The Prostate 66:557^566 (2006)

Prostatic FluidConcentrations of Isoflavonoids inSoyConsumersAre Sufficientto InhibitGrowthof

BenignandMalignantProstaticEpithelialCells InVitro

Tammy E. Hedlund,1* Adrie van Bokhoven,1 Widya U. Johannes,1

Steven K. Nordeen,1 and Lorraine G. Ogden2

1Departmentof Pathology,TheUniversityof Colorado atDenverandHealth Sciences Center,SchoolofMedicine, Aurora,Colorado

2Departmentof PreventiveMedicineandBiometrics,TheUniversityof ColoradoatDenverandHealth SciencesCenter,SchoolofMedicine, Aurora,Colorado

BACKGROUND. The differential intestinal metabolism of the soy isoflavones is likely toinfluence the ability of soy to prevent prostate cancer.While daidzein, genistein, and equol havedirect antiproliferative effects on prostatic epithelial cells in vitro, there are no such data forthe isoflavone glycitein, or seven metabolites: O-desmethylangolensin (ODMA), 6-hydro-xyODMA (6H-ODMA), dihydrodaidzein (DHD), cis-4-hydroxyequol (C4HE), 30-hydroxydaid-zein (3HD), 6-hydroxydaidzein (6HD), and 8-hydroxydaidzein (8HD). In the current study, thein vitro activities of these compounds were elucidated, and the active ranges of concentrationswere compared to that found in Caucasian prostatic fluid (PF) and plasma samples.METHODS. The effects of isoflavonoids on cell growth, cell cycle distribution, and apoptosis(active Caspase 3) were examined on benign prostatic epithelial cells (PrEC), and the prostatecancer cell line LNCaP.RESULTS. PF concentrations of genistein, equol, and daidzein (but not ODMA or DHD) wereoften within the ranges that reduce PrEC growth in vitro. Profound differences in sensitivitieswere observed with LNCaP. The hydroxydaidzeins, C4HE, and 6H-ODMA had significantinhibitory effects at 10�5MonPrECgrowth (but not LNCaP).Glycitein had significant effects onboth. Reductions in cell growth were typically associated with both changes in cell cycledistribution and Caspase 3 activation. When five isoflavonoids were used in combination atconcentrations present in PF samples, synergistic effects were observed.CONCLUSION. The profound differences in sensitivities of prostatic epithelial cells to thesecompounds alongwith their synergistic effects suggest thatmultiplemetabolites in vivomaybeoptimal for preventing prostate cancer. Prostate 66: 557–566, 2006. # 2005 Wiley-Liss, Inc.

KEY WORDS: daidzein; equol; O-desmethylangolensin; dihydrodaidzein; hydroxy-daidzein; hydroxyodma; cis-4 hydroxyequol; prostate cancer prevention;apoptosis

INTRODUCTION

Soy consumption has been linked to reduced risksfor prostate cancer in multiple studies [1–3]. Severallines of evidence suggest that the isoflavones genisteinand daidzein are involved. These compounds demon-strate many properties that make them good candi-dates for anticancer agents including their antioxidantcapacity [4], their favorable effects on steroid hormonebiosynthesis [5], transport [6], and metabolism [7], and

Grant sponsor: Department of Defense; Grant number: #PC020275;Grant sponsor: Grove Foundation and the San Francisco Foundation;Grant number: # PN0112-084.

*Correspondence to: Tammy E. Hedlund, Department of Pathology,Mail Stop 8104, RC1-South, Rm. L18-5118, 12801 East 17th Ave., P.O.Box 6511, Aurora, CO 80045. E-mail: tammy.hedlund@UCHSC.eduReceived 23 September 2005; Accepted 17 October 2005DOI 10.1002/pros.20380Published online 21 December 2005 in Wiley InterScience(www.interscience.wiley.com).

� 2005 Wiley-Liss, Inc.

their direct antiproliferative effects on many cell types[8–10]. However, nutritional studies reveal greatercomplexity to the paradigm that dietary soy protectsagainst prostate cancer. The isoflavone daidzein isdifferentially metabolized by the intestinal flora in dif-ferent individuals [11]. Three well-known metabolitesare dihydrodaidzein (DHD), O-desmethylangolensin(ODMA), and equol [12–14]. Several less-studiedcompounds have also been reported, including thepresumed intestinal reductive metabolites 6-hydroxy-ODMA (6H-ODMA) and cis-4-hydroxyequol (C4HE)[14], as well as the oxidative metabolites 30hydro-xydaidzein (3HD), 6-hydroxydaidzein (6HD), and8-hydroxydaidzein (8HD) that are likely produced bythe liver [15]. However, to date there are no dataaddressing thenatural ranges of concentrations of thesemore recently discovered compounds in humans.

Some of the first intestinal metabolites discoveredhave remarkable differences in their antioxidantcapacities, estrogenicities, and abilities to alter cellgrowth [4,16,17]. This is likely to be true for the morerecently discovered metabolites as well. An importantconcern is that each person consuming soy may beproducing different metabolites, and this is likely tocreate significant variation in the potential healthbenefits. For this reason, it is important to determinewhich metabolites are likely to be the most biologicallyactive, and perhaps the most potent at preventingprostate cancer.

Research from our laboratory and others suggestthat daidzein, genistein, and equol have direct anti-proliferative effects on prostate cancer cells in vitro[17–19]. However, many gaps in the current literatureexist. First, there are no reports of the effects of ODMAand DHD on prostatic epithelial cell growth orapoptosis. This is important given how commonlythese twometabolites are produced [11,20]. Second, it isnot clear if the active concentrations in vitro are withinthe natural ranges of concentrations that exist in theprostate gland after soy consumption. While isoflavo-noid concentrations have been measured in tissuehomogenates from surgical prostatectomy specimensfromprostate cancer patients [21], this is not possible instudies with healthy men. Morton et al. [22] were thefirst to use prostatic fluid (PF) as a less-invasive meansfor estimating prostatic levels of daidzein and equol. PFis obtained by the rectal massage of the prostate, and isconsidered to be a relatively pure source of glandularsecretions, unlike ejaculate. Any compound in PFmusthave passed through the basal and secretory epithelialcells that are prone to malignancy, and thus may betterrepresent what these epithelial cells are exposed to.

In our recent nutritional study on soy metabolismin healthy Caucasian men, the natural ranges of fivesoy isoflavonoids were simultaneously determined in

plasma and PF. Of importance, daidzein and itsintestinal metabolites (DHD, equol, and ODMA) werepresent at significantly higher concentrations in PFthan in plasma after soy consumption [20]. In contrast,mostmenhad lower concentrations of genistein in theirPF than plasma. The reason for this remains unclear.The ranges of isoflavonoid concentrationswere used inthe current study to determine if the antiproliferativeeffects observed in vitro occur near the ranges ofconcentrations that occur in vivo. Additionally, bycombining individual isoflavonoids at the concentra-tions present in several PF samples, we were able toaddress the possibility of synergy in terms of growth-inhibition.

There were four major goals for the current study:(1) to compare the dose-dependent effects of five isofla-vonoids on the growth of benign prostatic epithelialcells (PrEC) to the in vivo concentrations present inplasma and PF of Caucasianmen consuming soy; (2) toassess for the first time the relative potencies of severaladditional isoflavonoids on PrEC and LNCaP cellgrowth (including glycitein, 6H-ODMA, 3HD, 6HD,8HD, and C4HE); (3) to determine if these compoundsact by blocking cell cycle progression or by inducingapoptotic cell death; (4) to determine if these com-pounds are likely to act in an additive or synergisticmanner with each other when used at concentrationspresent in actual PF samples.

MATERIALSANDMETHODS

Plasma and PFSamples

Werecently completed a nutritional study in healthyCaucasian men [20] to determine how routine dietaryhabits affect the intestinal metabolism of daidzein, aswell as the ability of the prostate to concentrate thesemetabolites. This study was approved by the ColoradoMultiple Institutional Review Board and was compli-ant with the Health Insurance Portability andAccount-ability Act of 1998. Twenty-five men had routinelyconsumed high levels of soy (�30 mg/day isoflavonesfor at least 2 year), and 20 had routinely consumed lowamounts of soy (�5 mg/day isoflavones for at least2 year). All men consumed a soymilk-based beveragedaily for 7 days containing 42–60 mg isoflavones.Approximately 5–6 hr after consumption of the last soydrink, blood was drawn and PF collected by prostaticmassage. Plasma was isolated and all samples werestored at�708C until analyzed for isoflavonoid quanti-tation using reverse-phase HPLC with electrosprayionization and mass spectrometry (performed at theUniversity of Alabama at Birmingham under thedirection of Dr. Stephen Barnes). Detailed methodshave been previously reported [20].

The Prostate DOI 10.1002/pros

558 Hedlundet al.

Isoflavonoids

Glycitein was purchased from MP Biomedicals(Irvine, CA). Equol, C4HE, DHD, genistein, anddaidzein were purchased from LC Laboratories,Woburn, MA. ODMA, 3HD, 6HD, 8HD, and 6H-ODMAwere purchased from Plantech, Inc. (Berkshire,England,UK). Since the hydroxydaidzeinswere shownto be unusually labile once resuspended (unpublishedHPLC-mass spectrometry data, Jeevan Prasain, D. RayMoore, and Stephen Barnes at the University ofAlabama at Birmingham), these were prepared freshwith each experiment. Stock solutions of C4HE andglycitein were prepared at 10�2M in dimethylsulfoxide(DMSO), andwerediluted 1:10 serially to obtain 1,000�stocks for dilution into tissue culture medium suchthat the final vehicle concentration was 0.1% (V:V).All other isoflavonoids (except daidzein) were pre-pared in ethanol at 10�2M, and diluted serially asabove. Daidzein was prepared in ethanol at 5� 10�3M(due to its reduced solubility) and was serially diluted1:10 to obtain 500� stock solutions with a finalethanol concentration in tissue culture medium of0.2% (V:V). All isoflavonoid stocks were stored at�208C in amber vials andwere usedwithin 2months ofpreparation.

Cell Cultures

The LNCaP human prostate cancer cell line waskindly provided by Dr. J. Horoszewicz, Roswell ParkMemorial Institute, Buffalo, NY [23,24]. This establish-ed cell line was derived from a lymph node metastasisfrom a patient with hormone resistant prostate cancer.LNCaP cells (passage #28–35) were split at a 1:10dilution weekly in RPMI 1640 (Invitrogen Life Tech-nologies, Carlsbad, CA) with 10% fetal bovine serumwithout the use of antibiotics. Cellswere kept at 378C inhumidified incubators in air supplemented with 5%CO2. Authenticity of these cells has been confirmedby cytogenetic analysis and DNA profiling [25], andperiodic testing confirms that they remain free frommycoplasma contamination. Primary cultures ofbenign human prostatic epithelial cells (PrEC) wereobtained from Cambrex Clonetics Corp. (Walkersville,MD) and were propagated as directed in PrEGMmedium, omitting the antibiotic gentamycin and thefungicide amphotericin. PrEC were derived fromtwo different Caucasian males (ages 39 and 28), andwere utilized between passages #3 and 7, while stillactively dividing. Very similar results were ob-tained from the two cultures. It is generally believedthat PrEC represent the basal epithelial cell populationand not the secretory/luminal cells of the prostate[26,27].

GrowthAssays

Cells were trypsinized and counted using a hema-cytometer. Tissue culture dishes (24-well) were seededwith 103 cells/well (PrEC) or 104 cells/well (LNCaP).After 24 hr, one row of four wells was harvested for the‘‘day 0’’ measurement of DNA content, as previouslydescribed [28]. For the remaining cells, media werereplaced with that containing the appropriate isoflavo-noids or ethanol vehicle. [For PrEC experiments,isoflavonoids were added in a final ethanol concentra-tion of either 0.1% or 0.2% (V:V). For PF simulationexperiments with LNCaP cells, the final ethanolconcentration varied for each participant’s profile.Thus each required their own vehicle control. Forexample, the combination of isoflavonoids in the PF ofparticipant #76 required 0.78% ethanol, and the 1:10dilution of PF #76 required 0.078% ethanol.] Mediawere replaced every other day and cellswere harvestedafter 7–9 days when the control wells were approxi-mately 85% confluent. The DNA contents of themonolayers were quantitated using a Hoechst 33258fluorescence assay with a Dynex Fluorolite 1000fluorescence plate reader (Dynex Technologies, Inc.,Chantilly, VA), as previously described [28]. MeanDNA concentrations, SEM, and IC50 values werecalculated for quadruplicate samples in each conditionusingGraphPad Prism 4 (GraphPad Software, Inc., SanDiego, CA). Data were normalized and expressed as %of control, such that 0% growth¼ the mean DNAcontent at day 0, and 100% growth¼ the mean DNAcontent of the vehicle control after treatment. Forstatistical comparisons, variance across groups wasmeasured using one-way ANOVA. When significantchanges were observed (P< 0.05) Dunnett’s post-testwas applied to determine which groups differedsignificantly from the control.

FlowCytometricAnalysis ofApoptosisViaActive Caspase 3

ACaspase 3 kit was purchased from Immunochem-istry Technologies, LLC (Bloomington, MN) whichutilizes a cell permeable carboxyfluoroscein-conju-gated DEVD peptide inhibitor for Caspase 3. SincePrEC cells could not tolerate the recommended proto-col, we developed our ownmethodology, as describedhere. PrEC cells (2� 105) were plated in T75 flasks for24 hr. Cellswere then treatedwith 10�5M isoflavonoidsor the appropriate volume of vehicle. Media werereplaced onmost samples after 3 days, except for C4HEand its control; these were harvested on day 3 sinceapoptosis was already at a maximum with C4HE. Allother cultures were harvested on day 6. Medium waspoured off of the cell monolayers into centrifuge

The Prostate DOI 10.1002/pros

Soy Isoflavonoids in Prostatic Fluid 559

tubes to avoid rupturing floating apoptotic cells. Themonolayers were trypsinized using the recommendedreagents for PrEC (Cambrex Clonetics Corp.), neutra-lized with trypsin inhibitor, and pooled with theappropriate media in the tubes from above. Cells(3� 105) from each sample were placed in 15 mlcentrifuge tubes, and centrifuged at 220g for 5 min.Supernatants were aspirated leaving approximately100 ml of liquid on the pellets, and cells wereresuspended by tapping to avoid rupturing apoptoticcells. The volumes of each sample were adjusted toexactly 250 ml with PrEGM medium, and 2.5 ml of the30� FLICA reagant were added to each (except for theNo FLICA unstained control). Samples were mixed bytapping, and caps loosened to allow for air exchangeduring the 30 min incubation at 378C (mixing gentlyafter 15 min). PrEGM medium (10 ml) was added toeach tube and cells were incubated at 378C for 1 hr toallow unbound FLICA reagant to diffuse out of cells.Samples were centrifuged as above and pellets resus-pended in 300 ml PBS with 30 ml of the formaldehydefixative supplied with the kit. Cells were analyzed on aBeckmann Coulter FC-500 flow cytometer for fluor-oscein detection. The unstained No FLICA control wasset at 0.1%positive and relative fluorescence of all othersamples was compared to it.

FlowCytometric Analysis of Cell Cycle

PrEC were plated at a density of 2� 104 cells/T-25flask. After 24 hr, fresh medium was added containing10�5M of the specified isoflavonoid, or the appropriatevehicle controls. Media were replaced after 3 days, andcells were harvested after 6 days and stained aspreviously reported in detail [19]. In brief, media werepoured into centrifuge tubes, monolayers trypsinized,and cells pooled with the paired media. Samples werecentrifuged at 220g for 5 min, supernatant aspirated,and the pellets resuspended in a propidium iodidesolution (25 mg/ml propidium iodide, 0.1 mM EDTA,10 mg/ml DNase-free RNase A, and 0.3% saponin inPBS, pH 7.4). After incubating overnight at 48C in thedark, samples were analyzed for red fluorescence on aCoulter XL flow cytometer (Coulter Corp., Hialeah,FL). Statistics were performed on 10,000 events persample using Modfit LT 3.0 software (Verity SoftwareHouse, Inc., Topsham, ME).

RESULTS

PlasmaandPFconcentrations ofmany isoflavonoidsare sufficient to inhibit the in vitro growth of benignPrEC. Figure 1 shows the dose-dependent effects of5 isoflavonoids on the growth of benign PrEC. Growth

wasmeasured by the change in DNA content of the cellmonolayers over the 9-day treatment. All values werenormalized to the ethanol vehicle control, and statisti-cally significant changes are marked with an asterisk(P< 0.05). Genistein had the most potent antiprolifera-tive effects (Fig. 1A) with an IC50¼ 5.08� 10�7M. It isapparent that both plasma and PF levels of nearly allmen in the study are within the ranges that reduceproliferation of benign PrEC in vitro. Thus, genisteinis likely to contribute significantly to the anticancerproperties of dietary soy.

Figure 1B shows the results with equol (IC50¼1.8� 10�6M). Since only 20% of the men in our parallelstudy produced equol, we chose to present the rangesof concentrations that existed specifically in those menwho produced it. Plasma levels were not typically highenough to be within the range that affects growthof PrEC cells. However, because equol is so highlyconcentrated within PF as compared to plasma(median 12.7-fold higher, min 3.8-fold higher, max51.1-fold higher) [20], all menwho produced equol hadsufficient levels in their PF to elicit potent in vitroantiproliferative effects on PrEC. This observationsupports the hypothesis that the intestinal conversionof daidzein to equol gives soy consumers additionalprotection from prostate cancer.

ODMAwas produced by nearly allmen (98%) in ourparallel study, but at relatively low levels in bothplasma and PF (Fig. 1C). ODMA is somewhat inter-mediate inpotencywith an IC50¼ 1.18� 10�5M. This isthe first report to our knowledge of ODMA havingdirect antiproliferative effects in vitro. While 4/36 PFsamples contained� 1 mMODMA (doses that producemeasurable antiproliferative effects), this was not truefor the majority of plasma and PF samples. Thus, ifODMA helps to prevent prostate cancer, it is probablydoing so in most men through other mechanisms or byacting additively or synergistically with other isofla-vonoids.

Daidzein (Fig. 1D) has an IC50¼ 2.19� 10�5M.While plasma levels were not typically in the rangewhere we observe direct effects on PrEC growth, themajority of men had sufficient levels within their PFto do so. In fact, 21/36 PF samples had daidzein levels>2 mM, and 3/36 had levels exceeding 10�5M, with thehighest level being 30 mM. Thus, while daidzein mayhave a lower IC50 thanmany isoflavonoids in vitro, thehigh concentrations in vivo suggest that it has a highpotential for bioactivity in the prostate.

Another common daidzein metabolite is DHD,produced in 89% of the men in our previous study.DHD has the lowest potency in terms of inhibitingPrEC growth (Fig. 1E). An IC50 value could not bedetermined for this compound, as the highest concen-tration tested only inhibited growth by 20%.

The Prostate DOI 10.1002/pros

560 Hedlundet al.

The Prostate DOI 10.1002/pros

Fig. 1. Dose-dependent inhibitory effects of 5 isoflavonoids (A^E) on growth of PrECprimary cultures after 9 days of treatment.Growth(y-axis) ispresentedasapercentageof thatobtainedwiththeethanolvehiclecontrol.Eachgraphedvaluerepresents themeanofquadruplicatesamples� SEM. Statistically significantdifferences from the control (P< 0.05) aremarkedwith an asterisk, andwere calculatedusingone-wayANOVAwithDunnett’spost-test.TheIC50valuesandcurve fit (R2) foreachcompoundappearoneachgraph.Forcomparison,overlaidarrowsindicate thenaturalrangesof isoflavonoidconcentrations foundinplasma(dashedarrow),andprostatic fluid (solidarrow) fromCaucasianmenafter consuming soy in a previous study.Tickmarks on the arrows indicatemedian concentrations.These arrows represent the full ranges ofconcentrationsobservedinallparticipants,with theexceptionofequol (see text).

Soy Isoflavonoids in Prostatic Fluid 561

Differential Effects of Less-Studied IsoflavonoidMetabolites on PrECand LNCaPGrowth

Since the natural ranges of concentrations of theseadditional metabolites have not been determined inany population, we initially tested their in vitrobioactivities at a single concentration of 10�5M.Previously tested isoflavonoids were included forcomparison. Cells were treated for 8 days and growthwas normalized to that obtained with the appropriatevehicle controls.As shown inFigure 2A, all compoundshad significant effects (P< 0.01) on PrEC. Profoundeffects were observed with five compounds, reducinggrowth to less than 10% of the control (3HD, glycitein,C4HE, 8HD, and genistein). LNCaP cells respondedquite differently to many compounds (Fig. 2B). DHDhad the greatest inhibitory effect, while C4HE, thehydroxydaidzeins, and 6H-ODMA did not havesignificant effects.

As stated in our methods section, preliminary datausing HPLC-mass spectrometry suggest that thehydroxydaidzeins are unusually labile once resus-pended, evenwhen stored at�208C in the dark (JeevanPrasain, RayMoore, and Stephen Barnes, University ofAlabama at Birmingham). In agreement with this, weobserved amajor loss of activity over time inour assays.Since preparing each compound fresh before each useis economically unfeasible on a long-term basis, we didnot include them in subsequent experiments.

Effects of Isoflavonoids onCaspase 3Activity in PrEC

To determine if the inhibitory effects of the iso-flavonoids from above were due to the induction ofapoptosis, an assay for active caspase 3 was employed.PrEC cells were treated with 10�5M of each of theisoflavonoids for 6 days (except for C4HE, which werecompletely apoptotic after 3 days andwere analyzed atthat time). PrECwere incubated with a cell membrane-permeable carboxyfluoroscein-labeled DEVD peptidethat binds with high affinity to active caspase 3. Cellswere analyzed by flow cytometry for green fluores-cence (Fig. 3). PrEC cells that were treated with 0.1%DMSO showed a low level of fluorescence due to activecaspase 3 (6.8% positive when compared to unstainedcells, see Table I). In contrast, cells treated with C4HEhad a mean fluorescence nearly 10 times greater thancontrol cells andwere 98.3% positive for active caspase3 compared to unstained cells. The results for allcompounds are summarized in Table I. C4HE andgenistein induce apoptosis of PrEC to the greatestextent, while daidzein and DHD are the least effective.These results are in agreement with the relativepotencies of these compounds in our growth assaysabove.

Effects of Isoflavonoids onPrECCell Cycle Distribution

To determine if the isoflavonoids reduce prolifera-tion in addition to affecting apoptosis, cells weretreated with 10�5M isoflavonoids for 6 days and werestained with propidium iodide. DNA contents of thenuclei were quantitated based on red fluorescenceusing a flow cytometer (Table I). These results revealthat C4HE causes an accumulation of cells in G2/M

The Prostate DOI 10.1002/pros

Fig. 2. Inhibition of PrEC (A) and LNCaP (B) growth after 8 daystreatmentwith11isoflavonoids at10�5M. All vehicle controls werenormalizedto100%,butonly0.1%ethanolis shown.Eachvaluerepre-sents the mean of quadruplicate samples� SEM. Significant differ-ences were determined using one-way ANOVA with Dunnett’spost-test, andaremarkedwithanasterisk (P< 0.05).

Fig. 3. Flow cytometric histogram showing active Caspase 3 inPrEC cells treated for 3 days with 0.1% DMSO control (white) or10�5M C4HE (gray). Cells were incubated with a carboxyfluores-cein-labeled DEVD peptide and green fluorescence of 5,000 cellswasquantitatedona flowcytometer.

562 Hedlundet al.

(compare 16.5% for control with 28% for C4HE).Genistein caused a similar accumulation of cells inG2/M (compare 14% for control with 25.5% withgenistein), as previously reported in several cell types[9,19,29,30]. Of interest is that genistein also caused asignificant increase in the percentage of cells in G0/G1,as did all other isoflavonoids to various extents(Table I). In general, the compounds that were mostpotent at inhibiting growthwere also themost potent atinducing apoptosis and blocking cell cycle progression.The only exception to this is glycitein, which appears toact predominantly by blocking cell cycle progressionand not through caspase 3 induction.

Simulating the IsoflavonoidCompositions of PFSamples

Next we simulated in vitro the isoflavonoid compo-sitions of PF from various participants from theprevious nutritional study [20]. We chose to modelthe isoflavonoid composition of participant #76 (whohad the highest total isoflavonoid levels of all men inthe study), participant #18 (who had the highest levelsof equol), and participant #103 and #34, who hadmuch

lower levels of isoflavonoids in their PF. The concen-trations of each isoflavonoid present in these samplesare shown in Table II. The combined effects of theseisoflavonoids were tested on LNCaP growth after7 days of treatment (Fig. 4). The combination of isofla-vonoids from participants #76 and #18 significantlyinhibited LNCaP growth when compared to theirappropriate ethanol vehicle controls (P< 0.01). Incontrast, there were no significant changes in growthwith the combination of isoflavonoids from partici-pants #103 and #34. Additional experiments with PrECcells revealed that the isoflavonoids from participants#76 and #18 reduced growth to less than 5% of thecontrols (data not shown), suggesting that this effect isnot limited to the malignant cell type. We did notdirectly test the effects of isoflavonoid compositions ofparticipant #103 and #34 onPrEC cells.However, basedon the genistein concentration alone (0.3 mM and0.2 mM, respectively) and the IC50 curve establishedin Figure 1A, significant antiproliferative effects areexpected.

To determine which isoflavonoids were contribut-ing the most to the effects observed in Figure 4,additional experiments were carried out focusing on

The Prostate DOI 10.1002/pros

TABLE II. Isoflavonoid Compositions of Prostate Fluid Samples From Four Participants in our Previous Soy MetabolismStudy

Equol (mM) Daidzein (mM) DHD (mM) ODMA (mM) Genistein (mM) Total (mM)

Participant #76 5 30 11 1 1 48Participant #18 10 12 3 1 0.65 26.7Participant #103 ND 1 2 ND 0.3 3.3Participant #34 ND 0.2 ND ND 0.2 0.4

ND, not detectable.The limits of quantitative detection for each isoflavonoid ranged from 0.13–0.50 mM starting with 15 ml PF [20].

TABLE I. Isoflavonoids (10�5M)DifferentiallyAffectCaspase 3ActivityandCell CycleDistributionof PrECCells asAssessedby FlowCytometry

Treatment

Apoptosis Cell cycle distribution

% Positive for caspase 3(change from control) % G0/G1 % S-phase % G2/M

0.1% DMSO vehicle control 6.8 62.6 20.9 16.5C4HE 98.3 (91.5) 63.0 9.0 28.0Glycitein 11.4 (4.6) 79.7 10.7 9.60.2% ethanol vehicle control 8.4 57.4 29.3 13.3Daidzein 12.2 (3.8) 70.2 17.8 11.90.1% ethanol vehicle control 6.5 48.3 37.7 14.0Genistein 47.8 (41.3) 68.9 5.7 25.5Equol 18.4 (11.9) 83.0 4.8 12.26H-ODMA 15.8 (9.3) 50.9 36.3 12.8ODMA 15.0 (8.5) 68.0 21.4 10.6DHD 13.1 (6.6) 59.7 25.7 14.7

Soy Isoflavonoids in Prostatic Fluid 563

the isoflavonoid composition of the PF sample fromparticipant #76. In these experiments, the effects ofindividual isoflavonoidswere tested onLNCaPgrowthat the actual concentrations found in vivo (Fig. 5, firstsix bars). Several significant effects were observed(P< 0.01). 11 mM DHD reduced growth to 11% of itscontrol, making it the single most active component.

Daidzein (30 mM) reduced growth to 43% of its control,and equol (5 mM) to 57% of its control. ODMA (1 mM)andgenistein (1mM)hadno significant effect onLNCaPgrowth. The combination of these isoflavonoidsreduced growth to 13% of the control in this experi-ment. While the absolute magnitude of inhibitionvaried some between experiments (compare to 32% ofcontrol found in Fig. 4), the patterns of effects werealways similar.

To determine if the isoflavonoids were actingadditively or synergistically, an experiment wasperformed using 1:10 dilutions of each isoflavonoid atthe concentrations present in the PF of participant #76(Fig. 5, last six bars). At these concentrations, therewereno significant changes in growth with any individualcompound. In contrast, when all five isoflavonoids at a1:10 dilution are added together, growth is reduced to52% of the control (P¼ 0.01). This suggests that thesecompounds are acting in a synergistic fashion. There-fore, in men who sustain lower levels of soy isoflavo-noids in their PF, the synergistic effects of severalisoflavonoids may offer greater protection againstprostate cancer than any one compound individually.

DISCUSSION

Data from the current study indicate that the con-centrations of several soy isoflavonoids (obtained natu-rally through diet) are within the ranges that directlyaffect the growth of normal and malignant prostaticepithelial cells in vitro. At lower concentrations, theseisoflavonoids appear to act synergistically, althoughadditional research will be necessary to elucidatewhich compounds contribute to this effect. Preliminaryexperiments in our laboratory suggest that the anti-proliferative effects of these isoflavonoids are dimin-ished or absent in primary cultures of prostatic stromalcells, consisting of mixtures of smooth muscle andfibroblasts. This implies that dietary soy may have agreater impact on preventing prostate cancer thanbenign prostatic hyperplasia. However, it has also beensuggested that the development of cancer withinprostate epithelial glands is secondary to aberrantchanges in adjacent stroma [31]. In light of thishypothesis, the possibility remains that soy isoflavo-noids alter the state of differentiation of stromal cells,which in turn helps to prevent the development ofprostate cancer. The samemaybe true for isoflavonoidsin the prevention of benign prostatic hyperplasia. Weare currently pursuing experiments that address thesepossibilities.

This is thefirst report to our knowledgeof the invitrobioactivities of 3HD, 6HD, 8HD, 6H-ODMA, andC4HEin any human cell line. Given the significant inhibitoryeffects on PrEC at 10�5M, it would be valuable to know

The Prostate DOI 10.1002/pros

Fig. 4. Growth-inhibitory effects of combined isoflavonoids atthe concentrations present in four prostate fluid samples. LNCaPcellsweregrown for 7dayswith the concentrations of isoflavonoidslistedinTable II,or theappropriateconcentrationofethanolvehicle.Each value represents the mean of quadruplicate samples� SEM.Significant differences were determined using one-way ANOVAwithDunnett’spost-test, andaremarkedwithanasterisk (P< 0.01).

Fig. 5. Effects of individual versus combined isoflavonoids onLNCaPgrowthusing actual concentrationspresentin theprostaticfluidofparticipant#76.The first setofgraybarsinvolved theactualconcentrations found in the prostatic fluid of participant#76.Thesecond set of gray bars used a1:10 dilution of the actual concentra-tions.Growthwith each treatmentwasnormalized to its appropri-atevehicle control.Eachvaluerepresents themeanofquadruplicatesamples� SEM. Significance was determined using one-wayANOVAwithDunnett’s post-test, and is indicatedwith an asterisk(P< 0.01).

564 Hedlundet al.

what ranges of concentrations exist in plasma and PFin men after soy consumption. It is possible that someof these more recently discovered metabolites play anactive role in reducing prostate cancer risk in menconsuming soy.Additional studywill also be necessaryto determine if these compounds contribute additivelyor synergistically to the effects on cell growth. How-ever, their apparent instability in solution within daysof preparation (perhapsdue to rapid oxidation) is likelyto hinder further experimentation. Likewise, prelimin-ary HPLC data with C4HE suggest that it is prone tooxidation aswell. After 6months of storage inDMSOat�208C in the dark, the purity decreased from >95% toapproximately 70%; the decomposition products werenot conclusively identified, but appeared to includeapproximately 11% DHD, 7.5% trans-4 hydroxyequol,and 3.5% daidzein among other impurities (PaulDriedger, LC Laboratories, Inc.). Additional researchis necessary to help identify ways of preparation orstorage that preserve purity. Until then, it will beimportant for researchers to be aware of the potentialinstability of some of these metabolites to ensure thatfuture experiments produce meaningful results.

The steroid hormones and many phytoestrogensexist in plasma as conjugates to sulfate and glucuronicacid. Adlercreutz et al. found that 3%–54% of iso-flavonoids in the plasmaof Japanesemenwere sulfated(largely present as monosulfates), and 24%–86% wereglucuronidated [32]. The sulfated fraction has a highpotential for bioactivity, as sulfatase enzymes arecommonly expressed by a variety of epithelial celltypes [33]. However, glucuronidation is often consid-ered an irreversible step towards inactivation, targetingcompounds for excretion. Measurements of isoflavo-noids from the PF and plasma samples do not take intoaccount the degree of sulfation and glucuronidation.However, our PF simulation experiments provideus with some insight. If glucuronidation leaves only1/10th of the isoflavonoids active, our results inFigure 5 suggest that LNCaP growth would still bereduced by nearly 50% when the combination of fiveisoflavonoids are used.

In a previous study, we examined the apoptoticpotentials of daidzein, equol, and genistein on PrECcells and, contrary to our results here, found noevidence of apoptosis [19]. However, the method ofdetecting apoptosis was quite different, and attemptedto identify a sub-G0/G1 peak based onDNA content ofthe nuclei. Since the time of that initial study, manymore sensitive apoptosis assays have become available,such as the caspase 3 assay used in the current study.Caspase 3 detection also has the advantage of beingable to detect apoptosis in any phase of the cell cycle,where the other method could only detect apoptoticcells originating fromG0/G1. Thus, we believe that the

current study better estimates the apoptotic potentialsof the soy isoflavonoids than ourprevious report. Itwasalso interesting that the most potent compounds withregards to inhibiting cell growth, were also the mostpotent compounds with respect to changing cell cycledistribution and caspase 3 activation. This suggests thatthese phenomena are linked. In support of this, ourtime course experiments (datanot shown) revealed thatthe changes in cell cycle could be detected several daysbefore caspase activation occurred.

It is clear from the current study that LNCaP cellsare far more sensitive to the effects of DHD than arePrEC primary cultures. In fact, our unpublished datasuggests that DHD induces classical apoptosis inLNCaP cells. We have observed similar heterogeneityin the responses of various prostate cancer cell lines toother isoflavonoids. Given the heterogeneous nature ofprostate cancer in humans, it is likely that eating a soy-rich diet to produce multiple natural metabolites maybe more effective at reducing prostate cancer risk thantaking a single purified isoflavonoid supplement.

CONCLUSIONS

Our data suggest that dietary soy provides sufficientlevels of isoflavonoids to reduce the proliferationof normal and malignant prostatic epithelial cells—aphenomenon that would be expected to reduceprostate cancer risk inmany soy consumers.Additionalresearch will be necessary to directly determine if menwho produce certain metabolites have a lower risk ofprostate cancer than men who produce other metabo-lites. Given the prevalence of this disease in the US andNorthern European countries, such research seemswarranted.

ACKNOWLEDGMENTS

This work was supported in part by awards to T.Hedlund from the Department of Defense (grant#PC020275), the Grove Foundation and the SanFrancisco Foundation (award #PN0112-084).

REFERENCES

1. Jacobsen BK, Knutsen SF, Fraser GE. Does high soy milk intakereduce prostate cancer incidence? The Adventist Health Study(United States). Cancer Causes Control 1998;9(6):553–557.

2. Kolonel LN, Hankin JH,Whittemore AS,WuAH, Gallagher RP,Wilkens LR, John EM, Howe GR, Dreon DM, West DW,Paffenbarger RS Jr. Vegetables, fruits, legumes and prostatecancer: A multiethnic case-control study. Cancer EpidemiolBiomarkers Prev 2000;9(8):795–804.

3. LeeMM,GomezSL,Chang JS,WeyM,WangRT,HsingAW.Soyand isoflavone consumption in relation to prostate cancer riskin China. Cancer Epidemiol Biomarkers Prev 2003;12(7):665–668.

The Prostate DOI 10.1002/pros

Soy Isoflavonoids in Prostatic Fluid 565

4. Arora A, Nair MG, Strasburg GM. Antioxidant activities ofisoflavones and their biological metabolites in a liposomalsystem. Arch Biochem Biophys 1998;356(2):133–141.

5. Adlercreutz H, Bannwart C, Wahala K, Makela T, Brunow G,Hase T, Arosemena PJ, Kellis JT Jr, Vickery LE. Inhibition ofhuman aromatase by mammalian lignans and isoflavonoidphytoestrogens. J Steroid BiochemMol Biol 1993;44(2):147–153.

6. Adlercreutz H, Hockerstedt K, Bannwart C, Bloigu S,Hamalainen E, Fotsis T, Ollus A. Effect of dietary components,including lignans and phytoestrogens, on enterohepatic circu-lation and liver metabolism of estrogens and on sex hormonebinding globulin (SHBG). J Steroid Biochem 1987;27(4–6):1135–1144.

7. Evans BA, Griffiths K, Morton MS. Inhibition of 5 alpha-reductase in genital skin fibroblasts and prostate tissue bydietary lignans and isoflavonoids. J Endocrinol 1995;147(2):295–302.

8. Kim H, Peterson TG, Barnes S. Mechanisms of action of the soyisoflavone genistein: Emerging role for its effects via transform-ing growth factor beta signaling pathways. Am J Clin Nutr1998;68(6 Suppl):1418S–1425S.

9. McCabe MJ Jr, Orrenius S. Genistein induces apoptosis inimmature human thymocytes by inhibiting topoisomerase-II.Biochem Biophys Res Commun 1993;194(2):944–950.

10. Zhou JR, Mukherjee P, Gugger ET, Tanaka T, Blackburn GL,Clinton SK. Inhibition of murine bladder tumorigenesis by soyisoflavones via alterations in the cell cycle, apoptosis, andangiogenesis. Cancer Res 1998;58(22):5231–5238.

11. Rowland IR, Wiseman H, Sanders TA, Adlercreutz H, BoweyEA. Interindividual variation in metabolism of soy isoflavonesand lignans: Influence of habitual diet on equol production bythe gut microflora. Nutr Cancer 2000;36(1):27–32.

12. Axelson M, Sjovall J, Gustafsson BE, Setchell KD. Soya—Adietary source of the non-steroidal oestrogen equol in man andanimals. J Endocrinol 1984;102(1):49–56.

13. Kelly GE, Nelson C, Waring MA, Joannou GE, Reeder AY.Metabolites of dietary (soya) isoflavones in human urine. ClinChim Acta 1993;223(1–2):9–22.

14. Heinonen S, Wahala K, Adlercreutz H. Identification ofisoflavone metabolites dihydrodaidzein, dihydrogenistein,60-OH-O-dma, and cis-4-OH-equol in human urine by gaschromatography-mass spectroscopy using authentic referencecompounds. Anal Biochem 1999;274(2):211–219.

15. Kulling SE, Honig DM, Metzler M. Oxidative metabolism of thesoy isoflavones daidzein and genistein in humans in vitro andin vivo. J Agric Food Chem 2001;49(6):3024–3033.

16. Markiewicz L, Garey J, Adlercreutz H, Gurpide E. In vitrobioassays of non-steroidal phytoestrogens. J Steroid BiochemMol Biol 1993;45(5):399–405.

17. Mitchell JH, Duthie SJ, Collins AR. Effects of phytoestrogens ongrowth and DNA integrity in human prostate tumor cell lines:PC-3 and LNCaP. Nutr Cancer 2000;38(2):223–228.

18. Adlercreutz H. Phyto-oestrogens and cancer. Lancet Oncol2002;3(6):364–373.

19. Hedlund TE, Johannes WU, Miller GJ. Soy isoflavonoid equolmodulates the growth of benign and malignant prostaticepithelial cells in vitro. Prostate 2003;54(1):68–78.

20. Hedlund TE, Maroni PD, Ferucci PG, Dayton R, Barnes S, JonesK, Moore R, Ogden LG, Wahala K, Sackett HM, Gray KJ. Long-term dietary habits affect soy isoflavone metabolism andaccumulation in prostatic fluid in caucasian men. J Nutr 2005;135(6):1400–1406.

21. Hong SJ, Kim SI, Kwon SM, Lee JR, Chung BC. Comparativestudy of concentration of isoflavones and lignans in plasmaand prostatic tissues of normal control and benign prostatichyperplasia. Yonsei Med J 2002;43(2):236–241.

22. Morton MS, Chan PS, Cheng C, Blacklock N, Matos-Ferreira A,Abranches-Monteiro L, Correia R, Lloyd S, Griffiths K. Lignansand isoflavonoids in plasma and prostatic fluid inmen: Samplesfrom Portugal, Hong Kong, and the United Kingdom. Prostate1997;32(2):122–128.

23. Horoszewicz JS, Leong SS, Chu TM, Wajsman ZL, FriedmanM, Papsidero L, Kim U, Chai LS, Kakati S, Arya SK, SandbergAA. The LNCaP cell line—A new model for studies onhuman prostatic carcinoma. Prog Clin Biol Res 1980;37:115–132.

24. Horoszewicz JS, Leong SS, Kawinski E, Karr JP, Rosenthal H,Chu TM, Mirand EA, Murphy GP. LNCaP model of humanprostatic carcinoma. Cancer Res 1983;43(4):1809–1818.

25. van Bokhoven A, Varella-Garcia M, Korch C, Hessels D, MillerGJ. Widely used prostate carcinoma cell lines share commonorigins. Prostate 2001;47(1):36–51.

26. Fry PM, Hudson DL, O’Hare MJ, Masters JR. Comparison ofmarker protein expression in benign prostatic hyperplasia invivo and in vitro. BJU Int 2000;85(4):504–513.

27. LiuAY, PeehlDM.Characterization of culturedhumanprostaticepithelial cells by cluster designation antigen expression. CellTissue Res 2001;305(3):389–397.

28. HedlundTE,MillerGJ.A serum-freedefinedmediumcapableofsupporting growth of four established human prostatic carci-noma cell lines. Prostate 1994;24(5):221–228.

29. Markovits J, Linassier C, Fosse P, Couprie J, Pierre J, Jacquemin-SablonA, Saucier JM, LePecq JB, LarsenAK. Inhibitory effects ofthe tyrosine kinase inhibitor genistein on mammalian DNAtopoisomerase II. Cancer Res 1989;49(18):5111–5117.

30. Matsukawa Y, Marui N, Sakai T, Satomi Y, Yoshida M,Matsumoto K, Nishino H, Aoike A. Genistein arrests cell cycleprogression at G2-M. Cancer Res 1993;53(6):1328–1331.

31. Cunha GR, Hayward SW, Wang YZ. Role of stroma in car-cinogenesis of the prostate. Differentiation 2002;70(9–10): 473–485.

32. Adlercreutz H, Markkanen H, Watanabe S. Plasma concentra-tions of phyto-oestrogens in Japanese men. Lancet 1993;342(8881):1209–1210.

33. Suzuki T,MikiY,NakataT, ShiotsuY,AkinagaS, InoueK, IshidaT, KimuraM,Moriya T, SasanoH. Steroid sulfatase and estrogensulfotransferase in normal human tissue and breast carcinoma.J Steroid BiochemMol Biol 2003;86(3–5):449–454.

The Prostate DOI 10.1002/pros

566 Hedlundet al.