Involvement of estrogen receptors in prostatic diseases

Authors


Hidenori Kawashima M.D., Ph.D., Department of Urology, Osaka City University Graduate School of Medicine, 1-4-3 Asahimachi, Abenoku, Osaka 545-8585, Japan. Email: hidenori@msic.med.osaka-cu.ac.jp

Abstract

Accumulating evidence shows that estrogens participate in the pathogenesis and development of benign prostatic hyperplasia and prostate cancer by activating estrogen receptor α. In contrast, estrogen receptor β is involved in the differentiation and maturation of prostatic epithelial cells, and thus possesses antitumor effects in prostate cancer. However, the natural ligands of estrogen receptor β are not fully understood, and its mode of action according to its ligands and the binding sites located in the promoter regions of downstream genes remains to be elucidated. Here, we review recent experimental investigations of estrogen receptors and their urological relevance. Estrogen receptor-mediated signaling in the prostate is essential together with the androgen receptor-mediated pathway, providing a new therapeutic target for prostatic diseases.

Abbreviations & Acronyms
3,4-QE2 =

estradiol-3,4,quinone

3β-adiol =

5α-androstane-3β, 17β-diol

4-OHE2 =

4-hydroxyestradiol

17β-HSD VII =

17β-hydroxysteroid dehydrogenase type 7

AR =

androgen receptor

ArKO =

aromatase-knockout

BP =

DL-2-[4-(2-piperidinoethoxy)phenyl]-3-phenyl-2 H-1-benzopyran

BPH =

benign prostatic hyperplasia

CRPC =

castration-resistant prostate cancer

DES =

diethylstilbestrol

DHT =

dihydrotestosterone

E2 =

estradiol

EBAG9 =

estrogen receptor-binding fragment-associated gene 9

EMT =

epithelial-mesenchymal transition

ER =

estrogen receptor

ERαKO =

estrogen receptor α knockout

ERβKO =

estrogen receptor β knockout

ERE =

estrogen-responsive element

ERG =

v-ets erythroblastosis virus E26 oncogene homolog (avian)

ERR =

estrogen receptor-related receptors

hERβ =

human estrogen receptor β

HGPIN =

high-grade prostatic intraepithelial neoplasia

HIF-1α =

hypoxia inducible factor 1α

IGF-1 =

insulin-like growth factor 1

KLF-5 =

Krüppel-like zinc finger transcription factor 5

NFκ =

nuclear factor κ

NFκB =

nuclear factor κB

OHT =

4-hydroxytamoxifen

PGC-1 =

peroxisome proliferator-activated receptor γ coactivator 1

PIN =

prostatic intraepithelial neoplasia

PrSC =

prostatic stromal cells

PSA =

prostate-specific antigen

RT–PCR =

reverse transcription polymerase chain reaction

SERM =

selective estrogen receptor modulators

TGFβ =

transforming growth factor β

TMPRSS2 =

transmembrane protease serine 2

VEGF-A =

vascular endothelial growth factor A

Introduction

Although androgen dependence is the most striking characteristic of the prostate gland, estrogen has been reported to be involved in the development of the prostate gland and prostatic disease pathogenesis.1 Maternal estradiol, for instance, causes squamous metaplasia within the developing prostatic epithelium in male offspring,2 whereas treatment of dogs with estrogens combined with androgens was used to experimentally induce BPH.2 Earlier evidence of estrogen as a tumor-promoting factor was first presented in a study that showed prostate cancer developed in Noble rats more rapidly when estrogens were given in addition to testosterone.3

Until recently, estrogens were thought to indirectly act on the prostate; for example, through negative feedback on the hypothalamo–pituitary–gonadal axis, leading to suppression of testosterone production. However, since the discovery of ERβ in rat prostate,4 much attention has been paid to the two classes of the receptor, ERα and ERβ, in terms of their expression and function in the prostate and their involvement in prostatic diseases.

In the present article, we review and summarize the recent accumulating data providing evidence for estrogen signaling in the prostate, which substantiates the direct action of estrogens through ER on prostate cells, including cancer. As a future perspective, the possibilities of ER-targeted therapy for BPH, prostate cancer and chemoprevention are discussed.

Differential localization of ERα and ERβ, and their general function in prostate

The expression of ERα and -β has been extensively studied using RT–PCR or immunohistochemistry in several prostate cancer cell lines, as well as in normal and malignant prostatic tissues.5–10 The consensus is that ERα is mainly expressed in stromal cells, and ERβ is localized mainly, but not exclusively, in epithelial cells.11–13

ERβ expression is reduced in high-grade PIN compared with non-malignant tissues, and declines during malignant progression, suggesting a role for ERβ as a putative tumor suppressor.2,12,14 However, conflicting results have been published regarding ERβ expression and its correlation to prostate cancer grade.15–18 This issue will be discussed in further detail in this article with respect to the different isoforms of ERβ. Generally, ERα expression in stroma stimulates cell proliferation, whereas ERβ mediates differentiation of epithelial cells19 and anti-proliferation/apoptosis of prostate cancer cells12,14 (Fig. 1).

Figure 1.

General roles of estrogen receptors in prostatic cells. ERα expression in stroma stimulates cell proliferation, wheras ERβ mediates differentiation of epithelial cells and antiproliferation/apoptosis of prostate cancer cells. Estrogen (especially E2) and ERα are related to prostate cancer susceptibility and development.

ERα as a promoting factor of BPH

One of the hypothesized pathogeneses of BPH is that stromal hyperplasia is associated with enhanced estrogenic status after middle age, with a decrease in androgen level and an increase in conversion of adrenal androgens to estrogen by aromatase.20 It is thought that the activation of ERα by estrogen in stroma causes cell proliferation.

Important studies on the role of estrogen in pathogenesis of BPH have been carried out using normal human stromal and epithelial prostate cells. King et al. showed that an increased estrogen/androgen ratio enhanced the proliferation of PrSC and stimulated the growth of prostate epithelial cells co-cultured with stromal cells,21 the latter observation implying a paracrine action through ER activation. Zhong et al. observed that ERK was activated by E2, causing the proliferation of PrSC,22 whereas Nomura et al. showed that toremifene, an ERα antagonist, strongly suppressed the ER activity and proliferation of primary PrSC.23

Primary cultures of normal human prostatic epithelial and stromal cells have been used as useful models for studying the pathogenesis of BPH experimentally. However, the limitations of the prostatic primary cultures should also be considered; their proliferation rate is slow and they have a limited lifespan in culture, eventually undergoing senescence.24,25 AR mRNA is expressed in cultured PrSC,26–28 whereas reduced expression of AR protein in the stroma of late passages of cells has been reported.28 Commercially-available prostatic epithelial cells show basal epithelia that lack the AR protein expression25,29,30 and are not sufficiently differentiated to be secretory epithelial cells.25

The expression of ERα and -β in normal prostatic primary culture varies according to some experimental reports. Ho et al. investigated the effect of E2 on proliferation of human prostatic primary culture cells derived from BPH. E2 moderately increased stromal cell proliferation, which was antagonized by an anti-estrogen, ICI182780, whereas proliferation of epithelial cells was not increased by E2. Expression of ERβ in stromal cells was shown by RT–PCR and western blot analysis, whereas that of ERα in stromal cells was weaker. Aromatase, an enzyme that coverts androgen to estrogen, was also expressed in stromal cells. This emphasized that ERβ and local estrogen converted by aromatase play an important role in prostatic stromal overgrowth, a characteristic of the pathogenesis of BPH.31 Thus, the distribution of the two subtypes of ER in prostatic cells is still controversial. The expression of ER in prostate primary culture cells might be altered during passage, in common with AR expression. In our findings, the ERα expression level was much higher than that of ERβ in the PrSC of a commercially-available primary culture (Lonza Walkersville, Walkersville, MD, USA) (Kawashima, 2009, unpubl. data).

A recent report investigating the role of ERα in the etiology of BPH is that E2 or IGF-1 enhanced the proliferation of mouse prostatic smooth muscle cells, which was suppressed by a small interfering RNA against ERα. The expression of IGF-1 and its receptor, IGF-1R, were also ERα-dependent, and cyclin D1 was upregulated by E2 or exogenous IGF-1 in the presence of ERα.32 This model emphasizes that ERα is essential for IGF-1 signaling in stromal proliferation.

The role of ERα in stromal proliferation is well supported by the observations made with ERαKO mice. The ventral and dorsal-lateral prostate of ERαKO mice showed reduced branching morphogenesis, decreased fibroblast proliferation and stromal content changes.33 Thus, ERα is required for the proliferation of stromal components and, through the paracrine mechanism, the prostatic branching morphogenesis.

Together, the studies summarized here substantiate the role of ERα as a factor involved in promoting BPH development.

ER and prostate cancer

ERα and carcinogenesis in prostate

The involvement of ER (especially ERα) in the pathogenesis of prostate cancer and tumor progression was suggested as a result of immunohistochemical observations that ERα protein expression progressively increased from primary Gleason grade 2 tumors to recurrent and metastatic lesions.34 The expression of the estrogen-inducible progesterone receptor was also examined immunohistochemically using pathological specimens from prostate cancer patients.35 Progesterone receptor expression was increased in metastatic and androgen-insensitive cancers compared with primary tumors, suggesting that advanced tumors are capable of exploiting the ERα-mediated pathway for their progression.

Experimentally, prostatic carcinoma has been generated in an in vivo cell/tissue recombination and renal graft model. Prostate stem/progenitor cells were isolated from normal human prostatic epithelial cells, mixed with rat urogenital sinus mesenchyme and inoculated under the renal capsule of male nude mice, resulting in the generation of normal human prostate-like tissue in which AR and ER were expressed. Then, the mice were exposed to elevated testosterone and E2, resulting in the development of epithelial hyperplasia, PIN and prostate cancer.36

Recent investigations have emphasized the relevance and importance of the role of ERα in prostatic carcinogenesis.37,38 An attempt has been made to explain the molecular mechanism by which administration of both E2 and testosterone causes carcinogenesis in Noble rats. Ricke et al. treated ArKO mice with E2 and testosterone to stimulate prostate cancer development in order to investigate whether local in situ production of E2 can induce prostatic carcinogenesis.38 Aromatase is an enzyme that catalyzes the conversion of androgen to estrogen and is reported to be expressed in the stroma of the prostate. ArKO mice had reduced incidences of PIN compared with wild-type mice, showing the importance of in situ production of E2 for tumorigenesis. To determine whether E2-mediated induction of PIN was through ERα or ERβ signaling, ERαKO and ERβKO mice were treated with both E2 and testosterone. Although ERβKO and wild-type mice treated with both E2 and testosterone developed atypical hyperplasia and PIN, ERαKO mice did not.

In terms of the involvement of ER in prostate cancer progression, of note is a study on TMPRSS2-ERG fusion.39 The majority of prostate cancers harbor a chromosomal translocation causing the fusion of the promoter region of the TMPRSS2 gene to the coding region of an erythroblast transformation-specific transcription factor family member, ERG. Prostate cancer with this fusion generally shows a more aggressive clinical course. NCI-H660 prostate cancer cells express this TMPRSS2-ERG fusion, and both ERα and -β, but lack AR. The proliferation of this cell line was suppressed by an ERβ agonist and increased by an ERα-selective agonist. Accordingly, the expression of TMPRSS2-ERG was induced by an ERα agonist and suppressed by an ERβ agonist. Thus, the prostate cancer-related gene TMPRSS2-ERG fusion is regulated by ER-mediated signaling, supporting the notion that ERα activation is an important factor in promoting prostate cancer. Another estrogen-responsive gene reported to be related to the aggressiveness of prostate cancer is EBAG9,40 which has an ERE in its promoter region, and its transcription is enhanced by E2. By immunohistochemical analyses of benign and malignant prostatic specimens, Takahashi et al. showed that EBAG9 was expressed at higher levels in prostate cancer than in benign tissues, and that its increased expression was associated with the aggressive phenotypes, although they could not correlate its expression with the specific ER subtype.

ERα also seems to be related to the promotion of prostate cancer invasion.41 Yu et al. showed experimentally that E2, as well as an ERα-specific agonist, 1,3,5-Tris(4-hydroxyphenyl)-4-propyl-1H-pyrazole, induced matrix metalloproteinase 2 production in PrSC through ERα-induced expression of TGFβ, resulting in an increased stimulation of prostate cancer cell invasion in a paracrine manner.42

In addition to these experimental studies, the correlation between prostate cancer risk and polymorphism in estrogen-related genes has also been investigated. Nicolaiew et al. examined polymorphisms in ERα and ERβ among Caucasian prostate cancer cases and control subjects.43 They found no association between any variant of the ERβ gene and prostate cancer risk. Instead, the GGGA repeat (5/4, 6/5 and 6/6 genotype) located in the first intron of ERα is related to the risk of prostate cancer. This polymorphism was associated with prostate cancer with a favorable Gleason score and cancer of a late onset. Sonoda et al. investigated the polymorphisms in ERα, ERβ and CYP19A1 (aromatase) between Japanese patients and matched controls with subdivision of the participants in accordance with intake of isoflavone.44 CYP19A1 converts androgens to estrogen, thus affecting estrogen levels. The genotype of ERβ was not related to prostate cancer risk, whereas the TTTA long repeats in the CYP19A1 gene were significantly associated with increased risk. The combination of the TTTA long repeats in CYP19A1, and the minor alleles of rs10046 in CYP19A1 and rs2077647 in ERα was associated with a high risk of prostate cancer despite higher isoflavone intake. These reports support that estrogen (especially E2) and ERα are related to prostate cancer susceptibility.

Suppressive role of ERβ in prostatic cell proliferation including cancer

Since ERβ was first isolated from rat prostate,4 much attention has been paid to its role and function in the prostate. Several reports have described the loss of ERβ expression in HGPIN and high-grade dysplasia compared with normal prostate epithelium expressing ERβ, suggesting a tumor-suppressive role for this subtype.15,16,45 In this section, we review experimental reports of in vitro and knockout animal studies used to show the antiproliferative or tumor-suppressive role of ERβ in the prostate.

Interestingly, Hurtado et al. showed that overexpression of ERβ1 (ERβ) caused cell cycle arrest in early G1 phase in LNCaP cells.46 In addition, Walton et al. showed that re-expression of ERβ using a DNA demethylating agent and a histone deacetylase inhibitor caused increased apoptosis in prostate cancer cells.47 Adenovirus-mediated expression of ERβ in DU145 prostate cancer cells treated with E2 resulted in reduced cellular invasion in Matrigel and inhibition of cell proliferation.48 Rossi et al. showed that raloxifene, a mixed estrogen agonist/antagonist, had antiproliferative and pro-apoptotic effects in prostate cancer cells expressing ERβ and ERαin vitro.49 In that study, they showed that ERβ, rather than ERα, was activated by raloxifene when growth suppression occurred, supporting the antitumor role of ERβ.

A few reports have shown the association of ERβ and its downstream genes in prostate cancer cells. The following two examples are illustrated in Figure 2. Leng et al. showed that ICI182780 induced several important genes by activating ERβ to bind to upstream NFκ consensus sequences.50 Matsumura et al. reported that genistein, a phytoestrogen extracted from soybeans, induced expression of p21 and ERβ in PC-3 prostate cancer cells; silencing of ERβ by siRNA suppressed p21 expression and the transcriptional activity of p21 promoter.51

Figure 2.

Association of ERβ and promoter regions of the downstream genes in prostate cancer cells. A few reports have shown the association of ERβ and its downstream genes in prostate cancer cells with its mode of action according to its ligands and the binding sites. (a) ICI182780 activates ERβ, which binds to NFκ consensus sequences resulting in suppression of DU145 cell growth. (b) Genistein, a phytoestrogen extracted from soybeans, activates ERβ to enhance the transcriptional activity of p21 promoter.

The histological observation of the prostate of ERβKO mouse suggests the ability of ERβ to suppress the growth of prostate cells or tumors. The prostate of the ERβKO mouse contained multiple hyperplastic lesions with epithelial cells stained positively with Ki67.52 However, another ERβKO model did not show findings suggestive of the antiproliferative role of ERβ. Antal et al. constructed an ERβ mouse mutant (ERβSTL−/L−) in which exon 3 of the ERβ gene was deleted and was devoid of any transcript downstream of exon 3.53 This ERβKO mouse was sterile, with mildly affected histology of the ovary, but showed no histological defects in other organs, including the prostate. It would be interesting to examine whether this ERβKO mouse is more susceptible to prostatic carcinogenesis.

The ArKO mouse is estrogen-deficient and has elevated androgen and 5α-DHT levels, presumably owing to lack of conversion of androgen to estrogen. This mouse develops prostatic hyperplasia and hypertrophy, but not malignancy.54 Administration of ERβ agonist abrogated developing hyperplasia in ArKO mice, the possible explanation being that a failure to activate ERβ in the absence of in situ estrogen production will result in developing prostatic hyperplasia, which is reversed by ERβ agonist.55,56

Overall, these reports substantiate the antiproliferative or tumor-suppressive role of ERβ in prostatic cells.

Differences in activation of ERβ depending on different ligands and binding sites

As with other nuclear receptors, ERβ is activated by binding its ligand. The binding site of both ERα and ERβ is the ERE. Other binding sites, such as AP1 and NFκB, are also known to be involved in cross-talk activation.50,57 A year after the discovery of ERβ, a study showed differences in ERβ activation by different ligands at AP-1 sites.57 Diethylstilbestrol and 17β-estradiol (E2) activated ERβ at the estrogen responsive element, whereas they inhibited the activity of ERβ at AP1 sites. Furthermore, SERM, such as tamoxifen, raloxifene and ICI164384, activated ERβ at AP1 sites. Thus, ERβ is activated differently depending on ligands and response elements (binding sites) in the promoter regions of the target genes.

In addition, Nakajima et al. showed that ICI182780-activated ERβ enhanced the expression of FOXO1, a transcription factor that regulates genes related to apoptosis or anoikis, thorough KLF5 in DU145 cells, resulting in suppression of cell proliferation.58 In contrast, E2 induced proteasome-dependent degradation of KLF5 through ERβ, resulting in promotion of DU145 cell proliferation (Fig. 3). Activation of ERβ by E2 is not required for tumor suppression by ERβ in the prostate where the physiological level of estrogen is low.58,59 Thus, the action of ERβ is different depending on the different ligands. Interestingly, they also reported that KLF5 and ERβ expression is positively correlated with cancer-specific survival.58

Figure 3.

Different action of ERβ depending on different ligands. ICI182780-activated ERβ enhances the expression of FOXO1 through KLF5, causing suppression of DU145 cell proliferation. In contrast, E2 induces proteasome-dependent degradation of KLF5 through ERβ, resulting in growth promotion of DU145 cells.58

The physiological ligand of ERβ has not yet been determined, and several reports have focused on 5α-androstane-3β,17β-diol (3β-adiol) as a putative natural ERβ-ligand, the tissue level of which is much higher than E2 in the prostate (Fig. 4).60,61 3β-Adiol is a metabolite of DHT, which is produced by 17β-hydroxysteroid dehydrogenase type 7 and is further metabolized to triols by CYP7B1.61 Dondi et al. showed that 3β-adiol significantly decreased cell proliferation, adhesion, migration and invasion of PC3 prostate cancer cells in vitro, as well as in a xenograft nude mice model.62 Because PC3 cells express ERβ, but not AR, and 3β-adiol is thought to be a putative endogenous ligand of ERβ, they attributed the tumor-suppressive properties of 3β-adiol to ERβ activation. Guerini et al. also reported that 3β-adiol exerted an inhibitory effect on DU145 cell migration through the activation of ERβ signaling, and that E2 was not involved in this suppressive effect.63

Figure 4.

3β-Adiol as a putative natural ERβ ligand. The tissue level of 3β-adiol is much higher than E2 in prostate. (a) Activation of ERβ by 3β-adiol causes several beneficial effects. (b) Metabolism of 3β-adiol.

Furthermore, 3β-adiol was also reported to be involved in hindering EMT. Mak et al. hypothesized that the architectural pattern of prostate cancer cells, that is, loss of glandular structure, used for the classification of Gleason grading is a manifestation of EMT.64 They showed that ERβ impedes prostate cancer EMT by destabilizing HIF-1α, an important enhancing factor for EMT like TGFβ and hypoxia. Loss of HIF-1α reduced the expression of VEGF-A. The receptor of VEGF-A is neuropilin-1, which was shown to promote nuclear localization of Snail, causing EMT. They also showed that TGFβ or hypoxia, enhancing factors of EMT, diminished ERβ expression, and reciprocally that loss of ERβ caused EMT. 3β-Adiol sustained an epithelial phenotype in PC3 prostate cancer cells that express ERβ. This report substantiates the role of ERβ in epithelial differentiation and supports the observations that the expression of ERβ is negatively related to the aggressiveness of prostate cancer.

The functions and activities of ERβ that are dependent on ligands and binding sites in the promoter regions are summarized in Table 1.

Table 1.  Functions and activities of ERβ that are dependent on ligands and promoter site binding regions
LigandsBinding sitesERβ (agonist/antagonist)FunctionReference number
DES, E2EREActivated 57
DES, E2AP1Inhibited 57
TamoxifenAP1Activated 57
Raloxifene
ICI164384
ICI182780NFκ (cross-talk)ActivatedInhibition of DU145 cell proliferation50
ICI182780 ActivatedEnhancement of FOXO158
Suppression of DU145 proliferation
E2 ActivatedDegradation of KLF558
3β-Adiol (Natural ligand?)Suppression of PC3 cell proliferation, adhesion and invasion62
3β-Adiol (not E2) (Natural ligand?)Inhibition of DU145 cell migration63
3β-Adiol (Natural ligand?)Impediment of EMT64
8β-VE2 (Selective ERβ agonist)Apoptosis of PC3 and LNCaP cells65

Splice variants of ERβ

Splice variants of hERβ1, such as hERβ2 (hERβcx), hERβ3, hERβ4 and hERβ5, have been identified and reported.66–68 An earlier report showed that hERβ2 (hERβcx) functioned as a dominant negative of ERα.69 In contrast, it has been reported that the splice variants hERβ2, hERβ4 and hERβ5 are capable of forming heterodimers with ERβ, and enhance the transcription activities of ERβ.68

An important point of the existence of these splice variants is that some have been shown to be expressed at higher levels in prostate cancer with poor prognosis.70 Leung et al. generated isoform-specific antibodies against ERβ1, ERβ2 and ERβ5, and carried out immunohistochemical studies using a tumor microarray that consisted of 144 clinical specimens from patients who had undergone radical prostatectomy; these results correlated with clinical follow-up data.71 Positive expression of nuclear ERβ2 was significantly related to PSA failure and postoperative metastases, whereas combined positive nuclear ERβ2 and cytoplasmic ERβ5 expression correlated with the shortest postoperative metastasis-free survival. The researchers infected PC3 cells with ERβ1, ERβ2 and ERβ5 using lentivirus, and cell mobility and invasiveness was examined. Cells infected with ERβ2 and ERβ5 showed more invasiveness, and the authors reported that ERβ2 and ERβ5 have metastasis-promoting action. If the infected ERβ2 or ERβ5 formed heterodimers with endogenous ERβ in PC3 cells, enhanced ERβ activity resulted; these observations are contradictory to the tumor-suppressive property of ERβ.

Another issue arising from the existence of the ERβ splice variants is that ERβ isoform-specific primers for PCR or isoform-specific antibodies for immunohistochemistry were not used in some studies, especially in earlier investigations. This could explain the conflicting results published regarding ERβ expression and its correlation to prostate cancer grade.15–18 For example, Walton et al. examined the differences in ERβ gene expression in cancer and normal epithelium using laser microdissection and quantitative RT–PCR, and found that ERβ expression was significantly elevated in cancer epithelium compared with the benign control.18 However, the ERβ primers used in that study were not isoform-specific and thus amplified all isoforms including the splice variants, some of which are reported to be increasingly expressed in more aggressive prostate cancer.

Estrogen receptor-related receptors

The ERR belong to the NR3B subgroup, which includes ERRα, ERRβ and ERRγ. ERRα was originally identified because of its nucleotide and amino acid sequence similarities with ERα. They are constitutively activated and their natural ligands are yet to be identified, thus they are called orphan receptors.72

The three isoforms are expressed at elevated levels in heart, skeletal muscle and the kidneys; that is, organs of high energy demand. They are thought to be involved in energy metabolism through directly interacting with PGC-1α and -β. ERRα knockout mice have reduced bodyweight and are resistant to high fat diet-induced obesity. Their brown adipose tissue has a reduced mitochondrial mass, and these mice fail to keep body temperature when exposed to cold.73

ERR are also involved in steroid signaling in the prostate. Teyssier et al. showed that ERRα transactivated androgen-responsive element-containing promoters, including that of PSA, under the existence of AR. The ERRα-specific inverse agonist, XCT790, downregulated the expression of androgen-responsive genes in LNCaP cells.74

A steroidal anti-estrogen, SR16388, was also reported to be a selective inverse agonist of ERRα, inhibiting ERR-responsive element-directed transcription in MCF7 and PC3 cells cotransfected with ERRα and PGC-1α. SR16388 inhibited the growth of xenografts of PC3 prostate cancer cells in nude mice; the effect was synergic, especially in combination with paclitaxel treatment.75 Overall, these studies showed the capabilities of ERR as one of the factors involved in androgen-independent growth in prostate cancer.

With respect to the clinical relevance of the ERR, Fujimura et al. investigated the correlation between the expression of each ERR in resected specimens and the cancer-specific survival of the patients treated with radical prostatectomy using isoform-specific antibodies.76,77 Increased expression of ERRα was significantly related to poor prognosis; expression levels of ERRβ and ERRγ were lower in cancerous lesions than in benign lesions, and patients with high ERRα expression and low ERRγ expression showed a significantly poorer cancer-specific survival rate.

ER and CRPC

An earlier study examining the expression of ERα in prostatic adenocarcinoma showed that ERα expression was progressively increased from primary Gleason grade 2 tumors to recurrent and metastatic lesions, and that adenocarcinomas recurring after hormonal therapy expressed nuclear ERα in 94% of cases.34 The estrogen-inducible progesterone receptor is also expressed strongly in androgen-insensitive tumors at high frequency.35 These reports imply the relevance of ERα in castration-resistant diseases.

Sissung et al. reported that polymorphisms in ERα and CYP19 were related to shorter progression-free survival among CRPC patients treated with docetaxel compared with those with a wild-type allele.78 They also showed that ERα polymorphisms were associated with an increased risk of developing CRPC. The possible explanation is that the ERα regulates CYP1B1, by which E2 is converted to 4-OHE2, and that 4-OHE2 is subsequently oxidized to 3,4-QE2. 3,4-QE2 inhibits docetaxel-mediated tubulin polymerization and covalently binds docetaxel.

With respect to ERα and -β in CRPC, Celhay et al. compared the expression of estrogen-related genes in hormone-sensitive prostate cancer tissue obtained before androgen deprivation therapy and that of hormone-refractory disease.79 Increased expression of AR, phosphorylated AR and breast cancer anti-estrogen resistance 1, and decreased expression of ERβ and 5α-reductase 2 were observed in hormone-refractory prostate cancer. The decrease in ERβ in CRPC is consistent with the supposed notion that ERβ is a tumor suppressor. In contrast, higher frequency of ERα staining in stromal cells was related to delayed hormonal relapse and longer overall survival. Thus, the role of ERα in CRPC is not straightforward.

ER as a therapeutic and chemopreventative target of prostatic diseases

Experimentally, the beneficial effects of phytoestrogens, especially genistein, for the prevention and treatment of prostate cancer have been investigated using in vitro and animal models.80,81 Although the clinical efficacy of the use of phytoestrogens has not been proven,80 complementary and alternative medicines, as well as a diet rich in those phytochemicals, are increasingly used. ERβ was shown to be involved in the anticancer effect of phytoestrogens through the enhancement of p21 transcription, or downregulation of AR,51,82 thus therapy/prevention using phytoestrogens is thought to be ERβ-targeted. A recent report showed a synergic effect of combining phytoestrogens in preventing the proliferation of PC3 prostate cancer cells. The best combination of phytoestrogens for synergic action was identified to be that of genistein, biochanin and quercetin, and the authors provided an experimental basis for the notion that phytoestrogens present in vegetarian diets play an important role in the low incidence of prostate cancer in Asian men.83 Very recently, a result of a randomized, double-blind, placebo-controlled trial of oral isoflavone for reducing the incidence of prostate cancer development has been published from a Japanese study group. The incidence of biopsy-detectable prostate cancer among Japanese men aged 65 years or more was significantly lower in the isoflavone group than that in the placebo group.84

From a chemopreventative point of view, a clinical trial was carried out that showed that toremifene decreased the incidence of prostate cancer development in men with HGPIN.85 Toremifene also strongly suppresses ER activity and cell proliferation in PrSC,23 and the ability of toremifene to decrease the incidence of prostate cancer might be attributed to the suppression of the stromal paracrine that stimulates the precancerous lesion. Experimentally, toremifene suppressed prostate cancer development in the transgenic adenocarcinoma of mouse prostate model,86 and delayed tumor formation and prolonged survival compared with placebo-treated animals.86

Kumar et al. focused on ER-mediated signaling in stromal cells in the pathogenesis of BPH, and examined the suppressive effects of SERM on proliferation of BPH-derived stromal cells.87 Two SERM, BP and ormeloxifene, were evaluated and compared with tamoxifen and OHT. BP, OHT and tamoxifen effectively reduced the proliferation of stromal cells. These four SERM, BP, ormeloxifene, tamoxifen and OHT, increased caspase-3 activity, whereas BP increased the expression of ERβ, TGFβ1, Fas and FasL, and decreased that of ERα, AR, EGFR and IGF-1 in human PrSC. Then, combined treatment with SERM and finasteride was used, reducing rat prostate weight more effectively than SERM alone. It also increased the ratio of Bax/Bcl-2 with diminished acinar diameter and reduced epithelial cell height in rat ventral prostate.

The possible use of a selective ERβ agonist, 8β-VE2, for treating BPH and prostate cancer was investigated using the ArKO mouse, thus avoiding any influence by internal estrogens.65 8β-VE2 induced apoptosis in prostatic stromal, luminal and basal epithelial cells in ArKO mice independent of androgen, but dependent on tumor necrosis factor signaling. The effects of 8β-VE2 on xenografted BPH specimens, and grafted LNCaP cells and androgen-independent PC3 prostate cancer cells were also examined, and it was shown to induce apoptosis in these cells. The authors postulated ERβ as a new therapeutic target for BPH and prostate cancer by a mechanism different from androgen-mediated signaling.

Conclusions and perspectives

It is evident that the direct action of estrogens through ER is essential to the development of the prostatic gland and the pathogenesis/progression of prostatic diseases. However, unresolved issues still remain, such as the proper diets recommended for cancer prevention, and identification of ligands with specific activities for therapies, binding sites for ERβ and its “cross-talk” activation, and genetic polymorphism of ERα and related genes affecting cancer risk.

In conclusion, identification of effective SERM that strongly suppress ERα, as well as specific ligands that promote antitumor activities through the ERβ pathway, would significantly contribute to the prevention and treatment of prostatic diseases by targeting mechanisms other than AR-mediated signaling.

Conflict of interest

None declared.

Ancillary