Endothelin-1 Production and Agonist Activitiesin Cultured Prostate-Derived Cells: Implications

for Regulation of Endothelin Bioactivity andBioavailability in Prostatic Hyperplasia

Paul D. Walden,1,2* Michael Ittmann,1,3,6 Marie E. Monaco,4,6 andHerbert Lepor1,5

1Department of Urology, NYU Medical Center, New York, New York2Department of Biochemistry, NYU Medical Center, New York, New York

3Department of Pathology, NYU Medical Center, New York, New York4Department of Physiology and Neuroscience, NYU Medical Center, New York, New York

5Department of Pharmacology, NYU Medical Center, New York, New York6DVA Medical Center, New York, New York

BACKGROUND. Endothelin-1 (ET-1) interacts with specific G-protein-coupled receptors toinitiate short-term (contraction) and long-term (mitogenesis) events in target cells. ET-1 is anabundant prostate secretory protein that, in its biologically active form, elicits prostaticsmooth muscle contraction. The present study was designed to determine the effects of ET-1on prostate cell growth and to examine the regulation of endogenous ET-1 activity andbioavailability.METHODS. Primary cultures of prostate secretory epithelial (PE) and prostate fibromuscularstromal (PS) cells were established from benign human prostate tissue.RESULTS. In culture, PE cells secrete immunoreactive ET-1 (38.5 ± 1.6 pg/ml/106 cells/24 hr)into the conditioned medium. Levels of immunoreactive ET-1 produced by PS cells were morethan 10-fold lower. Endothelin-converting enzyme-1 (ECE-1) mRNA was detected in PE cellsand not in PS cells; however, big ET-1 was the predominant immunoreactive ET-1 secretoryproduct of PE cells. The ETB endothelin receptor was the predominant subtype in both PE andPS cells. In PS cells, but not PE cells, ET-1 induced significant inositol phosphate accumulationand [3H]-thymidine uptake. Agonist activity was inhibited by the ETB receptor selectiveantagonist, BQ 788. Intact PE cell monolayers secrete ET-1 through the apical surface, con-sistent with secretion of ET-1 into the glandular lumen in vivo.CONCLUSIONS. On the basis of these findings, regulation of ET-1 activity and bioavailabil-ity appears to be tightly regulated. Such findings have important implications in the patho-physiology of prostate disease. Prostate 34:241–250, 1998. © 1998 Wiley-Liss, Inc.

KEY WORDS: endothelin-1; G-protein-coupled receptors; prostatic hyperplasia

INTRODUCTION

The endothelins comprise a family of three small(21-amino acid) peptides: ET-1, ET-2, and ET-3 [1].Originally identified on the basis of potent vasocon-strictive activity [2], the endothelins were subse-quently shown to be mitogenic in a number of cellsystems [3–5]. The endothelins signal through two ma-

jor seven-transmembrane-spanning, G-protein-coupled receptor subtypes classified on the basis ofdifferential affinity for ET-1 and ET-3 [6]. The best

*Correspondence to: Paul Walden, Skirball 10R, NYU Medical Cen-ter , 550 First Avenue, New York, NY 10016. E-mai l :paul.walden@nyu.eduReceived 10 December 1996; Accepted 28 January 1997

The Prostate 34:241–250 (1998)

© 1998 Wiley-Liss, Inc.

characterized signal transduction mechanism acti-vated by endothelin liganded receptors involves theGq mediated stimulation of phospholipase Cb andphosphoinositide hydrolysis producing two secondmessengers: inositol 1,4,5-trisphosphate and diacyl-glycerol [7–9]. InsP3 binds to specific receptors on theendoplasmic (sarcoplasmic) reticulum, releasingstored calcium. Voltage-dependent calcium channelsare also activated in excitable cells. Increased intracel-lular calcium elicits contraction in smooth musclecells. Diacylglycerol and calcium stimulate protein ki-nase C (PKC), which mediates mitogenic signaling byendothelin [10]. Mitogenic activation can also occur bya PKC-independent route that involves cross-talk be-tween G-protein signaling pathways and receptor ty-rosine kinase signaling pathways that signal throughRas [10,11].

The human prostate gland can undergo a benignenlargement in a condition termed benign prostatichyperplasia (BPH). Histological evidence of BPH ispresent at autopsy in the prostates of up to 90% ofmales older than age 80 years [12,13]. The hyperplasticlesion may cause bothersome urinary symptoms(symptomatic BPH) in patients. By tissue morphom-etry, it was shown that increased stromal smoothmuscle proliferation in the prostate correlated signifi-cantly with clinical BPH [14,15]. Human seminal fluidcontains immunoreactive ET-1 (iET-1) at a concentra-tion of 0.5–5 ng/ml [16], which is more than 500-foldgreater than in any other body fluid examined; bycontrast, plasma contains 1–20 pg iET-1/ml [17,18].There is evidence to indicate that the prostate contrib-utes to this large quantity of iET-1 in seminal plasma.Similar iET-1 levels are present in the seminal plasmaof vasectomized men where testicular ET-1 contribu-tion has been eliminated [16] and ET-1 immunoreac-tivity is present in secretory prostatic epithelium [19].Endothelin receptors exist in both stromal and epithe-lial compartments of the prostate [20] and ET-1 is apotent agonist of prostatic smooth muscle contraction[19] and a mitogen in other systems [3–5]. Taken to-gether, these findings suggested that endogenous ET-1could potentially contribute to basal smooth muscletone and/or to the growth of prostatic cells. Thesefindings also suggested that, under normal conditions,mechanisms presumably exist to regulate ET-1 activityin the prostate. The objective of this study was to de-termine the effects of ET-1 on prostate cell growth andto examine the regulation of endogenous ET-1 activityand bioavailability in the prostate.

MATERIALS AND METHODS

Materials

Tissue culture reagents were obtained from SigmaChemical Co. (St. Louis, MO) or Life Technologies

(Grand Island, NY), with the exception of bovine pi-tuitary extract, which was obtained from Upstate Bio-technology (Lake Placid, NY). [3H]methoxyinulin,methyl-[3H]thymidine, and [125I]-ET-1 were obtainedfrom New England Nuclear (Boston, MA). [3H]inositolwas obtained from ARC (Bowling Green, MO). ET-1,BQ123, and BQ788 were obtained from the AmericanPeptide Corporation (Sunnyvale, CA).

Tissue Isolation and Cell Culture

Human prostate tissue was obtained in accordancewith the Institutional Board of Research Associates’guidelines in effect at NYU Medical Center from malepatients undergoing prostatectomy for BPH or pros-tate cancer. Under aseptic conditions in a positive-flow tissue culture hood, histopathologically con-firmed benign tissue was dissected away from a por-tion of the gland. For RNA isolation, tissue was snapfrozen in liquid nitrogen. For culture of prostate se-cretory epithelial (PE) and prostate fibromuscularstromal (PS) cells, fresh tissue (approximately 0.5 g)was minced into 1-mm3 (approximately) cubes, rinsedin phosphate-buffered saline (PBS), and incubated for16–24 hr at 37°C on a rocking platform in MCDB 153medium (Sigma Chemical Co.) containing 10% fetalcalf serum (FCS) and collagenase (600 U/ml). The col-lagen-treated cell suspension was centrifuged at1,500g for 5 min. The cell pellet was washed twice inPBS, resuspended in 12 ml of the appropriate medium,and plated into two 60-mm tissue culture dishes (col-lagen-treated dishes were used for PE cells). The‘‘basal medium’’ for selection and culture of PE cellsconsisted of 90% MCDB 153, 10% RPMI 1640 supple-mented with insulin (5 mg/ml), oleic acid (0.025% w/v)/bovine serum albumin (BSA) (5% w/v), choleratoxin (10 ng/ml), dexamethasone (1 mg/ml), heparin(50 mg/ml), Fungizone (25 ng/ml), bovine pituitaryextract (10 mg/ml), and epidermal growth factor [EGF(10 ng/ml)]. The medium for selection and culture ofPS cells was 90% RPMI 1640, 10% FCS supplementedwith Fungizone (25 ng/ml), and gentamycin (100 mg/ml). Cells were cultured at 37°C in a humidified at-mosphere of 95% air/5% CO2. Cells were re-fed every2 days with the appropriate media and, at confluence,were passaged by splitting trypsinized cells 1–3 infresh medium. Pure cultures of PE and PS cells wereobtained after two passages. Cells were used for ex-perimentation after two to three passages (PE cells) orafter two to four passages (PS cells). The cell selectionprocess was validated by immunohistochemistry, ac-cording to several criteria, including histology and theexpression of cell-specific markers. The PE cell mark-ers used were prostate-specific antigen (PSA) andprostatic acid phosphatase (PAP). PS cells were veri-

242 Walden et al.

fied by positive staining for vimentin (fibroblasts) andsmooth muscle a-actin (smooth muscle cells).

Culture of Polarized PE Cells

To determine the polarity (apical vs basolateral) ofET-1 secretion, 1 × 106 PE cells were seeded in 24-mmcollagen-coated Transwells (Corning-Costar, 0.4-mmpore size; Costar, Cambridge, MA) and cultured untilconfluence with medium changes every 2 days. At thistime, the apical and basolateral chambers were replen-ished with fresh medium. The integrity of the conflu-ent PE layers was determined by adding the extracel-lular space marker, [3H]methoxyinulin (2 mCi/ml), tothe apical medium. After 1 hr at 37°C, a sample (50 ml)of the apical and basolateral media was taken for de-termination of radioactivity by scintillation counting.Cultures with 4% or less leakage from the apical tobasolateral compartments after 1 hr at 37°C were usedfor experimentation. ET-1 levels in the conditionedapical and basolateral media of intact PE monolayerswere determined as described below in the section,Measurement of Endothelin.

For determination of endothelin receptor localiza-tion, 30,000 cpm of [125I]-ET-1 (2,000 Ci/mmol) wasadded to either the apical or basolateral medium. Spe-cific versus nonspecific binding of the radioligand wasdetermined, respectively, both in the absence and inthe presence of 1 mM ET-1. After 1 hr at 37°C, [3H] and[125I]in the apical and basolateral medium was deter-mined by liquid scintillation and gamma counting, re-spectively, to determine the leakage of radionuclidesbetween compartments. The cell monolayer waswashed three times in PBS, and bound counts weredetermined by gamma counting.

Measurement of Endothelin

The conditioned medium was removed from sub-confluent cultured PE or PS cell monolayers after 24 hrand from confluent intact PE monolayers (in Tran-swells) after 1–24 hr. The number of cells in the mono-layers was determined by trypsinizing and countingin a Coulter model ZM counter (Coulter, Hialeah, FL).An equal volume of 20% glacial acetic acid was addedto the conditioned media samples. Endothelin pep-tides present in the acidified samples were concen-trated by binding to and elution from octadecyl C18columns (Bakerbond ‘‘spe’’) as described [19]. Recov-ery from these columns was determined by addingknown amounts of human big ET-1, ET-1, or ET-3 toculture medium. Levels of immunoreactive ET-1 andET-3 were determined using radioimmunoassay (RIA)kits. The RIA kit used for detection of ET-3 (RIK6911from Peninsula Laboratories, Belmont, CA) had lowcross-reactivities (0.1–0.2%) with human big ET-1 and

ET-1. Two different RIA kits were used for detectionof ET-1. The first kit (NEK-067 from New EnglandNuclear) detected both human ET-1 (100%) and hu-man big ET-1 (69.6%). The second kit (RIK6912 fromPeninsula Laboratories) was highly selective for hu-man big ET-1 (100%), having zero cross-reactivity withhuman ET-1.

Radioligand Binding to Cell Membranes

Membranes were prepared from cultured cells asdescribed [21]. Membrane protein was determined us-ing a kit (Bio-Rad Laboratories, Hercules, CA) withbovine plasma g-globulin as the standard. Saturationand competitive binding experiments were performedusing [125I]-ET-1 or [125I]-ET-3 as described [22]. Re-ceptor subtype determinations were performed by in-cluding ETA selective (BQ123) or ETB selective(BQ788) antagonists in the bindings assays. Threeseparate competitive experiments were performed foreach of the competitors. Data were analyzed by itera-tive curve fitting to a one- or two-site binding model,and Ki values were calculated from IC50 values [23].Linear transformation of the binding data was con-ducted as described [24].

Measurement of Phosphoinositide Hydrolysis

For assessment of phosphoinositide hydrolysis,cells were plated in 12-well dishes. At 80% confluence,the growth medium was replaced with regular growthmedium (PE cells) or Minimal Essential Medium(MEM) supplemented with 0.5% FBS (PS cells) thathad been supplemented with 5 mCi/ml [3H]inositol(sp. act., 15 Ci/mole), and the cell monolayers wereincubated at 37°C overnight. The medium was thenchanged to serum-free Ham’s F12 containing 10 mMlithium chloride and the indicated concentration ofagonist for a further 90 min. Phosphoinositides andinositol phosphates were extracted and analyzed aspreviously described [25].

Thymidine Uptake Assays

For measurement of thymidine uptake, cell mono-layers in 12-well plates were incubated for 72 hr inbasal medium lacking bovine pituitary extract andEGF (PE cells) or in 99.5% RPMI 1640, 0.5% FCS (PScells). At this time, ET-1 (0.01–100 nM) or FCS (10%)was added. In parallel experiments, the ETA selectiveantagonist BQ123 (1–100 nM) or the ETB selective an-tagonist BQ788 (1–100 nM) was added to cell mono-layers 5 min before the addition of ET-1 (1–10 nM).Following a 23 hr incubation with agonists or ago-nists/antagonists, 5 mCi/ml of methyl-[3H]thymidinewas added to the medium and the incubation at 37°Ccontinued for 1 hr. At this point cells were 80–90%

Endothelin-1 in the Prostate 243

confluent. Cellular DNA was extracted as described[10] and incorporated radioactivity was determinedby scintillation counting.

Determination of DNA Fragmentation

DNA fragmentation indicative of apoptosis was de-termined in PS cell monolayers after 5 days of incu-bation with RPMI 1640, 0.5% FCS containing ET-1 (1–100 nM) or no ET-1. Adherent and floating cells wereharvested and high- and low-molecular-weight DNAwas isolated and quantified as described [26]. In sepa-rate experiments, DNA fragmentation was deter-mined in situ in adherent monolayers by the TUNELmethod using a commercially available kit (MBL,MA). The proportion of nuclei with evidence of apop-tosis was scored.

RNA Isolation and RT-PCR

Total RNA was isolated from cell cultures and tis-sue samples by acid guanidinium thiocyanate phenolchloroform extraction [27]. Total RNA (1 mg) was re-verse transcribed into cDNA, using an oligo (dT)primer and Superscript reverse transcriptase (LifeTechnologies). An aliquot (10%) of the reverse tran-scription reaction was used for PCR with the variousprimer pairs listed in Table I. Control tubes containingno reverse transcriptase were also used for PCR, as acontrol for DNA contamination of the RNA samples.Primer pairs specific for human b-actin, endothelin-converting enzyme-1 (ECE-1), and the ETA and ETB

receptor mRNAs were designed using the GCG Primesoftware package (Table I). The ETA and ETB receptoramplimers were able to be used alone or in combina-tion in the same PCR reaction with no apparent effectson the yield of either product (as determined either byincorporation of a labeled nucleotide into the PCRproducts or by Southern blotting followed by densi-tometry). The inclusion of both primer pairs acted asan internal control for the amplification process. Al-though not designed as such, this reverse transcrip-tion-polymerase chain reaction (RT-PCR) assay turnedout to be semiquantitative inasmuch as there wasgood concordance between the data obtained by thismethod and the data obtained by solution hybridiza-tion (not shown). In all cases, the quality of cDNA wasconfirmed by amplification with actin primers. PCRconditions were 35 cycles of 94°C for 30 sec, 53°C for30 sec, and 72°C for 1 min, using a Perkin Elmer 9600thermal cycler (Perkin Elmer, Norwalk, CT). By mea-suring radioactive label incorporation into the cor-rectly sized products after various numbers of PCRcycles, we established that amplification for eachprimer pair (whether used alone or in combination)was exponential under the conditions used in thismanuscript. PCR products were resolved by electro-phoresis in a 5% polyacrylamide, 1 × TBE (0.09 M Tris,0.09 M boric acid, 2 mM EDTA) gel. Authenticity ofthe PCR products was verified by Southern hybridiza-tion to the corresponding fragments of the cloned en-dothelin receptor cDNAs (gift of Dr. M. Yanagisawa).

Statistical Analysis

Data are expressed as the mean ± standard error(SEM). Statistical significance was determined by Stu-dent’s t-test. The GraphPad Prism software packagewas used throughout to calculate radioligand bindingdata and statistical significance.

RESULTS

PE Cells Secrete Immunoreactive ET-1 and ET-3

By immunohistochemistry, Langenstroer et al. [19]showed that immunoreactive ET-1 (iET-1) was local-ized predominantly in the prostatic epithelium, sug-gesting that these cells are the primary source of pros-tatic ET-1 production. To determine whether endothe-lins were secreted by cultured prostate-derived cells,the 24-hr conditioned medium from six independentPE and PS cell lines was analyzed for iET-1 and iET-3production by RIA. The results are shown in Figure 1.The levels of iET-1 secreted by PE cells in 24 hr aremore than 10-fold higher than the levels produced by

TABLE I. PCR Primers Used*

Primer sequence Description

58-AATGCTTCTAGGCGGACT-38 Human b-actinforward primer

58-ACTCCCAGGGAGACCAAA-38 Human b-actinreverse primer

58-GAATCCGAGCCTATTGTG-38 Human ECE-1forward primer

58-TCCTCATCCATCCACTTC-38 Human ECE-1reverse primer

58-AGCTTGAGAATTGCCCTC-38 Human ETA receptorforward primer

58-TCTTCCACTGGATGCTTG-38 Human ETA receptorreverse primer

58-TGCTGCACATCGTCATTG-38 Human ETB receptorforward primer

58-TGAAGCAAGCAGATTCGC-38 Human ETB receptorreverse primer

*Primers were designed using the GCG Prime software packageand were chosen to span separate coding exons of the corre-sponding mRNAs.

244 Walden et al.

PS cells during the same time period. PE cells pro-duced 6-fold lower levels of iET-3 than iET-1. Levels ofiET-3 produced by PS cells were close to the lowerlimit of detection of the RIA. In BPH prostate tissue,we determined that the average levels of iET-1 to bewithin the range described previously, 580 ± 20 pg/gwet wt [19]. In addition, we determined that ET-3 waspresent in prostate tissue at an average concentrationof 95 ± 5 pg/g (wet wt). The amounts of endothelindetermined in prostate tissue samples presumablyrepresent average values and higher local concentra-tions of ET-1 and ET-3 would be expected dependingon the degree of epithelial hyperplasia.

Endothelin-Converting Enzyme-1 inProstate-Derived Cells

The endothelins are synthesized as large, biologi-cally inactive precursors called preproendothelins. Pre-pro ET-1 mRNA encodes a 203-amino acid precursor,which is subsequently cleaved by dibasic pair prote-ases [28] to yield ‘‘big endothelin’’ (39 amino acids).Big endothelin is two orders of magnitude less potentthan the 21-amino acid ET-1 in eliciting contraction inisolated blood vessels [29]. Conversion of big endothe-lin to the biologically active 21-amino acid ET-1 re-quires the membrane metalloproteinase, endothelinconverting enzyme-1 (ECE-1), which cleaves at the se-quence Trp21–Val22 [2,30]. The RIA kit used for de-termination of iET-1 secretion by PE and PS cells (Fig.1) reacted with both human ET-1 (100%) and humanbig ET-1 (70%). We performed the following experi-ments to determine whether ECE-1 was present andactive in the PE and PS cells. Initially, we examinedthe expression of ECE-1 mRNA in PE and PS cells by

RT-PCR. ECE-1 PCR products were amplified fromRNA obtained from subconfluent (passage 2 or 3) orintact PE cell monolayers (Fig. 2). Under the same con-ditions, using RNA obtained from PS cell monolayers,a PCR product was not visible on an ethidium bro-mide-stained gel after 35 cycles (Fig. 2). Using a RIAthat detects ET-1 and big ET-1 (total iET-1) and a RIAspecific for big ET-1 (see Methods) we determined thatPE cells secreted predominantly (>95%) big ET-1.These results indicate that ECE-1 mRNA is present inPE cells, but is either not efficiently translated or theactivity of the translation product is regulated.

Endothelin Receptor Expression inProstate-Derived Cell Lines

The endothelins signal through two major G-protein-coupled receptor subtypes classified on thebasis of differential affinity for ET-1 and ET-3 [6]; theETA receptor has greater affinity for ET-1 than for ET-3, whereas the ETB receptor has equal affinity for ET-1and ET-3. The expression of the ETA and ETB receptormRNAs in prostate samples was determined using areverse transcription polymerase chain reaction (RT-PCR) assay in which ETA and ETB receptor cDNAswere coamplified in the same PCR tube. Representa-tive results of this assay are shown in Figure 3. Al-though not designed as such, this RT-PCR assayturned out to be semiquantitative, as the relativeabundance of ETA to ETB receptor RNAs determinedby this method was in good agreement with solutionhybridization data (data not shown) and with theabundance of the receptor subtypes as determinedpharmacologically (see below). Both ETA and ETB re-ceptor mRNAs were detected in BPH prostate tissue(Fig. 3, lanes 1–3), a finding consistent with the greaterabundance of ETA relative to ETB receptor in BPH tis-sue determined by pharmacology [20]. The ETB recep-tor mRNA product was the predominant amplifica-tion product seen in both cultured PE cells and PS cellsusing this RT-PCR assay (Fig. 3, lanes 4–7). To rule outthe possibility that endothelin receptor subtypemRNA expression was related to the age or density ofthe cell cultures, receptor expression was analyzed inPE cultures that had been passaged two or three times(Fig. 3, lanes 4 and 6) and in PS cultures that had beenpassaged two, three, or four times (Fig. 3, lanes 5, 7,and 8). Furthermore, the ratio of the receptor mRNAsdid not change in 3 day postconfluent PS monolayers,although under these conditions there was a concomi-tant loss of expression of mRNAs for both receptors(data not shown).

To determine whether the pattern of receptormRNA expression determined by RT-PCR was re-flected in receptor protein levels, saturation and com-

Fig. 1. ET-1 and ET-3 production by cultured prostate-drivedcells. Levels of ET-1 and ET-3 in the 24-hr conditioned mediumfrom cultured PE and PS cell monolayers were determined by RIA.The data represent ±SEM for six independent experiments per-formed in quadruplicate.

Endothelin-1 in the Prostate 245

petition binding experiments were performed. BothPE and PS cell membranes demonstrated high-affinityand saturable binding for both [125I]-ET-1 and [125I]-ET-3 (Table II). [125I]-ET-1 binding reflects total ET re-ceptor binding since ET-1 binds both ETA and ETBreceptors with equal affinity. By contrast, [125I]-ET-3binds ETB receptors with high affinity (Ki 1 nM) and toETA receptors with low affinity (Ki @ 100 nM). Thus, atthe concentration used in this experiment (0.5 nM),[125I]-ET-3 would bind to <1% of the available ETAreceptors. Thus, the identification of the ETB subtypeas the predominant endothelin receptor in cultured PEcells and PS cells by competition binding (Table III)was consistent with the a greater relative abundanceof ETB receptor mRNA compared to ETA receptormRNA in these cells determined by RT-PCR (Fig. 3).Autoradiography of slide-mounted prostate tissuesections following incubation with [125I]-ET-1 and re-ceptor subtype selective antagonists revealed thatboth endothelin receptor subtypes were present instroma and epithelium, with ETA predominating instroma and ETB predominating in epithelium [31]. Thepredominance of the ETB receptor subtype in culturedPS cells is inconsistent with the abundance of the ETAreceptor in prostatic stroma revealed by autoradiog-

raphy [31]. This inconsistency presumably arises fromthe inevitable dedifferentiation of PS cells in cultureresulting in a loss of ETA receptor expression insmooth muscle like cells, the predominant cell type inPS cultures as determined by a-actin staining (data notshown). However, the functional significance of the

Fig. 2. ECE-1 mRNA is abundant in PE cells but not in PS cells.Total RNA (1 mg) isolated from prostate tissue (lane 1), culturedprostate-derived PE cells (lanes 2–4) and PS cells (lanes 5, 6) wasreverse transcribed into cDNA and used as a template for PCRamplification with primer pairs specific for human ECE-1 mRNA.For the cultured prostate-derived cell lines, RNA was isolated

after two (lanes 2, 5) or three (lane 3) passages and in cultures 3days postconfluence (lanes 4, 6). PCR products were resolved byelectrophoresis in a 5% acrylamide, 1 × TBE gel and stained withethidium bromide. The size of the authentic PCR product (asdetermined by Southern hybridization) is indicated. Lane 7, shamRT-PCR reaction performed with no reverse transcriptase.

Fig. 3. ET receptor mRNA expression by prostate-derived celllines. Total RNA (1 mg) isolated from prostate tissue (lanes 1, 2and 3), cultured prostate-derived PE cells (lanes 4, 6) and PS cells(lanes 5, 7, 8) was reverse transcribed into cDNA and used as atemplate for PCR amplification with primer pairs specific for hu-man ETA (lane 1), ETB (lane 2), or ETA and ETB (lanes 3–9) re-ceptor mRNAs (see Table I). For the cultured prostate-derived

cell lines, RNA was isolated after 2 (lanes 4, 5), 3 (lanes 6, 7), and4 (lane 8) passages. PCR products were resolved by electropho-resis in a 5% acrylamide, 1 × TBE gel and stained with ethidiumbromide. The size of the authentic PCR products (as determinedby Southern hybridization) are indicated. Lane 9, sham RT-PCRreaction performed with no reverse transcriptase.

TABLE II. [125I]-ET-1 and [125I]-ET-3 Saturation andCompetition Binding Experiments Using PE and

PS Cell Membranes*

PE cells PS cells

Kd (M) for ET-1 6 ± 2 × 10−11 7 ± 3 × 10−11

Kd (M) for ET-3 8 ± 3 × 10−11 8 ± 3 × 10−11

Bmax (fmol/mg protein)for ET-1 50 ± 10 300 ± 70

Bmax (fmol/mg protein)for ET-3 60 ± 10 330 ± 80

*Competitive binding experiments using cell membranes wereperformed in the presence of a constant final concentration of aradioligand (0.5 nM [125I]-ET-1 or [125I]-ET-3) and increasingconcentrations of unlabeled ET-1 or ET-3. Linear transformationof the binding data was conducted as described [24]. Data rep-resent the mean ± SEM for three independent experiments per-formed in triplicate.

246 Walden et al.

expression of both endothelin receptor subtypes inprostate is unknown, as there is an apparent redun-dancy in receptor function: Contraction in prostatemuscle strips was shown to be mediated by both ETAand ETB receptors [19] and in PS cells contraction ismediated by the predominant ETB receptor [22].

Agonist Activities of ET-1 in CulturedProstate-Derived Cells

Since ET-1 had previously been shown to inducemitogenic signaling in vascular smooth muscle cells[3], fibroblasts [4], and certain endothelial cells [5], weinvestigated the possibility that ET-1 could inducesimilar signaling mechanisms in cultured prostate-derived cells. We initially looked at ET-1 induced sec-ond messenger generation in cultured PE and PS cells.ET-1 stimulated inositol phosphate accumulation incultured PS cells in a concentration dependent manner(Fig. 4). By contrast, ET-1 did not result in any detect-able increase in inositol phosphate accumulation in PEcells.

Most studies have indicated that ET-1 is a co-mitogen that requires low concentrations of serum formaximal activity. ET-1 induced a concentration de-pendent increase in [3H]-thymidine uptake by quies-cent PS cells in the presence of 0.5% serum (Fig. 5A).Under these conditions, 10 nM ET-1 produced a 1.4 ±0.1-fold increase in thymidine uptake and 10% FCSproduced a 5.4 ± 0.4-fold increase in thymidine up-take. Although ET-1 induces modest growth effects invitro, these growth affects might be expected to besignificant in vivo when considered within the contextof the etiology of BPH which has been estimated tohave a doubling time of 20 years [32]. ET-1 induced[3H]-thymidine uptake by PS cells was significantlyinhibited by the ETB receptor selective antagonist,BQ788 (Fig. 5B). Therefore, signaling in cultured PScells appears to mediated by the most abundantly ex-pressed receptor in PS cells, the ETB receptor (see Fig.3). ET-1 did not increase [3H]-thymidine uptake by PEcells. Uncoupling of endothelin receptors from mito-

genic signaling in PE cells may be a biological mecha-nism to prevent continued autocrine stimulation.

ET-1 Does Not Affect PS Cell Apoptosis

We investigated the possibility that ET-1 might af-fect programmed cell death (apoptosis) in quiescentcultures of PS cells. Intranucleosomal DNA fragmen-tation occurs as a consequence of the endonucleaseactivity that accompanies apoptosis [26]. DNA iso-lated from quiescent PS cells or PS cells stimulatedwith ET-1 (1–100 nM) displayed no significant differ-ences in the fragmentation pattern (data not shown).Also, the number of nuclei showing nuclear evidenceof apoptosis in situ using the TUNEL method was notsignificantly different in control and ET-1-treated cells.Thus, ET-1 stimulated thymidine uptake by PS cellsdoes not appear to be accompanied by a significantlydecreased rate of apoptosis.

Regulation of Endothelin Bioavailability

Cultured PE cell monolayers are organized in amanner reminiscent of columnar epithelial cells invivo; the cells are polarized and possess distinct apicaland basolateral surfaces formed by tight junctions be-tween the cells [33–36]. Production of ET-1 by PE cells,and not by PS cells, therefore identifies another pos-sible component of the endothelin axis, since the po-

TABLE III. Inhibition of Specific [125I]-ET-1 Binding toPS Cell Membranes*

Competitor Ki (nM)

ET-1 0.4 ± 0.1ET-3 0.1 ± 0.02BQ123 >10,000BQ788 0.7 ± 0.2

*Ki values calculated from IC50 values [23] are in nM. Data arethe mean ± SEM for three independent experiments performedin triplicate.

Fig. 4. ET-1 induces inositol phosphate accumulation in PS cells.PS cell monolayers were labeled with [3H]inositol overnight. Thefollowing day cells were stimulated in serum-free medium withvehicle or ET-1 (0.01–100 nM) for 90 min. Incorporation of [3H]into inositol phosphates was determined by ion-exchange chro-matography as described (25). Data are represented as the meanfold stimulation over control ±SEM for four independent experi-ments performed in sextuplicate. Filled circles, significant increase(P > 0.05) over control.

Endothelin-1 in the Prostate 247

larity (apical vs basolateral) of endothelin secretioncould determine normal biological availability. Thelevels of iET-1 in the conditioned apical and basolat-eral media of intact PE cell monolayers were mea-sured. The results are shown in Figure 6A. Our resultsindicate a distinct polarization of ET-1 secretionthrough the apical surface of intact PE monolayers,consistent with secretion into the glandular lumen invivo. This finding suggests that ET-1 is primarily asecretory protein and not a paracrine regular of pros-tate smooth muscle. By contrast, ET-1 secretion by cul-tured vascular endothelial cells is distinctly basolat-eral [37], providing a mechanism for paracrine stimu-lation of vascular smooth muscle cells. These resultssuggest that epithelial cell tight junctions serve to seg-regate ET-1 from stromal smooth muscle cells.

To determine the polarity of endothelin receptorlocalization on PE cells, [125I]-ET-1 was added to eitherthe apical or basolateral media of intact cell monolay-ers. Binding to the monolayer was determined after 1hr at 37°C. Nonspecific binding was determined in thepresence of 1 mM ET-1. Specific ET-1 binding to themonolayers is shown in Figure 6B. The results indicatespecific high-affinity [125I]-ET-1 binding to both apicaland basolateral surfaces.

DISCUSSION

The results of this paper demonstrate the existenceof a tightly regulated endothelin axis in the humanprostate. Consistent with the much greater immuno-histochemical staining of ET-1 in epithelium com-pared to stroma in prostate tissue sections [19], wehave shown that prostate-derived PE cells secretemore than 10 times iET-1 into the culture mediumthan do PS cells. Since ET-1 is a potent contractileagonist of prostatic smooth muscle, epithelial ET-1production would be expected to be tightly regulated.ECE-1 mRNA was not detected in PS cells. In PE cells,ECE-1 mRNA was present but ECE-1 activity was lowas judged by the abundance of big ET-1 secretion.Seminal plasma contains proteases such as prostatespecific antigen (PSA) and cathepsin D [38]. CathepsinD has ECE-1 like activity under certain conditions [39]which could release ET-1 from big ET-1. The normalphysiological function of ET-1 in the genitourinarytract remains elusive. It has been suggested that ET-1may promote sperm transport either in the male re-productive tract or in the uterine cavity by stimulatingmyometrial contractions [16]. In addition to the regu-lating ET-1 secretion, prostate secretory epithelial cellsare polarized and possess distinct apical and basolat-eral surfaces formed by tight junctions between thecells [33–36]. Cultured endothelial cells secrete ET-1through the basolateral surface [37] providing amechanism for the paracrine action of ET-1 on vascu-lar smooth muscle cells. By contrast, we have shownthat intact monolayers of PE cells secrete ET-1 throughthe apical surface, consistent with luminal secretion invivo and that epithelial cell tight junctions serve tosegregate endogenous ET-1 and stromal endothelinreceptors. These results suggest that in the prostateET-1 is intended as a secretory product and not as aparacrine effector of stromal smooth muscle. Disrup-tion of the compartmentalization of ET-1 may occur indisease states such as prostatitis [40]. Cellular tightjunction permeability can be altered by many factorsin vitro including insulin, disassembly of actin micro-filaments and phorbol esters [41–43]. In addition,physiological changes in steroids can elicit ultrastruc-tural changes in the intracellular junctions betweenprostatic epithelial cells in experimental animals [44].Atrophy of epithelial cells such that occurs during an-drogen deprivation [45,46] could also result in a‘‘leak’’ of normally segregated secretory products.

The intracellular signaling mechanisms evoked bybinding of ET-1 to endothelin receptors can result inshort-term and long-term biological events in the tar-get cell. In this study we showed that ET-1 can stimu-late phosphoinositide hydrolysis and thymidine up-take by cultured PS cells. The physiological conse-

Fig. 5. ET-1 induces [3H]-thymidine uptake by PS cells. Culturedprostate-derived PS cell monolayers were incubated in RPMI 1640,0.5% serum for 72 hr. Cells were then exposed to vehicle or ET-1(0.01–100 nM) for 23 hr and pulsed for 1 hr with [3H]-thymidine.Incorporation of label into DNA was determined. Data are rep-resented as the mean fold stimulation over control ±SEM for sixindependent experiments performed in sextuplicate. Filled circles,significant increase (P > 0.05) over control. Cultured prostate-derived PS cell monolayers were incubated in RPMI 1640, 0.5%serum for 72 hr as described under Materials and Methods.BQ123 (10 nM) or BQ788 (10 nM) were then added to mono-layers. After 5 min, ET-1 (1 nM) was added for 23 hr; cells werethen pulsed for 1 hr with [3H]thymidine. Incorporation of labelinto DNA was determined. Data are the mean ±SEM for fourindependent experiments performed in sextuplicate. The inhibi-tion produced with BQ788 was significant (P < 0.001).

248 Walden et al.

quences of endothelin receptor expression in PE cellsremain to be determined, in light of observations thatPE cells express endothelin receptors but do not re-spond to ET-1 in terms of phosphoinositide turnoveror mitogenesis. Uncoupling of endothelin receptorsfrom mitogenic signaling in PE cells, perhaps by re-duced expression of a specific G-protein, may be abiological mechanism to prevent continued autocrinestimulation of this cell type. Alternatively, the endo-thelin receptors may elicit as yet unidentified biologi-cal functions in PE cells, such as feedback regulation ofET synthesis.

In summary, these results demonstrate the exis-tence of a tightly regulated endothelin axis in the pros-tate. Presumably under normal circumstances suffi-cient quantities of ET-1 are not biologically active oravailable to elicit and effect on stromal cells. Alter-ations in the regulation of endothelin activity and bio-availability could have important consequences in thepathophysiology of prostate disease.

REFERENCES

1. Inoue A, Yanagisawa M, Kimura S, Kasuya Y, Miyauchi T, GotoK, Masaki T: The human endothelin family: Three structurallyand pharmacologically distinct isopeptides predicted by threeseparate genes. Proc Natl Acad Sci USA 1989;86:2863–2867.

2. Yanagisawa M, Kurihara H, Kimura S, Tomobe Y, Kobayashi M,Mitsui Y, Yazaki Y, Goto K, Masaki T: A novel potent vasocon-strictor peptide produced by vascular endothelial cells. Nature1988;332:411–415.

3. Bobik A, Grooms A, Millar JA, Mitchell A, Grinpukel S: Growth

factor activity of endothelin on vascular smooth muscle. Am JPhysiol 1990;258:C408–C415.

4. Takuwa N, Takuwa Y, Yanagisawa M, Yamashita K, Masaki T:A novel vasoactive peptide endothelin stimulates mitogenesisthrough inositol lipid turnover in Swiss 3T3 fibroblasts. J BiolChem 1989;264:7856–7861.

5. Vigne P, Marsault R, Breittmayer JP, Frelin C: Endothelin stimu-lates phosphatidylinositol hydrolysis and DNA synthesis inbrain capillary endothelial cells. Biochem J 1990;266:415–420.

6. Sokolovsky M: Endothelins and sarafotoxins: Physiologicalregulation, receptor subtypes and transmembrane signaling.Trends Biochem Sci 1991;16:261–264.

7. Kramer BK, Nishida M, Kelly RA, Smith TW: Endothelins: Myo-cardial actions of a new class of cytokines. Circulation 1992;85:350–356.

8. Bogoyevitch MA, Parker PJ, Sugden PH: Characterization ofprotein kinase C isotype expression in adult rat heart. Proteinkinase C-e is a major isotype present, and it is activated byphorbol esters, epinephrine, and endothelin. Circ Res 1993;72:757–767.

9. Simonson MS, Dunn MJ: Cellular signaling by peptides of theendothelin gene family. FASEB J 1990;4:2989–3000.

10. Simonson MS, Herman WH: Protein kinase C and protein tyro-sine kinase activity contribute to mitogenic signaling by endo-thelin-1. Cross-talk between G protein-coupled receptors andpp60c-src. J Biol Chem 1993;268:9347–9357.

11. van Biesen T, Hawes BE, Luttrell DK, Krueger KM, Touhara K,Porfiri E, Sakaue M, Luttrell LM, Lefkowitz RJ: Receptor-tyrosine-kinase and Gbg-mediated MAP kinase activation by acommon signaling pathway. Nature 1995;376:781–784.

12. Walsh PC: Benign prostatic hyperplasia. In Walsh PC, Retik AB,Stamey TA, Vaughan ED (eds): ‘‘Cambell’s Urology,’’ Vol 6.New York: WB Saunders, 1992, pp 1009–1027.

13. Berry SJ, Coffey DS, Walsh PC, Ewing LL: The development ofhuman benign prostatic hyperplasia with age. J Urol 1984;132:474–479.

14. Bartsch G, Muller HR, Oberholzer M, Rohr HP: Light micro-

Fig. 6. Polarity of ET-1 secretion and ET receptor localization inPE cells. PE cells (106) were seeded in 24-mm collagen-coatedTranswells, cultured until 100% confluent with medium changesevery 2 days and tested for monolayer integrity using the extra-cellular space marker [3H]methoxyinulin. Monolayers with lessthan 4% leak per hr from apical to basolateral compartments wereused for experimentation. A: ET-1 secretion into the apical andbasolateral media of intact monolayers was determined by RIA

after1 and 24 hr. Data are represented as the mean ±SEM forthree independent experiments performed in quadruplicate.B: [125I]-ET-1 (30,000 cpm) was added to either the apical orbasolateral medium of intact PE monolayers. The percentage ofthe label bound to the monolayer after 1 hr was determined. Dataare represented as the mean ±SEM for three independent experi-ments performed in quadruplicate. For controls, 1 mM unlabeledET-1 was added before the addition of [125I]-ET-1.

Endothelin-1 in the Prostate 249

scopic stereological analysis of the normal human prostate andof benign prostatic hyperplasia. J Urol 1979;122:487–491.

15. Shapiro E, Becich MJ, Hartanto V, Lepor H: The relative pro-portion of stromal and epithelial hyperplasia is related to thedevelopment of symptomatic benign prostate hyperplasia. JUrol 1992;147:1293–1297.

16. Casey ML, Byrd W, MacDonald PC: Massive amounts of im-munoreactive endothelin in human seminal fluid. J Clin Endo-crinol Metab 1992;74:223–225.

17. Hartter E, Woloszczuk W: Radioimmunoassay of endothelin.Lancet 1989;1:909.

18. Saito Y, Nakao K, Itoh H, Yamada T, Mukoyama M, Arai H,Hosoda K, Shirakami G, Shin-Ichi S, Jougasaki M, Morichika S,Imura H: Endothelin in human plasma and culture medium oraortic endothelial cells–detection and characterization with ra-dioimmune assay using monoclonal antibody. Biochem BiophysRes Commun 1989;161:320–326.

19. Langenstroer P, Tang R, Shapiro E, Divish B, Opgenorth T,Lepor H: Endothelin-1 in the human prostate: Tissue levels,source of production and isometric tension studies. J Urol 1993;150:495–499.

20. Kobayashi S, Tang R, Wang B, Opgenorth T, Langenstroer P,Shapiro E, Lepor H: Binding and functional properties of endo-thelin receptor subtypes in the human prostate. Mol Pharmacol1994;45:306–311.

21. Forray C, Bard JA, Wetzel JM, Chiu G, Shapiro E, Tang R, LeporH, Hartig PR, Weinshank RL, Branchek TA, Gluchowski C: Thealpha 1-adrenergic receptor that mediates smooth muscle con-traction in human prostate has the pharmacological propertiesof the cloned human alpha 1c subtype. Mol Pharmacol 1994;45:703–708.

22. Webb ML, Chao CC, Rizzo M, Shapiro RA, Neubauer M, LiuEC, Aversa CR, Brittain RJ, Treiger B: Cloning and expression ofan endothelin receptor subtype B from human prostate thatmediates contraction. Mol Pharmacol 1995;47:730–737.

23. Cheng YC, Prusoff WH: Relationship between the inhibitionconstant (Ki) and the concentration of inhibitor which causes 50per cent inhibition (I50) of an enzymatic reaction. Biochem Phar-macol 1973;22:3099–3109.

24. Scatchard G: The attractions of proteins for small molecules andions. Ann NY Acad Sci 1949;51:660–672.

25. Koreh K, Monaco ME: The relationship of hormone-sensitiveand hormone-insensitive phosphatidylinositol to phosphati-dylinositol 4,5-bisphosphate in the WRK-1 cell. J Biol Chem1986;261:88–91.

26. Wyllie AH: Glucocorticoid-induced thymocyte apoptosis is as-sociated with endogenous endonuclease activation. Nature1980;284:555–556.

27. Chomczynski P, Sacchi N: Single-step method of RNA isolationby acid guanidinium thiocyanate–phenol–chloroform extrac-tion. Anal Biochem 1987;162:156–159.

28. Seidah NG, Day R, Marcinkiewicz M, Chretien M: Mammalianpaired basic amino acid convertases of prohormones and pro-proteins. Ann NY Acad Sci 1993;680:135–146.

29. Kimura S, Kasuya Y, Sawamura T, Shinimi O, Sugita Y, Yanag-isawa M, Goto K, Masaki T: Conversion of big endothelin-1 to21-residue endothelin-1 is essential for expression of full vaso-

constrictor activity: Structure–activity relationships of big endo-thelin-1. J Cardiovasc Pharmacol 1989;13(suppl 5):S5–S7.

30. Opgenorth TJ, Wu-Wong JR, Shiosaki K: Endothelin-convertingenzymes. FASEB J 1992;6:2653–2659.

31. Kobayashi S, Tang R, Wang B, Opgenorth T, Stein E, Shapiro E,Lepor H: Localization of endothelin receptors in the humanprostate. J Urol 1994;151:763–766.

32. Coffey DS, Walsh PC: Clinical and experimental studies of be-nign prostatic hyperplasia. Urol Clin North Am 1990;17:461–475.

33. Douglas WH, Terracio L, Glass H: Isolation, culture and char-acterization of epithelial cells derived from rat ventral prostate.Anat Rec 1980;197:239–256.

34. Kachar B, Reese TS: Formation of misplaced and reflexive tightjunction strands in prostate epithelial cells. J Ultrastruct Res1983;82:90–95.

35. Kachar B, Pinto de Silva P: Freeze-fracture study of rat ventralprostate: The columnar epithelial cell. Am J Anat 1981;161:49–69.

36. Terracio L, Douglas WH, Pennachio D, Vena RL, Ofner P: Pri-mary epithelial cell cultures derived from canine prostate: Iso-lation, culture, and characterization. Am J Anat 1982;164:311–332.

37. Wagner OF, Christ G, Wojta J, Vierhapper H, Parzer S, NowotnyPJ, Schneider B, Waldhausl W, Binder BR: Polar secretion ofendothelin-1 by cultured endothelial cells. J Biol Chem 1992;267:16066–11608.

38. Yang Y, Chisholm GD, Habib FK: The distribution of PSA, ca-thepsin-D, and pS2 in BPH and cancer of the prostate. Prostate1992;21:201–228.

39. Sawamura T, Shinmi O, Kishi N, Sugita Y, Yanagisawa M, GotoK, Masaki T, Kimura S: Analysis of big endothelin-1 digestionby cathepsin D. Biochem Biophys Res Commun 1990;172:883–889.

40. Hasui Y, Marutsuka K, Asada Y, Ide H, Nishi S, Osada Y: Re-lationship between serum prostate specific antigen and histo-logical prostatitis in patients with benign prostatic hyperplasia.Prostate 1994;25:91–96.

41. Engle MJ, Grove ML, Becich MJ, Mahmood A, Alpers DH: Ap-pearance of surfactant-like particles in apical medium of Caco-2cells may occur via tight junctions. Am J Physiol 1995;268:C1401–C1432.

42. Marmorstein AD, Mortell KH, Ratcliffe DR, Cramer EB: Epithe-lial permeability factor: A serum protein that condenses actinand opens tight junctions. Am J Physiol 1992;262:C1403–C1410.

43. McRoberts JA, Aranda R, Riley N, Kang H: Insulin regulates theparacellular permeability of cultured intestinal epithelial cellmonolayers. J Clin Invest 1990;85:1127–1134.

44. Merk FB, Leav I, Kwan PW, Ofner P: Effects of estrogen andandrogen on the ultrastructure of secretory granules and inter-cellular junctions in regressed canine prostate. Anat Rec 1980;197:111–132.

45. Wendel EF, Brannen GE, Putong PB, Grayhack JT: The effect oforchiectomy and estrogens of benign prostatic hyperplasia. JUrol 1972;108:116–119.

46. Uke E, Lee C, Grayhack JT: The effect of CIS hydroxyproline onventral prostatic growth in rats. J Urol 1983;129:171–174.

250 Walden et al.