J Sex Med 2005;2:759–770

759

Blackwell Science, LtdOxford, UKJSMJournal of Sexual Medicine1743-6095Journal of Sexual Medicine 2004200526759770Original ArticleTestosterone in Erectile FunctionTraish and Kim

ORIGINAL RESEARCH—ENDOCRINOLOGY

The Physiological Role of Androgens in Penile Erection: Regulation of Corpus Cavernosum Structure and Function

Abdulmaged Traish, PhD, and Noel Kim, PhD

Boston University School of Medicine, Department of Urology, Boston, MA, USA

DOI: 10.1111/j.1743-6109.2005.00094.x

A B S T R A C T

It is generally accepted that androgens are critical for development, growth, and maintenance of penile erectiletissue. However, their role in erectile function, especially in humans, remains controversial. Clinical and preclini-cal studies have suggested that venoocclusion is modulated by the tone of the vascular smooth muscle of theresistance arteries and the cavernosal tissue and a balance between trabecular smooth muscle content and connec-tive tissue matrix. In men with erectile dysfunction, venous leakage is thought to be a common condition amongnonresponders to medical management and is attributed to penile smooth muscle atrophy. In the animal model,androgen deprivation produces penile tissue atrophy concomitant with alterations in dorsal nerve structure,endothelial morphology, reduction in trabecular smooth muscle content, and increased deposition of extracellularmatrix. Further, androgen deprivation results in accumulation of fat-containing cells (adipocytes) in the subtunicalregion of the corpus cavernosum. Androgen deficiency diminishes protein expression and enzymatic activity ofnitric oxide synthases (eNOS and nNOS) and phosphodiesterase type 5 (PDE5). The androgen-dependent loss oferectile response is restored by androgen administration but not by administration of PDE5 inhibitors alone.These data suggest that androgens regulate trabecular smooth muscle growth and connective tissue protein syn-thesis in the corpus cavernosum. Further, androgens may stimulate differentiation of progenitor cells into smoothmuscle cells and inhibit their differentiation into adipocytes. Thus, we conclude that androgens exert a directeffect on penile tissue to maintain erectile function and that androgen-deficiency produces a metabolic and struc-tural imbalance in the corpus cavernosum, resulting in venous leakage and erectile dysfunction. Traish A, Kim N.The physiological role of androgens in penile erection: regulation of corpus cavernosum structure andfunction. J Sex Med 2005;2:759–770.

Key Words. Adipocytes; Anatomy (Gross and Microscopic); Androgens; Animal Models; Arginase; EndocrinologicStudies of Sexual Function; Nitric Oxide Synthase; Sexual Arousal; Smooth Muscle; Vascular Physiologic Studiesof Genital Arousal

Introduction

ormal erectile function is dependent uponthe health of the penile vascular tissues and

the perineal and ischiocavernosus muscles thatsupport the proximal penis. Adequate arterialinflow and trapping of blood within the cavernosalbodies (venoocclusion) is critical for the develop-ment of increasing pressure and volume expansion.In addition to arterial blood pressure, contractionof the perineal and ischiocavernosus muscles

Nenhances penile rigidity. The venoocclusivemechanism depends on the integrity of neural,vascular, and endocrine systems, as well as on thefibroelastic properties of the cavernosal tissue. Ithas been noted that cavernosal tissue from menwith erectile dysfunction of various etiologies––whether hormonal, neurological, or vascular––exhibited reduced smooth muscle content and con-comitant increase in connective tissue deposition[1,2]. It is likely that such changes in penile tissuestructure contribute to venoocclusive dysfunction.

760 Traish and Kim

J Sex Med 2005;2:759–770

Clinical studies have suggested that surgical ormedical castration results in loss of libido anderectile function [3–10]. In a double-blind corre-lation analysis, Aversa et al. [11,12] studied menwith erectile dysfunction without vascular riskfactors. The authors noted a strong direct corre-lation between resistive index values and freetestosterone, a relationship that was maintainedafter adjusting for age, sex hormone bindingglobulin, and estradiol. They concluded thatmen with erectile dysfunction and low free test-osterone may have impaired relaxation of penilesmooth muscle, thus providing clinical evidencefor the importance of androgen in regulatingerectile function. In selected men with total test-osterone below 10–13 nmol/L and/or free test-osterone below 200–250 pmol/L, androgensupplementation improved therapeutic efficacy ofphosphodiesterase type 5 (PDE5) inhibitors [13].In addition, hypogonadal men with confirmedlack of response to sildenafil monotherapyshowed greater improvement in erectile functionwhen treated with testosterone [14]. A relation-ship between restoration of serum testosteroneconcentrations and improvement in sexual func-tion has been proposed by Seftel et al. [15]. Insevere hypogonadal men, testosterone treatmentfor 6 months induced normalization of nocturnalpenile tumescence activity [16]. The authors sug-gested that testosterone has a key role in thecentral and peripheral modulation of erectilefunction even if the specific threshold con-centration of plasma testosterone remains to beestablished.

In laboratory studies using an animal model,Mills et al. [17–21] proposed that androgens arecritical for maintaining erectile function and mayact specifically to support the responsiveness of thevascular smooth muscle to vasoactive drugs. Babaet al. [22,23] showed that the intracavernosal pres-sure decreased significantly in castrated animals(vs. control) both after pelvic nerve stimulationand after intracavernosal papaverine injection.More importantly, testosterone replacementrestored penile hemodynamics. Testosterone hasalso been shown to be critical for maintaining themass of skeletal muscles in the perineum, as wellas neuron size [24–26]. In addition to these trophiceffects, which are presumed to be due to mecha-nisms involving transcriptional regulation, andro-gens may also regulate vascular smooth musclecontractility by nongenomic mechanisms. Whilethese mechanisms have not been investigated inpenile cavernosal tissue, testosterone has been

shown to relax coronary arteries by activatingpotassium channels and inhibiting calciumchannels [27–29].

One of the least understood aspects of erectilefunction is the role of androgens in maintainingpenile structural and functional integrity. To date,research efforts on the mechanisms by whichandrogens regulate penile erectile physiology havemainly focused on the role of the nitric oxide/cyclic guanosine monophosphate (NO/cGMP)pathway. However, androgen-dependent mecha-nisms that regulate tissue remodeling have beenpoorly defined. The objectives of this review areto summarize the existing data on the effects ofandrogen deficiency on specific biochemical path-ways and tissue structure in the corpus caverno-sum and to critically evaluate the potential effectsof androgen treatment on penile tissue remodelingand erectile function.

Role of Androgens in Penile Tissue DevelopmentIt has been generally recognized that growthand development of the penis is an androgen-dependent process [30,31]. Clinical and preclinicalstudies indicate that testosterone undergoes con-version to 5a-dihydrotestosterone (5a-DHT) inorder to modulate penile erectile growth or phys-iology. Consistent with this is the observation thatthe decline in erectile activity in orchiectomizedrats is prevented by testosterone alone or 5a-DHT alone, but not by coadministration of test-osterone and 5a-reductase inhibitor [32]. Further,Choi et al. [33] and Charmandari et al. [34]showed that percutaneous administration of 5a-DHT was successful in promoting phallic growthin infants and children with microphallus due to5a-reductase deficiency. Gonzalez-Cadavid et al.[35] and Lugg et al. [36] have suggested that 5a-DHT is the active androgen that promotes penilesmooth muscle growth and prevents erectile fail-ure in castrated rats. These investigators postu-lated that these beneficial effects are mediated, atleast partially, by modulation of nitric oxide syn-thase (NOS) via the androgen receptor.

In our studies using male New Zealand whiterabbits, we have observed that larger, older ani-mals (4.9 ± 0.2 kg and >8 months of age) exhibitbetter erectile responses to pelvic nerve stimula-tion at low frequencies than smaller, younger ani-mals (2.9 ± 0.5 kg and <6 months of age), while nodifference was observed at higher frequencies (seeFigure 1). This shift in the frequency response istypical for inhibitory modulation. In addition, theerectile tissue from the younger animals had sinu-

Testosterone in Erectile Function 761

J Sex Med 2005;2:759–770

soids that were smaller than those of the olderanimals (see Figure 2). The New Zealand whiterabbit begins sexual maturation at approximately2 months of age and is considered a mature adultat 6 months when skeletal growth is complete [37–

39]. Presumably, the “smaller younger” group inour study consists of adolescent animals, while the“larger older” group consists of fully mature ani-mals. Thus, it may be inferred that penile tissuecontinues to develop over a prolonged period oftime after the initial increase in testosterone levelsassociated with adolescence and puberty. Cavern-osal tissue structure and trabecular smooth musclefunction may continue to change during this timeto attain optimal function when the animal reachesmaturity. These observations support a role forandrogens in penile tissue development and func-tional response to sexual stimulation.

Role of Androgens in Maintaining Erectile Tissue InnervationMeusburger and Keast [40] have suggested thatcirculating androgens have potent effects onmaintaining the structure and function of manypelvic ganglion neurons. Recently, Keast et al.[41] demonstrated that testosterone is critical formaturation and maintenance of terminal axondensity and neuropeptide expression in the vasdeferens. However, limited data are available onthe role of androgens in maintaining nerve fiberstructure and regulating the synthesis anddistribution of neurotransmitters in the corpuscavernosum.

Using transmission electron microscopy, Rog-ers et al. [42] have shown that castration results inultrastructural changes in the dorsal nerve of therat. The dorsal nerve from sham-operated animalswas filled with both myelinated and nonmyeli-nated nerve bundles. In tissue from castrated rats,the diameter of both the myelinated and nonmy-elinated axons appeared smaller than those of thesham-operated rats. Many nonmyelinated nervefibers became indistinct and smaller. There wasalso an increase in the number of nucleatedSchwann cells. Testosterone treatment of castratedanimals restored the nerve fibers and myelinsheath structure, which were similar to that foundin the sham (control) group, suggesting thatandrogens are important in maintaining periph-eral autonomic and sensory nerve structure andfunction in the penis.

Role of Androgens in Regulating NOS Isoforms in Penile TissueThe NOS/cGMP pathway has been deemed crit-ical for erectile function. NO mediates relaxationof the vascular smooth muscle of the resistancearteries and the trabeculae in the corpus caverno-sum to facilitate penile erection. Numerous stud-

Figure 2 Cavernosal tissue structure in adolescent andmature rabbits. Penile cavernosal tissues from small (A)and large (B) rabbits (see Figure 1) were fixed in bufferedformalin, embedded in paraffin, sectioned (6 mm), and sub-jected to Masson’s trichrome staining procedure. Cellsappear red, while extracellular material appears blue-purple. Note the difference in the sizes of the trabecularsmooth muscle bundles and the cavernosal sinusoids.

Figure 1 Erectile response in adolescent and mature rab-bits. New Zealand white rabbits were categorized by totalbody weight into small adolescent (<3.5 kg) and largemature (>4.5 kg) groups. Erectile response to pelvic nervestimulation was assessed by determining the area underthe curve (AUC) of intracavernosal pressure recordings foreach stimulation frequency.

762 Traish and Kim

J Sex Med 2005;2:759–770

ies have reported that androgens regulate theexpression of NOS isoforms in the corpus caver-nosum of various species [22,32,43–53]. Orchiec-tomy results in a rapid decrease in serumtestosterone and a decreased erectile response tocavernous nerve stimulation. A marked decreasein nicotinamide adenine dinucleotide phosphate,reduced form (NADPH)-diaphorase staining, amarker for neural NOS (nNOS) expression, wasalso observed. Testosterone administration toorchiectomized animals restored the erectileresponse and normalized the staining pattern ofNADPH-positive nerve fibers [23,45,52].

Western blot analyses and biochemical assaysshowed that castration of adult male rats reducesby half the activity of penile nNOS and endo-thelial NOS. These metabolic and physiologicalchanges were restored by androgen administration[47,50,54]. Reilly et al. [55] also demonstratedthat the erectile response in the rat penis isandrogen-dependent and may be mediated byNO-independent, as well as NO-dependent,pathways. Interestingly, both of these pathwaysappear to involve the synthesis of cGMP. In otherstudies, analyses of mRNA for NOS isoforms inpenile tissue from intact and castrated rats treatedwith or without testosterone suggested that andro-gens enhance nNOS gene expression in the penilecorpus cavernosum of rats and that DHT wasmore potent than testosterone [53].

Aging-related impairment of penile erectioncan be corrected by administration of exogenousandrogens [46]. However, enzymatic and semi-quantitative Western blot assays found no signifi-cant variations in either NOS activity or NOSlevels between untreated aged rats or those treatedwith testosterone. Thus, the authors suggestedthat aging-related erectile dysfunction in theintact rat may be responsive to androgen therapy,but this correction is not associated with anincrease in the basal levels of penile nNOS, incontrast with that observed in castrated rats.

Role of Androgens in the Regulation of PDE5 Expression in Penile TissuePDE5 is the predominant enzyme responsible forcGMP hydrolysis in trabecular smooth muscle.Activation of PDE5 terminates NO-induced,cGMP-mediated smooth muscle relaxation,resulting ultimately in restoration of basal smoothmuscle contractility and penile flaccidity. Follow-ing sexual stimulation, PDE5 inhibitors act toenhance cGMP-mediated smooth muscle relax-

ation, resulting in improved penile erection inmen with erectile dysfunction.

We have reported that castration resulted inreduced expression and activity of PDE5 in therabbit model and that androgen treatment upreg-ulated the expression of PDE5 activity [56]. Thisobservation was also confirmed in the rat modelin our laboratory (A. Traish, unpublished obser-vation). Using real-time reverse transcript poly-merase chain reaction (RT-PCR) and Westernblot analyses, Morelli et al. [57,58] have alsoobserved that androgen deprivation reducedPDE5 expression, while testosterone supplemen-tation restored PDE5 gene and protein expressionin rabbit corpus cavernosum. In addition, organbath assays with isolated cavernosal tissue stripsshowed that androgen deprivation reduced thefacilitative effect of sildenafil on neurogenic relax-ation, and this facilitation was restored in tissuesof castrated animals treated with testosterone.These data suggest that androgens are critical formaintaining normal expression of PDE5 in rabbitpenis.

In a separate study, we tested if administrationof PDE5 inhibitor to animals that were surgicallyor medically (luteinizing hormone–releasing hor-mone agonist treated) castrated enhances erectilefunction in response to pelvic nerve stimulation[59]. Both surgical and medical castration signifi-cantly decreased plasma androgen concentrationcompared with intact control animals. Administra-tion of PDE5 inhibitor 10 minutes before nervestimulation enhanced erectile function in intactanimals. However, administration of PDE5 inhib-itor to surgically or medically castrated animalsdid not enhance erectile function. Taken together,these observations suggest that androgens arecritical not only for NOS activity but also formodulating PDE5 activity.

Role of Androgens in Maintaining Penile Trabecular Smooth Muscle Structure and FunctionTrabecular smooth muscle is an important com-ponent of the penis, regulating detumescence anderection [60]. In surgically or medically castratedanimal models, we have demonstrated that andro-gen deprivation results in a significant reductionin trabecular smooth muscle content and markedincrease in connective tissue deposition [56,59].These structural alterations are also associatedwith loss of erectile function as demonstratedby the reduction in intracavernosal pressure inresponse to electrical stimulation of the pelvicnerve.

Testosterone in Erectile Function 763

J Sex Med 2005;2:759–770

In preliminary studies using transmission elec-tron microscopy, we have noted marked differ-ences in the trabecular smooth muscle fromcastrated animals compared with that of intact(sham) animals. In castrated animals, the smoothmuscle appeared disorganized, with a large num-ber of cytoplasmic vacuoles, whereas in the intactanimals, the smooth muscle cells exhibited normalmorphology and were arranged in clusters(Figure 3). These observations are supported bythe studies of Rogers et al. [42], who assessed theeffects of androgen deprivation and supplementa-tion on cavernosal tissues, using quantitative

analysis of the smooth muscle content (by alphasmooth muscle actin staining) and transmis-sion electron microscopy. The authors noted thatthere was no significant difference in the smoothmuscle content between intact, castrated, andtestosterone-treated castrated animals. However,transmission electron microscopy showed markeddifferences in the morphology of the smooth mus-cle cells. In sham-operated rats, the smooth mus-cle cells were arranged in clusters and separatedby fine strands of fibroconnective tissue. Theintercellular spaces between myocytes were usu-ally quite narrow, with many gap junctionsconnecting individual cells, and the cytoplasmcontained abundant contractile myofilamentsand dense bodies. In contrast, myocytes in caver-nosal tissue from castrated animals exhibitedlarger intercellular spaces and decreased amountsof cytoplasmic myofilaments. Tissue fromtestosterone-treated castrated rats had myocytesthat were similar in appearance to those ofsham-operated animals.

These results suggest that androgens have aprofound effect on the ultrastructure of the corpuscavernosum and that these alterations may beresponsible for the loss of physiological function.However, limited investigations have been madeto define the molecular and cellular mechanismspertaining to changes in penile nerve fibers, tra-becular smooth muscle, or connective tissueproteins in response to androgen deficiency andsupplementation. In androgen-dependent prostatecancer cells, androgens have been shown to inhibitapoptosis by blocking the activation of caspases, alarge family of proteases that play a central role inapoptosis [61]. In other studies using cell lines thatexpress androgen receptor, androgens activatepathways involving phosphatidylinositol 3-kinaseand Akt, which can inhibit apoptosis [62]. How-ever, the effect of androgens on these pathways inpenile corpus cavernosum have not been exam-ined. A more complete discussion of apoptosis canbe found in recent publications [63,64]. In addi-tion to programmed cell death, androgen defi-ciency may cause changes in smooth muscle cellsthat pertain to their morphology, orientation,organelle content/function, contacts with connec-tive tissue proteins, and responsiveness to vasoac-tive substances.

Role of Androgens in Maintenance of Penile Corpus Cavernosum Fibroelastic PropertiesThe extracellular matrix (ECM) is a dynamic cel-lular scaffold that plays an important role in mod-

Figure 3 Effects of androgen deprivation on smooth mus-cle organization and ultrastructure in corpus cavernosum.Mature New Zealand white rabbits were left intact ororchiectomized. After 2 weeks, penile cavernosal tissuesfrom intact (A) or orchiectomized (B) animals were fixed inglutaraldehyde, postfixed with osmium tetroxide, embeddedin epoxy resin, and sections (1 mm) were stained with uranylacetate/lead citrate and examined by transmission electronmicroscopy.

764 Traish and Kim

J Sex Med 2005;2:759–770

ulating tissue physiological function. The ECMregulates cell morphology, movement, growth,differentiation, and survival by regulating celladhesion, cytoskeletal machinery, and intracellularsignaling. The ECM in the corpus cavernosumconsists of a network of fibrillar collagen which isintimately connected to the trabecular smoothmuscle. Fibrillar collagen types I and III are themajor components of the corpus cavernosum col-lagen matrix [65–67]. Collagen type I representsmost of the total collagen protein, while type IIIcollagen is present in lower proportions [65,66].Collagen types I and III exhibit high tensilestrength which plays an important role in thefunction of the corpus cavernosum during erectionand flaccidity [68–70].

Changes in the fibroelastic properties of peniletissue may alter penile compliance and hemody-namics, resulting in erectile dysfunction. Histo-logical studies have shown an association betweenvasculogenic erectile dysfunction and alteredECM deposition [2,67,71–73]. The severity oferectile dysfunction correlated with a reducedsmooth muscle/ECM ratio [73–75]. Furthermore,in animal studies, androgen ablation resulted in asignificant reduction of trabecular smooth musclecontent and increased connective tissue deposi-tion. These marked structural alterations are asso-ciated with loss of erectile function [56,59]. Todate, limited studies have investigated the role ofandrogens in modulating the synthesis, accumula-tion, and deposition of connective tissue proteinsin the penis (e.g., collagens, fibronectin, and elas-tin). In addition, no data are available on the reg-ulation of growth factors that modulate ECM ormetalloproteases by androgens.

The amount, composition, and organization ofECM is maintained or altered by at least threedifferent biochemical processes involving connec-tive tissue proteins (collagens, elastin, fibronectin,etc.): (i) synthesis and post-translational modifi-cation; (ii) assembly and cross-linking; and (iii)degradation. ECM deposition is the net resultof increased synthesis and reduced degradation(turnover). Gene expression of ECM proteins isregulated by multiple signals including hormones,cytokines, and growth factors. TGF-b1, connec-tive tissue growth factor (CTGF), and vascularendothelial growth factor (VEGF) are thought toplay a critical role in ECM remodeling that takesplace in normal physiological processes such asembryogenesis, implantation, and wound healing,as well as in fibrosis and scarring by diverse auto-crine and paracrine mechanisms in many cell types

[76]. A regulatory interplay has been proposed toexist between TGF-b1 and CTGF as well asVEGF and CTGF in remodeling of the ECM[77–83]. Further, cyclic AMP, forskolin, prostag-landin E2, and tumor necrosis factor alpha havebeen shown to inhibit the action of CTGF at thetranscriptional level [76]. However, to date, nodata are available on regulation of these growthfactors by androgens in penile tissue.

We postulate that in our androgen-deprivedanimal model, the changes in penile erectile func-tion may be due to (i) reduced synthesis of para-crine growth factors (e.g., VEGF, fibroblastgrowth factor, and insulin-like growth factor-I)necessary to maintain the structural and functionalintegrity of the smooth muscle, endothelium, andnerves; (ii) upregulation of paracrine factors (e.g.,CTGF and TGF-b1) which increase expression ofconnective tissue proteins; and (iii) downregula-tion of metalloproteinases and upregulation oftissue inhibitors of metallproteinases, resultingin increased ECM deposition.

Thus, we suggest that androgen deprivation, inthe animal model, produces alterations in cellularsignaling, structure and function of penile nerves,trabecular smooth muscle, endothelium, and con-nective tissue. These alterations are manifestedas changes in tissue fibroelastic properties thatimpede expandability of the corpus cavernosum,causing reduced blood inflow and failure of thevenoocclusive mechanism, contributing to erectiledysfunction (Figure 4).

Role of Androgens in Regulating Adipocyte Accumulation in the Subtunical Region of the Corpus CavernosumIn addition to the changes in connective tissue andsmooth muscle, histological examination of peniletissue sections from orchiectomized animalsstained with hematoxylin and eosin or Masson’strichrome revealed clusters of “empty” cellularstructures in the subtunical region of the corpuscavernosum that were distinct from cavernousspaces. These hollow cells resembled adipocytesand were consistently present in tissue fromorchiectomized animals but absent in tissue fromcontrol animals. Because normal processing ofparaffin-embedded tissue with organic solventsresults in removal of fat droplets, we fixed peniletissue in glutaraldehyde and postfixed withosmium tetroxide, which binds to unsaturated lip-ids and results in a brown or gray-black stain.When tissues were embedded in epoxy resin andsectioned, we observed large lipid droplets that

Testosterone in Erectile Function 765

J Sex Med 2005;2:759–770

were stained with osmium tetroxide, verifying thesubtunical cells were indeed adipocytes (Figure 5).These alterations in cavernosal tissue compositionand structure are accompanied by a reduced erec-tile response to pelvic nerve stimulation. It isinteresting to note that because venoocclusiondepends on the compression of the subtunicalvenules to impede blood outflow during sexualstimulation, it is possible that the presence of fatcells in the subtunical region of the corpus caver-nosum may contribute to venous leak in theorchiectomized animal [84].

Previous studies in the intact rabbit have notedthat administration of the endocrine disrupters,bisphenol A and tetrachlorodibenzodioxin (TCDD),which may act as antiandrogens in the corpuscavernosum, resulted in an abnormal deposition

of fat-containing cells in the subtunical regionof the corpus cavernosum [85,86]. Organ bathstudies using cavernosal tissue from bisphenol Aand TCDD-treated animals showed a reducedrelaxation response to nitroprusside and acetyl-

Figure 4 Working model of androgen action in the corpuscavernosum. This framework proposes that androgens arecritical for maintaining the structure and function of penileerectile tissue. Thus, androgen deprivation may result insmooth muscle atrophy and apoptosis, altered nerve andendothelial growth, and increased deposition of extracellu-lar matrix (ECM). We suggest that androgens may berequired for maintaining an appropriate balance of para-crine and autocrine growth factors (CTGF, VEGF, TGF-b,NO). In addition, androgens may regulate the proteasesthat are involved in the dynamics of ECM accumulationand deposition (MMPs and TIMPs). Androgen-sensitivechanges in the production of growth factors and ECM con-stituents most likely occur early in relation to changes incell growth/atrophy/apoptosis and ECM deposition, whichoccur later in the remodeling process. Thus, androgen dep-rivation may lead to overexpression of ECM and atrophy ofsmooth muscle, dysfunction of some or all of the cellularcomponents (nerve, smooth muscle, endothelium), result-ing in erectile dysfunction. Furthermore, these adversechanges may be reversible with androgen supplementation.

Figure 5 Appearance of adipocytes after androgen depri-vation. Mature New Zealand white rabbits were left intactor orchiectomized. After 2 weeks, penile cavernosal tissuesfrom intact (A) or orchiectomized (B) animals were fixed,embedded, and sectioned for light microscopy, asdescribed in Figure 2, and stained with hematoxylin andeosin. Tissue samples from orchiectomized animals werealso processed in parallel for transmission electron micros-copy (C), as described in Figure 3. Arrows in panel Bdenote “hollow” cellular structures in the subtunical region.

766 Traish and Kim

J Sex Med 2005;2:759–770

choline. In other studies, neonatal rats exposed tothe estrogen receptor agonist, diethylstilbestrol,showed accumulation of fat-containing cells inthe penile corpus cavernosum, whereas animalstreated with vehicle exhibited no fat-containingcells [87,88]. The authors suggested that estrogentreatment, coupled with low plasma androgen lev-els, may have contributed to alterations in penilemorphology, infertility, and erectile dysfunction[87,88]. Because estrogens are known to act asantiandrogens, these studies point to the potentialrole of androgens in maintaining penile corpuscavernosum structural integrity.

The mechanism by which androgens regulategrowth and differentiation of trabecular smoothmuscle cells and/or adipocytes in the penis remainspoorly understood. Bhasin et al. [89] proposed thatandrogens promote the commitment of pluripo-tent stem cells into a muscle lineage and inhibittheir differentiation into an adipocyte lineage. Ina recent study, Singh et al. [90] have shown thatdifferentiation of pluripotent cells is androgen-dependent. Both testosterone and DHT decreasedthe number of adipocytes and downregulated theexpression of the adipogenic markers PPAR-g2 andC/EBPa. However, these mechanisms have yet tobe investigated in tissue or cells from the corpuscavernosum. It is possible that pluripotent stemcells are present in the corpus cavernosum and thatthese cells respond to androgen deprivation bydifferentiation into an adipogenic lineage.

Another possibility is the dedifferentiation ofthe corpus cavernosum trabecular smooth musclecells into other phenotypes. Corradi et al. [91]have shown that inhibition of 5a-reductase activ-ity induces stromal remodeling and smooth mus-cle dedifferentiation in the prostate, suggestingthat 5a-DHT deficiency promotes smooth musclededifferentiation. While in several experimentalsystems vascular smooth muscle was shown toundergo dedifferentiation into other phenotypes[92,93], there are no data in the literature on thededifferentiation of the trabecular smooth musclein the corpus cavernosum. Future studies usingexpression of muscle-specific biochemical markersas well as changes in ultrastructure, as determinedby electron microscopy, will be needed to test thispossibility in the corpus cavernosum under andro-gen deprivation and supplementation.

Summary and Perspective

The effects of androgens on erectile physiologyare complex. We suggest that in the corpus caver-

nosum, androgens regulate (i) the expression and/or activity of NOS isoforms, phosphodiesterases,and ion channels; (ii) the growth and state of dif-ferentiation of smooth muscle cells; (iii) connec-tive tissue metabolism; and (iv) the differentiationof progenitor stromal cells into myogenic and adi-pogenic lineages. Much of the data on mechanisticalterations influenced by androgens are necessarilyobtained through animal studies. Many of thesestudies rely on castrated animal models that causealmost immediate and severe androgen depletionvs. the gradual decline that occurs in most men.Interestingly, a recent study demonstrated a dose-dependent effect of testosterone on changes inoverall sexual function and self-reported wakingerections in men between the ages of 60 and75 years [94]. While some investigators have usedaging animal models in laboratory studies, furtherinvestigation of the effects of multiple doses ofandrogens (ranging from subphysiological tophysiological) is required. In summary, androgendeficiency is likely to affect multiple cellular com-ponents and molecular pathways to adverselychange the structural and functional integrity ofpenile corpus cavernosum.

Acknowledgments

We wish to thank our colleagues, Drs. Irwin Goldstein,Ricardo Munarriz, Seong Choi, Kweonsik Min, SooWoong Kim, Seong-Joo Jeong, Kwangsung Park, andPaul Toselli for their helpful suggestions and contribu-tions to this work.

Corresponding Author: Abdulmaged M. Traish,PhD, Boston University School of Medicine—Urology,700 Albany St. W607, Boston, MA 02118, USA.Tel: (617) 638-4578; Fax: (617) 638-5412; E-mail:atraish@bu.edu

Conflict of Interest: None.

References

1 Karadeniz T, Topsakal M, Aydogmus A, Gulgun C,Aytekin Y, Basak D. Correlation of ultrastructuralalterations in cavernous tissue with the clinical diag-nosis vasculogenic impotence. Urol Int 1996;57:58–61.

2 Mersdorf A, Goldsmith PC, Diederichs W, PadulaCA, Lue TF, Fishman IJ, Tanagho EA. Ultrastruc-tural changes in impotent penile tissue: A compari-son of 65 patients. J Urol 1991;145:749–58.

3 Ellis WJ, Grayhack JT. Sexual function in agingmales after orchiectomy and estrogen therapy. JUrol 1963;89:895–8.

Testosterone in Erectile Function 767

J Sex Med 2005;2:759–770

4 Peters CA, Walsh PC. The effect of nafarelinacetate, a luteinizing-hormone-releasing hormoneagonist, on benign prostatic hyperplasia. N Engl JMed 1987;317:599–604.

5 Rousseau L, Dupont A, Labrie F, Couture M.Sexuality changes in prostate cancer patientsreceiving antihormonal therapy combining theantiandrogen flutamide with medical (LHRHagonist) or surgical castration. Arch Sex Behav1988;17:87–98.

6 Marumo K, Baba S, Murai M. Erectile functionand nocturnal penile tumescence in patients withprostate cancer undergoing luteinizing hormone-releasing hormone agonist therapy. Int J Urol 1999;6:19–23.

7 Greenstein A, Plymate SR, Katz PG. Visuallystimulated erection in castrated men. J Urol1995;153:650–2.

8 Hirshkowitz M, Moore CA, O’Connor S, BellamyM, Cunningham GR. Androgen and sleep-relatederections. J Psychosom Res 1997;42:541–6.

9 Eri LM, Tveter KJ. Safety, side effects andpatient acceptance of the luteinizing hormonereleasing hormone agonist leuprolide in treatmentof benign prostatic hyperplasia. J Urol 1994a;152:448–52.

10 Eri LM, Tveter KJ. Safety, side effects and patientacceptance of the antiandrogen Casodex in thetreatment of benign prostatic hyperplasia. Eur Urol1994b;26:219–26.

11 Aversa A, Isidori AM, De Martino MU, Caprio M,Fabbrini E, Rocchietti-March M, Frajese G, FabbriA. Androgens and penile erection: Evidence for adirect relationship between free testosterone andcavernous vasodilation in men with erectile dysfunc-tion. Clin Endocrinol 2000;53:517–22.

12 Aversa A, Isidori AM, Spera G, Lenzi A, Fabbri A.Androgens improve cavernous vasodilation andresponse to sildenafil in patients with erectile dys-function. Clin Endocrinol 2003;58:632–8.

13 Aversa A, Isidori AM, Greco EA, Giannetta E,Gianfrilli D, Spera E, Fabbri A. Hormonal supple-mentation and erectile dysfunction. Eur Urol2004;45:535–8.

14 Shabsigh R, Kaufman JM, Steidle C. Padma-Nathan H. Randomized study of testosterone gel asadjunctive therapy to sildenafil in hypogonadal menwith erectile dysfunction who do not respond tosildenafil alone. J Urol 2004;172:658–63.

15 Seftel AD, Mack RJ, Secrest AR, Smith TM.Restorative increases in serum testosterone levelsare significantly correlated to improvements insexual functioning. J Androl 2004;25:963–72.

16 Foresta C, Caretta N, Rossato M, Garolla A, FerlinA. Role of androgens in erectile function. J Urol2004;171:2358–62.

17 Mills TM, Lewis RW. The role of androgens in theerectile response: A 1999 perspective. Mol Urol1999;3:75–86.

18 Mills TM, Lewis RW, Stopper VS. Androgenicmaintenance of inflow and veno-occlusion dur-ing erection in the rat. Biol Reprod 1998;59:1413–8.

19 Mills TM, Stopper VS, Reilly CM. Sites ofandrogenic regulation of cavernosal blood pressureduring penile erection in the rat. Int J ImpotRes 1996;8:29–34.

20 Mills TM, Stopper VS, Wiedmeier VT. Effects ofcastration and androgen replacement on the hemo-dynamics of penile erection in the rat. Biol Reprod1994;51:234–8.

21 Mills TM, Wiedmeier VT, Stopper VS. Androgenmaintenance of erectile function in the rat penis.Biol Reprod 1992;46:342–8.

22 Baba K, Yajima M, Carrier S, Morgan DM, NunesL, Lue TF, Iwamoto T. Delayed testosteronereplacement restores nitric oxide synthase-containing nerve fibres and the erectile responsein rat penis. BJU Int 2000;85:953–8.

23 Baba K, Yajima M, Carrier S, Akkus E, Reman J,Nunes L, Lue TF, Iwamoto T. Effect of testoster-one on the number of NADPH diaphorase-stainednerve fibers in the rat corpus cavernosum and dorsalnerve. Urology 2000;56:533–8.

24 Nanasaki Y, Sakuma Y. Perineal musculature andits innervation by spinal motoneurons in the malerabbit: Effects of testosterone. J Nippon Med Sch2000;67:164–71.

25 Matsumoto A. Androgen stimulates neuronal plas-ticity in the perineal motoneurons of aged male rats.J Comp Neurol 2001;430:389–95.

26 Rand MN, Breedlove SM. Androgen alters thedendritic arbors of SNB motoneurons by actingupon their target muscles. J Neurosci 1995;15:4408–16.

27 Yue P, Chatterjee K, Beale C, Poole-WilsonPA, Collins P. Testosterone relaxes rabbit coro-nary arteries and aorta. Circulation 1995;91:1154–60.

28 Deenadayalu VP, White RE, Stallone JN, Gao X,Garcia AJ. Testosterone relaxes coronary arteries byopening the large-conductance, calcium-activatedpotassium-channel. Am J Physiol 2001;281:H1720–7.

29 Scragg JL, Jones RD, Channer KS, Jones TH, PeersC. Testosterone is a potent inhibitor of 1-type Ca2+

channels. Biochem Biophys Res Commun 2004;318:503–6.

30 Dorfman RI, Shipley RA. Androgens: biochemistry,physiology, and clinical significance. New York:John Wiley & Sons, Inc; 1956.

31 Shen R, Lin MC, Sadeghi F, Swerdloff RS, RajferJ, Gonzalez-Cadavid NF. Androgens are not majordown-regulators of androgen receptor levels duringgrowth of the immature rat penis. J SteroidBiochem Mol Biol 1996;57:301–13.

32 Seo SI, Kim SW, Paick JS. The effects of androgenon penile reflex, erectile response to electrical stim-

768 Traish and Kim

J Sex Med 2005;2:759–770

ulation and penile NOS activity in the rat. Asian JAndrol 1999;1:169–74.

33 Choi SK, Han SW, Kim DH, de Lignieres B.Transdermal dihydrotestosterone therapy and itseffects on patients with microphallus. J Urol1993;150:657–60.

34 Charmandari E, Dattani MT, Perry LA, HindmarshPC, Brook CG. Kinetics and effect of percutaneousadministration of dihydrotestosterone in children.Horm Res 2001;56:177–81.

35 Gonzalez-Cadavid N, Vernet D, Fuentes NA,Rodriguez JA, Swerdloff RS, Rajfer J. Up-regulationof the levels of androgen receptor and its mRNAby androgens in smooth-muscle cells from rat penis.Mol Cell Endocrinol 1993;90:219–29.

36 Lugg JA, Rajfer J, Gonzalez-Cadavid NF. Dihy-drotestosterone is the active androgen in the main-tenance of nitric oxide-mediated penile erection inthe rat. Endocrinology 1995;136:1495–501.

37 Berger M, Jean-Faucher C, de Turckheim M,Veyssiere G, Blanc MR, Poirier JC, Jean C. Test-osterone, luteinizing hormone (LH) and folliclestimulating hormone (FSH) in plasma of rabbitfrom birth to adulthood: Correlation with sexualand behavioural development. Acta Endocrinol-Cop 1982;99:459–65.

38 Gilsanz V, Roe TF, Gibbens DT, Schulz EE,Carlson ME, Gonzalez O, Boechat MI. Effectof sex steroids on peak bone density of growingrabbits. Am J Physiol 1988;1:255:E416–21.

39 Masoud I, Shapiro F, Kent R, Moses A. A longitu-dinal study of the growth of the New Zealand whiterabbit: Cumulative and biweekly incrementalgrowth rates for body length, body weight, femorallength, and tibial length. J Orthop Res 1986;4:221–31.

40 Meusburger SM, Keast JR. Testosterone and nervegrowth factor have distinct but interacting effects onstructure and neurotransmitter expression of adultpelvic ganglion cells in vitro. Neuroscience2001;108:331–40.

41 Keast JR, Gleeson RJ, Shulkes A, Morris MJ. Mat-urational and maintenance effects of testosterone onterminal axon density and neuropeptide expressionin the rat vas deferens. Neuroscience 2002;112:391–8.

42 Rogers RS, Graziottin TM, Lin CM, Kan YW, LueT. Intracavernosal vascular endothelial growth fac-tor (VEGF) injection and adeno-associated virus-mediated VEGF gene therapy prevent and reversevenogenic erectile dysfunction in rats. Int J ImpotRes 2003;15:26–37.

43 Lugg J, Ng C, Rajfer J, Gonzalez-Cadavid N. Cav-ernosal nerve stimulation in the rat reverses castra-tion-induced decrease in penile NOS activity. Am JPhysiol 1996;271:E354–61.

44 Muller SC, Hsieh JT, Lue TF, Tanagho EA. Cas-tration and erection. An animal study. Eur Urol1988;15:118–24.

45 Zvara P, Sioufi R, Schipper HM, Begin LR, BrockGB. Nitric oxide mediated erectile activity is a tes-tosterone dependent event: A rat erection model.Int J Impot Res 1995;7:209–19.

46 Garban H, Marquez D, Cai L, Rajfer J, Gonzalez-Cadavid NF. Restoration of normal adult penileerectile response in aged rats by long-term treat-ment with androgens. Biol Reprod 1995;53:1365–72.

47 Penson DF, Ng C, Cai L, Rajfer J, Gonzalez-Cadavid NF. Androgen and pituitary control ofpenile nitric oxide synthase and erectile function inthe rat. Biol Reprod 1996;55:567–74.

48 Penson DF, Ng C, Rajfer J, Gonzalez-Cadavid NF.Adrenal control of erectile function and nitricoxide synthase in the rat penis. Endocrinology1997;138:3925–32.

49 Shen Z, Chen Z, Lu Y, Chen F, Chen Z. Relation-ship between gene expression of nitric oxide syn-thase and androgens in rat corpus cavernosum. ChinMed J (Engl) 2000;113:1092–5.

50 Marin R, Escrig A, Abreu P, Mas M. Androgen-dependent nitric oxide release in rat penis correlateswith levels of constitutive nitric oxide synthaseisoenzymes. Biol Reprod 1999;61:1012–6.

51 Reilly CM, Zamorano P, Stopper VS, Mills TM.Androgenic regulation of NO availability in ratpenile erection. J Androl 1997;18:110–5.

52 Schirar A, Bonnefond C, Meusnier C, Devinoy E.Androgens modulate nitric oxide synthase messen-ger ribonucleic acid expression in neurons of themajor pelvic ganglion in the rat. Endocrinology1997;138:3093–102.

53 Park KH, Kim SW, Kim KD, Paick JS. Effects ofandrogens on the expression of nitric oxide synthasemRNAs in rat corpus cavernosum. BJU Int1999;83:327–33.

54 Reilly CM, Stopper VS, Mills TM. Androgensmodulate the alpha-adrenergic responsiveness ofvascular smooth muscle in the corpus cavernosum.J Androl 1997;18:26–31.

55 Reilly CM, Lewis RW, Stopper VS, Mills TM.Androgenic maintenance of the rat erectile responsevia a non-nitric-oxide-dependent pathway. J Androl1997;18:588–94.

56 Traish AM, Park K, Dhir V, Kim NN, MorelandRB, Goldstein I. Effects of castration and androgenreplacement on erectile function in a rabbit model.Endocrinology 1999;140:1861–8.

57 Morelli A, Filippi S, Mancina R, Luconi M,Vignozzi L, Marini M, Orlando C, Vannelli GB,Aversa A, Natali A, Forti G, Giorgi M, Jannini EA,Ledda F, Maggi M. Androgens regulate phosphod-iesterase type 5 expression and functional activity incorpora cavernosa. Endocrinology 2004;145:2253–63.

58 Zhang XH, Morelli A, Luconi M, Vignozzi L, Fil-ippi S, Marini M, Vannelli GB, Mancina R, Forti G,Maggi M. Testosterone regulates PDE5 expression

Testosterone in Erectile Function 769

J Sex Med 2005;2:759–770

and in vivo responsiveness to tadalafil in rat corpuscavernosum. Eur Urol 2005;47:409–16.

59 Traish AM, Munarriz R, O’Connell L, Choi S, KimSW, Kim NN, Huang YH, Goldstein I. Effects ofmedical and surgical castration on erectile functionin an animal model. J Androl 2003;24:381–7.

60 Saenz de Tejada I. Molecular mechanisms for theregulation of penile smooth muscle contractility. IntJ Impot Res 2002;14(Suppl. 1):S6–10.

61 Kimura K, Markowski M, Bowen C, GelmannEP. Androgen blocks apoptosis of hormone-dependent prostate cancer cells. Cancer Res 2001;61:5611–8.

62 Sun M, Yang L, Feldman RI, Sun XM, Bhalla KN,Jove R, Nicosia SV, Cheng JQ. Activation of phos-phatidylinositol 3-kinase/Akt pathway by androgenthrough interaction of p85alpha, androgen receptor,and Src. J Biol Chem 2003;278:42992–3000.

63 Hengartner MO. The biochemistry of apoptosis.Nature 2000;407:770–6.

64 Zimmermann KC, Bonzon C, Green DR. Themachinery of programmed cell death. PharmacolTher 2001;92:57–70.

65 Moreland RB, Traish A, McMillin MA, Smith B,Goldstein I, Saenz de Tejada I. PGE1 suppresses theinduction of collagen synthesis by transforminggrowth factor-beta 1 in human corpus cavernosumsmooth muscle. J Urol 1995;153:826–34.

66 Raviv G, Kiss R, Vanegas JP, Petein M, Danguy A,Schulman C, Wespes E. Objective measurement ofthe different collagen types in the corpus caverno-sum of potent and impotent men: An immunohis-tochemical staining with computerized-imageanalysis. World J Urol 1997;15:50–5.

67 Luangkhot R, Rutchik S, Agarwal V, Puglia K,Bhargava G, Melman A. Collagen alterations in thecorpus cavernosum of men with sexual dysfunction.J Urol 1992;148:467–71.

68 Udelson D, Nehra A, Hatzichristou DG, AzadzoiK, Moreland RB, Krane J, Saenz de Tejada I,Goldstein I. Engineering analysis of penile hemo-dynamic and structural-dynamic relationships: PartI—Clinical implications of penile tissue mechanicalproperties. Int J Impot Res 1998;10:15–24.

69 Udelson D, Nehra A, Hatzichristou DG, AzadzoiK, Moreland RB, Krane RJ, Saenz de Tejada I,Goldstein I. Engineering analysis of penile hemo-dynamic and structural-dynamic relationships: PartII—Clinical implications of penile buckling. Int JImpot Res 1998;10:25–35.

70 Udelson D, Nehra A, Hatzichristou DG, AzadzoiK, Moreland RB, Krane RJ, Saenz de Tejada I,Goldstein I. Engineering analysis of penile hemo-dynamic and structural-dynamic relationships: PartIII—Clinical considerations of penile hemodynamicand rigidity erectile responses. Int J Impot Res1998;10:89–99.

71 Persson C, Diederichs W, Lue TF, Yen TSB,Fishman IJ, Mclin P, Tanagho EA. Correlation

of altered penile ultrastructure with clinical arterialevaluation. J Urol 1989;142:1462–8.

72 Wespes E, Goes PM, Schiffmann S, Deprierreux M,Vanderhaeghen JJ, Schulman CC. Computerizedanalysis of smooth muscle fibers in potent andimpotent patients. J Urol 1991;146:1015–7.

73 Jevtich M, Khawand NY, Vidic B. Clinical sig-nificance of ultrastructural findings in the corpuscavernosa of normal and impotent men. J Urol1990;143:289–93.

74 Nehra A, Goldstein I, Pabby A, Nugent M, HuangYH, de las Morenas A, Krane RJ, Udelson D, Saenzde Tejada I, Moreland RB. Mechanisms of venousleakage: A prospective clinicopathological correla-tion of corporeal function and structure. J Urol1996;156:1320–9.

75 Wespes E, Schiffmann S, Depierreux M, Vander-haegan JJ, Schulman CC. Is cavernovenous leakagerelated to a reduction of intracavernous smoothmuscle fibers? Int J Impot Res 1990;2:30.

76 Moussad EE, Brigstock DR. Connective tissuegrowth factor: What’s in a name? Mol Genet Metab2000;71:276–92.

77 Rageh MA, Moussad EE, Wilson AK, BrigstockDR. Steroidal regulation of connective tissuegrowth factor (CCN2; CTGF) synthesis in themouse uterus. Mol Pathol 2001;54:338–46.

78 Hong HH, Uzel MI, Duan C, Sheff MC, TrackmanPC. Regulation of lysyl oxidase, collagen, and con-nective tissue growth factor by TGF-beta1 anddetection in human gingiva. Lab Invest 1999;79:1655–67.

79 Ihn H. Pathogenesis fibrosis role TGF-beta andCTGF. Curr Opin Rheumatol 2002;14:681–5.

80 Blom IE, Goldschmeding R, Leask A. Gene regu-lation of connective tissue growth factor: Newtargets for antifibrotic therapy? Matrix Biol 2002;21:473–82.

81 Hashimoto G, Inoki I, Fujii Y, Aoki T, Ikeda E,Okada Y. Matrix metalloproteinases cleave connec-tive tissue growth factor and reactivate angiogenicactivity of vascular endothelial growth factor 165. JBiol Chem 2002;277:36288–95.

82 Inoki I, Shiomi T, Hashimoto G, Enomoto H,Nakamura H, Makino K, Ikeda E, Takata S,Kobayashi K, Okada Y. Connective tissue growthfactor binds vascular endothelial growth factor(VEGF) and inhibits VEGF-induced angiogenesis.FASEB J 2002;16:219–21.

83 Grotendorst GR. Connective tissue growth factor:A mediator of TGF-beta action on fibroblasts.Cytokine Growth Factor Rev 1997;8:171–9.

84 Traish AM, Toselli P, Jeong SJ, Kim NN. Adipocyteaccumulation in penile corpus cavernosum of theorchiectomized rabbit: A potential mechanism forveno-occlusive dysfunction in androgen deficiency.J Androl 2005;26:242–8.

85 Moon DG, Sung DJ, Kim YS, Cheon J, Kim JJ.Bisphenol A inhibits penile erection via alteration

770 Traish and Kim

J Sex Med 2005;2:759–770

of histology in the rabbit. Int J Impot Res2001;13:309–16.

86 Moon DG, Lee KC, Kim YW, Park HS, Cho HY,Kim JJ. Effect of TCDD on corpus cavernosumhistology and smooth muscle physiology. Int JImpot Res 2004;16:224–30.

87 Goyal HO, Braden TD, Williams CS, Dalvi P,Williams JW, Srivastava KK. Exposure of neonatalmale rats to estrogen induces abnormal morphologyof the penis and loss of fertility. Reprod Toxicol2004;18:265–74.

88 Goyal HO, Braden TD, Williams CS, Dalvi P,Mansour MM, Mansour M, Williams JW, BartolFF, Wiley AA, Birch L, Prins GS. Abnormal mor-phology of the penis in male rats exposed neonatallyto diethylstilbestrol is associated with altered profileof estrogen receptor-alpha protein, but not ofandrogen receptor protein: A developmental andimmunocytochemical study. Biol Reprod 2004;70:1504–17.

89 Bhasin S, Taylor WE, Singh R, Artaza J, Sinha-Hikim I, Jasuja R, Choi H, Gonzalez-Cadavid NF.The mechanisms of androgen effects on bodycomposition: Mesenchymal pluripotent cell as thetarget of androgen action. J Gerontol 2003;58A:1103–10.

90 Singh R, Artaza JN, Taylor WE, Gonzalez-CadavidNF, Bhasin S. Androgens stimulate myogenic dif-ferentiation and inhibit adipogenesis in C3H 10T1/2 pluripotent cells through an androgen receptor-mediated pathway. Endocrinology 2003;144:5081–8.

91 Corradi LS, Goes RM, Carvalho HF, Taboga SR.Inhibition of 5a-reductase activity induces stromalremodeling and smooth muscle de-differentiationin adult gerbil ventral prostate. Differentiation2004;72:198–208.

92 Rucker-Martin C, Pecker F, Godreau D, HatemSN. Dedifferentiation of atrial myocytes duringatrial fibrillation: Role of fibroblast proliferation invitro. Cardiovasc Res 2002;55:38–52.

93 Johnson JL, van Eys GJ, Angelini GD, George SJ.Injury induces dedifferentiation of smooth musclecells and increased matrix-degrading metallopro-teinase activity in human saphenous vein. Arterio-scler Thromb Vasc Biol 2001;21:1146–51.

94 Gray PB, Singh AB, Woodhouse LJ, Storer TW,Casaburi R, Dzekov J, Dzekov C, Sinha-Hikim I,Bhasin S. Dose-dependent effects of testosterone onsexual function, mood and visuospatial cognition inolder men. J Clin Endocrinol Metab 2005;doi:10.1210/jc.2005-0247.