell growth (Bolli et al., 2003). Recently, accumulating evidence indicated that IL-6 is indispensable for activation of JAK/   STAT pathway, which is involved in the oncogenesis of prostate cancer (Liu X. et al., 2012). Compared with men with normal prostates, benign prostatic hyperplasia, prostatitis and localized disease, approximately 50% of patients with advanced prostate cancer have increased levels of serum IL-6 (Twillie et al., 1995). Upon the binding of IL-6 to the IL-6 receptor, JAK-1 and STAT-3 become activated in sequence (C Schindler and Jr, 1995). L Tam et al. reported that cytoplasmic expression of IL-6 receptor and pSTAT3 Tyr705 are associated with the shortened biochemical recurrence time and death time from hormone relapse, respectively. Therefore, it is reasonable to target this pathway in hormone-refractory prostate cancer treatments (Tam et al., 2007).

Ca 2+ Signaling Pathway
Ca 2+ signaling is also involved in prostate cancer progression (Figiel et al., 2019;Chalmers and Monteith et al., 2018). Increased calcium intake from dairy products has been considered as a risk factor for prostate cancer (Foradori et al., 2007;Flourakis and Prevarskaya, 2009). As a primary signaling molecule, extracellular Ca 2+ works through the Ca 2+ -sensing receptor (CaR, a G protein coupled receptor) which directly regulates cell signal transduction and the Ca 2+ channels (Vaz et al., 2015). Depletion of intracellular Ca 2+ stores serves as a signal for the activation of Ca 2+ influx across the plasma membrane. The proteins STIM1 and ORAI1 were identified as the key components of store-operated Ca 2+ entry (SOCE). When Ca 2+ is released from intracellular Ca 2+ pool, Ca 2+ dissociates from a luminal EF hand motif of STIM1. As a consequence, STIM1 proteins aggregate and recruit Orai1 Ca 2+ channels, which then mediate SOCE (Kilch et al., 2016).
Recently, Huang et al. found that Ca 2+ via CaR-mediated signaling induces filamin A cleavage, which is an actin-binding protein, and promotes the migration of AR-deficient and highly metastatic prostate cancer cells . In one additional study, Christian demonstrated that transient receptor potential melastatin 4 channel (TRPM4) is activated by a rise in intracellular Ca 2+ in prostate cancer cells. Upon activation, a Na + influx via TRPM4 depolarizes the membrane potential, which reduces the driving force for Ca 2+ and limits SOCE, and thus promotes migration of androgen-insensitive prostate cancer cells (Christian et al., 2015).

Other Genomic Alterations in Castration-Resistant Prostate Cancer
Prostate cancer is characterized by a high genetic heterogeneity due to genomic alterations and instabilities associated with diverse PCa risk factors (Squire et al., 2011;Yap et al., 2016;Ciccarese et al., 2017;Rodrigues et al., 2017), which was evidenced by extensive genomic profiling analysis conducted on primary tumors (Network T. C. G. A., 2015) and on metastatic samples (Dan et al., 2015).
Speckle-type POZ protein (SPOP) is the substrate-binding subunit of a cullin-3 (CUL3)-based E3 ubiquitin ligase complex, which mediates the ubiquitylation and degradation of many target proteins. SPOP binds to the substrates through its N-terminal meprin and traf homology (MATH) domain, whereas it interacts with cullin-3 via BTB domain on its C terminal (Pintard et al., 2003;Xu et al., 2003;Zhuang et al., (Barbieri et al., 2012;Network T. C. G. A., 2015). Interestingly, all SPOP somatic mutations identified in prostate cancer are clustered in its substrate binding MATH domain, thus having a dominantnegative effect on substrate binding and degradation (Theurillat et al., 2014). Recent studies have unanimously reported that SPOP interacts with bromodomain and extraterminal (BET) proteins that largely act as transcriptional coactivators and play vital roles in cell cycle, apoptosis, migration and invasion in physiological conditions. In addition, SPOP also promotes the ubiquitylation and proteasomal degradation of bromodomain-containing protein 2 (BRD2), BRD3 and BRD4, (Dai et al., 2017;Janouskova et al., 2017;Zhang et al., 2017). Pathologically, BET proteins are frequently overexpressed and are clinically linked to various types of human cancer (French et al., 2003;Crawford et al., 2008;Belkina and Denis, 2012). Recently, Janouskova et al. reported prostate cancer-associated SPOP mutants impaired its binding to BET proteins, leading to the reduced proteasomal degradation and accumulation of these proteins in prostate cancer cell lines and patient specimens, which subsequently causes resistance to BET inhibitors (Janouskova et al., 2017). Similar study has also demonstrated that SPOP-mutated prostate cancer cell lines and patient-derived organoids were intrinsically resistant to BET inhibitor-induced growth arrest and apoptosis (Dai et al., 2017). Furthermore, Dai et al. provided that stabilization of BRD4 may be a molecular mechanism for resistance to BET inhibitors in patients with prostate cancer bearing SPOP mutations (Dai et al., 2017). Taken together, these findings offer mechanistic insights into how SPOP mutations influence prostate cancer. The TMPRSS2-ERG fusion gene arising from genetic rearrangement (fusion of encoding transmembrane protease serine 2, TMPRSS2 gene, and EST-related gene, ERG) has also been a central focus in prostate cancer, which leads to aberrant expression of the ETS transcription factor ERG (Tomlins et al., 2005;Kandoth, 2013). TMPRSS2-ERG is the most common gene rearrangement in prostate cancer and is present in approximately 50% of prostate cancer tissues in Western countries (Cary and Cooperberg, 2013). Previous studies indicated that ERG overexpression was driven by hijacking of androgenresponsive elements within the TMPRSS2 promoter (Tomlins et al., 2005;Wang et al., 2007;Thangapazham et al., 2014). However, Kron et al. found that the molecular process is more complex. Their study indicated that the frequent deletion allows a cluster of regulatory elements (CORE) in the TMPRSS2 promoter to expand into the rearranged ERG allele. This expanded CORE contains some CREs within the ERG locus that can promote ERG overexpression. Studies also revealed that overexpressed ERG co-opts prostate-specific master regulatory transcription factors, including AR, HOXB13 and FOXA1, in a process facilitated by their physical interaction with ERG and actives NOTCH signaling in primary prostate cancer (Kron et al., 2017). ERG overexpression is now an instrumental indicator in the diagnosis of prostate cancer. In addition, Graff et al. recently found that obesity and height might be correlated with the development of TMPRSS2-ERG-positive prostate cancer (Graff et al., 2018). Collectively, the functions and mechanisms of TMPRSS2-ERG increase the opportunities for finding new therapeutic targets for prostate cancer Kong et al., 2020).

NATURAL COMPOUNDS THAT EXERT ANTI-PROSTATE EFFECTS
Natural compounds that have been found to inhibit prostate cancer cells proliferation/tumor growth, promote prostate cancer cells apoptosis, or modulate specific signaling pathways involved in prostate cancer in vivo and in vitro are categorized and presented according to their source of isolation (marine organisms, microorganisms and plants) and the structural scaffolds. Besides the effects on prostate cancer cells growth or apoptosis, special emphasis was given to the mechanism of action of a compound interfering specific signaling pathways involved in prostate cancer.

Natural Compounds Obtained From Marine Organisms or Microorganisms
As is well known, marine organisms or microorganisms possess the capacity to produce a large amount of diverse secondary metabolites with unique structural features and biological properties. Thus, marine and microbial organisms represent interesting and important sources of single molecules with promising skeletons and significant antiprostate cancer activity. Up to now, a total of 24 natural compounds (Figures 3) isolated from marine organisms have been found to exhibit significant anti-prostate cancer activity either in vivo or in vitro. Detailed information about the compounds origin, activity and mechanism of action is listed in Table 1. Most of them exhibit antiproliferative, apoptosis inducing or metastasis inhibitory activities, with various acting mechanisms such as induction of autophagy, inhibition of AR activation, PI-3K/AKT/mTOR or JAK/STAT signaling pathways (Senderowicz et al., 1995;Liu et al., 2006;Wang WL. et al., 2008;Hellsten et al., 2008;Gantar et al., 2012;Meimetis et al., 2012;Shin et al., 2013;Liberio et al., 2015;Liu et al., 2016). Especially, frondoside A not only caused cell type Frontiers in Pharmacology | www.frontiersin.org September 2021 | Volume 12 | Article 738235 specific cell cycle arrest and induction of caspase-dependent or -independent apoptosis in vivo but also significantly inhibited the cell growth of PC-3 and DU145 with a notable reduction of lung metastasis and decrease of circulating tumor cells in the peripheral blood (Dyshlovoy et al., 2016). In addition, gliotoxin, chaetocin and chetomin exhibited antiangiogenic effects in vitro and attenuated tumor growth mainly by disrupting the HIF-1α/p300 complex, which makes them attractive molecules for the design of future chemotherapeutic agents (Cook et al., 2009).

Natural Compounds Isolated From Plants
Medicinal plants have always been a very good source of drugs, which could produce plenty of secondary metabolites with high structural diversity and versatile bioactivities. Many candidates with promising anti-prostate activity have been reported, including 7 alkaloids, 23 flavanoids, 25 terpenoids, 13 polyphenols, 10 lignans and 48 other compounds (Figures 4-9). Almost all these candidates show anti-prostate cancer activities via anti-proliferation, apoptosis induction or metastasis and invasion inhibition, involved in canonical AR signaling and non-AR signaling like caspase cascades, AKT/mTOR pathway, MAPKs pathway, NF-κB pathway, Ca2+ pathway and JAK/STATs pathway. Additionally, there exist other acting mechanisms, for example, anibamine exhibited anti-prostate cancer activity by binding to the chemokine receptor CCR5; fisetin inhibited tumor growth by downregulating the expression of NudC protein, MMP-2 and MMP-9; lycopene showed anti-prostate cancer effects by inhibiting androgen receptor element and signaling of insulin-like growth factor-1 Khan et al., 2008;Bureyko et al., 2009;Wertz, 2009;Zhang et al., 2010b;Chien et al., 2010;Suh et al., 2010;Tang et al., 2011;Holzapfel et al., 2013;Mukhtar et al., 2015). Detailed information is provided in Tables 2-7.

EXTRACTS WITH ANTI-PROSTATE CANCER ACTIVITY
Extracts consist of a group of bioactive natural compounds, which may exert and possess the advantages of synergistic effects against diseases. Recently, nutraceuticals have also received increasing attention as the agents or dietary supplements for cancer prevention and treatment, as well as some extracts derived from edible sources. Thus in this section we will respectively review those extracts and nutraceuticals that have the potential effects against prostate cancer either in vitro or in prostate cancer mice models. Chinese herbal compound preparations of more than one medicinal plants that have been reported to inhibit prostate cancer are also presented in this review.

Herbal Extracts
Traditional and folk herbal medicines from medicinal plants offer great potential for the discovery of novel anti-prostate cancer drugs. The plant extracts listed in Table 8 are complex mixtures, which need further investigations to reveal their bioative constituents through bioguided isolation and to clarify the roles of these different compounds in agaisnt prostate cancer when used alone or in combination. Also, the synergistic effect of the individual active components of these extracts and molecular mechanisms involved need further elucidation in order to evaluate the potential of these compounds as antineoplastic agents.

Chinese Herbal Compound Preparations
There are four traditional Chinese medical formulations reported to display significant anti-prostate cancer properties, that is, Zyflamend, PC-SPES and LCS101, which are composed of different medicinal plants (Table 9; Bemis et al., 2005;Hsieh et al., 1997;Cohen et al., 2015). Especially, PC-SPES significantly inhibited prostate tumor growth in tumor-bearing mouse model, mainly through cell cycle arrest and apoptosis induction, which is already clinically utilized for the treatment of clincal patients with prostate cancer (Hsieh et al., 1997).

Nutraceuticals and Extracts Derived From Edible Sources
Nowadays, dietary factors play an increasingly important role in the chemopreventive and/or therapeutic management of cancer (  (Lippert et al., 1999;Nam et al., 1999). There are strong evidences that nutraceuticals and extracts derived from edible spices, vegetables or fruits such as vitamin D, pomegranate and tea polyphenols have demonstrated significant anti-prostate cancer activity when tested either in vitro and/or in vivo (Kasimsetty et al., 2009;Gregory et al., 2010;Koyama et al., 2010;Mordanmccombs et al., 2010;Hsu et al., 2011;Xiao et al., 2011;Turan et al., 2017). Especially, dietary phytochemicals that can selectively interfere cellular pathways involved in prostate

CONCLUSION AND PERSPECTIVES
Prostate cancer is the second most frequently diagnosed tumor and the fifth leading cause of cancer-related deaths in men in the worldwide (McEleny et al., 2002). And the mortality of prostate cancer mainly occurs as a result of the castrate resistant ones. Up to date, different kinds of drugs have been employed to improve the treatment condition, mainly including LHRH antagonists, antiandrogen (androgen receptor antagonists and androgen synthesis inhibitors), tyrosine kinase inhibitors, angiogenesis inhibitors, endothelin antagonists, matrix metalloproteinase inhibitors, antioxidants, and cell cycle inhibitors. However, as mentioned above, there is no effective therapy for CRPC at present, except for docetaxel, which is the only chemotherapeutic agent that has been proven to prolong the overall survival in CRPC patient population though with many adverse effects reported (Eyben et al., 2015). Hence, it is urgent for us to explore an effective treatment for prostate cancer, especially for CRPC. In recent years, many natural products and extracts have been scientifically investigated in vitro and/or in vivo and proved as potential anti-prostate cancer agents, which are currently scattered across various publications. So a systematic summary and knowledge of future prospects are necessary to facilitate further chemical and pharmacological studies for antiprostate cancer agents.
In our review, we provided a comprehensive overview of the molecular basis of the incidence and development of prostate cancer, especially for castration-resistant prostate cancer (CRPC), which mainly including canonical AR signaling (AR

Dietary agents
Cell type Observation Activity

Du145
In vivo Antiproliferation Downregulation of ID2; increase of Bax/Bcl-2; inducing cell death via an ID2-related mechanism; Antiproliferation by increasing the activity of ERK 1/2 through a MEKindependent and PI3K-dependent mechanism.
Xiao et al. (2011) In vitro Antiproliferation Increase of p53 expression; apoptosis induction by superoxide generation through the mitochondrial-dependent pathway amplification, over-expression, mutation, and unconventional activation), and non-nuclear AR signaling (PI3K/AKT, Src, MAPKs, JAK-STAT3, and Ca 2+ signaling pathways). So most components involved in above-mentioned pathways represent potential targets for screening natural compounds and/or extracts with anti-prostate cancer activity. And natural compounds or Frontiers in Pharmacology | www.frontiersin.org September 2021 | Volume 12 | Article 738235 30 extracts that could function as modulators of canonical AR or non-nuclear AR signalling pathways thus can be regarded as promising candidates for anti-prostate interventions.
So far, a great amount of natural products isolated from diverse sources have been found to significantly inhibit prostate cancer cell proliferation/tumor growth or affected cellular signaling pathways in prostate cancer. As shown in our paper, the majority of natural compounds with direct relevance to prostate cancer are primarily derived from plants, with comparatively few molecules from marine and microbial sources. For these reported bioactive constituents, there is still plenty of room for improvement regarding the studies focused on efficacy enhancement and side effects amelioration by semi-synthetic modifications based on quantitative structural activity relationship elucidation. Since marine and microbial organisms represent important sources for single molecules exploit, more available and improved approaches should be included in finding novel natural products with significant anti-prostate cancer activity from these resource. Especially, engineering bacteria or fungus with novel gene clusters, currently used mainly for the identification of antibiotics or anti-tumor drugs, would be another promising approach for discovering natural compounds with anti-prostate activity. Extracts are another applicable option for anti-prostate purposes, in which case the chemical profile should be further elucidated, possibly affording a pure bioactive compound with precise mechanism of action. Also clinically used Chinese herb preparations should be profiled using techniques such as HPLC-MS to standardize the complex system to make it more controllable, stable, and reproducible in prostate cancer treatment. Furthermore, drug combination of these reported natural compounds with conventional chemotherapeutic agents may also be a promising way in finding solution for prostate cancer treatment. Finally, safty large-scale studies are needed to evaluate promising compounds or extracts and determing non-toxic doses for treating prostate cancer in mammals.
In conclusion, tackling prostate cancer (especially CRPC) is a much needed task that requires not only the great progress in understanding the genetic basis of prostate cancer, but also the significant technological improvements in tracking of bioactive natural compounds and structural characterization, which will facilitate the identification of novel natural compounds with significant antiprostate cancer properties for drug development and therefore can be translated into significant health benefits for humans.

AUTHOR CONTRIBUTIONS
CZ, WX, and QY conceived and designed the outline of the manuscript; BB, QC, and RJ collated and summarized the literatures; XQ, HW and YB analyzed the data; RJ, BB, and CZ wrote the manuscript, QC, WX, and QY revised the manuscript and provided technical assistance to RJ and BB.

FUNDING
The work was supported by the National Nature Science Foundation of China (No. 82174081, 81770604), the Program of Shanghai Health System Subject Chief Scientist (2017BR004), and Shanghai Natural Science Fund (21ZR1478000).