TheProstate 71:1033^1046 (2011)

5a -ReductaseType 3 Expression inHumanBenign andMalignant Tissues: AComparativeAnalysisDuring

ProstateCancer Progression

Alejandro Godoy,1 Elzbieta Kawinski,1 Yun Li,1 Daizo Oka,1 Borislav Alexiev,2

Faris Azzouni,1 Mark A. Titus,1 and James L. Mohler, MD1,3,4,5*

1DepartmentofUrology,Roswell ParkCancer Institute,Buffalo,NewYork142632Departmentof Pathology,Roswell ParkCancer Institute,Buffalo,NewYork14263

3DepartmentofUrology,Universityat Buffalo SchoolofMedicineandBiotechnology, Buffalo,NewYork142614Lineberger Comprehensive Cancer Center,UniversityofNorthCarolina SchoolofMedicine,ChapelHill,

North Carolina 275995Departmentof Surgery,DivisionofUrology,UniversityofNorthCarolina SchoolofMedicine,ChapelHill,

North Carolina 27599

BACKGROUND. A third isozyme of human 5a-steroid reductase, 5a-reductase-3, was ident-ified in prostate tissue at the mRNA level. However, the levels of 5a-reductase-3 proteinexpression and its cellular localization in human tissues remain unknown.METHODS. A specific monoclonal antibody was developed, validated, and used to charac-terize for the first time the expression of 5a-reductase-3 protein in 18 benign and 26 malignanthuman tissue types using immunostaining analyses.RESULTS AND CONCLUSIONS. In benign tissues, 5a-reductase-3 immunostaining washigh in conventional androgen-regulated human tissues, such as skeletal muscle and prostate.However, high levels of expression also were observed in non-conventional androgen-regulated tissues, which suggest either multiples target tissues for androgens or differentfunctions of 5a-reductase-3 among human tissues. In malignant tissues, 5a-reductase-3 immu-nostaining was ubiquitous but particularly over-expressed in some cancers compared to theirbenign counterparts, which suggests a potential role for 5a-reductase-3 as a biomarker ofmalignancy. In benign prostate, 5a-reductase-3 immunostaining was localized to basal epi-thelial cells, with no immunostaining observed in secretory/luminal epithelial cells. In high-grade prostatic intraepithelial neoplasia (HGPIN), 5a-reductase-3 immunostaining was local-ized in both basal epithelial cells and neoplastic epithelial cells characteristic of HGPIN. Inandrogen-stimulated and castration-recurrent prostate cancer (CaP), 5a-reductase-3 immunos-taining was present in most epithelial cells and at similar levels, and at levels higher thanobserved in benign prostate. Analyses of expression and functionality of 5a-reductase-3 inhuman tissuesmayproveuseful for development of treatment for benignprostatic enlargementand prevention and treatment of CaP. Prostate 71: 1033–1046, 2011. # 2010 Wiley-Liss, Inc.

KEY WORDS: 5a-reductases; androgen metabolism; dihydrotestosterone; benignprostatic hyperplasia; castration-recurrent prostate cancer

INTRODUCTION

Androgen target cells use testosterone (T) as a pro-hormone, where T is converted by intracrine pathwaysinto dihydrotestosterone (DHT), the most potentandrogen receptor (AR) ligand. The enzyme steroid5a-reductase (EC1.3.99.5) [1] localizes DHT biosyn-thesis to androgen responsive tissues. 5a-reductase

Conflict of interest statement: The authors have no conflicts of interest.

*Correspondence to: James L. Mohler, MD, Associate Director forTranslational Research and Chair, Department of Urology, RoswellPark Cancer Institute, Buffalo, NewYork 14263.E-mail: james.mohler@roswellpark.orgReceived 9 November 2010; Accepted 10 November 2010DOI 10.1002/pros.21318Published online 28 December 2010 in Wiley Online Library(wileyonlinelibrary.com).

� 2010Wiley-Liss,Inc.

isozymes 1 and 2 are well characterized [2] and DHTlevels in target tissues are reduced using 5a-reductaseinhibitors.However,male pattern baldness, that affectsan estimated 70% of men by age 80 [3] responds poorlyto 5a-reductase-2 inhibition (finasteride, PropeciaTM)[4]. Lower urinary tract symptoms from benign pros-tate enlargement affect 50% of men age 50 or older [5],and 20% ofmen treated with the 5a-reductase-2 inhibi-tor, finasteride (ProscarTM), or the bispecific inhibitor of5a-reductase-1 and 2, dutasteride (AvodartTM), requireoperative treatment within 4 years [6]. CaP is the mostcommon non-skin cancer in American men; approxi-mately 217 730 Americans will be diagnosed with CaPand approximately 32 050 men will die from CaP in2010 [7]. Finasteride treatment of men at increasedrisk of CaP decreased the diagnosis of CaP by 25%,however, the CaP that developed often was morepoorly differentiated than observed in controls [8].Dutasteride decreased the incidence of diagnosis ofCaP anddidnot affect differentiation [9], butmay causedelay of diagnosis that is concerning especially foraggressive CaP [10].

Menwho fail curative therapy for clinically localizedCaP or are diagnosed with advanced disease usuallyreceive androgen deprivation therapy (ADT) thatcauses regression of androgen-stimulatedCaP throughprogrammed cell death [11]. However, ADT is pallia-tive since CaP almost always recurs and causes death.A molecular role for AR in the transition from andro-gen-stimulated CaP to castration-recurrent CaP is sup-ported by continuous expression of AR [12–14] andandrogen-regulated genes [15]. Many alternativemechanisms allow AR-mediated transactivation ofgene expression despite castrate levels of circulatingtesticular androgens (reviewed by Refs. 16–23).However, we and others have demonstrated thatcastration-recurrent CaP maintains tissue levels ofDHT sufficient for activation of even wild-type AR[24–26]. Limited clinical experience has not demon-strated efficacy for finasteride added to ADT foradvanced CaP [27] or for dutasteride or LY320236,another bi-specific 5a-reductase inhibitor [28], forsecondary treatment of castration-recurrent CaP.Consequently, castration-recurrent CaP must becapable of biosynthesis of intraprostatic T from circu-lating adrenal androgens [29] or cholesterol [30] andmetabolism of T to DHT at levels sufficient for ARactivation by incompletely characterized mechanisms.

Laboratory observations and clinical experience ledto speculation about the existence of an additional 5a-reductase isozyme. A candidate sequence consistentwith a 5a-reductase isozyme that mapped to chromo-some 4q12 was identified by mining the HumanGenome Project database [31]. Uemura et al. [32] con-firmed the existence of a novel 5a-steroid reductase

(SRD5A3, type-3) in human tissues but only at themRNA level. Our study is the first to analyze theexpression of 5a-reductase-3 protein in human benignand malignant tissues. In this analysis, special empha-sis was given to the expression of 5a-reductase-3during prostate carcinogenesis and progression.

MATERIALSANDMETHODS

CellCulture

PWR-1E, a benign human prostate epithelial cellline, was purchased from American Type CultureCollection (ATCC, Manassas, VA) and maintained inKerationocyte Serum Free media (Invitrogen,Carlsbad, CA) supplemented with 5 ng/ml recombi-nant EGF and 0.05 mg/ml bovine pituitary extract(Gibco, Carlsbad, CA). LNCaP and C4-2 cells weremaintained in RPMI-1640 supplemented with 2 mMglutamine and 10% fetal bovine serum (FBS,Mediatech, Inc, Herndon, VA). LAPC-4 was main-tained in RPMI-1640 supplemented with 10 nMR1881 (Perkin-Elmer, Boston,MA).CWR-R1wasmain-tained in Richter’s Improved MEM (Mediatech)supplemented with 20 mM HEPES, 5 mg/ml insulin,5 mg/ml transferin, 5 ng/ml sodium selenite (ITS,Roche, Indianapolis, IN), 0.1 mg/ml epidermal growthfactor (Invitrogen), 1 mg/ml licoleic acid, 10 mM nic-otinamide (Sigma–Aldrich, St. Louis,MO), and2%FBS.CHO-K1, a subclone from the parental CHO cellline, was purchased from ATCC and maintained inNutrient Mixture F12K Kaighn’s modification media(Invitrogen) supplemented with 2 mM L-glutamineand 10% FBS.

Anti-5a -reductase-3PolyclonalandMonoclonalAntibodies

A commercially available polyclonal antibody pro-duced by Sigma (SRD5A3, Sigma–Aldrich) and a totalof four rabbit polyclonal antibodies produced commer-cially (OpenBiosystems,Huntsville,AL) against aminoacids 1–16, 3–18, 46–62, and 48–65 of the N-terminus ofthe 5a-reductase-3 protein were evaluated usingWestern blot and immunohistochemistry and thepolyclonal antibody targeting the first epitope,M1APWAEAEHSALNPLR16 performed best. Twohybridomas were produced in our laboratory againstthe same epitopeswithin theN-terminus and theRPCI-5aR3monoclonal antibodyperformedbest. RPCI-5aR3antibody resulted from fusion of murine myeloma(P3X63 Ag8U.1, ATCC CRL 1597) with spleen cellsof mouse hosts immunized with synthetic peptide(M1APWAEAEHSALNPLR16) using a standardprocedure [33]. Each peptide was coupled to keyholelimpet hemocyanin using the 1-Ethyl-3-[3-dimethyla-minopropyl] carbodiimide hydrochloride method

1034 Godoyet al.

The Prostate

(Pierce). Hybridomas were cloned twice using the lim-iting dilution method [34]. Stable hybridomas that pro-duced monoclonal antibodies were expanded andcryopreserved. Hybridoma cells were injected(1 � 107) into the peritoneal cavity of female SCIDmouse hosts to produce ascites fluid.All animal studieswere performed in compliance with US Department ofHealth andHumanServicesGuide for theCare andUseof Laboratory Animals and approved by Roswell ParkCancer Institute Animal Care and Use Committee.Ascites fluidwas harvested 3–4weeks after hybridomacells were injected when it contained 1.2–4 mg/ml ofmonoclonal antibody. Monoclonal antibodies werepurified from ascites on Affi-gel protein A agarose(Bio-Rad) following manufacturers instructions.RPCI-5aR3 antibody was of the IgG-1 subclass, asdetermined using an ImmunoPure MonoclonalAntibody IsotypingKit (Pierce). The immunoreactivityof the RPCI-5aR3 antibody with prostate cells wasconfirmed; crude extracts (50 mg) prepared fromLNCaP, C4-2, LAPC-4, and CWR-R1 cells were reactedwith synthetic peptide (20 ng) at 48C for 2 hr. Theproteins were separated using SDS-Page and electro-phoretically transferred onto a polyvinylidene fluoridemembrane. Membranes were developed using theRPCI-5aR3 monoclonal or polyclonal antibodies,horseradish peroxidase (HRP), and electrochemolumi-nescence. The specificity of each antibody also wastested using an enzyme-linked immunosorbent assay(ELISA) and Western blot. The detection antigen waspeptide conjugated with bovine serum albumin (BSA,EMD Chemicals, Gibbstown, NJ).

ReverseTranscriptionandQuantitativeReal-TimePCR

Total RNA from CWR-R1 cells was isolated usingRNeasy Mini kits (Qiagen). Total RNA (400 ng) wasreverse-transcribed into cDNA using random primers(Invitrogen). Primers and probes for 5a-reductase-3were purchased from Applied Biosystems. PCR reac-tions were performed using the 7300 Real-Time PCRsystem (Applied Biosystems) with a total reactionmixture volume of 25 mL containing 8.0 ng cDNA,1x TaqMan Universal PCR Master Mix (AppliedBiosystems) and 1x primers and probe mix. The PCRconditions for all reactions were: 958C for 10 min, 40amplification cycles at 958C for 15 sec and 608C for1 min. The relative mRNA levels of 5a-reductase-3 inLNCaP, C4-2, LAPC-4, and CWR-R1 were expressedrelative to PWR-1E.

WesternBlot

Total cell extracts (50 mg of total protein) from PWR-1E, LNCaP, C4-2, LAPC-4, CWR-R1, and CHO-K1,

were separated electrophoretically using SDS-poly-acrylamide gel electrophoresis (10% w/v; Bio-RadLaboratories,Hercules, CA) under reducing conditionsand the separated proteins transferred to nitrocellulosemembranes using standard procedures [35,36].Western analyses utilized as primary antibody theRPCI-5aR3 (1 mg/ml) or the SRD5A3 (1 and 5 mg/ml)antibody. HRP-conjugated anti-mouse or anti-rabbit IgG (Dako) was used as secondary antibody,respectively. Antibody localization was visualizedusing enhanced chemi-luminescence (PierceBiotechnology). CHO-K1 cell line was used as negativecontrol for expression of 5a-reductase-3 [37]. Equalloading of total cell extracts was verified visually byimmunohistochemical staining of membranes withanti-b-actin (Santa Cruz Biotechnology).

5a -Reductase-3 Immunostaining

CWR-R1 cells were fixed in situ for 30 min at roomtemperature using 4% w/v paraformaldehyde.Endogenous peroxidase activity was inhibited with3.0% v/v H2O2 in methanol. Antibody specificitywas evaluated by pre-incubating the monoclonal anti-body with the inhibitor peptide for 2 hr before immu-nostaining. Tissue microarrays from benign andmalignant human tissues were obtained from theRoswell Park Cancer Institute Department ofPathology. In addition, prostate tissue sections werecut from formalin-fixed, paraffin-embedded clinicalspecimens of androgen-stimulated benign prostate(BP) (n ¼ 8), high-grade prostatic intraepithelial neo-plasia (HGPIN), (n ¼ 8) and androgen-stimulated CaP(n ¼ 8) obtained from the Department of Pathologyarchives or from a tissue microarray constructed fromformalin-fixed, paraffin-embedded clinical specimensof androgen-stimulated BP (n ¼ 18), androgen-stimu-lated CaP (n ¼ 21) and castration-recurrent CaP(n ¼ 19), which has been studied previously[14,22,38–45]. All tissues and the images obtained fromthese tissues for analysis were reviewed by one path-ologist (Borislav Alexiev). In tissue sections, antigenswere retrieved usingmicrowave irradiation and citratebuffer pH 6.0 for 15 min. Endogenous peroxidaseactivity was inhibited using 3% v/v H2O2 in methanol.Sections and CWR-R1 cells were incubated overnightwith the RPCI-5aR3 antibody diluted 1:100 (optimumby checkerboard analysis of 1:50–1:1000 dilutions) in100 mM Tris-HCl buffer (pH 7.8) that contained8.4 mM sodium phosphate, 3.5 mM potassium phos-phate, 120 mMNaCl, and 1%w/v BSA. After washing3 times for 10 min in Tris-HCl buffer (pH 7.8), sectionswere incubated with HRP-conjugated anti-mouseIgG (1/100, Dako, Carpinteria, CA) secondary anti-body for 2 hr at room temperature. Peroxidaseactivity was developed using 100 mM Tris-HCl buffer

5a -Reductase-3 ExpressioninHumanTissues 1035

The Prostate

containing 3,3-diaminobenzidine tetrahydrochloride(1 mg/ml, Sigma–Aldrich) and H2O2 (1 ml/ml, VWRInternational, West Chester, PA). Colocalizationanalyses of 5a-reductase-3/p63 and 5a-reductase-3/alpha-methylacyl-coenzyme-A racemase (AMACR)were performed using the EnVisionTMG/2 DoubleStain System according to the manufacturer’s instruc-tions (Dako) [46]. 5a-Reductase-3 was detected usinganHRP-conjugated secondary antibody and 3,3-diami-nobenzidine tetrahydrochloride as substrate (brownprecipitate) and p63 and AMACR protein expres-sion were visualized using an alkaline phosphatase-conjugated secondary antibody and Permanent Redsubstrate (red precipitate). Sections were counter-stained using hematoxylin. Immunohistochemistry inabsence of primary antibody provided negativecontrols.

Digital ImageCollectionandAnalysis

For cell lines and tissue sections (including tissuemicroarrays), images were collected using anHamamatsu Color Chilled 3CCD camera(Hamamatsu, Bridgewater, NJ) mounted on anAxioskop microscope (Carl Zeiss, Thornwood, NY).For the prostate tissue microarray, five images werecollected from each 2.0 mm core at total magnification400�. The images were arranged in random order in adigital albumusingAdobePhotoshop (AdobeSystems,San Jose, CA) to facilitate visual scoring. Images werevisually scored by three experienced observers (M. A.Titus, O. H. Ford and J. L.Mohler) whowere blinded totissue type. The observers scored the nuclear and cyto-plasmic immunostaining on a scale ranging from 0 (noimmunostaining) to 3 (strong immunostaining) in eachof 100 nuclei or cells, respectively, to yield a visual scorefor each observor ranging from 0 to 300 for each featurein each specimen as described by Miyamoto et al. [47].The intensity of cytoplasmic immunostaining also wasassessed by quantitative image analysis using ImagePro Plus software (MediaCybernetics, Bethesda, MD).A circular region of fixed area was used to extract theintensity information for cytoplasmic immunostainingof all epithelial cells in each image scored visually.Mean optical density (MOD) was derived from meanintensity (Ii) by setting the background intensity towhite (Io ¼ 255) and using the equation [48]:

MOD ¼ �logIiIo

� �

Mean visual scores and MOD were analyzed usingMinitab (StateCollege, PA).ANOVA(Tukey’s honestlysignificant differences test) was used to compare visualscores of nuclear and cytoplasmic immunostaining andautomated cytoplasmic immunostaining MOD [44].

Differences were considered statistically significant ifP < 0.05.

RESULTS

ValidationoftheRPCI-5aR3MonoclonalAntibody

Expression of 5a-reductase-3 at themRNA levelwasanalyzed in the BP epithelial cell line PWR-1E and theandrogen-sensitive CaP cell lines, LNCaP, and LAPC-4, and the castration-recurrent CaP cell lines, C4-2 andCWR-R1 (Fig. 1A). 5a-reductase-3 mRNA expressionin androgen-stimulated CaP and castration-recurrentCaP cell lines were expressed relative to PWR-1E cellline, which expressed very low levels of 5a-reductase-3mRNAas determined using standard PCR.High levelsof 5a-reductase-3 mRNA expression compared toPWR-1E were observed in the LNCaP, LAPC-4, andCWR-R1 cell lines. However, the castration-recurrentC4-2 cell line showed 5a-reductase-3 mRNA levelscomparable to the PWR-1E cell line. These results wereconfirmed at the protein level using western blotanalysis and theRPCI-5aR3 antibody (Fig. 1B). A singleband for 5a-reductase-3 of in the range of 25–37 kDawas observed in total protein extracts of LAPC-4 andCWR-R1 cells (Fig. 1B), which corresponds to theexpected size for 5a-reductase-3. However, noimmuno-reactive band was observed in LNCaP cells,even though LNCaP expressed similar 5a-reductase-3mRNA levels as LAPC-4 and CWR-R1. No immuno-reactive band in the range of 25–37 kDa was observedin PWR-1E, C4-2, and CHO-K1 cell lines (Fig. 1B).CWR-R1 cells were chosen to compare sensitivity ofthe RPCI-5aR3 and SDR5A3 (Sigma) antibodies. Noimmuno-reactive band in the range of 25–37 kDawas detected using the SDR5A3 antibody, even whenconcentrations were 5-fold higher than suggested bythe manufacturer (Fig. 1C). A weak band was appreci-ated using SDR5A3 antibody with longer times ofexposure of the films (over 5 min). Both antibodies(RPCI-5aR3 and SDR5A3) recognized a single bandover 50 kDa, which does not correspond to theexpected size for 5a-reductase-3 (data not shown).Sub-cellular localization of 5a-reductase-3 proteinand specificity of RPCI-5aR3 antibody were con-firmed using immunostaining analyses. RPCI-5aR3antibody showed a cytoplasmic immunostaining pat-tern (Fig. 1D); no nuclear immunostaining wasobserved. Incubation of RPCI-5aR3 antibody withthe inhibitor peptide at increasing concentration(1 and 10 mg/ml) for 2 hr before immunostaining pro-duced a peptide concentration-dependent inhibitionof 5a-reductase-3 immunostaining in CWR-R1 cells(Fig. 1D). Complete inhibition of 5a-reductase-3 immu-nostaining was achieved at peptide concentration of10 mg/ml.

1036 Godoyet al.

The Prostate

5a -Reductase-3ExpressioninBenignHumanTissues

Semi-quantitative analysis of 5a-reductase-3 expres-sionwas performed in apanel of human benign tissues,which included androgen-responsive (liver [n ¼ 20],skeletal muscle [n ¼ 2], skin [n ¼ 1] testes [n ¼ 20],thyroid [n ¼ 20]), and androgen-insensitive tissues(brain [n ¼ 20], breast [n ¼ 20], colon [n ¼ 20], endo-metrium [n ¼ 20], kidney [n ¼ 20], lung [n ¼ 20],myometrium [n ¼ 20], ovary [n ¼ 20], pancreas[n ¼ 15], spleen [n ¼ 20], stomach [n ¼ 20], and tonsil[n ¼ 15]) (Figs. 2 and 3, Table I). High levels of 5a-reductase-3 protein expressionwere observed in a sub-set of human samples, which included skin (stratumbasale and stratum spinosum), kidney (mostly proximaland some distal convoluted tubules), liver, skeletalmuscle, myometrium, and pancreas (secretory epi-thelial cells). Moderate levels of expression of 5a-

reductase-3 protein were observed in testis (Leydigcells), brain (neurons), breast (myoepithelial cells),colon (epithelial cells from colonic glands), andstomach (epithelial cells from the base of the gastricglands). Low levels of 5a-reductase-3 immunostainingwere observed in lung (bronchial epithelium), andthyroid (cuboidal epithelium from thyroid follicles).5a-reductase-3 was immunohistochemically undetect-able in ovary, spleen, endometrium, and tonsil. Inmost cases, 5a-reductase-3 immunostaining was dis-tributed homogeneously within the cytoplasm.However, perinuclear localization was observed when5a-reductase-3 immunostaining was intense. 5a-reductase-3 protein expression was observed both inconventional androgen-responsive benign tissues,such as skin and skeletalmuscle, and non-conventionalandrogen-responsive tissues, such as kidney, colon andpancreas.

Fig. 1. Validationof theRPCI-5aR3 antibody.A: 5a -reductase-3mRNAexpression level in LNCaP,C4-2,LAPC-4, andCWR-R1cells. 5a -reductase-3mRNAexpression levels in theCaPcell lineswerenormalized to thelevelof expressionof5a -reductase-3 in thePWR-1Ebenignprostate epithelial cell line (discontinuous red line).B: PWR-1E, LNCaP,C4-2, LAPC-4,CWR-R1, and CHO-K1protein lysates (50 mg) wereimmunoblottedand5a -reductase-3proteinexpressionwasdetectedusing theRPCI-5aR3antibody.C:CWR-R1lysates (50 mg)wereimmu-noblotted and 5a -reductase-3 protein expression was analyzed using RPCI-5aR3 antibody at a concentration of 1 mg/ml or SRD5A3(Sigma) antibody atconcentrations1 mg/ml and 5 mg/ml.D: Immunostaining analyseswereperformedinCWR-R1using theRPCI-5aR3 anti-body.Preincubationof theRPCI-5aR3antibodywith theinhibitorpeptide (1and10 mg/ml) confirmedspecificity.Absenceofprimaryantibodyprovidednegative control (Control).

5a -Reductase-3 ExpressioninHumanTissues 1037

The Prostate

5a -Reductase-3ExpressioninMalignantHumanTissues

As was observed in benign tissues, 5a-reductase-3immunostaining in malignant tissues showed inter-organ and inter-patient variability (Figs. 4 and 5,Table II). High levels of 5a-reductase-3 immunostain-ing were observed in malignant tissues from kidney(chromophobe carcinoma), liver (hepatocellular carci-noma), stomach (adenocarcinoma), thyroid (papillarycarcinoma), colon (adenocarcinoma), and uterus

(endometrioid adenocarcinoma). Moderate to lowlevels of 5a-reductase-3 immunostaining wereobserved in adrenal (pheochromocytoma), bladder(high grade urothelial carcinoma), breast (lobularand metaplastic carcinomas), esophagus (adenocarci-noma), kidney (clear cell renal cell carcinoma andpapillary carcinoma), lung (adenocarcinoma), ovary(mucinous adenocarcinoma), testis (seminoma, embry-onal and yolk sac carcinomas), and thyroid (medullarycarcinoma). 5a-reductase-3 was immunocytochemi-cally undetectable in bladder (low grade urothelial

Fig. 2. 5a -reductase-3 immunostaining in benign human tissues.High level expressing tissues (representative sections that demonstratesemi-quantitative data inTable I). In kidneyA,B: 5a -reductase-3 expressionwas localized to the cytoplasm of the epithelial cells from someof the proximal and distal convoluted tubules (B, arrow). In myometriumC,D: 5a -reductase-3 immunostaining was located preferentiallyto the cytoplasmof the smoothmuscle cells (D: arrow, fibers in transverse). InpancreasE,F: 5a -reductase-3 immunostainingwas observedin some of the secretory cells from the acini (F, arrow).Bars: 30 mm.

1038 Godoyet al.

The Prostate

carcinoma and small cell carcinoma), lung (meso-thelioma), leiomyosarcoma, and ovary (serous adeno-carcinoma). 5a-reductase-3 immunostaining wasconfined to the cytoplasm of malignant epithelial cells,which exhibited perinuclear localization when intense.Among the malignant human tissues analyzed, onlybreast (lobular and metaplastic carcinomas), testis(seminoma, embryonal, and yolk sac carcinomas), lung(adenocarcinoma), and thyroid (papillary carcinoma)clearly showed over-expression of 5a-reductase-3protein compared to their benign counterpart (compareTables I and II). However, in the majority of cases,

similar levels of 5a-reductase-3 immunostaining wereobserved between the benign and malignant tissuepairs (compare Tables I and II).

5a -ReductaseExpressionDuringCarcinogenesis andProgressionof ProstateCancer

5a-Reductase-3 immunostaining was analyzed inclinical specimens of different types of prostate tissues,since the prostate is a highly androgen-responsiveorgan (Fig. 6). In BP, 5a-reductase-3 immunostainingwas observed primarily at the periphery of the benign

Fig. 3. 5a -reductase-3immunostaininginbenignhumantissues.Moderatetolowlevelexpressing tissues(representativesectionsthatdem-onstrate semi-quantitative data inTable I). In testesA,B: 5a -reductase-3 expressionwas localized to the cytoplasmofLeydigcells (B, arrow).InbrainC: 5a -reductase-3immunostainingwaslocalizedto thecytoplasmofneurons(D, arrow).Lowlevelsof5a -reductase-3 immunostain-ingwere observedin theglandularepitheliumofendometrium(E, arrow). 5a -reductase-3wasimmunohistochemicallyundetectablein tonsil(F).Bars: 30 mm.

5a -Reductase-3ExpressioninHumanTissues 1039

The Prostate

glands, which suggests 5a-reductase-3 expression inthe basal cell compartment. This observation was con-firmed by co-localization of 5a-reductase-3 and p63,a basal cell marker. Low to undetectable levels of5a-reductase-3 immunostaining were observed inluminal epithelial cells. In HGPIN, 5a-reductase-3

immunostaining localized to both basal cells and neo-plastic luminal epithelial cells. However, 5a-reductase-3 was low in luminal epithelial cells of adjacent benignglands. In androgen-stimulated CaP and castration-recurrent CaP, 5a-reductase-3 immunostaining wasfound in the cytoplasm of most malignant epithelial

Fig. 4. 5a -reductase-3 immunostaininginmalignanthumantissues.Highlevelexpressing tissues(representativesections thatdemonstratesemi-quantitativedatainTableII).5a -reductase-3expressionwaslocalizedmainly to thecytoplasmof themalignantepithelialcells fromadeno-carcinomaof the stomach (A), hepatocellularcarcinomaof the liver (B), papillarycarcinomaof the thyroid (C), endometroidadenocarcinomaof theuterus (D), and adenocarcinoma of the colon (E).Bar: 30 mm.

TABLE I. 5a -Reductase-3 Immunostaining in BenignHumanTissues

TissueNumber of positive cases/

total casesStainingintensity Preferential localization

Brain 6/20 �/þ NeuronsBreast 4/20 �/þ Myoepithelial cellsColon 7/20 �/þþ Epithelial cells from colonic glandsEndometrium 10/20 �/þ Glandular epitheliumKidney 20/20 þþþ Epithelial cells (PCT, DCT)Liver 20/20 þþþ HepatocytesLung 2/20 �/þ Bronchial epitheliumMyometrium 11/20 þþ Smooth muscle cellsOvary 0/20 �Pancreas 15/15 þþ/þþþ Secretory cells (GA)Prostate 23/25 þþ Basal cellsSkeletal muscle 2/2 þ/þþ Skeletal muscle cellsSkin 1/1 þ/þþ Stratum basale, stratum spinosumSpleen 0/20 �Stomach 12/20 þ/þþ Epithelial cells from the base of the gastric glandsTestes 8/20 �/þ Leydig cellsThyroid 13/20 �/þ Cuboidal epithelium from thyroid folliclesTonsil 0/15 �

Criteria for immunohistochemical analysis:�, negative staining; þ, weak staining;þþ, moderate staining; þþþ, intense staining; PCT,proximal convoluted tubule; DCT, distal convoluted tubule; GA, glandular acini.

1040 Godoyet al.

The Prostate

cells. 5a-reductase-3 immunostaining wasmostly peri-nuclear when intense. Cytoplasmic 5a-reductase-3immunostaining intensity was similar in androgen-stimulated CaP and castration-recurrent CaP, andwas higher than androgen-stimulated BP, when com-pared using visual scores (ANOVA, P < 0.00001) orimage analysis (ANOVA, P < 0.00001) (Table III).Nuclear immunostaining was minimal and variedmarginally among androgen-stimulated BP, andro-gen-stimulated CaP, and castration-recurrent CaP(ANOVA, P ¼ 0.205) (Table III). 5a-reductase-3 immu-nostaining in malignant epithelial cells was confirmedby co-localization of 5a-reductase-3 and AMACRimmunostaining.

DISCUSSION

Several pieces of evidence suggested the presence ofa third isozyme of 5a-reductase: a candidate genesequence [31], low response rate of men with

castration-recurrent CaP to dutasteride in a clinical trial[49], and failure of bi-specific 5a-reductase inhibitors toeliminate tissue levels of DHT (dutasteride reducedtissue levels of DHT by 94% in androgen-stimulatedBP [50] and LY320236 failed to decrease serum DHTlevels in intact men [51]). Uemura et al. [32] confirmedthe existence of a novel 5 alpha-steroid reductase(SRD5A3, type-3) in human tissues but only at themRNA level. They reported that 5a-reductase-3 wasoverexpressed in castration-recurrent CaP cells.Gaining insight into the localization and function of5a-reductase-3 has proven difficult because commer-cially available antibodies against 5a-reductase-3became available only recently and characterizing5a-reductase-3 enzymatic activities is challenging,because the enzymes appear embedded within theendoplasmic reticulum, are hydrophobic, and forminclusion bodies when expressed in vitro. Our labora-tory produced a highly specific monoclonal antibody(RPCI-5aR3) against the amino-terminal portion of the

Fig. 5. 5a -reductase-3 immunostaininginmalignanthuman tissues.Moderate to lowlevel expressing tissues (representative sections thatdemonstrate semi-quantitative data inTable II).Moderate to lowlevels of 5a -reductase-3 expressionwere observedin the cytoplasmofneo-plastic epithelial cells from adenocarcinoma of the esophagus (A), clear cell renal cell carcinoma of the kidney (B), papillary carcinoma of thekidney (C), embryonal testicular carcinoma (D), adenocarcinoma of the lung (E), andmedullary carcinoma of the thyroid (F).Negative 5a -reductase-3 immunostainingwasobservedin smallcellcarcinomaof thebladder (G),mesothelioma(H), andleiomyosarcoma (I).Bar: 30 mm.

5a -Reductase-3ExpressioninHumanTissues 1041

The Prostate

5a-reductase-3 protein and analyzed expression of thisprotein in benign and malignant human tissues, withspecial emphasis on prostate carcinogenesis and CaPprogression. During the course of these studies, acommercially available polyclonal antibody against5a-reductase-3 was released from Sigma. The Sigmaantibody at concentrations that ranged from 1 mg/ml(concentration suggested by the manufacturer) to5 mg/ml revealed no immuno-reactivity usingwesternblot. The RPCI-5aR3 antibody, however, detected 5a-reductase-3 protein using immunocytochemistry,immunohistochemistry, and western blot. Furtheranalyses are required to validate the RPCI-5aR3 anti-body for ELISA and immunoprecipitation.

5a-Reductase-3 expression analyses in humantissues performed by Uemura et al. [32] were basedon northern blot analysis. Their tissue expressionanalyses revealed very low levels of expression of5a-reductase-3 mRNA in most benign adult organs,with benign pancreas showing the highest levels ofexpression of 5a-reductase-3 mRNA. These results,however, contrast with the data reported by Yamanaet al. [52] using QRT-PCR, which indicated a broaderand higher expression of 5a-reductase-3 mRNA com-pared to 5a-reductase-1 and 2 mRNAs in human tis-sues. Our immunostaining analyses support Yamana’sresults and showed variable levels of expression of 5a-reductase-3 protein in several benign human tissues.The differences in expression of 5a-reductase-3 at themRNA and protein levels between Uemura’s studyand our immunostaining analysis could be explained,at least in part, by differences in the sensitivity of thetechniques utilized (northern blot versus immunos-taining). Alternatively, presence of mRNA does notalways correlate with expression of protein, as wasdemonstrated clearly in this study for 5a-reductase-3expression in LNCaP cells, which highlights theimportance of our immunostaining analysis.

5a-Reductase-3 protein was expressed in both con-ventional androgen-regulated human tissues, such asskeletal muscle, skin, and prostate, and in non-conven-tional androgen-regulated tissues, such as pancreas,colon, and kidney. These data suggest either multipletarget tissues for androgensordifferent functions of 5a-reductase-3 amongbenignhuman tissues. In support ofthe second hypothesis, Cantagrel et al. [53], reportedthat mutations in 5a-reductase-3 gene caused congen-ital disorders of glycosylation, which induce mentalretardation and ophthalmologic and cerebellar defects.In this study, 5a-reductase-3 proteinwas found necess-ary for the reduction of the a-isoprene unit of polypre-nols to form dolichols, which are required for theN-glycosylation process. Unpublished data from ourlaboratory using protein assays of recombinant 5a-reductase-3 expressed in CHO cells and bacteriasuggest conversion of T, androstenedione, andprogesterone to DHT, androstanedione, and 5a-preg-nan-3-20-dione, respectively. Broad substrate speci-ficity of this enzyme is consistent with ubiquitousexpression of 5a-reductase-3 protein in human tissues.

In tumor tissues, 5a-reductase-3 protein expressionwas observed in several different types of malignan-cies. Among those tissues that expressed moderateto high levels of 5a-reductase-3, only cancers of breast,testis, lung, thyroid, and prostate showed over-expression of 5a-reductase-3 compared to their benigncounterpart. Neither 5a-reductase-1 nor 5a-reductase-2 has been advocated as a potential biomarker ofmalignancy. However, Uemura et al. [32] suggested

TABLE II. 5a -Reductase-3 Immunostaining in MalignantHumanTissues

Tissue

Number ofpositive cases/

total casesStainingintensity

Adrenal Pheochromocytoma 1/2 �/þBladderLow grade TCC 0/2 �High grade TCC 2/2 þSmall Cell CA 2/2 �

BreastLobular CA 2/2 þMetaplastic CA 2/2 þ

Colon Adeno CA 2/2 þþEsophagus Adeno CA 2/2 þ/þþKidneyChromophobe CA 4/4 þþþPapillary CA 1/2 þþRenal Clear Cell CA 2/2 þþ

Liver HCC 2/2 þþ/þþþLungAdeno CA 2/2 þþMesothelioma 2/2 �

Leiomyosarcoma 2/2 �OvaryMucinous CA 4/4 þ/þþSerous CA 0/2 �

Stomach Adeno CA 2/2 þþ/þþþTestisSeminoma 2/2 þþEmbryonal CA 2/2 þYolk Sac CA 1/2 þþ

ThyroidPapillary CA 4/4 þþþMedullary CA 2/2 þ

Uterus Endometroid CA 2/2 þþ

Criteria for immunohistochemical analysis:�, negative staining;þ, weak staining;þþ, moderate staining;þþþ, intense staining;TCC, transitional cell carcinoma; CA, carcinoma; HCC, hepato-cellular carcinoma.

1042 Godoyet al.

The Prostate

over-expression of 5a-reductase-3 at themRNA level incastration-recurrent CaP compared to BP. Even thoughthe number of clinical specimens analyzed in this studyis limited, these results warrant further investigation ofa putative role for 5a-reductase-3 as a biomarker ofmalignancy in the prostate, breast, testis, lung, andthyroid tissues.

Analysis of archival specimens suggested that 5a-reductase-3 may provide a biomarker of malignancy inthe prostate. 5a-reductase-3 was expressed in the basalcells of androgen-stimulatedBPbutwas not detected inthe benign luminal epithelium. In contrast, in CaP thatlacks basal cells, 5a-reductase-3 was found in the

malignant epithelium. HGPIN is hypothesized to bethe precursor lesion for CaP [54]. In all six specimens ofHGPIN examined, 5a-reductase-3 was present in thebasal cell layer of both benign glands and HGPIN, andalthough 5a-reductase-3 was not detected in luminalepithelial cells in benign glands adjacent to HGPINlesions, 5a-reductase-3 was expressed in the neoplasticcells of HGPIN lesions. These findings suggest that acharacteristic of the development of CaP may be thetransition of 5a-reductase-3 expression from the basalcell compartment to the cancer epithelial cell compart-ment; maintenance of 5a-reductase-3 expression maybe a characteristic of progression.

Fig. 6. 5a -reductase-3 immunostaining in androgen-stimulated benign prostate (AS-BP), androgen-stimulated high grade intraepithelialneoplasia (AS-HGPIN), androgen-stimulatedCaP (AS-CaP), and castration-recurrentCaP (CR-CaP) tissue sections (representative sectionsthatdemonstrate quantitative data inTable III). 5a -reductase-3 immunostaining in AS-BPwas observedprimarily at theperipheryof benignglands, which suggests 5a -reductase-3 expression in the basal cell compartment (B,C [arrows]) that was confirmed by co-localization of5a -reductase-3(browncytoplasmic stain)andp63(rednuclear stain)immunostaining(D).No tolowlevelsof5a -reductase-3immunostainingwere observed in benign luminal epithelial cells (B [arrowhead]). In HGPIN, 5a -reductase-3 immunostaining was located in both basal cellsandhyperproliferativemalignant luminal epithelial cells (F,G [arrow]) butno to low levels of 5a -reductase-3 immunostainingwere observedin luminal epithelial cells of adjacent benign glands (F,G [arrowhead]). In AS-CaP and CR-CaP, 5a -reductase-3 immunostaining was locatedin most malignant epithelial cells (I,J,N). 5a -reductase-3 immunostaining was mostly perinuclear when intense (J,N [arrows]. 5a -reductase-3 immunostaining inmalignant epithelial cells was confirmedby co-localization of 5a -reductase-3 (brown cytoplasmic stain) andAMACR (red cytoplasmic stain) immunostaining (K). Incubation in the absence of primary antibody provided negative controls (A,E,H,L).Blackbars: 30 mM.Images reduced from200�magnification (columns1and 2) or 400�magnification (columns 3 and 4).

5a -Reductase-3 ExpressioninHumanTissues 1043

The Prostate

We [44] and others [55] reported previously that 5a-reducing capacity shifted from 5a-reductase-2 inandrogen-stimulated BP to 5a-reductase-1 in andro-gen-stimulated CaP and castration-recurrent CaP.However, some if not most, of the 5a-reducingcapability of castration-recurrent CaP may be due to5a-reductase-3. However, the presence of high levelsof 5a-reductase-3 protein, and new knowledgegained of the potential importance of 5a-reductionin preclinical models of castration-recurrent CaP[29,30,56] suggest an important role for 5a-reductase-3 in the lethal phenotype of CaP. The potential clinicalimportance of 5a-reductase-3 in CaP requires furtherinvestigation but tri-specific 5a-reductase-3 inhibitorsmay prove useful for treatment of benign prostaticenlargement and prevention and treatment ofCaP, especially advancedCaP. These common diseasesall depend on tissue levels of DHT for AR trans-activation that may persist in spite of currently avail-able treatments due to unblocked 5a-reductase-3activity.

CONCLUSIONS

The results of this study allow one to conclude that:(i) expression of 5a-reductase-3 in ‘‘classical’’ aswell as‘‘non-classical’’ androgen-regulated tissues is consist-entwith 5a-reductase-3 enzymehaving functions otherthan converting T to DHT in human tissues, suchas participation in the N-glycosylation process; (ii)over-expression of 5a-reductase-3 in breast, testis,lung, thyroid, and particularly prostate cancer, com-pared to their benign counterparts, suggests apotential role for 5a-reductase-3 as a biomarkerof malignancy; and (iii) over-expression of 5a-reductase-3 inAS-CaPandCR-CaP suggests apotentialrole for this enzyme in synthesizing DHT in both anandrogen-stimulated and an androgen-deprivedhuman prostate microenvironment.

ACKNOWLEDGMENTS

The CaP cell lines LNCaP, LAPC-4, LNCaP-C4-2,and CWR-R1, were provided by Drs. Allen Gao,Charles Sawyers, and Christopher Gregory. Theauthors thank Carol Wrzosek for technical assistance.This work was supported by the National CancerInstitute at the National Institutes of Health PO1-CA77739 (J. L. Mohler), National Cancer InstituteCancer Center Support Grants CA016156 andCA034026 to Roswell Park Cancer Institute andUniversity of North Carolina, respectively, andDepartment of Defense Grant W81XWH-08-1-0330(A. Godoy).

REFERENCES

1. Labrie F. Intracrinology. Mol Cell Endocrinol 1991;78(3):C113–118.

2. Russell DW, Wilson JD. Steroid 5 alpha-reductase: two genes/two enzymes. Annu Rev Biochem 1994;63:25–61.

3. Hawk E, Breslow RA, Graubard BI. Male pattern baldness andclinical prostate cancer in the epidemiologic follow-upof the firstNational Health and Nutrition Examination Survey. CancerEpidemiol Biomarkers Prev 2000;9(5):523–527.

4. Olsen EA, Hordinsky M, Whiting D, Stough D, Hobbs S, EllisML, Wilson T, Rittmaster RS. The importance of dual 5alpha-reductase inhibition in the treatment of male pattern hair loss:Results of a randomized placebo-controlled study of dutasterideversus finasteride. J Am Acad Dermatol 2006;55(6):1014–1023.

5. Napalkov P, Maisonneuve P, Boyle P. Worldwide patterns ofprevalence and mortality from benign prostatic hyperplasia.Urology 1995;46(3 Suppl A):41–46.

6. Roehrborn CG, Marks LS, Fenter T, Freedman S, Tuttle J,Gittleman M, Morrill B, Wolford ET. Efficacy and safety ofdutasteride in the four-year treatment of men with benign pro-static hyperplasia. Urology 2004;63(4):709–715.

7. Jemal A, Siegel R, Xu J, Ward E. Cancer Statistics, 2010.CA Cancer J Clin 2010;60(5):277–300.

8. Thompson IM, Goodman PJ, Tangen CM, Lucia MS, Miller GJ,Ford LG, Lieber MM, Cespedes RD, Atkins JN, Lippman SM,Carlin SM, Ryan A, Szczepanek CM, Crowley JJ, Coltman CA,The influence of finasteride on the development of prostatecancer. N Engl J Med 2003;349(3):215–224.

9. Andriole GL, Bostwick DG, Brawley OW, Gomella LG,Marberger M, Montorsi F, Pettaway CA, Tammela TL,Teloken C, Tindall DJ, Somerville MC, Wilson TH, Fowler IL,Rittmaster RS. Effect of dutasteride on the risk of prostate cancer.N Engl J Med 2010;362(13):1192–1202.

10. Walsh PC. Chemoprevention of prostate cancer. N Engl J Med2010;362(13):1237–1238.

11. Denmeade SR, LinXS, Isaacs JT.Role of programmed (apoptotic)cell death during the progression and therapy for prostate can-cer. Prostate 1996;28(4):251–265.

12. Culig Z,HobischA,HittmairA, PeterzielH, CatoAC, BartschG,Klocker H. Expression, structure, and function of androgenreceptor in advanced prostatic carcinoma. Prostate 1998;35(1):63–70.

13. de Vere White R, Meyers F, Chi SG, Chamberlain S, Siders D,Lee F, Stewart S, Gumerlock PH. Human androgen receptor

TABLE III. 5a -Reductase-3 ImmunostainingofAS-BP, AS-CaP, andCR-CaP

Mean � SE

AS-BP AS-CaP CR-CaP

Visual scoreNuclear 14.7 � 4.4 25.6 � 3.5 29.2 � 8.4Cytoplasmic 115.6 � 10.5 200.5 � 9.1a 163 � 11.7a

MODCytoplasmic 0.32 � 0.02 0.50 � 0.03a 0.49 � 0.03a

aStatistically different than androgen-stimulated BP. AS-BP,androgen-stimulated BP; AS-CaP, androgen-stimulated CaP;CR-CaP, castration-recurrent CaP; SE, standard error.

1044 Godoyet al.

The Prostate

expression in prostate cancer following androgen ablation. EurUrol 1997;31(1):1–6.

14. Mohler JL, GregoryCW, FordOH, , KimD,Weaver CM, PetruszP,WilsonEM,FrenchFS.The androgenaxis in recurrentprostatecancer. Clin Cancer Res 2004;10(2):440–448.

15. Mousses S,Wagner U, Chen Y, Kim JW, Bubendorf L, BittnerM,Pretlow T, Elkahloun AG, Trepel JB, Kallioniemi OP. Failure ofhormone therapy in prostate cancer involves systematic restor-ation of androgen responsive genes and activation of rapamycinsensitive signaling. Oncogene 2001;20(46):6718–6723.

16. Arnold JT, Isaacs JT. Mechanisms involved in the progression ofandrogen-independent prostate cancers: It is not only the cancercell’s fault. Endocr Relat Cancer 2002;9(1):61–73.

17. Feldman BJ, Feldman D. The development of androgen-independent prostate cancer. Nat Rev Cancer 2001;1(1):34–45.

18. Gelmann EP. Molecular biology of the androgen receptor. J ClinOncol 2002;20(13):3001–3015.

19. Agoulnik IU, Vaid A, Nakka M, Alvarado M, Bingman WE,Erdem H, Frolov A, Smith CL, Ayala GE, Ittmann MM,Weigel NL. Androgens modulate expression of transcriptionintermediary factor 2, an androgen receptor coactivator whoseexpression level correlates with early biochemical recurrence inprostate cancer. Cancer Res 2006;66(21):10594–10602.

20. Gregory CW,He B, Johnson RT, Ford OH,Mohler JL, French FS,WilsonEM.Amechanism for androgen receptor-mediatedpros-tate cancer recurrence after androgen deprivation therapy.Cancer Res 2001;61(11):4315–4319.

21. Guo Z, Dai B, Jiang T, Xu K, Xie Y, Kim O, Nesheiwat I, Kong X,Melamed J,HandrattaVD,NjarVC, BrodieAM,YuLR,VeenstraTD, Chen H, Qiu Y. Regulation of androgen receptor activity bytyrosine phosphorylation. Cancer Cell 2006;10(4):309–319.

22. MahajanNP, LiuY,Majumder S,WarrenMR, ParkerCE,MohlerJL, EarpHS,WhangYE.ActivatedCdc42-associatedkinaseAck1promotes prostate cancer progression via androgen receptortyrosine phosphorylation. Proc Natl Acad Sci USA 2007;104(20):8438–8443.

23. Ponguta LA, Gregory CW, French FS, Wilson EM. Site-specificandrogen receptor serine phosphorylation linked to epidermalgrowth factor-dependent growth of castration-recurrent pros-tate cancer. J Biol Chem 2008;283(30):20989–21001.

24. Belanger B, Belanger A, Labrie F, Dupont A, Cusan L, MonfetteG. Comparison of residual C-19 steroids in plasma and prostatictissue of human, rat and guinea pig after castration: uniqueimportance of extratesticular androgens in men. J SteroidBiochem 1989;32(5):695–698.

25. Geller J, Albert J, Loza D. Steroid levels in cancer of theprostate–markers of tumour differentiation and adequacy ofanti-androgen therapy. J Steroid Biochem 1979;11(1B):631–636.

26. Titus MA, Schell MJ, Lih FB, Tomer KB,Mohler JL. Testosteroneand dihydrotestosterone tissue levels in recurrent prostate can-cer. Clin Cancer Res 2005;11(13):4653–4657.

27. Fleshner NE, Trachtenberg J. Combination finasteride and flu-tamide in advanced carcinoma of the prostate: Effective therapywith minimal side effects. J Urol 1995;154(5):1642–1645 ; discus-sion 1645-1646.

28. Gregory CW, Fei X, Ponguta LA, He B, Bill HM, French FS,Wilson EM. Epidermal growth factor increases coactivation ofthe androgen receptor in recurrent prostate cancer. J Biol Chem2004;279(8):7119–7130.

29. StanbroughM,BubleyGJ,RossK,GolubTR,RubinMA,PenningTM,FebboPG,BalkSP. Increasedexpressionof genes converting

adrenal androgens to testosterone in androgen-independentprostate cancer. Cancer Res 2006;66(5):2815–2825.

30. Locke JA, Guns ES, Lubik AA, Adomat HH, Hendy SC, WoodCA, Ettinger SL, Gleave ME, Nelson CC. Androgen levelsincrease by intratumoral de novo steroidogenesis during pro-gression of castration-resistant prostate cancer. Cancer Res2008;68(15):6407–6415.

31. Ota T, Suzuki Y, Nishikawa T, Otsuki T, Sugiyama T, Irie R,WakamatsuA,HayashiK, SatoH,NagaiK,KimuraK,MakitaH,Sekine M, Obayashi M, Nishi T, Shibahara T, Tanaka T, Ishii S,Yamamoto J, Saito K, Kawai Y, Isono Y, Nakamura Y, NagahariK, Murakami K, Yasuda T, Iwayanagi T, Wagatsuma M,Shiratori A, Sudo H, Hosoiri T, Kaku Y, Kodaira H, Kondo H,Sugawara M, Takahashi M, Kanda K, Yokoi T, Furuya T,Kikkawa E, Omura Y, Abe K, Kamihara K, Katsuta N, Sato K,TanikawaM, Yamazaki M, Ninomiya K, Ishibashi T, YamashitaH, Murakawa K, Fujimori K, Tanai H, Kimata M, Watanabe M,Hiraoka S, Chiba Y, Ishida S, Ono Y, Takiguchi S, Watanabe S,Yosida M, Hotuta T, Kusano J, Kanehori K, Takahashi-Fujii A,Hara H, Tanase TO, Nomura Y, Togiya S, Komai F, Hara R,Takeuchi K, Arita M, Imose N, Musashino K, Yuuki H, OshimaA, Sasaki N, Aotsuka S, YoshikawaY,MatsunawaH, Ichihara T,Shiohata N, Sano S, Moriya S, MomiyamaH, SatohN, Takami S,Terashima Y, Suzuki O, Nakagawa S, Senoh A, Mizoguchi H,Goto Y, Shimizu F, Wakebe H, Hishigaki H, Watanabe T,Sugiyama A, Takemoto M, Kawakami B, Yamazaki M,Watanabe K, Kumagai A, Itakura S, Fukuzumi Y, Fujimori Y,Komiyama M, Tashiro H, Tanigami A, Fujiwara T, Ono T,Yamada K, Fujii Y, Ozaki K, Hirao M, Ohmori Y, KawabataA, Hikiji T, Kobatake N, Inagaki H, Ikema Y, Okamoto S,Okitani R, Kawakami T, Noguchi S, Itoh T, Shigeta K, SenbaT,MatsumuraK, Nakajima Y,Mizuno T,MorinagaM, SasakiM,Togashi T, Oyama M, Hata H, Watanabe M, Komatsu T,Mizushima-Sugano J, Satoh T, Shirai Y, Takahashi Y,Nakagawa K, Okumura K, Nagase T, Nomura N, Kikuchi H,Masuho Y, Yamashita R, Nakai K, Yada T, Nakamura Y, OharaO, IsogaiT, SuganoS.Complete sequencingandcharacterizationof 21,243 full-length human cDNAs.Nat Genet 2004;36(1):40–45.

32. Uemura M, Tamura K, Chung S, Honma S, Okuyama A,Nakamura Y, Nakagawa H. Novel 5 alpha-steroid reductase(SRD5A3, type-3) is overexpressed in hormone-refractory pros-tate cancer. Cancer Sci 2008;99(1):81–86.

33. Kohler G, Milstein C. Continuous cultures of fused cells secret-ing antibody of predefined specificity. Nature 1975;256(5517):495–497.

34. Coller HA, Coller BS. Statistical analysis of repetitivesubcloning by the limiting dilution technique with a viewtoward ensuring hybridoma monoclonality. Hybridoma1983;2(1):91–96.

35. Godoy A, Ormazabal V, Moraga-Cid G, Zuniga FA, SotomayorP, Barra V, Vasquez O, Montecinos V, Mardones L, Guzman C,Villagran M, Aguayo LG, Onate SA, Reyes AM, Carcamo JG,Rivas CI, Vera JC. Mechanistic insights and functional determi-nants of the transport cycle of the ascorbic acid transporterSVCT2. Activation by sodium and absolute dependence onbivalent cations. J Biol Chem 2007;282(1):615–624.

36. Godoy A, Watts A, Sotomayor P, Montecinos VP, Huss WJ,Onate SA, Smith GJ. Androgen receptor is causally involvedin the homeostasis of the human prostate endothelial cell.Endocrinology 2008;149(6):2959–2969.

37. Thigpen AE, Cala KM, Russell DW. Characterization ofChinese hamster ovary cell lines expressing human steroid 5alpha-reductase isozymes. J Biol Chem 1993;268(23):17404–17412.

5a -Reductase-3ExpressioninHumanTissues 1045

The Prostate

38. Ford OH, , Gregory CW, Kim D, Smitherman AB, Mohler JL.Androgen receptor gene amplification and protein expression inrecurrent prostate cancer. J Urol 2003;170(5):1817–1821.

39. Karpf AR, Bai S, James SR, Mohler JL, Wilson EM. Increasedexpression of androgen receptor coregulator MAGE-11 in pros-tate cancer by DNA hypomethylation and cyclic AMP. MolCancer Res 2009;7(4):523–535.

40. MahajanNP,WhangYE,Mohler JL, EarpHS.Activated tyrosinekinase Ack1 promotes prostate tumorigenesis: Role of Ack1 inpolyubiquitination of tumor suppressor Wwox. Cancer Res2005;65(22):10514–10523.

41. Majumder S, Liu Y, Ford OH, , Mohler JL, Whang YE.Involvement of argininemethyltransferase CARM1 in androgenreceptor function and prostate cancer cell viability. Prostate2006;66(12):1292–1301.

42. Singh SS, Qaqish B, Johnson JL, Ford OH, , Foley JF, MaygardenSJ, Mohler JL. Sampling strategy for prostate tissue microarraysfor Ki-67 and androgen receptor biomarkers. Anal Quant CytolHistol 2004;26(4):194–200.

43. Smitherman AB, Mohler JL, Maygarden SJ, Ornstein DK.Expression of annexin I, II and VII proteins in androgen stimu-lated and recurrent prostate cancer. J Urol 2004;171(2 Pt 1):916–920.

44. Titus MA, Gregory CW, Ford OH, , Schell MJ, Maygarden SJ,Mohler JL. Steroid 5alpha-reductase isozymes I and II in recur-rent prostate cancer. Clin Cancer Res 2005;11(12):4365–4371.

45. Titus MA, Tan JA, Gregory CW, Ford OH, Subramanian RR, FuH, Wilson EM, Mohler JL, French FS. 14-3-3{eta} AmplifiesAndrogen Receptor Actions in Prostate Cancer. Clin CancerRes 2009;15(24):7571–7581.

46. Sotomayor P, Godoy A, Smith GJ, Huss WJ. Oct4A is expressedby a subpopulation of prostate neuroendocrine cells. Prostate2009;69(4):401–410.

47. Miyamoto KK, McSherry SA, Dent GA, Sar M, Wilson EM,French FS, Sharief Y, Mohler JL. Immunohistochemistry of theandrogen receptor in human benign and malignant prostatetissue. J Urol 1993;149(5):1015–1019.

48. Singh SS, Kim D, Mohler JL. Java Web Start based softwarefor automated quantitative nuclear analysis of prostate cancer

and benign prostate hyperplasia. Biomed Eng Online 2005;4(1):31.

49. Shah SK, Trump DL, Sartor O, Tan W, Wilding GE, Mohler JL.Phase II study of Dutasteride for recurrent prostate cancerduring androgen deprivation therapy. J Urol 2009;181(2):621–626.

50. Wurzel R, Ray P, Major-Walker K, Shannon J, Rittmaster R. Theeffect of dutasteride on intraprostatic dihydrotestosterone con-centrations in men with benign prostatic hyperplasia. ProstateCancer Prostatic Dis 2007;10(2):149–154.

51. EisenbergerMA, LauferM,VogelzangNJ, SartorO, ThorntonD,NeubauerBL, SinibaldiV, LieskovskyG,CarducciMA,ZahurakM, Raghavan D. Phase I and clinical pharmacology of a type Iand II, 5-alpha-reductase inhibitor (LY320236) inprostate cancer:Elevation of estradiol as possible mechanism of action. Urology2004;63(1):114–119.

52. Yamana K, Labrie F, Luu-The V. Human type 3 5a-reductase isexpressed in peripheral tissues at higher levels than type 1 andtype 3 and its activity is potently inhibited by finasteride anddutasteride. Horm Mol Biol Clin Invest 2010;2(3):293–299.

53. Cantagrel V, Lefeber DJ, Ng BG, Guan Z, Silhavy JL, Bielas SL,Lehle L, Hombauer H, Adamowicz M, Swiezewska E,De Brouwer AP, Blumel P, Sykut-Cegielska J, Houliston S,Swistun D, Ali BR, Dobyns WB, Babovic-Vuksanovic D, vanBokhoven H, Wevers RA, Raetz CR, Freeze HH, Morava E,Al-Gazali L, Gleeson JG. SRD5A3 is required for convertingpolyprenol to dolichol and is mutated in a congenital glycosyl-ation disorder. Cell 2010;142(2):203–217.

54. McNeal JE, BostwickDG. Intraductal dysplasia: A premalignantlesion of the prostate. Hum Pathol 1986;17(1):64–71.

55. Thomas LN, Lazier CB, Gupta R, Norman RW, Troyer DA,O’Brien SP, Rittmaster RS. Differential alterations in 5alpha-reductase type 1 and type 2 levels during development andprogression of prostate cancer. Prostate 2005;63(3):231–239.

56. Montgomery RB, Mostaghel EA, Vessella R, Hess DL, KalhornTF, Higano CS, True LD, Nelson PS. Maintenance of intratu-moral androgens inmetastatic prostate cancer: Amechanism forcastration-resistant tumor growth. Cancer Res 2008;68(11):4447–4454.

1046 Godoyet al.

The Prostate