The Prostate 50:4 ^14 (2002)

PhospholipaseA2Degradation ProductsModulateEpithelialand Stromal 5a -ReductaseActivityofHumanBenign ProstaticHyperplasia InVitro

Heike Weisser,* Tjalf Ziemssen, and Michael Krieg

InstituteofClinicalChemistry,TransfusionandLaboratoryMedicine,UniversityClinicBergmannsheil,Bochum,Germany

BACKGROUND. Recent studies have demonstrated the inhibition of 5a-reductase activity inhuman prostate by phospholipases. Among those phospholipases, phospholipase A2 cleavesone of the acyl chains from phospholipids, thereby producing fatty acids and lysopho-spholipids such as LPC, LPS, and LPE. Therefore, we were interested in the effect of thoselysophospholipids on 5a-reductase activity in human benign prostatic hyperplasia (BPH).METHODS. In a ®rst set of experiments, cell homogenates were incubated withphospholipase A2 either in the presence or absence of albumin, which is known to bindfatty acids and lysophospholipids. Thereafter, the effect of lysophospholipids of knownstructure on 5a-reductase activity was investigated.RESULTS. In epithelium and stroma of human BPH, 5a-reductase activity was inhibited in adose-dependent manner by phospholipase A2. In the presence of albumin, this inhibition wasenhanced. In epithelium, LPC at low concentration yielded a dose-dependent stimulation of5a-reductase activity up to 167%. At higher concentrations, epithelial as well as stromal 5a-reductase activity was inhibited signi®cantly. As indicated by results of enzyme kineticanalyses, the LPC-mediated activation in the epithelium results from an increase of the activepopulation of 5a-reductase. In contrast, LPC reduces the af®nity of epithelial 5a-reductase totestosterone. LPE had no effect on epithelial 5a-reductase, whereas stromal 5a-reductase wasinhibited in a dose-dependent manner up to 46%. Finally, LPS stimulated epithelial andstromal 5a-reductase activity; this stimulation was signi®cantly stronger in epithelium (296%)than in stroma (163%). The LPC-mediated effects could be neutralized by the addition ofalbumin.CONCLUSIONS. The present data on BPH tissue suggest that lysophospholipids may playa speci®c and structure-related role in the posttranslational regulation of human prostatic5a-reductase. Prostate 50: 4±14, 2002. # 2002 Wiley-Liss, Inc.

KEY WORDS: 5a-reductase; modulation; lipid environment; lysophospholipid; humanprostate

INTRODUCTION

The irreversible conversion of testosterone to itspotent metabolite 5a-dihydrotestosterone (DHT), cat-alyzed by 5a-reductase, represents a crucial step forandrogen-mediated action in human prostate [1±3].Based on results of enzyme kinetic analyses inseparated epithelium and stroma of human prostate,the existence of isoenzymes of 5a-reductase has beenpostulated [3±5]. Recent molecular and genetic studiesdemonstrate the existence of at least two 5a-reduc-

Abbreviations: BPH, benign prostatic hyperplasia; DHT,5a-dihydrotestosterone, 5a-androstan-17b-ol-3-one; LPC, lysopho-sphatidylcholine; LPE, lysophosphatidylethanolamine; LPS, lyso-phosphatidylserine; Phospholipase A2, phosphatide acylhydrolase,EC 3.1.1.4; Testosterone, 17b-hydroxy-4-androsten-3-one

*Correspondence to: Priv.-Doz. Dr. med. Heike Weisser, Institute ofClinical Chemistry, Transfusion and Laboratory Medicine, Univer-sity Clinic Bergmannsheil, BuÈ rkle-de-la-Camp-Platz 1, D-44789Bochum, Germany. E-mail: heike.weisser@ruhr-uni-bochum.de

Received 25 September 2000; Accepted 30 August 2001

ß 2002 Wiley-Liss, Inc.DOI10.1002/pros.10027

tases, termed 5a-reductase type 1 and type 2 [6±8].Studies on the intraprostatic expression of these twoisoenzymes have yielded con¯icting results [9±11].Moreover, the biochemical characteristics (Km, Vmax,Ki) of 5a-reductase measured in epithelium andstroma of human BPH [3,5,12,13] are only in part inaccordance with those of recombinant 5a-reductaseisoenzymes expressed in various transfected cells[8,9]. Thus, it is conceivable that the activity of 5a-reductase in epithelium and stroma of human prostateis not only dictated by cell-type speci®c expression of5a-reductase isoenzymes, but also by posttranslationalevents. In this context, it is already known thatmembrane-bound 5a-reductase loses nearly all of itsactivity if removed from the membrane environment[14]. Thus, membrane components of the prostaticcell are probably involved in the regulation of 5a-reductase activity.

One main constituent of such cellular membranesare lipids. Changes in the composition of those lipidsgive rise to a variety of altered membrane properties.Those alterations could lead to posttranslationalmodulation of the 5a-reductase activity. In this con-text, we have already demonstrated that there aresigni®cant differences between epithelium and stromain both the lipid and phospholipid composition and inthe fatty acid composition of the phospholipids[15,16]. Moreover, in epithelium the fatty acid compo-sition seems to be affected by aging [16]. These datasupport the hypothesis that lipids could be somehowinvolved in the modulation of 5a-reductase activityin human prostate. This is underlined by results ofstudies in rats indicating that epididymal 5a-reductaseactivity can be modulated in vitro by the addition ofphospholipases and phospholipids [17,18]. Veryrecently, we also demonstrated such modulation forepithelial and stromal 5a-reductase of human BPH[19]. In those studies, 5a-reductase activity wassigni®cantly inhibited by phospholipase A2 and C,the pattern of inhibition being strikingly differentbetween epithelial and stromal 5a-reductase.

Phospholipase A2 cleaves one of the acyl chainsfrom phospholipids, thereby producing degradationproducts, i.e., free fatty acids and lysophospholipids.Previous studies have already demonstrated that va-rious fatty acids are capable of inhibiting 5a-reductaseactivity in human BPH [20,21] and prostate cancercells in culture [22]. This study was conducted toevaluate the impact of lysophospholipids on theactivity of epithelial and stromal 5a-reductase inhuman BPH. This impact was investigated in twoways. First, epithelial and stromal cell homogenateswere treated with phospholipase A2 either in thepresence or absence of albumin, which is known tobind fatty acids and lysophospholipids [23,24]. Sec-

ond, the impact of various lysophospholipids of knownstructure on 5a-reductase activity was investigated.

MATERIALSANDMETHODS

Chemicals

[1a,2a-3H]-Testosterone (S.A. 2.08 TBq/mmol) waspurchased from Amersham Buchler (Braunschweig,Germany), and the unlabeled steroids from SigmaChemical Co. (St. Louis, MO). The eluant for high-performance liquid chromatography (HPLC) and thescintillation solution Rialuma were obtained fromMallinckrodt Baker (Griesheim, Germany). Phospho-lipase A2 from Apis mellifera, L-a-LPC, L-a-LPE type II,and L-a-lysophosphatidyl-L-serine were bought fromSigma. All other chemicals used were from Merck AG(Darmstadt, Germany), Serva (Heidelberg, Germany),and Boehringer (Mannheim, Germany).

Tissue Preparation

Specimens of BPH tissue were obtained frompatients undergoing suprapubic prostatectomy. Ineach instance, written consent for the use of tissue inthis study was given. The age of the patients rangedfrom 66 to 82 years (mean, 74 years). The clinicaldiagnosis of BPH was made by an experienced urolo-gist and based on subjective and objective symptomsof infravesical obstruction with reduced urinary peak¯ow (< 12 ml/sec), increased postvoidal urine volume,and urinary retention. In case of uncertainty, uro-dynamic evaluation was performed. After surgicalextirpation, the tissue was immediately chilled inice-cold 150 mmol/L NaCl. All tissue specimenswere divided into small pieces and stored in plastictubes at ÿ1968C. Representative sections from alltissue specimens were ®xed in phosphate-bufferedformalin, embedded in paraf®n, and stained withhematoxylin and eosin. The stained sections wereexamined by light microscopy. The histology of allexamined prostates was that of glandular±stromalhyperplasia.

The prostatic epithelium and stroma wereseparated as described previously [25]. Using thisseparation procedure, the relative purity of theepithelial and the stromal fraction was more than83%, as estimated by measuring acid phosphatase as amarker for epithelial cells and hydroxyproline as amarker for stromal elements in both tissue fractions[5,26±28].

For enzyme measurement, aliquots of the frozenhomogenates were pulverized in a porcelain mortarchilled with liquid nitrogen. The powder was allowedto thaw in small tubes which were kept in an ice-bathfor approximately 1 hr.

Modulation of 5a -Reductaseby PLA2Degradation Products 5

Measurementof 5a -ReductaseActivity

The 5a-reductase activity was measured underoptimized incubation conditions. Brie¯y, the incuba-tion mixtures (®nal volume 202 ml), each preparedin duplicate, were tissue homogenate (300±700 mgprotein) diluted in 10 mmol/L Tris buffer (2 mmol/LEDTA, 5 mmol/L NaN3, 10 mmol/L MgCl2 �H2O, pH7.5 at 48C), a NADPH-generating system (5 mmol/Lglucose-6-phosphate, 0.6 U glucose-6-phophate dehy-drogenase), varying concentrations (14±585 nM) oftestosterone (either as [3H]testosterone alone or [3H]testosterone plus unlabeled testosterone), and varyingconcentrations of phospholipase A2 or lysophospholi-pids. Because of the known calcium dependence ofphospholipase A2, the reaction buffer contained3 mmol/L CaCl2 when the effect of phospholipaseA2 was examined. When included, lysophospholipidsin chloroform:methanol (1:1, v:v) were placed inincubation tubes and the solvent removed under astream of nitrogen. After the addition of all otherreagents, except the cell homogenate, testosterone andNADPH, the tubes were vortexed and sonicated untilno further decrease in turbidity of the sample wasobserved visually. The reaction was started by adding0.5 mmol/L NADPH, and the mixtures were incu-bated at 378C for 15 min. The reactions were stoppedby adding 3 ml of ether, and the steroids wereextracted twice with ether (2� 60 sec). The etherphases were evaporated to dryness (Vortex Evapora-tor; Haake Buchler, Saddle Brook, NJ), redissolved in500 ml ether, and again evaporated to dryness. Thedried steroids were redissolved in 50 ml acetonitrilecontaining 100 mg of the following steroids as tracer:

testosterone, DHT, 5a-androstane-3a,17b-diol, and5a-androstane-3b,17b-diol. The steroids were sepa-rated by reverse-phase HPLC as previously described[27]. The unlabeled tracers were included for thedetection of the HPLC separated metabolites by arefractive index detector. The eluant was composed ofa ®ltered and helium degassed mixture of acetoni-trile:methanol:H2O (53:12:35, v/v/v). On average,the recovery was 80% of the starting material. The5a-reductase activity was calculated from the percen-tage of radioactively labeled 5a-reduced metabolites,taking into consideration recovery, blank values, thespeci®c activity of [3H]testosterone, and the ratio ofadded [3H]testosterone to unlabeled testosterone.

OtherMethods

Protein content was determined according to themethod of Lowry et al. [29], using bovine serumalbumin as standard. Acid phosphatase activity (EC3.1.3.2) was measured by the method of Walter andSchuÈ tt [30]. Michaelis-Menten constant (Km) andmaximal velocity (Vmax) values were derived fromLineweaver±Burk plots [31], from which regressionlines were computed by the method of least squares.The goodness of ®t was described by the correlationcoef®cient, r. The mean correlation coef®cients (� SD)of the regression lines for untreated epitheium andstroma were 0.994� 0.005 and 0.996� 0.006, res-pectively. The respective data for LPC (500, 1000,5000 mg/ml) treated homogenates were 0.996�0.004, 0.999�0.001, and 0.992� 0.010, respectively, inepithelium and 0.995� 0.006, 0.992� 0.009, and0.996� 0.004, respectively, in stroma. There were no

Fig. 1. Mean 5a -reductase activity (%� SEM) in (a) epithelium and (b) stroma (n� 8) in the presence of phospholipase A2.The activities(pmol/mgprotein �h) of 5a -reductase (epithelium: 53�17; stroma:193�22) in the absence ofphospholipasewere taken as100%.The concen-tration of testosteronewas 585 nM.

6 Weisseret al.

Fig. 2. Mean5a -reductase activity (%� SEM) in (a) epitheliumand (b) stroma (n� 6) in thepresence ofphospholipaseA2with andwithoutthe addition of albumin (100 mg/ml); (c) Ratio (mean� SEM) between 5a -reductase activity with and without albumin (n� 6). *P< 0.05,epithelium vs. stroma. The activities (pmol/mg protein �h] of 5a -reductase (epithelium: 52� 5 (with albumin), 45� 4 (without albumin);stroma: 77�11 (with albumin), 84� 22 (without albumin)) in the absence of phospholipasewere taken as100%.The concentration of testo-steronewas 585 nM.

Fig. 3 (overleaf). Mean5a -reductase activity (%� SEM) in (a) epithelium and (b) stroma (n� 6) in thepresence of LPC,LPE, and LPS.Theactivities (pmol/mgprotein �h) of 5a -reductase (epithelium: 40� 6 (LPC), 22� 3 (LPE), 29� 4 (LPS); stroma:186�20 (LPC),189� 22 (LPE),181�18 (LPS)) in the absence of lysophospholipidwere taken as100%.The concentration of testosteronewas 585 nM.

Modulation of 5a -Reductaseby PLA2Degradation Products 7

8 Weisseret al.

signi®cant differences between the r values in LPCtreated and untreated homogenates ((ANOVA) fol-lowed by the Bonferroni post test; P< 0.05 wasconsidered signi®cant). As described previously, theef®ciency of testosterone 5a-reductase (Vmax/Km) wasestimated by the equation v�Vmax � S/(Km� S), whereS is the endogenous testosterone concentration [5,26±28]. Because the Km value of the 5a-reductase aremuch higher than S under in vivo conditions [32,33],it follows that Km� S&Km and v�Vmax/Km � S.Assuming identical substrate concentrations, thepotential enzymatic ef®ciency, previously calledpotential activity, is given by the Vmax/Km ratio. Thestatistical signi®cance of the mean values was deter-mined by means of a Student t-test and a P value lessthan 0.05 was considered signi®cant.

RESULTS

Bindingof the PhospholipaseA2DegradationProductsbyAlbumin

In epithelium and stroma of human BPH, 5a-reductase activity was inhibited in a dose-dependentmanner by phospholipase A2 (Fig. 1). Albumin itselfhad no signi®cant effect on 5a-reductase activity.However, the inhibition of 5a-reductase activity byphospholipase A2 was signi®cantly enhanced by theaddition of albumin, which is known to bind fattyacids and lysophosphospholipids (Fig. 2a and b).Moreover, as indicated by the ratio of 5a-reductaseactivity with to that without albumin (Fig. 2c), en-hancement was signi®cantly stronger in the epi-thelium than in the stroma. Additionally, this

albumin-mediated effect increased proportianate tothe increase in phospholipase A2 concentration both inepithelium and stroma.

Direct IncubationWithVarious LysophospholipidsDose^ResponseCurves

The direct effects of LPC, LPE, and LPS on 5a-reductase activity have been investigated (Fig. 3). Inthe epithelium, LPC showed a signi®cant biphasicmodulation. At relatively low concentrations, 5a-reductase activity (relative activity� SEM) was stimu-lated in a dose-dependent manner by as much as167%� 6 (P< 0.001) of the control. At higher concen-trations (> 1000 mg/ml), epithelial and stromal 5a-reductase activity decreased signi®cantly (P< 0.005).LPE had no effect on epithelial 5a-reductase, whereasin the stroma a signi®cant (P< 0.002) and dose-dependent inhibition of as much as 46%� 8 of thecontrol was observed. LPS stimulated epithelial andstromal 5a-reductase activity in a dose-dependentmanner, and this stimulation was signi®cantly(P< 0.03) stronger in epithelium (296%� 26) than instroma (163%� 9).

As shown for LPC (Fig. 4), the lysophospholipid-mediated stimulatory and inhibitory effects on epithe-lial and stromal 5a-reductase activity could be almostcompletely neutralized by addition of albumin.

EnzymeKinetic Analyses

To describe in more detail the LPC-mediatedmodulation of 5a-reductase activity, the Vmax, Km,and Vmax/Km of 5a-reductase were determined using

Fig. 4. 5a -reductaseactivity (%) in (a)epitheliumand (b) stromaofonehumanBPHin thepresence ofLPCwith andwithout theadditionofalbumin (100mg/ml).The activities of 5a -reductasein the absence ofLPCwere taken as100%.The concentrationof testosteronewas 585nM.

Modulation of 5a -Reductaseby PLA2Degradation Products 9

Lineweaver±Burk plots for various concentrations oftestosterone and LPC. Examples of such plots areshown in Figure 5. In epithelium, LPC showed asigni®cant biphasic modulation of Vmax (Fig. 6). Atrelatively low concentrations, the mean Vmax valueincreased signi®cantly (P< 0.005), to 237% comparedwith control values. Then, an increase in the LPCconcentration to 5,000 mg/ml caused a decrease in theVmax to 85% of control values. In contrast, in stromathe mean Vmax decreased at all LPC concentrations.Moreover, only in the epithelium, mean Km valuesincreased signi®cantly (P< 0.05) with addition of LPC.As a result of these effects on Vmax and Km, both in

epithelium and stroma, the potential ef®ciency (Vmax/Km) of 5a-reductase was decreased by LPC (Fig. 7). Instroma, this decrease was already signi®cant at aconcentration of 1,000 mg/ml, whereas in epitheliumthe decrease became signi®cant at 5,000 mg/ml. More-over, using a concentration of 500 mg/ml LPC in theepithelium, even a slight increase of Vmax/Km wasobserved. Therefore, at this concentration the potentialef®ciency was signi®cantly (P< 0.05) higher in theepithelium as compared to the stroma.

DISCUSSION

The present study demonstrates for the ®rst timethat epithelial and stromal 5a-reductase activity inhuman BPH can be modulated by phosphalipase A2

degradation products such as lysophsopholipids. Inthis context, we already have shown that phospholi-pase A2 and C type III led to a strong inhibition of the5a-reductase activity in the epithelium and stroma ofhuman BPH. Based on the estimation of IC50 values(50% inhibitory concentration), the inhibitory potencyof phospholipase C type III was signi®cantly strongeras compared to phospholipase A2. Moreover, incontrast to phospholipase C type III, a short timeincubation (5 min) by phospholipase A2 (5 U/ml)induced a slight, but signi®cant, stimulation of 5a-reductase activity [19]. In the present study, in order tocharacterize in more detail this phospholipase A2

mediated modulation of 5a-reductase activity, theimpact of the degradation products (being generatedthrough the action of phospholipase A2) on theepithelial and stromal 5a-reductase activity in humanBPH was investigated. The study was carried out onso-called cell-free homogenates of epithelium andstroma. Thus, the 5a-reductase enzyme pool mightconsist of 5a-reductase type I and type II. Previousstudies based on the Vmax/Km ratio, which representsthe potential ef®ciency of 5a-reductase, indicate thatthe epithelial and stromal 5a-reductase is ef®cient inmetabolizing testosterone at neutral pH [34]. More-over, investigations on the sensitivity to ®naste-ride, undertaken at a neutral pH, demonstrated a Ki

value (7 nM) in epithelium that is similar to the®nasteride sensitivity of the cDNA-encoded 5a-reduc-tase type II [12]. However, based on other studiesconcerning the in¯uence of pH on 5a-reductaseactivity [7,8], it has to be considered that in the enzymeactivities measured herein, a neutral pH might pre-dominantly re¯ect that of the 5a-reductase type I.

To clarify whether the inhibition of 5a-reductaseactivity by phospholipase A2 is due to the phospho-lipase A2 degradation products, a simultaneousincubation with albumin was performed knowingthat albumin binds and thus neutralizes those

Fig. 5. Lineweaver^Burk plots of the LPC mediated effect on5a -reductase activity in epithelium and stroma of humanBPH.Theeffect was determined throughout in a contemporaneous assay,carried out at varying LPC aswell as testosterone concentrations.

10 Weisseret al.

products [23,24]. If the degradation products them-selves actually are responsible for the 5a-reductaseinhibition, the inhibitory potency of phospholipaseA2 should be diminished by the addition of albumin.However, just the opposite was found. The experi-ments with albumin led to an enhancement of thephospholipase A2 mediated 5a-reductase inhibition.So it seems that the inhibition of 5a-reductase byphospholipase A2 is not primarily mediated by thedegradation products. Rather, the phospholipase A2

degradation products seem to counterbalance the directaction of the phospholipase. This conclusion issupported by a study of Cooke and Robaire, demon-strating by a sequential addition of the phospholipasesthat lysophospholipids and diacylglycerols can inpart maintain the activity of rat microsomal 5a-reductase [35].

In a second experimental step, we focussed on thedirect effect of various lysophospholipids on 5a-reductase activity. In principle, studies on the effect

Fig. 6. Effect of LPC on (a) Vmax and (b) Km of 5a -reductase in epithelium (upper panels) and stroma (lower panels) (n� 5). Data areexpressed asmean� SEM.The concentration of testosteronewas14 ^585 nM. *P< 0.05; **P< 0.01; ***P< 0.005 vs. control.

Modulation of 5a -Reductaseby PLA2Degradation Products 11

of phospholipids can be carried out with solubilizedor with native, membrane-bound 5a-reductase asdone in the present study. At present, merely apartial puri®cation of 5a-reductase could be obtained[14±38]. In this context, it could be demonstratedthat after partial solubilization of epididymal 5a-reductase, a phospholipase A2 or C treatment resultsin a further 50% inhibition of 5a-reductase activityindicating that 5a-reductase remains embedded in aphospholipid matrix after solubilization. Therefore,investigations on so-called solubilized 5a-reductaseare in actual fact studies on membrane-bound 5a-reductase which is present in reduced amount. Thisconception is supported by studies of Cooke andRobaire [17] demonstrating a dilauroyl phosphatidyl-choline-mediated activation of 5a-reductase whichwas similar for native and solubilized 5a-reductase.Moreover, Ichihara and Tanaka [39] discussed that thephospholipid speci®city of 5a-reductase activationdepends on the solubilization procedure used. Allin all, investigations with native, membrane-bound5a-reductase seem to re¯ect the in vivo situation betterthan studies with solubilized enzyme.

It was found that the various lysophospholipids didnot give rise to a uniform, unspeci®c modulation of 5a-reductase activity. Rather, the kind of modulation, i.e.,stimulation or inhibition, depends on the lysopho-spholipid class used. Moreover, the impact of thelysophospholipids was signi®cantly different betweenthe epithelium and stroma. In studies elsewhere,the effect of lysophospholipids on 5a-reductase activ-ity in rat and guinea pig has been investigated[18,38,39]. All in all, those studies also demonstrate

that the lysophospholipid mediated 5a-reductasemodulation is not a uniform event, but rather species-and organ-dependent.

As yet, little is known about the molecularmechanism by which lysophospholipids modulatehuman prostatic 5a-reductase. In principle, it couldbe a detergent effect. In this context, Enderle-Schmitt[40] demonstrated that the nuclear 5a-reductase of therat ventral prostate can in fact be partially solubilizedby LPC. However, concerning the human prostatic5a-reductase, this explanation seems unlikely sinceclassical detergents such as Triton X-100 always giverise to a uniform and pronounced inhibition, but neverto a stimulation of 5a-reductase activity [18,38,40,41].Moreover, if lysophospholipids act simply as deter-gents, the effect on 5a-reductase should always bethe same, irrespective of the lysophospholipid classused.

Furthermore, in this respect also the signi®cantdifferences between the epithelium and the stromacannot be explained by a detergent effect. It is morelikely that the difference is due to the differentphospholipid environment of epithelium and stroma,where the 5a-reductase is embedded. In accordancewith this assumption are studies of Oishi et al. [42].They showed that the effect of LPC on protein kinase Cwill be enhanced by the addition of phosphatidyl-serine or phosphatidic acid.

It is conceivable that lysophospholipids may begenerated from the corresponding phospholipidsthrough phospholipase A2, inasmuch as recentlyphospholipase A2 has been detected in the epithelialcells of the human prostate [43]. In this context, it

Fig. 7. Effect of LPC onVmax/Km of 5a -reductase in (a) epithelium and (b) stroma (n� 5).Data are expressed asmean� SEM.The concen-tration of testosteronewas14 ^585 nM. *P< 0.05; **P< 0.005 vs. control.

12 Weisseret al.

is remarkable that various studies indicated thatphospholipase A2 activity is androgen-dependent[44,45]. Thus, the impact of phospholipase A2 on the5a-reductase activity, and thereby, indirectly on theDHT milieu in the prostate is only one thinkablecellular event. Alternatively, it cannot be ruled out thatthe prostatic hormonal milieu itself in¯uences theactivity of the phospholipase A2. However, the genera-tion of lysophospholipids under physiological condi-tions, which results in activation or inhibition ofprostatic 5a-reductase, deserves further investiga-tions, especially with respect to the relatively highlysophospholipid levels necessary to affect 5a-reduc-tase activity in-vitro.

CONCLUSIONS

The present data on BPH tissue suggest thatlysophospholipids could have a regulatory impact onhuman prostatic 5a-reductase activity. The impactseems to be speci®c and structure-related. It isconceivable that lysophospholipids may be generatedfrom the corresponding phospholipids through phos-pholipase A2.

ACKNOWLEDGMENTS

We thank Prof. Dr. Th. Senge (Department ofUrology, University Clinic Marienhospital, Herne,Germany) for supplying the prostatic tissue, and Mrs.T. SchluÈ ter, C. Rohrbach, and P. Fritz for their excellenttechnical assistance.

REFERENCES

1. Coffey DS, Pienta KJ. New concepts in studying the control ofnormal and cancer growth of the prostate. In: Coffey DS,Gardner WA Jr., Bruchovsky N, Resnick MI, Karr JP, editors.Current concepts and approaches to the study of prostatecancer. New York: Alan R. Liss Inc.; 1987. p 1±73.

2. Krieg M, Tunn S. Androgens and human benign prostatichyperplasia (BPH). In: Nieschlag E, Behre HM, editors.Testosterone. Action, de®ciency, substitution. Berlin: SpringerVerlag; 1990. p 219±244.

3. Krieg M, Weisser H, Tunn S. Potential activities of androgenmetabolizing enzymes in human prostate. J Steroid BiochemMolec Biol 1995;53:395±400.

4. Rennie PS, Bruchovsky N, McLoughlin MG, Batzold FH,Dunstan-Adams EE. Kinetic analysis of 5a-reductase isoen-zymes in benign prostatic hyperplasia (BPH). J Steroid Biochem1983;19:169±173.

5. Tunn S, Hochstrate H, Grunwald I, FluÈ chter St H, Krieg M.Effect of aging on kinetic parameters of 5a-reductase inepithelium and stroma of normal and hyperplastic humanprostate. J Clin Endocrinol Metab 1988;67:979±985.

6. Andersson S, Russell DW. Structural and biochemicalproperties of cloned and expressed human and rat steroid 5a-reductases. Proc Natl Acad Sci USA 1990;87:3640±3644.

7. Andersson S, Berman DM, Jenkins EP, Russell DW. Deletion ofsteroid 5a-reductase 2 gene in male pseudohermaphroditism.Nature 1991;354:159±161.

8. Russell DW, Wilson JD. Steroid 5a-reductase: Two genes/twoenzymes. Ann Rev Biochem 1994;63:25±61.

9. Thigpen AE, Cala KM, Russell DW. Characterization of Chinesehamster ovary cell lines expressing human steroid 5a-reductaseisozymes. J Biol Chem 1993;268:17404±17412.

10. Bonnet P, Reiter E, Bruyninx M, Sente B, Dombrowicz D, deLeval J, Closset J, Heenen G. Benign prostatic hyperplasiaand normal prostate aging: Differences in types I and II 5a-reductase and steroid hormone receptor messengerribonucleic acid (mRNA) levels, but not in insulin-likegrowth factor mRNA levels. J Clin Endocrinol Metab 1993;77:1203±1208.

11. Bruchovsky N, Sadar MD, Akakura K, Goldenberg SL,Matsuoka K, Rennie PS. Characterization of 5a-reductasegene expression in stroma and epithelium of human prostate.J Steroid Biochem Molec Biol 1996;59:397±404.

12. Weisser H, Tunn S, Debus M, Krieg M. 5a-reductase inhibitionby ®nasteride (Proscar1) in epithelium and stroma of humanbenign prostatic hyperplasia. Steroids 1994;59:616±620.

13. Weisser H, Krieg M. Kinetic analysis of androstenedione 5a-reductase in epithelium and stroma of human prostate. Steroids1997;62:589±594.

14. Sargent NSE, Habib FK. Partial puri®cation of human prostatic5a-reductase (3-oxo-5a-steroid:NADP� 4-ene-oxido-reductase;EC 1.3.1.22) in a stable and active form. J Steroid Biochem MolBiol 1991;38:73±77.

15. Weisser H, Krieg M. Lipid composition in epithelium andstroma of human benign prostatic hyperplasia. Prostate 1997;30:41±46.

16. Weisser H, Krieg M. Fatty acid composition of phospholipids inepithelium and stroma of human benign prostatic hyperplasia.Prostate 1998;36:235±243.

17. Cooke MG, Robaire B. Modulation of epididymal D4-steroid 5a-reductase activity in vitro by the phospholipid environment.J Biol Chem 1985;260:7489±7495.

18. Kawai C, Ichihara K. Phospholipid requirement of epididymaltestosterone 5a-reductase and phospholipid composition ofepididymal microsomes. Steroids 1993;58:472±477.

19. Weisser H, Ziemssen T, Krieg M. In vitro modulation of steroid5a-reductase activity by phospholipases in epithelium andstroma of human benign prostatic hyperplasia. Steroids 2001;66:521±528.

20. Weisser H, Tunn S, Behnke B, Krieg M. Effects of Sabal serrulataextract and its subfractions on 5a-reductase in human benignprostatic hyperplasia. Prostate 1996;28:300±306.

21. NiederpruÈ m HJ, Schweikert HU, ThuÈ roff JW, ZaÈnker KS.Inhibition of steroid 5a-reductase activity by aliphatic fattyacids. Candidates for chemoprevention of prostate cancer. AnnNY Sci 1995;768:227±230.

22. Liang T, Liao S. Inhibition of steroid 5a-reductase byspeci®c aliphatic unsaturated fatty acids. Biochem J 1992;285:557±562.

23. Meissner G, Fleischer S. The role of phospholipid in Ca2�-stimulated ATPase activity of sarcoplasmic reticulum. BiochimBiophys Acta 1972;255:19±23.

24. Pikula S, Epstein L, Martonosi A. The relationship be-tween phospholipid content and Ca2�-ATPase activity inthe sarcoplasmic reticulum. Biochim Biophys Acta 1994;1196:1±13.

Modulation of 5a -Reductaseby PLA2Degradation Products 13

25. Krieg M, KloÈtzl G, Kaufmann J, Voigt KD. Preferential tissue forandrogen metabolism and oestrogen binding. Acta Endocrinol(Copenh) 1981;96:422±432.

26. Tunn S, Haumann R, Hey J, FluÈ chter St H, Krieg M. Effect ofaging on kinetic parameters of 3a(b)-hydroxysteroid oxidor-eductases in epithelium and stroma of human normal andhyperplastic prostate. J Clin Endocrinol Metab 1990;71:732±739.

27. Tunn S, Claus S, Schulze H, Braun B-E, Krieg M. 5a-androstane-3û,17b-diol hydroxylating enzymes in stroma and epithelium ofhuman benign prostatic hyperplasia (BPH). J Steroid Biochem1987;28:257±265.

28. Tunn S, Schulze H, Krieg M. 17b-hydroxysteroid oxidoreduc-tase in epithelium and stroma of human prostate. J SteroidBiochem 1993;46:91±101.

29. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Proteinmeasurement with the folin phenol reagent. J Biol Chem 1951;193:265±275.

30. Walter K, SchuÈ tt C. Saure und alkalische Phosphatase in Serum(Zwei-Punkt- Methode). In: Bergmeier HU, editor. Methodender enzymatischen Analysen. Weinheim: Verlag Chemie; 1970.p 818±822.

31. Lineweaver H, Burk D. The determination of enzyme dissocia-tion constants. J Am Chem Soc 1934;56:658±666.

32. Bolton NJ, Lahtonen R, Hammond GL, Vihko R. Distributionand concentrations of androgens in epithelial and stromalcompartments of the human benign hypertrophic prostate.J Endocr 1981;90:125±131.

33. Krieg M, Nass R, Tunn S. Effect of aging on endogenous level of5a-dihydrotestosterone, testosterone, estradiol, and estrone inepithelium and stroma of normal and hyperplastic humanprostate. J Clin Endocrinol Metab 1993;77:375±381.

34. Weisser H, Stachon A, Krieg M. pH-dependency of 5a-reductase kinetic characteristics in epithelium and stroma ofhuman BPH. Eur Urol 1999;36:492±493.

35. Cooke MG, Robaire B. Differential effects of combinations ofphospholipase A2 and phospholipase C on the activity of ratepididymal nuclear and microsomal 4-ene steroid 5a-reductase.J Steroid Biochem 1987;26:581±588.

36. Watkins WJ, Goldring CEP, Gower DB. Properties of 4-ene-5a-reductase and studies on its solubilization from porcinetesticular microsomes. J Steroid Biochem 1988;29:325±331.

37. Quemener E, Amet Y, di Stefano S, Fournier G, Floch HH,Abalain JH. Puri®cation of testosterone 5a-reductase fromhuman prostate by a four-step chromatographic procedure.Steroids 1994;59:712±718.

38. Ichihara K, Tanaka C. Phospholipid requirement of progester-one 5a-reductase from gastric mucosa microsomes of guineapig. Lipids 1991;26:531±535.

39. Ichihara K, Tanaka C. Some properties of progesterone 5a-reductase solubilized from rat liver microsomes. Biochem Int1987;15:1005±1011.

40. Enderle-Schmitt U, Neuhaus C, AumuÈ ller G. Solubilization ofnuclear steroid 5a-reductase from rat ventral prostate. BiochimBiophys Acta 1989;987:21±28.

41. Houston B, Chisholm GD, Habib FK. Solubilization of humanprostatic 5a-reductase. J Steroid Biochem 1985;22:461±467.

42. Oishi K, Raynor L, Charp PA, Kuo JF. Regulation of proteinkinase C by lysophospholipids. Potential role in signaltransduction. J Biol Chem 1988;263:6865±6871.

43. RoÈnkkoÈ S. Immunohistochemical localization of phospholipaseA2 in human and bovine reproductive organs. Comp BiochemPhysiol 1995;110:503±509.

44. Bjerve KS, Revitan LJ. The presence of an androgen controlledphospholipase A in rat epididymis. Int J Androl Suppl 1978;2:574±591.

45. Ellis LC, Boccio JR, Cunningham MJ, Croesbeck MD, CosentinoMJ. Rat testicular phospholipase A2 activity: pH optima, itscellular and subcellular distribution in the gonade and somefactors that may modulate its activity. J Androl 1981;2:94±102.

14 Weisseret al.