Volume 124, Issue 9 p. 2050-2059
Cancer Cell Biology
Free Access

Inhibition of prostaglandin synthesis and actions by genistein in human prostate cancer cells and by soy isoflavones in prostate cancer patients

Srilatha Swami

Srilatha Swami

Department of Medicine, Stanford University School of Medicine, Stanford, CA

Search for more papers by this author
Aruna V. Krishnan

Aruna V. Krishnan

Department of Medicine, Stanford University School of Medicine, Stanford, CA

Search for more papers by this author
Jacqueline Moreno

Jacqueline Moreno

Department of Medicine, Stanford University School of Medicine, Stanford, CA

Search for more papers by this author
Rumi S. Bhattacharyya

Rumi S. Bhattacharyya

Department of Medicine, Stanford University School of Medicine, Stanford, CA

Search for more papers by this author
Christopher Gardner

Christopher Gardner

Department of Medicine, Stanford University School of Medicine, Stanford, CA

Search for more papers by this author
James D. Brooks

James D. Brooks

Department of Urology, Stanford University School of Medicine, Stanford, CA

Search for more papers by this author
Donna M. Peehl

Donna M. Peehl

Department of Urology, Stanford University School of Medicine, Stanford, CA

Search for more papers by this author
David Feldman

Corresponding Author

David Feldman

Department of Medicine, Stanford University School of Medicine, Stanford, CA

Fax: 650-725-7085.

Stanford University School of Medicine, 300 Pasteur Drive, Room S025, Stanford, CA 94305-5103, USASearch for more papers by this author
First published: 24 February 2009
Citations: 49

Abstract

Soy and its constituent isoflavone genistein inhibit the development and progression of prostate cancer (PCa). Our study in both cultured cells and PCa patients reveals a novel pathway for the actions of genistein, namely the inhibition of the synthesis and biological actions of prostaglandins (PGs), known stimulators of PCa growth. In the cell culture experiments, genistein decreased cyclooxygenase-2 (COX-2) mRNA and protein expression in both human PCa cell lines (LNCaP and PC-3) and primary prostate epithelial cells and increased 15-hydroxyprostaglandin dehydrogenase (15-PGDH) mRNA levels in primary prostate cells. As a result genistein significantly reduced the secretion of PGE2 by these cells. EP4 and FP PG receptor mRNA were also reduced by genistein, providing an additional mechanism for the suppression of PG biological effects. Further, the growth stimulatory effects of both exogenous PGs and endogenous PGs derived from precursor arachidonic acid were attenuated by genistein. We also performed a pilot randomised double blind clinical study in which placebo or soy isoflavone supplements were given to PCa patients in the neo-adjuvant setting for 2 weeks before prostatectomy. Gene expression changes were measured in the prostatectomy specimens. In PCa patients ingesting isoflavones, we observed significant decreases in prostate COX-2 mRNA and increases in p21 mRNA. There were significant correlations between COX-2 mRNA suppression, p21 mRNA stimulation and serum isoflavone levels. We propose that the inhibition of the PG pathway contributes to the beneficial effect of soy isoflavones in PCa chemoprevention and/or treatment. © 2008 Wiley-Liss, Inc.

Prostate cancer (PCa) is one of the most common forms of cancer affecting men in the US. Environmental factors, especially diet, have been shown to play a role in PCa risk. Epidemiological studies as well as in vitro and animal data suggest that an increase in dietary isoflavones lowers the risk of several cancers including PCa.1 Soy-derived isoflavones have a wide spectrum of biochemical activities including anti-cancer effects.1 Asian men who consume large quantities of soy-based foods have reduced mortality and lower incidence of PCa than their Western counterparts.2, 3 Furthermore, a rise in the incidence of PCa was reported among Asian immigrants to the US who adopted a Western diet, low in soy products.4-6 Genistein, an isoflavone that structurally resembles estradiol, is found in higher concentrations in soy than any other food.4 Genistein inhibits the growth of several types of cancer cells including PCa cells.7-9 Data from animal studies also suggest that genistein may be useful in the prevention and treatment of PCa.10-12 Several investigators have conducted pilot clinical trials examining the beneficial effect of the administration of soy isoflavones to PCa patients and some of these studies report decreases in the rate of rise in serum prostate specific antigen (PSA) levels due to soy isoflavones.13-15 At the cellular level genistein exerts multiple actions to mediate its anti-cancer effects including: binding to estrogen receptors, inhibition of protein tyrosine kinases, inhibition of both NF-kappa B activation and Akt signaling, induction of apoptosis, anti-oxidant effects and suppression of metastasis and angiogenesis.16, 17

Experimental evidence suggests that prostaglandins (PGs) are pro-inflammatory agents that play a key role in the carcinogenic process of many cancers including PCa18 through the stimulation of cell proliferation, inhibition of differentiation and apoptosis as well as the augmentation of tumour cell invasiveness, metastasis and mutagenesis.19 Genistein has been shown to decrease the synthesis of PGs in several normal and malignant cells.20-24 In the current study we report that in established human PCa cell lines and in cultured primary human prostate epithelial cells, genistein decreases the synthesis and biological actions of the PGs. Our results reveal that the major action of genistein on the PG pathway is the inhibition of the expression of prostaglandin G/H synthase/cyclooxygenase-2 (COX-2), the enzyme that catalyzes the synthesis of PGs leading to a decrease in the synthesis and secretion of PGs by the prostate cells. Importantly, our study shows that in vivo administration of moderate doses of soy isoflavones to PCa patients in the neo-adjuvant setting for 2–4 weeks before prostatectomy results in the suppression of COX-2 expression in prostatectomy specimens. The inhibition of PG synthesis and actions provides an additional mechanism for the growth repressive effects of genistein in PCa. Our observations suggest that genistein or soy may play a therapeutic role in the chemoprevention and/or treatment of PCa.

Abbreviations

COX-2, cyclooxygenase 2; EP4, prostaglandin receptor subtype EP4; EGF, epidermal growth factor; FP, prostaglandin receptor subtype FP; FBS, fetal bovine serum; GAPDH, glyceraldehyde 3 phosphate dehydrogenase; IGF-1, insulin-like growth factor 1; KGF, keratinocyte growth factor; PG, prostaglandin; PGE, prostaglandin E; PGF, prostaglandin F; 15-PGDH, 15-prostaglandin dehydrogenase; PMA, phorbal 12-myristic 13-acetate.

Material and methods

Materials

PGE2, PGF, and arachidonic acid were obtained from Cayman Chemical Co. (Ann Arbor, MI). Genistein, daidzein and phorbol 12-myristate 13-acetate (PMA) were from Sigma Aldrich (St.Louis, MO). Equol (a racemic mixture) was obtained from LC laboratories (Woburn, MA). Epidermal growth factor (EGF), insulin-like growth factor (IGF-1) and keratinocyte growth factor (KGF) were obtained from PeproTech (Rocky Hill, NJ). PD153035 and AG99 were from Calbiochem/EMD Biosciences (La Jolla, CA). Tissue culture media, supplements and fetal bovine serum (FBS) were obtained from GIBCO BRL (Grand Island, NY). The COX-2 promoter-luciferase plasmid containing approximately7 kb fragment of the human COX-2 promoter (−7140 to +123) cloned into the luciferase reporter vector, pGL3-basic, was a kind gift from Dr. Stephen Prescott (University of Utah, Salt Lake City, UT).

Cell culture

LNCaP and PC-3 cells were grown in RPMI 1640 medium supplemented with 5% FBS, 100 IU/ml of penicillin and 100 μg/ml streptomycin. Cells were maintained at 37°C with 5% CO2 in a humidified incubator. Primary cells were derived from radical prostatectomy specimens from men undergoing operation to treat PCa and propagated in culture as described previously.25 The normal cell cultures (E-PZ-1 to-5) were derived from peripheral zone tissue with no histological evidence of cancer in adjacent sections. The cancer cell cultures used (E-CA-1 (Gleason grade 3/4), E-CA-2 (Gleason grade 4/3) and E-CA-3 (Gleason grade 3) were derived from adenocarcinoma specimens. None of the patients had prior therapy or soy ingestion and were not the subjects in the pilot clinical trial for PCa.

Cell proliferation assays

LNCaP cells were seeded at an initial density of 1.5 × 105 cells/well in 6-well tissue culture plates and allowed to attach overnight in RPMI 1640 medium with 5% FBS. Cell cultures were shifted to medium containing 2% FBS and treated with either 0.1% ethanol vehicle or the indicated concentrations of drugs. Fresh media and drugs were replenished every other day. At the end of 6 days, cells were collected and processed for measurement of DNA content.26 The proliferation of primary epithelial cells was analysed in a high-density growth assay as described.27

RNA isolation and Real-Time RT-PCR

Total RNA was isolated from vehicle- or drug-treated cells using Trizol reagent (Invitrogen, Lifetechnologies, Carlsbad, CA) as previously described.28 Five microgram of total RNA were used in reverse-transcription (RT) reactions using the SuperScript III first strand synthesis kit (Invitrogen) and gene expression was analysed by real-time PCR using gene-specific primers.28 The reactions were carried-out with the DyNamo SYBR green qPCR kit (Finnzymes, New England Biolabs, Ipswich, MA) as described previously28 using an Opticon 2 DNA engine (Bio Rad, Hercules, CA). Melt curves were run with the PCR product to ascertain presence of a single peak. Relative changes in mRNA expression levels were assessed by the 2−ΔΔC(T) method.29 Changes in mRNA expression of the different genes were normalised to either TATA binding-protein (TBP) gene or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene as described previously.28

Measurement of PGE2 secretion

Subconfluent cell cultures were treated with vehicle or genistein for 72 hr. Conditioned media were collected and secreted PGE2 levels were quantitated using a PGE2 monoclonal Enzyme Immunoassay kit (Cayman Chemical) according to the manufacturer's protocol.

Western blot analysis

Cell lysates were prepared from vehicle- or genistein-treated cells and subjected to Western blot analysis as described earlier.28 The COX-2 monoclonal antibody (1:1,000 dilution) was purchased from Cayman Chemicals. β-actin monoclonal antibody (dilution 1:500) was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Immunoreactive bands were visualised using the enhanced chemiluminescence Western blot detection system (Amersham, Piscataway, N.J.) according to the manufacturer's instructions.

Transient transfections and luciferase assay

Approximately 7-kb fragment of the human COX-2 promoter fused to the luciferase reporter was transiently transfected into LNCaP cells using LipofectAmine reagent (Life Technologies). A renilla luciferase plasmid (pRLnull, Promega, Madison, WI) was co-transfected to control for transfection efficiency. After transfections the cells were treated with 0.1% ethanol vehicle control (Con) or 10 μM genistein (G) for 6 hr either in the presence or absence of EGF (100 ng/ml), IGF-1 (100 ng/ml), KGF (100 ng/ml) or PMA (100 ng/ml) added to the culture medium (RPMI + 5% FBS). Reporter and renilla luciferase activities were measured using the Dual luciferase assay kit (Promega) following the instructions of the manufacturer.

Soy isoflavone administration to PCa patients

Under a protocol approved by the Internal Review Board at Stanford, men diagnosed with PCa scheduled to undergo prostatectomy were enrolled in the study after providing informed consent. Eligibility criteria included: newly diagnosed adenocarcinoma by prostate needle biopsy, clinically localised disease, and no hormone therapy or radiation before operation. The study was a double-blinded randomised placebo controlled trial in which participants were asked to take study tablets for a minimum of 2 weeks before operation. Patients were randomly assigned to receive either a soy isoflavone supplement or a placebo. The isoflavone (Novasoy) tablets (27.2 mg isoflavone aglycones per tablet, 3 tablets/day) were provided and analysed for isoflavone content by Archer Daniels Midland Co (Decatur, IL). The isoflavones were present at a ratio of 1.0:1.3 of genistein to daidzein. The placebo pills were manufactured to appear identical to the isoflavone tablets. The tablets were distributed at clinic visits and residual tablets were counted to assess compliance. The clinical characteristics of the patients in the placebo and Novasoy group are given in Table I. Aliquots of the prostatectomy tissue (approximately 50 mg) were snap frozen and stored at −80°C. The prostatectomy specimens contained both cancer and normal areas analysed by histological analysis. RNA was isolated from the frozen tissue using the RecoverAll total nucleic acid isolation kit (Ambion, Austin, TX) and used for measurement of COX-2, 15-PGDH, p21/waf1 and GAPDH mRNA levels by real-time RT-PCR. At the end of the study, serum isoflavone levels were measured by Dr. Adrian Franke (Cancer Research Centre of Hawaii, Honolulu, HI) by liquid chromatography-mass spectrometry analysis.30

Table I. Patient Characteristics and Clinical Parameters
PLACEBO NOVASOY
Number of patients 12 13
Age (yr) 58 ± 2 57 ± 2
PSA (ng/ml) 6.6 ± 0.58 5.25 ± 0.63
Clinical tumor stage
T1c (n = 7) T1c (n = 9)
T2a (n = 4) T2a (n = 4)
T2b (n = 1)
Gleason score
6 (n = 3) 6 (n = 2)
7 (n = 5) 7 (n = 11)
9 (n = 4)
Recurrences (at 3 years) 5 1
Serum isoflavones (μM)
 Genistein 0.004 ± 0.0022 0.424 ± 0.110
 Daidzein 0.011 ± 0.0044 0.164 ± 0.036
 Equol 0.001 ± 0.0005 0.093 ± 0.065
Total isoflavones 0.016 ± 0.0033 0.785 ± 0.214

Statistical analysis

Statistical analyses were performed using GraphPad Prism 5. Data were evaluated by ANOVA with Scheffé's F test as the post-hoc analysis. In the patient trial, unpaired samples were assessed using both Mann Whitney test (nonparametric variables) and Student t tests with Welch correction (parametric variables). Comparisons between paired samples were made using linear regression analysis and determination of the Pearson's correlation coefficient.

Results

Regulation of COX-2 expression by genistein and other soy components

RT-PCR analysis revealed a significant inhibition of COX-2 mRNA expression in both PCa cell lines and primary prostate epithelial cells (Fig. 1a1c). COX-2 mRNA levels were suppressed in LNCaP cells (approximately 50% decrease) exposed to 10 μM genistein for 24 hr, whereas in PC-3 cells a more modest but statistically significant (approximately 38%) decrease was seen (Fig. 1a). We also examined the COX-2 mRNA expression in 4 different primary epithelial cell cultures derived from normal human peripheral zone prostate tissue (E-PZ-1-4, Fig. 1b) and 3 cell cultures derived from human PCa specimens (E-CA-1-3, Fig. 1c). Genistein at 10 μM caused significant decreases in COX-2 mRNA in 3 of the 4 normal cell cultures tested (E-PZ-1, -2 and -4) and in 2 of the 3 cancer-derived cell cultures (E-CA-1 and -3) after 24 hr of treatment. Panel D shows the results of a Western blot analysis of COX-2 protein expression in the normal primary cell cultures E-PZ-3-5. The data reveal appreciable decreases in immunoreactive COX-2 protein after 72 hr of treatment with 10 μM genistein in E-PZ-4 and -5 cells. We also examined the effect of 2 other major soy components, daidzein and its metabolite equol on COX-2 expression in LNCaP cells. When compared with vehicle-treated controls, 24-hr treatment of cells with 10 μM daidzein or equol also resulted in a significant (p < 0.01) down-regulation of COX-2 mRNA levels (data not shown). 15-PGDH mRNA levels did not appreciably change in LNCaP or PC-3 cells after 24 hr of genistein treatment. However, significant up-regulation of 15-PGDH mRNA was seen in 2 of the normal (E-PZ-1 and -3) and 2 of the cancer-derived (E-CA-2 and -3) primary cell cultures after 24 hr of genistein treatment (Fig. 2, Panels b and c).

Details are in the caption following the image

Effect of genistein on COX-2 expression in PCa cell lines and primary prostatic epithelial cells. Subconfluent cultures of LNCaP and PC-3 cells were treated with 0.1% ethanol vehicle (control, Con) or 10 μM genistein (G) for 24 hr. Total RNA was extracted and COX-2 mRNA levels were analysed by real-time RT-PCR analysis as described in Material and Methods. COX-2 mRNA were normalised to TBP mRNA levels and the ratios are given as fold change over control set at 1 (Panel a). Values represent Mean ± SE of at least 3 individual experiments. *p < 0.05 as compared with control. Primary cultures of prostatic epithelial cells obtained from the peripheral zone of normal prostate tissue (E-PZ) and adenocarcinoma (E-CA) were treated with either 0.01% ethanol vehicle (control, Con) or 10 μM genistein (G) for 24 hr. Total RNA was extracted and the expression of COX-2 mRNA was analysed. COX-2 mRNA levels were normalised to GAPDH mRNA levels and the ratios are given as fold change over control set at 1. Values represent Mean ± SE for measurements conducted in triplicate. *p < 0.05, **p < 0.01 and ***p < 0.001 as compared with control. COX-2 mRNA expression in normal prostatic epithelial cells is shown in panel b and that of cancer-derived epithelial cells is shown in panel c. Panel d is a representative Western blot showing COX-2 protein expression in primary normal prostatic epithelial cell cultures (E-PZ- 3–5) after 72 hr of treatment with 0.01% ethanol vehicle (control, C) or genistein at 10 μM (G). The expression of β-actin was used as a control.

Details are in the caption following the image

Effect of genistein on 15-PGDH mRNA expression in PCa cell lines and primary prostatic epithelial cells. The various PCa cell lines and primary cell strains were treated with genistein and analysed as described in Material and methods section. 15-PGDH mRNA levels were normalised to TBP or GAPDH mRNA levels and the ratios are given as fold change over control set at 1. Values represent Mean ± SE for measurements conducted in triplicate. *p < 0.05, **p < 0.01 and ***p < 0.001 as compared with control. 15-PGDH mRNA levels in PCa cell lines are shown in panel a, normal prostatic epithelial cells in Panel b and that of cancer-derived epithelial cells in panel c.

Regulation of EP and FP prostaglandin receptor mRNA expression by genistein

We examined the effects of genistein on the mRNA expression of the PG receptors EP2, EP4 and FP. Genistein significantly decreased the EP4 and FP mRNA levels by approximately 65% and approximately 45%, respectively, in LNCaP cells at the end of 24 hr of treatment (Fig. 3a). The reduction in EP2 mRNA was not statistically significant. Genistein did not significantly alter EP or FP receptor mRNA levels in PC-3 cells (Fig. 3b). In the case of primary cells, both EP2 and EP4 receptor mRNA could be detected in normal and cancer-derived cultures whereas FP mRNA was undetectable. Significant decreases after genistein treatment were seen in EP4 mRNA levels in 3 of the 4 normal cell cultures, and 2 of the 3 cancer-derived cell cultures tested (Fig. 3, Panels c and d). Genistein treatment did not change EP2 mRNA expression significantly in the primary cells (data not shown).

Details are in the caption following the image

Regulation of PG receptor expression by genistein in LNCaP, PC-3 and primary prostatic epithelial cells. Subconfluent cultures of LNCaP and PC-3 cells were treated with 0.1% ethanol (control, Con) or 10 μM genistein (G) for 24 hr. Total RNA was extracted and EP2, EP4 and FP mRNA expression were analysed using real-time RT-PCR as described in Material and methods. Values were normalised to TBP mRNA levels and the ratios are given as fold change over control set at 1 (Panel a and b). Values represent Mean ± SE of at least 3 individual experiments. *p < 0.05 and ***p < 0.001 as compared with control. The primary cell cultures were treated with genistein and analysed for EP4 receptor levels as described earlier. EP4 PG receptor mRNA levels were normalised to GAPDH mRNA levels and the ratios are given as fold change over control set at 1. Values represent Mean ± SE from 3 experiments. *p < 0.05, **p < 0.01 and ***p < 0.001 as compared with vehicle. EP4 mRNA expression in normal epithelial cells is shown in panel c and that in cancer-derived epithelial cells is shown in panel d.

Effect of genistein on prostaglandin levels

As a reflection of the effects of genistein on the expression levels of the PG synthesizing enzyme COX-2 and the PG catabolizing enzyme 15-PGDH, we examined PGE2 production and secretion into the conditioned media from cell cultures exposed to 10 μM of genistein for 48 hr. As shown in Figure 4a, genistein caused significant reductions in PGE2 secretion in LNCaP cells (approximately 60%), PC-3 cells (approximately 75%) and in the primary epithelial cell cultures (approximately 40–60%).

Details are in the caption following the image

Effect of genistein on PGE2 levels and cell growth. Panel a: Subconfluent cultures of LNCaP, PC-3 and the primary normal prostatic epithelial cells E-PZ-1 and E-PZ-5 were treated with 0.01% ethanol vehicle (control, Con) or 10 μM genistein (G) for 72 hr. Conditioned media from vehicle and genistein-treated cultures were collected and PGE2 levels were analysed using an EIA kit as described in Material and methods. Values are given as a percent of PGE2 levels in vehicle-treated control set at 100%, which corresponded to 0.262 ± 0.04 pg/μg protein in LNCaP, 5.9 ± 1.2 pg/μg in PC-3, 175 ± 8.6 pg/μg in E-PZ-1 and 0.029 ± 0.003 pg/μg in E-PZ-5 cells. Values represent Mean ± SE from 3 experiments. *p < 0.05 and **p < 0.01 as compared with control. Panel b: Primary prostatic epithelial cells derived from normal human prostate tissue were seeded and cultured in defined growth media containing vehicle (0.01% ethanol) or 10 μM genistein. Cell growth at the end of 10 days was analysed by a high-density growth assay as described in Material and methods. Values represented are Mean ± SE from 3 experiments. ***p < 0.001 as compared with control. Panel c: Semiconfluent LNCaP cells cultures in RPMI + 2% FBS were treated with 0.1% ethanol vehicle (control, Con), arachidonic acid (AA, 3 μM), PGE2 (10 μM) or PGF (10 μM) in the absence or presence of 10 μM genistein (G) added to the media. Fresh media and the various agents were replenished at the end of 3 days. Cell growth was analysed at the end of 6 days by measuring DNA content as described in Material and methods. Values are expressed as μg of DNA/well and represent Mean ± SE of at least 3 experiments. *p < 0.05, **p < 0.01 and ***p < 0.001 when compared with vehicle. +++p < 0.001 when compared with AA, PGE2 or PGF alone.

Effects of genistein on basal and prostaglandin-stimulated cell growth

The effects of genistein on the growth of the 3 normal primary prostate epithelial cell cultures E-PZ -1, -3 and -5 were analysed. As shown in Figure 4b, addition of 10 μM genistein to the cultures caused highly significant inhibition of the growth of all the cell cultures tested (approximately 70–85% inhibition as compared with control). We also analysed the effect of genistein on the stimulation of LNCaP cell growth by exogenous PG addition and by endogenous PGs derived from the substrate arachidonic acid added to the culture medium. LNCaP cells were treated with arachidonic acid (3 μM), PGE2 or PGF (10 μM each) in the absence or presence of 10 μM genistein for 6 days and growth was assessed by measuring DNA. Our results revealed a moderate but statistically significant growth stimulation (p < 0.05) because of the addition of arachidonic acid and a more pronounced growth stimulation (p < 0.01–0.001) by the PGs (Fig. 4c). Genistein (10 μM) had a marked growth inhibitory effect (p < 0.001) on the basal (vehicle-treated) cell growth. In addition, genistein completely blocked the growth stimulation due to endogenous PGs derived from the added arachidonic acid and exogenous PG addition (Fig. 4c).

Modulation of COX-2 promoter-luciferase activity by genistein

To analyse whether genistein decreases COX-2 gene expression by a direct transcriptional repression, we transiently transfected into LNCaP cells a plasmid containing approximately 7-kb fragment (−7140 to +123) of the human COX-2 promoter cloned into the pGL3-basic vector.31 Luciferase activity was analysed both under basal conditions and after treatment with known stimulators of COX-2 expression with and without genistein co-treatment and the findings are shown in Figure 5a. Genistein had no effect on basal COX-2-luciferase activity. Addition of the growth factors EGF, KGF and IGF-1 or the phorbol ester PMA caused significant increases in COX-2-luciferase activity (approximately 2-fold and approximately 3-fold over basal levels). Co-addition of genistein completely abolished EGF-, KGF- and IGF-1-stimulated COX-2-luciferase. However, the PMA-induced increase in COX-2-luciferase was not significantly affected by genistein co-treatment. As the receptors for EGF, KGF and IGF-1 exhibit tyrosine kinase activity and genistein is an inhibitor of protein tyrosine kinases, we examined the effect of known tyrosine protein kinase inhibitors on basal and EGF-stimulated COX-2-luciferase activity. As shown in Fig. 5b, addition of EGF caused a approximately 2-fold increase in COX-2-luciferase activity over basal levels. The tyrosine protein kinase inhibitors AG99 and PD153035 abolished EGF stimulation of COX-2-luciferase. We also analysed the effect of PD153035 on COX-2 mRNA expression in LNCaP cells in growth medium containing 5% FBS. After treatment with 10 μM PD153035 for 24 hr, a highly significant decrease in COX-2 mRNA levels was seen in treated cells compared with cells exposed to the vehicle (Fig. 5c).

Details are in the caption following the image

(a) Effect of genistein on COX-2 promoter activity in LNCaP cells. A reporter construct containing approximately 7-kb fragment of the human COX-2 promoter fused to the luciferase reporter was transiently transfected into semiconfluent cultures of LNCaP cells as described in Material and methods. A renilla luciferase plasmid was co-transfected to control for transaction efficiency. After transfections, the cells were treated with 0.1% ethanol vehicle (control, Con) or 10 μM genistein (G) either in the presence or absence of EGF (100 ng/ml), PMA (1 μg/ml) for 6 hr and KGF (100 ng/ml) or IGF-1 (100 ng/ml) for 24 hr. Luciferase activities were measured using the Dual luciferase assay kit. Results are expressed as relative luciferase units (RLU), which represent the ratio of reporter luciferase to the renilla luciferase. Values are given as percent of control RLU set at 100%. *p < 0.05, **p < 0.01 and ***p < 0.001 when compared with vehicle and +p < 0.05, ++p < 0.01 as compared with EGF, IGF or KGF alone. (b) Effect of tyrosine kinase inhibitors on COX-2 promoter activity. After transfections (as described in Fig 4), cells were treated with 0.1% ethanol vehicle (control, Con) or 100 ng/ml EGF in the absence or presence of the tyrosine kinase inhibitors AG99 (10 μM) and PD153035 (10 μM) for 6 hr. Values are given as percent of control RLU set at 100%. ***p < 0.001 when compared with vehicle and +p < 0.05 as compared with EGF treatment. (c) Effect of PD153035 on COX-2 mRNA expression. LNCaP cells in culture media containing 5% FBS were treated with 10 μM PD153035 for 24 hr. RNA was isolated and the expression of COX-2 mRNA was analysed as described in Figure 1. COX-2 mRNA was normalised to TBP mRNA levels and the ratios are given as fold change over vehicle treatment (control, Con) set at 1. ***p < 0.001 as compared with control. All values represent Mean ± SE of 3–6 experiments.

Effect of soy isoflavone administration on gene expression in prostate surgical specimens from PCa patients

To analyse whether the effects we demonstrated in cell culture were also relevant to human subjects, PCa patients scheduled for prostatectomy were given placebo or Novasoy tablets (27.2 mg isoflavone/tablet, 3 tablets/day) for 2 weeks before operation. The assignment of patients to the placebo or isoflavone-treated groups was randomised and double blinded. The characteristics of the patients and serum isoflavone levels are shown in Table I. Serum total isoflavone levels in the group taking the Novasoy tablets were approximately 50-fold higher than the placebo group (Table I). We measured the mRNA expression of COX-2, 15-PGDH and the cell cycle inhibitor p21/Waf1 in the prostatectomy specimens from the 2 groups. The expression of GAPDH mRNA was used as a control. A statistically significant decrease in COX-2 mRNA levels (p < 0.01) was seen in the isoflavone-treated group (Fig. 6a) when compared with the placebo group. In this group of patients, COX-2 mRNA levels exhibited a significant (p < 0.05) negative correlation with serum isoflavone levels achieved by Novasoy administration (Fig. 6b). The expression of 15-PGDH mRNA did not show a statistically significant change when placebo and soy isoflavone-treated samples were compared (data not shown). However, when compared with placebo-treated patients (Fig. 6c), the soy isoflavone-treated patients showed a statistically significant increase in p21 mRNA levels (p < 0.01). In the isoflavone-treated group, p21 mRNA levels exhibited a significant (p < 0.05) positive correlation with serum isoflavone levels (Fig. 6d). Similar analyses examining the correlation of COX-2 or p21 mRNA expression in the prostate to serum isoflavone levels was not possible in the placebo group because of the very low levels of isoflavones in the serum samples in these patients. We also measured COX-2 protein levels by Western blot in extracts from some of the prostatectomy specimens from the placebo- and soy isoflavone-treated groups. As shown in Figure 6e, COX-2 protein expression in prostate tissue extracts from patients ingesting Novasoy (samples S1–S4 exhibiting COX-2 mRNA levels of 0.04–0.5 arbitrary units) was appreciably lower than from patients on placebo (samples P1 and P2 exhibiting COX-2 mRNA levels of 2 and 9 arbitrary units).

Details are in the caption following the image

Analysis of gene expression in prostatectomy samples from PCa patients ingesting placebo or Novasoy. Panel a. RNA was isolated from frozen tissues of prostatectomy specimens from PCa patients ingesting placebo (placebo, closed squares, n = 12) or Novasoy tablets (Soy, 30 mg/tablet, 3 tablets/day, closed triangles, n = 13) for 2–4 weeks before operation. COX-2 mRNA levels were analysed by real-time RT-PCR, using specific primers as described in Material and methods. The expression of GAPDH gene is used as the control. Patients who had tumour recurrence at the end of 3 years are denoted with a circle around the symbol. Panel b. Correlation of prostate COX-2 mRNA levels to serum isoflavone levels in patients ingesting Novasoy. Panel c. p21 mRNA levels in the prostatectomy samples. Panel d. Correlation of prostate p21 mRNA levels to serum isoflavone levels in patients ingesting Novasoy. Panel e. Western blot showing COX-2 expression in extracts of tissue samples from patients ingesting placebo (P1 and P2) or soy (S1–S4). Positive and negative controls include extracts from COS-7 cells transfected with COX-2 expression plasmid and COX-1 expression plasmid or vector plasmid respectively. The expression of β actin was used as a control.

Discussion

Genistein inhibits growth and induces apoptosis in several human PCa cells.7, 8 Multiple molecular pathways appear to be involved in the anti-cancer effects of genistein.17 These include the inhibition of cell cycle progression through the up-regulation of the cyclin-dependent kinase inhibitor p21,7, 9 down-regulation of cyclin expression,7 inactivation of NF-κB16, 22 and modulation of the androgen receptor.32, 33 Inhibition of tyrosine phosphorylation, particularly the expression and phosphorylation of EGF receptors in the prostate, may also play a role.34 Our study shows that the regulation of the PG pathway is an additional mechanism by which genistein exerts its anti-proliferative effects on human prostate cells.

The main action of genistein on the PG pathway appears to be the suppression of COX-2 expression as seen in the PCa cell lines (both LNCaP and PC-3) and in most of the primary cell cultures analysed. We examined the effects of genistein in multiple isolates of primary prostate epithelial cells to account for the variations between the cell cultures as each culture is derived from an individual prostatectomy sample from a PCa patient. Because these cells are not immortalised, the cell cultures used were different in the various experiments depending on the availability at the time of analysis. In 1 of the normal primary cell cultures (E-PZ-1), we were able to carry out a complete assessment of the changes in all the key PG pathway genes, PGE2 secretion and cell growth. However, it should be noted that the primary cell cultures used in the in vitro experiments were from untreated patients and were not derived from the patients who participated in the pilot clinical study described in this report.

COX-2 basal expression is usually low to negligible in nearly all normal tissues, however, it can be induced by multiple factors including hormones, growth factors, cAMP, phorbol esters, inflammatory factors and cytokines.35 In our studies we found appreciable COX-2 expression in normal primary cell cultures presumably because the cultures were grown in defined media containing hormones, growth factors and cAMP inducers.36 Genistein treatment caused significant decreases in both COX-2 mRNA and protein levels in normal as well as cancer-derived primary cells and established PCa cell lines. The magnitude of COX-2 mRNA suppression was also higher in LNCaP cells, which represent a less aggressive and earlier stage of cancer when compared with the more aggressive and invasive PC-3 cells. There also seemed to be a greater degree of response among the primary cultures derived from normal prostate tissue than those derived from cancerous prostate tissue. Interestingly, a significant increase in 15-PGDH mRNA was observed only in the primary cultures and not in the established PCa cell lines. The down-regulation of the PG receptor mRNA was seen in LNCaP cells and most of the primary cell cultures but was not seen in PC-3 cells. Taken together our data revealed that the primary cultures (both normal and cancer) and LNCaP cells (representative of a more differentiated or an earlier stage cancer) were more responsive to genistein treatment than the more aggressive and invasive PC-3 cells (representative of later stage cancer). The findings suggest that the use of genistein might be more beneficial earlier in the course of PCa.

The decrease in secreted PGE2 levels seen in these cells mostly reflects the effect of genistein to inhibit PG synthesis by decreasing COX-2 expression. The ability of genistein to inhibit arachidonic acid- stimulated LNCaP cell growth is due to its ability to inhibit endogenous PG synthesis from arachidonic acid by decreasing COX-2 expression. Furthermore genistein abolished the growth stimulatory effects of added PGs indicating that the down-regulation of the PG receptors played a role in its action to inhibit the biological actions of PGs. Genistein has been shown to decrease PGE2 synthesis through COX-2 suppression in a variety of normal and malignant cells.20-24 Our study is the first demonstration of the inhibition of COX-2 expression by genistein in prostate cells. Decreases in COX-2 mRNA and protein were seen in all the cell models that we studied. Importantly, we also showed significant decreases in COX-2 expression in prostatectomy specimens after isoflavone administration to PCa patients.

COX-2 is regarded as a pro-inflammatory molecule that catalyzes the synthesis of PGs, which in many cancer cells exhibit growth stimulation, increased adhesion to extracellular matrix, resistance to apoptosis and stimulation of angiogenesis.19 Several studies investigating COX-2 expression in human PCa reported increases in COX-2 expression in PCa and in high-grade prostate intraepithelial neoplasia (PIN).37, 38 Although 2 other studies did not find a consistent increase in COX-2 expression in established PCa,39, 40 they did find COX-2 over-expression in proliferative inflammatory atrophy (PIA) lesions, which are thought to be the precursors of PIN.40 More recently, increased COX-2 expression in prostate was reported in multiple studies of PCa patients.41-43 We also found measurable levels of COX-2 mRNA and protein in human PCa cells.28 Most investigators agree that local production of PGs by inflammatory cells increases the risk of prostate carcinogenesis and/or PCa progression.39, 40, 43, 44 In our study of prostatectomy specimens we used represent active tissue that was a heterogeneous mixture of both normal and tumour cells infiltrated with inflammatory cells. We demonstrate that in these prostatectomy specimens, COX-2 is elevated in most samples and that genistein suppresses COX-2 mRNA and protein expression. The data suggest that genistein, through its ability to down-regulate COX-2 expression, may exhibit a beneficial chemopreventive and /or therapeutic effect in PCa regardless of whether the COX-2 is being produced by the tumour cells or the infiltrating inflammatory cells.

Genistein is a phytoestrogen that can bind to the estrogen receptors (ER) α and β and regulate gene transcription.45 Prostate tissue is known to express ERβ45 and it has been shown that genistein has greater affinity for ERβ than ERα.45 However, an investigation of the effects of genistein on ERβ is beyond the scope of this article.

Our studies on the effect of genistein on COX-2 promoter activity revealed that genistein abolished the activation of COX-2 promoter by growth factors with known receptor tyrosine kinase activities. Further, tyrosine kinase inhibitors (AG99 and PD153035) mimicked the effects of genistein to suppress EGF-stimulated COX-2 mRNA and promoter activity. These data suggest a contributing role for protein tyrosine kinase inhibition in the down-regulation of COX-2 expression by genistein.

In vitro studies using genistein have shown that concentrations between 1 and 10 μM are required to inhibit growth of PCa cell lines.46-48 In our study we found a dose dependent decrease in COX-2 mRNA with 1 μM genistein showing a approximately 30% decrease (data not shown) and 10 μM showing a 50% decrease. The decrease with the lower concentration of genistein, however, was not statistically significant suggesting that concentrations >1 μM are needed to see a significant effect on COX-2 gene expression. Based on studies of isoflavone administration to human subjects the highest serum genistein levels achievable by ingesting soy-rich diets appears to be 1–5 μM.49-51 However, studies in mice suggest that the concentration achieved within the prostate after genistein administration is about 10-fold higher than serum genistein levels.52 In patients receiving oral isoflavone supplements the prostate tissue levels of the individual isoflavones were at least 2 to 3-fold higher than their levels in plasma.51 In the current study we investigated the effects of administering moderate amounts of soy isoflavones (Novasoy) to PCa patients in a neo-adjuvant setting for approximately 2 weeks between diagnosis and prostatectomy. Soy administration for this short period of time resulted in serum total isoflavone levels of approximately 0.785 μM. Although these levels were lower than those reported in some human studies cited above49-51 and the sample size was modest, our analysis revealed a statistically significant decrease in COX-2 expression in the isoflavone-treated prostate samples (p < 0.01), probably indicating achievement of higher concentrations within the prostate. In the current study soy administration to patients resulted in an approximately 6-fold increase in prostate tissue levels of total isoflavones and an approximately 4-fold increase in genistein levels when compared with their concentrations in plasma.53 These observations suggest that although the achievable range for serum isoflavones is approximately 1–5 μM, the intraprostatic concentration is much higher approaching levels at which genistein exerts its biological effects on the PG pathway. Additionally, it is also important to note that although our in vitro studies examined the effects of genistein as a single agent, the effects seen in prostatectomy samples might reflect the synergistic and/or cumulative effect of all the isoflavones whose levels were elevated in the patients after Novasoy administration. Importantly, we found a significant negative correlation between prostate COX-2 mRNA and serum isoflavone levels in the isoflavone-treated group, suggesting that modest doses of soy supplements exhibit an anti-inflammatory effect in PCa patients.

As shown in Table I, 5 patients in the placebo group and 1 patient in the isoflavone group exhibited disease recurrence at the end of 3 years of study. Further analyses of the patient characteristics revealed that although the assignment of patients to the placebo or treated groups was double-blinded and randomised, it was unfortunate that all the patients with Gleason scores >7 were in the placebo group offering a plausible explanation for the higher recurrence rate in the placebo-treated patients. Interestingly, however, prostate COX-2 mRNA levels were very high in the 3 of 4 patients in the placebo group who exhibited disease recurrence and 4 of 4 with Gleason scores of 9. This observation is in agreement with the findings of Cohen et al.41 who showed that elevated COX-2 expression was associated with high Gleason score, poor prognosis and high rate of recurrence in PCa patients.

Isoflavone administration also resulted in significant increases in the mRNA expression of the cell cycle inhibitor p21 compared with the placebo group. The gene expression levels, however, varied among the individual patients. The observed positive correlation between prostate p21 mRNA and serum isoflavone levels is supportive of an antiproliferative effect of the administered isoflavones. In the placebo group, we generally found an inverse association between COX-2 mRNA and p21 mRNA levels. A similar trend was seen in the soy treated group, where all patients had a suppressed COX-2 and elevated p21 levels compared with the placebo group (See Fig. 6). Interestingly, the patient who had the highest p21 mRNA had undetectable levels of COX-2 mRNA.

The variations in expression of genes such as COX-2 and p21 within the same experimental group (placebo or soy isoflavone group) may arise due to a number of factors such as the differences in the cellular composition of the tissue samples we studied (epithelial versus stromal cells and normal versus malignant epithelial cells) and the presence or absence of inflammatory cells. However, it is important to note that in spite of these and other possible variations, COX-2 expression was consistently and significantly lower in the samples from patients ingesting soy isoflavones, suggesting that soy isoflavones are likely to suppress COX-2 expression in more than 1 cell type within the prostate. The results indicate that ingesting modest amounts of soy can achieve adequate concentrations of genistein and other isoflavones in the prostate to mediate beneficial actions such as suppression of inflammation and proliferation by modulating gene expression within the prostate. Overall, the findings suggest that by decreasing the synthesis and actions of PGs, genistein/soy may exert a beneficial effect to inhibit the development and/or progression of PCa.

Acknowledgements

This work was supported by NIH grant DK 42482 and AICR grant 06A114 (to D.F.), NIH grant AT00486 (to C.G.) and DAMD grant W81XWH-05-1-0111 (to D.M.P.). J.M. was supported by the Department of Army fellowship PC040120. We thank Dr. Stephen Prescott, University of Utah, for providing the COX-2 promoter-luciferase plasmid, and Dr. Adrian Franke (Cancer Research Centre of Hawaii, Honolulu, HI) for serum isoflavone measurements and analysis of NovaSoy tablets.

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.