jou rn al hom epage: www.elsev ier .com/ locate /neule t
ndrogen regulates development of the sexually dimorphicastrin-releasing peptide neuron system in the lumbar spinal cord:vidence from a mouse line lacking androgen receptor in theervous system
Ushimado Marine Institute (UMI), Graduate School of Natural Science and Technology, Okayama University, Ushimado, Setouchi, Okayama 701-4303,apanCentre National de la Recherche Scientifique (CNRS), Unité Mixte de Recherche (UMR) 7224, 9 quai St Bernard, 75005 Paris Cedex 05, FranceInstitut National de la Santé et de la Recherche Médicale (INSERM) U952, 9 quai St Bernard, 75005 Paris Cedex 05, FranceUniversité Pierre et Marie Curie, 75005 Paris Cedex 05, France
i g h l i g h t s
Androgen regulates masculine features primarily through androgen receptors (ARs).The gastrin-releasing peptide (GRP) system also mediates male sexual behavior.Testosterone may differentiate and maintain the activation of the spinal GRP system.We show that AR deletion attenuates GRP neurons and potentially sexual behavior in males.
r t i c l e i n f o
rticle history:eceived 10 October 2013eceived in revised form 27 October 2013ccepted 30 October 2013
Androgens including testosterone, organize the nervous system as well as masculine external and inter-nal genitalia during the perinatal period. Androgen organization involves promotion of masculine bodyfeatures, usually by acting through androgen receptors (ARs). We have recently demonstrated that thegastrin-releasing peptide (GRP) system in the lumbar spinal cord also mediates spinal centers promot-ing penile reflexes during male sexual behavior in rats. Testosterone may induce sexual differentiationof this spinal GRP system during development and maintain its activation in adulthood. In the presentstudy, we examined the role of ARs in the nervous system regulating the development of the sexuallydimorphic GRP system. For this purpose, we used a conditional mouse line selectively lacking the ARgene in the nervous system. AR floxed males carrying (mutants) or not (controls) the nestin-Cre trans-gene were castrated in adulthood and supplemented with physiological amounts of testosterone. Loss
utant mouse of AR expression in the nervous system resulted in a significant decrease in the number of GRP neuronscompared to control littermates. Consequently, the intensity of GRP axonal projections onto the lowerlumbar and upper sacral spinal cord was greater in control males than in mutant males. These resultssuggest that ARs expressed in the nervous system play a significant role in the development of the GRPsystem in the male lumbar spinal cord. The AR-deletion mutation may attenuate sexual behavior and
Early in life, androgens such as testosterone (T) induce exter-nal and internal genitalia to develop into a masculine form andalso masculinize the developing nervous system. This results in a
permanent organization of neural populations and synaptic con-nections underlying male sexual behavior [8,10,13]. In adulthood,T acts to activate these neural areas including brain regions (e.g.,medial amygdala, bed nucleus of stria terminalis, and medial
reoptic area) and spinal nuclei. In the spinal cord, androgen orga-ization and activation of neural populations usually occurs bycting through androgen receptors (ARs) [8,10]. However, whetherhese processes involve the peripheral or central AR remain to berecised. We have recently demonstrated that the gastrin-releasingeptide (GRP) system in the lumbar spinal cord also mediates spinalenters promoting penile reflexes during male sexual behavior inats [17,19]. We have also reported androgenic effects on the GRPystem in the lumbar spinal cord in two animal models, one involv-ng castration and T replacement in adult male rats and the secondsing genetically XY male rats carrying the testicular feminizationTfm) allele of the AR gene [20]. Both animal models indicate thatndrogen signaling plays a pivotal role in the development of thepinal GRP system and in the regulation of GRP expression in theale lumbar spinal cord. However, in these animal models, it is
ifficult to distinguish between central and peripheral effects ofndrogens, including behavioral modulations. In the present study,e therefore used a conditional mouse line selectively lacking theR gene in the nervous system [14] to elucidate whether centralR is important for the sex differentiation of the spinal GRP sys-
em, which plays a crucial role in male sexual behavior [19]. Usinghis mouse model in the study of the spinal GRP system is of greatnterest because this conditional mutation in mice interferes withhe expression of male sexual behavior [14].
. Materials and methods
.1. Animals and genotyping
Control males with a floxed AR allele (ARfl/Y) and their mutantittermates (ARfl/Y, Nes-Cre; ARNesCre) expressing Cre recombinasender the control of the promoter and the nervous system enhancerf nestin (Nes) gene were obtained as previously described [14]. TheRNesCre mouse line was then backcrossed for at least nine genera-
ions into strain C57BL6J. Control and mutant males obtained in theame litters were weaned at 24–26 days of age and group housednder a controlled photoperiod (12 h light, 12 h dark cycle; lightsn at 07:00 h) and temperature (22 ◦C) and given free access toood and water. The Cre transgene and floxed or excised AR allelesere identified by PCR analysis as previously described [4,14]. All
tudies were performed in accordance with the guidelines for carend use of laboratory animals [National Institutes of Health (NIH)uide] and French and European legal requirements (Decree 87-48, 86/609/ECC). The experimental procedures have also beenuthorized by the Committee for Animal Research, Okayama Uni-ersity, Japan.
.2. Gonadectomy and T supplementation
Adult males (8–10 weeks of age) of control (ARfl/Y) and ARNesCre
ere castrated under general anesthesia (xylazine/ketamine). Athe time of castration, all males received 1 cm subcutaneous SILAS-IC implants (3.18 mm outer diameter × 1.98 mm inner diameter;ow Corning, Midland, MI, USA), either empty or filled with 10 mg T
Sigma–Aldrich, St.-Quentin Fallavier, France) and sealed at eachnd with SILASTIC adhesive. Animals were killed 3–4 weeks later.irculating levels of T under these conditions were previouslyublished elsewhere [7,14,15]. Circulating T was in the highesthysiological range and similar comparable between controls andutants treated with T.
.3. Immunohistochemistry and immunofluorescence
Mice were anesthetized and transcardially perfused withhysiological saline followed by 4% paraformaldehyde in 0.1 Mhosphate buffer (PB) (pH 7.4). Lumbar spinal cords were quickly
etters 558 (2014) 109– 114
removed and immersed in the same fixative for 4 h at room tem-perature. After immersion in 25% sucrose in 0.1 M PB at 4 C forcryoprotection until they sank, the preparations were quicklyfrozen using powdered dry ice and cut into 30 �m-thick crossor horizontal sections on a cryostat (CM3050 S, Leica, Nussloch,Germany). We performed immunohistochemical analysis accord-ing to our established methods [18–20]. In brief, endogenousperoxidase activity was eliminated from the sections by incuba-tion in a 1% H2O2 absolute methanol solution for 30 min followedby three 5-min rinses with phosphate buffered saline (PBS) (pH7.4). These processes were omitted for the immunofluorescencemethod. After blocking nonspecific binding components with 1%normal goat serum and 1% BSA in PBS containing 0.3% Triton X-100 for 1 h at room temperature, sections were incubated withprimary rabbit antiserum against GRP (1:5000) (Phoenix Pharma-ceuticals, Burlingame, CA, USA) for 48 h at 4 ◦C. Immunoreactive(ir) products were detected with a streptavidin-biotin kit (Nichirei,Tokyo, Japan), followed by diaminobenzidine development accord-ing to our previous method [19]. The specificity of the antiserumwas published previously [19]. GRP-ir cells in the spinal cord werelocalized using an Olympus Optical (Tokyo, Japan) BH-2 micro-scope. To determine the effect of central ARs on the projection siteof GRP-ir axons, double-immunofluorescence staining of GRP andneuronal nitric oxide synthase (nNOS) (A-11; mouse monoclonalantibody, Santa Cruz Biotechnology, Santa Cruz, CA, USA) (1:5000dilution), a marker protein for neurons in the sacral parasympa-thetic nucleus (SPN), was performed as described previously [19].Alexa Fluor 546-linked anti-mouse IgG (Molecular Probes, Eugene,OR, USA) and Alexa Fluor 488-linked anti-rabbit IgG, both raised ingoats (Molecular Probes), were used at a 1:1000 dilution for detec-tion. Immunostained sections were imaged with a confocal laserscanning microscopy (Fluoview 1000, Olympus, Tokyo, Japan). Forquantitative analysis, GRP-ir cells with clearly visible round nuclearprofiles were counted in the anterior part of the lumbar spinal cord(L3–L4 level). To determine the optical density (OD) of positive GRP-ir fibers in the SPN, dorsal gray commissure (DGC) and dorsal horn(DH), at least six sections per animal were analyzed using ImageJsoftware (ImageJ 1.36b) with a set threshold level. The GRP-ir fiberpixel density was quantified as the average pixel density in threeregions of each animal (SPN, DGC and DH), and were calculated asthe ratio to the density seen in the DH in control males. At least fiveanimals were used for these analyses in each group. The numbersand optical density of the GRP-ir neurons in the lumbar spinal cordwere expressed as the mean ± standard error of the mean (S.E.M.)in each group, and were analyzed by a Student’s t-test. P < 0.05 wasconsidered statistically significant.
3. Results
Using immunohistochemical staining for GRP, we first exam-ined the number of GRP-ir neurons located in the lumbar spinalcord (L3–L4 level; containing somata and dendrites of the GRPneurons) of control and ARNesCre male mice (Fig. 1). The numbersof GRP-positive somata in the lumbar spinal cord of mice were277.3 ± 14.1 and 134.0 ± 10.4 in the control and ARNesCre males,respectively (Fig. 1). The number of GRP-ir neurons was signifi-cantly fewer in ARNesCre males than in control males in the upperlumbar spinal cord (n = 5, t = 8.24, P < 0.05) (Fig. 1). Because all themice in this study were castrated and received long-term (3–4weeks) T supplementation, levels of circulating T were maintainedat a high physiological range, comparable to that between controls
and mutants as previously reported [7,14,15]. Then, we exam-ined the projection of GRP-ir fibers to the SPN in the lower spinalcord (L5–L6 and S1 level; containing GRP neuronal axon termi-nals) of male mice because SPN provides autonomic preganglionic
H. Sakamoto et al. / Neuroscience Letters 558 (2014) 109– 114 111
Fig. 1. Immunohistochemical analysis for gastrin-releasing peptide (GRP) in the upper lumbar spinal cord (L3–L4 level) of adult male mice. GRP-immunoreactive neuronswere found in the upper lumbar spinal cord of control males (A) and mutant males, selectively lacking the androgen receptor (AR) gene in the nervous system (ARNesCre) (B).The number of GRP-immunoreactive neurons was fewer in ARNesCre males than in control males in the upper lumbar spinal cord. Arrows indicate possible GRP-positive cellb
fiatafccafiats
4
tsiiasbtcmTm[m[c[asnnImtgGP
odies. Dashed lines indicate the level of the mid-line. Scale bar = 50 �m.
bers to the genitalia [9]. Double immunofluorescence for GRPnd nNOS clearly showed that GRP-ir fibers projected densely intohe SPN (Fig. 2). Consistent with the difference in the GRP somarea (L3–L4 level), double immunofluorescence for GRP and nNOSurther revealed that the intensity of GRP-ir fibers was greater inontrol than in ARNesCre males in the autonomic center SPN whereashanges were observed for nNOS-ir intensity (Fig. 2). Quantificationnalysis of GRP-ir clearly confirmed that the intensity of GRP-irbers in the lower lumbar and upper sacral spinal cord (L5–L6nd S1 level) was greater in control than in ARNesCre males in bothhe autonomic centers of the SPN and DGC but not in the somaticensory layers of the DH (n = 5 each) (Fig. 3).
. Discussion
The goal of the present study was to determine whether cen-ral AR expression directly regulates the development of the GRPystem in the lumbar spinal cord because such a spinal regions involved in the androgenic modulation of male specific behav-ors [17,19]. There is substantial evidence of androgenic effects onggressive and sexual behaviors, although direct effects of the ARignaling on the GRP system in the lumbar spinal cord have noteen fully investigated. Therefore, the present study focused onhe effects of central AR deletion on the spinal GRP system, whichontrols male copulatory behavior [18–20] by using geneticallyodified mutant mice lacking AR in the nervous system [7,14,15].
he male urogenital tract of ARNesCre mutant develops normally andice are able to produce offspring, although with reduced fertility
14,15]. Regardless of attenuated sexual motivation and perfor-ance, ARNesCre males exhibited a sexual behavior against females
14]. ARNesCre males also exhibit a very low aggressive behaviorompared to their control littermates as we have previously shown7,14]. During mating, the mean length of grooming erection islso significantly reduced in mutant males [14,15]. In the lumbarpinal cord of ARNesCre males, we found a significant decrease in theumber of GRP neurons compared to control littermates, despiteormalization of circulating levels of T between the two genotypes.
n addition, GRP-ir fibers were much more prominent in controlales than in ARNesCre males in both the SPN and DGC surrounding
he central canal. Both the SPN and DGC provide autonomic pregan-lionic fibers to genitalia [9]. Fig. 2 showed many overlaps betweenRP-ir fibers and SPN neuron somata with the proximal dendrites.reviously, using double immunoelectron microscopy, we have
demonstrated that GRP-containing presynaptic boutons innervatenNOS-positive dendrites in the SPN in rats [12,19]. Therefore, theseoverlaps might be possible sites of synaptic contacts also in mice. Onthe other hand, orchiectomy of adult male rats significantly reducedthe expression of GRP in the lumbar spinal cord after 1 month andthis reduction was prevented by long-term androgen replacementtreatment [19,20]. Conversely, long-term T treatment of castratedadult female rats did not fully masculinize GRP expression [19],suggesting that, at least in rats, androgens exert organizationaleffects on GRP neuronal cell number only during the perinatal criti-cal period [20]. These GRP-expressing neurons in the lumbar spinalcord of males also express AR but do not express estrogen receptora in rats [19] and mice (Tamura et al., our unpublished observation).Collectively, a male-dominant sexual dimorphism in the spinal GRPsystem may be developed through AR-mediated mechanisms bythe androgen surge during the critical period. This male specificGRP neuronal cell number is perhaps maintained with considerablylow plasticity in adulthood. However, we cannot rule out the pos-sibility that not only the number of GRP neurons but also ir of GRPin the GRP neurons are decreased in ARNesCre males in adulthood.
The Tfm rodent is a unique model for examining the role of ARsin the central nervous system because a point mutation in the ARgene renders the protein dysfunctional [23]. Because, these dys-functional ARs in Tfm models are expressed in the whole bodyincluding the brain, spinal cord, skeletal muscles, reproductiveorgans and so on [1,2], we could not distinguish tissue-specific ARfunctions. Additionally, rat and mouse Tfm models differ in termsof circulating T levels [23]. Tfm male mice have significantly lessendogenous T compared to their wild-type littermates [6,22], whileTfm male rats and ARNesCre males have circulating T levels in thehigh male range [14–16,20]. We have recently reported that thespinal cord of Tfm male rats is hyperfeminine in terms of the num-ber of GRP-expressing neurons in the upper lumbar region (L3–L4level), although circulating T levels tended to be high [19,20]. In thepresent study, we found a similar feminization in the spinal GRPsystem of ARNesCre male mice. Taken together, the present resultsfrom ARNesCre mice further imply that ARs expressed in the periph-eral muscles and/or reproductive organs could not directly opposethe organizational effects on the sexually dimorphic GRP system in
the lumbar spinal cord.
In rodents, the spinal nucleus of the bulbocavernosus (SNB) islocated in the lower lumbar and upper sacral spinal cord. It is alsoa sexually dimorphic nucleus that innervates the perineal muscles
112 H. Sakamoto et al. / Neuroscience Letters 558 (2014) 109– 114
Fig. 2. Distribution of GRP-immunoreactivity in the lower lumbar and upper sacral spinal cord (L5–L6 and S1 level) of adult male mice. In the sacral parasympathetic nucleus(SPN) located in the lower lumbar and upper sacral spinal cord, GRP-immunoreactive fibers (green) were observed both in control (A) and ARNesCre (B) males. ARNesCre malespresent decreased GRP-immunoreactive fibers compared with control littermates in this region. The neuronal nitric oxide synthase (nNOS) serves as a marker for SPNneurons, and the immunoreactivity for nNOS in the SPN neurons (magenta) showed no difference between control (C) and ARNesCre (D) males. Double immunofluorescencef (E) ann in this
iarAnPrrtre[Attt
or GRP and nNOS revealed closely appositions of GRP-containing fibers of controleurons in the SPN. Scale bar = 50 �m. (For interpretation of the references to color
nvolved in male copulatory behavior [3,5,21]. Male rats have morend larger SNB motoneurons than do females, a dimorphism thatesults from differences in perinatal androgen signaling through anR-mediated mechanism [3,21]. The sex-related difference in theumber of SNB motoneurons develops perinatally in rats and mice.rior to birth, the number of motoneurons in the SNB increases andeaches a maximum in both sexes; at this point, functional neu-omuscular junctions have been established in the SNB system athis time. However, in females, these components (both motoneu-ons and muscles) die near the time of birth unless the animals arexposed to testosterone during the critical period (androgen surge)11]. If an androgen surge occurs, it results in higher expression of
R in the perineal muscles and spinal motoneurons. Interestingly,
he number of SNB motoneurons is unrelated to ARNesCre muta-ion status and adult changes in T levels [15], although we foundhat the ARNesCre mutation caused a significant reduction in the
d ARNesCre (F) males with the cell bodies and proximal dendrites of nNOS-positive figure legend, the reader is referred to the web version of the article.)
number of the spinal GRP neurons in adult males. In the SNB, thecentral AR participates in the developmental regulation of somasize and dendritic length but not the survival of SNB motoneurons.The AR immunohistochemistry in the lumbosacral spinal cord alsodemonstrated the expression of AR in the cellular nuclei of SNBmotoneurons in controls but not in ARNesCre males [15]. It is sug-gested that AR could be differently involved in the organizationof cell populations in the spinal cord. Recently, we have shownthat high-voltage electron microscopy (HVEM) analysis provides aclear three-dimensional visualization of multiple synaptic contactsfrom the GRP system in the lumbar spinal cord onto the dendritesof SNB motoneurons in rats [18]. Further investigations of the AR
dependent male specific organizational effects on these GRP andSNB systems located both in the lower spinal cord and the physi-ological approaches for their functional correlations are needed todraw firm conclusions.
H. Sakamoto et al. / Neuroscience L
Fig. 3. Quantification analysis for the intensity of GRP-immunoreactive fibers inthe lower lumbar and upper sacral spinal cord (L5-L6 and S1 level) in control andARNesCre males. (A) A schematic drawing of the lower lumbar and upper sacral spinalcord, indicating the intra-spinal location of the SPN (A1), dorsal gray commissure(DGC) (A2) and the dorsal horn (DH) (A3) in this spinal cord level. The blockedareas are enlarged in (B). These higher magnification images show the distributionof GRP-immunoreactive fibers in the SPN (B1), DGC (B2) and DH (B3), respectively.The lower panels in (B) show the black-and-white images with a set threshold levelcorresponding to each upper panel. The optical density of GRP fibers in the DGC, SPNand DH of control males versus ARNesCre males were analyzed in (C). The intensity ofGRP-immunoreactive fibers in the lower lumbar and upper sacral spinal cord wasgreater in control than in ARNesCre males both in the autonomic centers of the DGCand SPN but not in the somatic sensory layers of the DH. The optical density of theGRP-ir in the lumbar spinal cord was expressed as the mean ± standard error oftm
5
iocudaam
C
[
[
[
[
[
[
[
[
[
[
11 (2008) 634–636.
he mean (S.E.M.) in each group. CC; central canal. **P < 0.01 compared with controlales (n = 5 mice in each group). Scale bars = 100 �m.
. Conclusions
Here we show, for the first time in mice, that ARs expressedn the nervous system play a significant role in the developmentf the GRP system in the male lumbar spinal cord. Thus, theentral but not peripheral ARs are critical for developmental reg-lation and maintenance of the male dominant cell number andendritic arborizations in the spinal GRP system in rodents. Alter-tion by the ARNesCre mutation may attenuate sexual behavior andctivity of mutant males via spinal GRP system-mediated neuralechanisms.
onflict of interest
The authors declare no potential conflicts of interest.
[
etters 558 (2014) 109– 114 113
Acknowledgements
This work was supported in part by KAKENHI from the Min-istry of Education, Science, Sports, Culture and Technology (MEXT),Japan (to H.S.), by the Research Grant from the Senri Life ScienceFoundation, Japan (to H.S.), by the French Agence Nationale de laRecherche (to S. M.-K.) and by the Réseau Santé EnvironnementToxicologie of the Région Ile de France (to S. M.-K.). Takumi Oti issupported by a Research Fellowship of the Japan Society for the Pro-motion of Science (JSPS) for Young Scientists. We thank Prof. GuidoVerhoeven (Laboratory of Experimental Medicine and Endocrinol-ogy, Katholieke Universiteit Leuven, Leuven, Belgium) for providingthe floxed AR mouse line.
References
[1] C.W. Bardin, L. Bullock, W.R. Blackburn, R.J. Sherins, T. Vanha-Perttula, Testos-terone metabolism in the androgen-insensitive rat: a model for testicularfeminization, Birth Defects Orig. Artic. Ser. 7 (1971) 185–192.
[2] F.A. Beach, M.G. Buehler, Male rats with inherited insensitivity to androgenshow reduced sexual behavior, Endocrinology 100 (1977) 197–200.
[3] S.M. Breedlove, A.P. Arnold, Hormone accumulation in a sexually dimorphicmotor nucleus of the rat spinal cord, Science 210 (1980) 564–566.
[4] K. De Gendt, J.V. Swinnen, P.T. Saunders, L. Schoonjans, M. Dewerchin, A. Devos,K. Tan, N. Atanassova, F. Claessens, C. Lecureuil, W. Heyns, P. Carmeliet, F. Guil-lou, R.M. Sharpe, G. Verhoeven, A Sertoli cell-selective knockout of the androgenreceptor causes spermatogenic arrest in meiosis, Proc. Natl. Acad. Sci. U.S.A. 101(2004) 1327–1332.
[5] N.G. Forger, S.M. Breedlove, Sexual dimorphism in human and canine spinalcord: role of early androgen, Proc. Natl. Acad. Sci. U.S.A. 83 (1986) 7527–7531.
[6] R.D. Jones, P.J. Pugh, J. Hall, K.S. Channer, T.H. Jones, Altered circulating hormonelevels, endothelial function and vascular reactivity in the testicular feminisedmouse, Eur. J. Endocrinol./Eur. Federation Endocr. Soc. 148 (2003) 111–120.
[7] C. Marie-Luce, K. Raskin, M. Bolborea, M. Monin, M. Picot, S. Mhaouty-Kodja,Effects of neural androgen receptor disruption on aggressive behavior, argininevasopressin and galanin systems in the bed nucleus of stria terminalis andlateral septum, Gen. Comp. Endocrinol. 188 (2013) 218–225.
[8] K.I. Matsuda, H. Sakamoto, M. Kawata, Androgen action in the brain and spinalcord for the regulation of male sexual behaviors, Curr. Opin. Pharmacol. 8 (2008)747–751.
[9] C. Morgan, I. Nadelhaft, W.C. de Groat, The distribution of visceral primaryafferents from the pelvic nerve to Lissauer’s tract and the spinal gray matterand its relationship to the sacral parasympathetic nucleus, J. Comp. Neurol. 201(1981) 415–440.
10] J.A. Morris, C.L. Jordan, S.M. Breedlove, Sexual differentiation of the vertebratenervous system, Nat. Neurosci. 7 (2004) 1034–1039.
11] E.J. Nordeen, K.W. Nordeen, D.R. Sengelaub, A.P. Arnold, Androgens preventnormally occurring cell death in a sexually dimorphic spinal nucleus, Science229 (1985) 671–673.
12] T. Oti, K. Satoh, K. Saito, K. Murata, M. Kawata, T. Sakamoto, H. Sakamoto,Three-dimensional evaluation of the spinal local neural network revealed bythe high-voltage electron microscopy: a double immunohistochemical study,Histochem. Cell Biol. 138 (2012) 693–697.
13] C.H. Phoenix, R.W. Goy, A.A. Gerall, W.C. Young, Organizing action of prena-tally administered testosterone propionate on the tissues mediating matingbehavior in the female guinea pig, Endocrinology 65 (1959) 369–382.
14] K. Raskin, K. de Gendt, A. Duittoz, P. Liere, G. Verhoeven, F. Tronche, S. Mhaouty-Kodja, Conditional inactivation of androgen receptor gene in the nervoussystem: effects on male behavioral and neuroendocrine responses, J. Neurosci.29 (2009) 4461–4470.
15] K. Raskin, C. Marie-Luce, M. Picot, V. Bernard, P. Mailly, H. Hardin-Pouzet, F.Tronche, S. Mhaouty-Kodja, Characterization of the spinal nucleus of the bul-bocavernosus neuromuscular system in male mice lacking androgen receptorin the nervous system, Endocrinology 153 (2012) 3376–3385.
16] C.E. Roselli, R.L. Salisbury, J.A. Resko, Genetic evidence for androgen-dependentand independent control of aromatase activity in the rat brain, Endocrinology121 (1987) 2205–2210.
17] H. Sakamoto, Brain-spinal cord neural circuits controlling male sexual functionand behavior, Neurosci. Res. 72 (2012) 103–116.
18] H. Sakamoto, T. Arii, M. Kawata, High-voltage electron microscopy revealsdirect synaptic inputs from a spinal gastrin-releasing peptide system to neu-rons of the spinal nucleus of bulbocavernosus, Endocrinology 151 (2010)417–421.
19] H. Sakamoto, K.-I. Matsuda, D.G. Zuloaga, H. Hongu, E. Wada, K. Wada, C.L. Jor-dan, S.M. Breedlove, M. Kawata, Sexually dimorphic gastrin releasing peptidesystem in the spinal cord controls male reproductive functions, Nat. Neurosci.
20] H. Sakamoto, K. Takanami, D.G. Zuloaga, K.I. Matsuda, C.L. Jordan, S.M.Breedlove, M. Kawata, Androgen regulates the sexually dimorphic gastrin-releasing peptide system in the lumbar spinal cord that mediates male sexualfunction, Endocrinology 150 (2009) 3672–3679.
21] D.R. Sengelaub, N.G. Forger, The spinal nucleus of the bulbocavernosus:firsts in androgen-dependent neural sex differences, Horm. Behav. 53 (2008)596–612.
22] D.G. Zuloaga, J.A. Morris, C.L. Jordan, S.M. Breedlove, Mice with the testicularfeminization mutation demonstrate a role for androgen receptors in the
[
etters 558 (2014) 109– 114
regulation of anxiety-related behaviors and the hypothalamic-pituitary-adrenal axis, Horm. Behav. 54 (2008) 758–766.
23] D.G. Zuloaga, D.A. Puts, C.L. Jordan, S.M. Breedlove, The role of androgen recep-tors in the masculinization of brain and behavior: what we’ve learned from thetesticular feminization mutation, Horm. Behav. 53 (2008) 613–626.
