Anabolic steroids, testosterone-precursors and virilizing androgens induce distinct activation profiles of androgen responsive promoter constructs

Journal of Steroid Biochemistry & Molecular Biology 82 (2002) 269–275

Anabolic steroids, testosterone-precursors and virilizing androgens induce distinct activation profiles of androgen responsive promoter constructs夽 P.M. Holterhus∗ , S. Piefke1 , O. Hiort Department of Pediatrics, Medical University of Lübeck, Lübeck, Germany Received 8 April 2002; accepted 21 August 2002

Abstract Different androgens, e.g. virilizing androgens such as testosterone and its precursors as well as synthetic anabolic steroids, respectively, induce diverse biological effects. The molecular basis for this variety in biological actions, however, is not well understood. We hypothesized that this variability of actions may be due to steroid-specific target gene expression profiles following androgen receptor (AR)-activation. Therefore, we investigated androgen receptor dependent transactivation of three structurally different androgen responsive promoter constructs ((ARE)2 TATA-luc, MMTV-luc, GRE-OCT-luc) in co-transfected Chinese hamster ovary (CHO)-cells as an artificial model simulating different natural target genes. Three virilizing androgens (dihydrotestosterone, testosterone, methyltrienolone), three anabolic steroids (oxandrolone, stanozolol, nandrolone) and two testosterone-precursors of gonadal and adrenal origin (dehydroepiandrosterone, androstenedione) were used as ligands (0.001–100 nM). All steroids proved to be potent activators of the AR. Remarkably, anabolic steroids and testosterone-precursors showed characteristic promoter activation profiles distinct from virilizing androgens with significantly lower (ARE)2 TATA-luc activation. Hierarchical clustering based on similarity of activation profiles lead to a dendrogram with two major branches: first virilizing androgens, and second anabolics/testosterone-precursors. We conclude that steroid-specific differences in gene transcription profiles due to androgen receptor activation could contribute to differences in biological actions of androgens. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Androgen; Anabolic; Androgen receptor; Promoter; Activation profile; Testosterone; Dihydrotestosterone; Dehydroepiandrosterone; Androstenedione; Oxandrolone; Nandrolone; Methyltrienolone

1. Introduction Androgens induce biological effects via activation of the androgen receptor (AR) in androgen responsive target tissues. Following binding of the ligand to the AR, the activated androgen-AR-complex enters the nucleus and binds to hormone responsive elements in the promoter of androgen responsive target genes. Then, either transcriptional activation or repression occurs, resulting in expression of target genes which in turn is responsible for androgen mediated biological effects [1]. Interestingly, different androgens can show very diverse profiles of biological actions. Clinical entities such 夽 Data have been presented in part on the 40th, 6th Joint LWPES/ESPE Meeting (Lawson Wilkins Pediatric Endocrine Society/European Society for Pediatric Endocrinology), Montreal, Canada, 6–10 July, 2001 and on the 46th Annual Meeting of the German Society of Endocrinology (DGE), Göttingen, Germany, 27 February–2 March, 2002. ∗ Corresponding author. Tel.: +49-451-500-2595; fax: +49-451-500-2184. E-mail address: [email protected] (P.M. Holterhus). 1 This work is part of the medical thesis by S. Piefke.

as defects of androgen biosynthesis or overall resistance to androgenic steroids in androgen insensitivity prove in vivo examples for the biological potency of these steroids. Dihydrotestosterone is essential for the virilization of the external genitalia and the development of the prostate. This is clinically obvious in patients with 5␣-reductase-II deficiency as an example of severe under-masculinization of the external genitalia and limited prostate development despite unaffected synthesis of testosterone [2,3]. In contrast, 5␣-reductase-II deficient patients show well differentiated Wolffian derivates while the latter are usually absent in complete androgen insensitivity syndrome due to inactivating mutations of the AR [1]. This was seen as an indication for a particular role of testosterone for Wolffian duct development. Testosterone-precursors like dehydroepiandrosterone or androstenedione originate from either the adrenals or the gonads. Remarkably, 46,XY patients with 17␤-hydroxysteroid dehydrogenase deficiency and consequently increased serum levels of androstenedione but reduced concentrations of testosterone [4,5] may have nearly completely female external genitalia at birth associated with normal Wolffian duct development [6]. This suggests

0960-0760/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 0 - 0 7 6 0 ( 0 2 ) 0 0 2 2 0 - 0

270

P.M. Holterhus et al. / Journal of Steroid Biochemistry & Molecular Biology 82 (2002) 269–275

a limited biological relevance of testosterone-precursors in male sexual differentiation. Synthetic anabolic steroids are structurally related to natural androgens. They represent an intriguing biological model of steroid hormones because of the variable degree of artificial reduction of androgenic (virilizing) androgen actions [7]. Virilizing actions comprise male sexual differentiation, phallus growth, beard growth and deepening of the voice during puberty while anabolic effects are characterized by a positive nitrogen balance and the rise of body muscle mass [8]. The molecular background for the different biological profiles of natural androgens and also the separation of anabolic and virilizing actions in anabolic steroids is poorly understood. Differences in the metabolism in target tissues [9] or interactions with other steroid hormone receptors or intracellular binding sites than the AR [10–12] have been discussed. Only sparse data exist on activation of androgen regulated promoters due to testosterone-precursors and anabolic steroids, respectively [13,14]. We hypothesized the existence of distinct steroid-specific target gene transcription profiles following AR-activation. We tested this hypothesis in an artificially generated target gene model comprising three structurally different androgen responsive promoters. Our data suggest that steroid-specific modulation of AR-activated target gene expression profiles may be a relevant factor for differences in the biological effects of androgens. 2. Materials and methods 2.1. Hormones Methyltrienolone (R1881) (17␤-hydroxy-17␣-methyl-4, 9,11-estrotrien-3-one) was purchased from New England Nuclear (NEN, Boston, MA). Testosterone (17␤-hydroxy-4androsten-3-one), dihydrotestosterone (17␤-hydroxy-5␣-androstan-3-one), androstenedione (4-androstene-3,17-dione), dehydroepiandrosterone (5-androsten-3␤-ol-17-one), nandrolone (17␤-hydroxy-4-estren-3-one), oxandrolone (17␤hydroxy-17␣-methyl-2-oxa-5␣-androstan-3-one) and stanozolol (17␤-hydroxy-17␣-methylandrostano(3,2-c)pyrazole) were obtained from Sigma (St. Louis, MO). 2.2. Cell culture CHO-cells were cultured at 37 ◦ C and 5% CO2 in Dulbecco’s modified Eagle medium with the nutrient mix F12 (DMEM-F12, Gibco), 10% (v/v) fetal calf serum (FCS), and penicillin (200 IU/ml)/streptomycin (0.2 mg/ml). For transactivation studies 10% dextran/charcoal treated FCS was used. 2.3. Plasmids The expression plasmid for the human androgen receptor pSVAR0 was a gift from A.O. Brinkmann, Erasmus

University, Rotterdam [15]. Three androgen responsive reporter genes each encoding for firefly luciferase were used. They were selected because they were characterized by clear differences in the molecular structures of their androgen responsive promoter regions and thus could serve as a model for three structurally different natural AR target gene promoters. GRE-OCT-luc consists of a promoter with two consensus glucocorticoid response elements (GRE), the TATA-box and a SP1 site upstream of the luciferase gene. It was obtained from A.O. Brinkmann. (ARE)2 TATA-luc is characterized by two androgen response elements followed by the TATA-box preceding the luciferase gene. It was obtained from G. Jenster, also at Erasmus University, Rotterdam [16]. The MMTV-luc reporter gene was purchased form Organon (West Orange, NJ). The hormone responsive region of the more complex MMTV-promoter is well described [17]. In brief, it consists of four hormone responsive elements (one imperfect palindrome, one anomalous palindrome with only two separating base pairs between the half sites and two half palindromes) followed by an NF1 binding site, two octamer motifs and finally the TATA-box preceding the luciferase gene. Transfection efficiency was monitored by co-transfection of the constitutively expressed, SV40 driven pRLSV40 (Promega Corp., Madison, WI). 2.4. Transactivation studies For transient transfections, cells were seeded in 10 cm2 six-well multi-dishes. Transfections were performed by classical Ca2+ -phosphate precipitation method [18] with only minor changes as previously described [19]. Per well, 12.5 ng of AR expression plasmid pSVAR0 and either 0.5 ␮g MMTV-luc, or 0.25 ␮g (ARE)2 TATA-luc or 1.0 ␮g GRE-OCT-luc reporter plasmids were used. Half a nanogram of the constitutively active pRL-SV40 Renilla luciferase expression plasmid (Promega Corp., Madison, WI) were added to monitor transfection efficiency. Final DNA amount was adjusted to 5 ␮g plasmid DNA per well with pTZ19 plasmid. Activation of the androgen responsive luciferase reporter plasmids due to pSVAR0 was investigated by treating the cells 24 h after transfection, for another 24 h with 0.001, 0.01, 0.1, 1.0, 10.0 or 100.0 nM of respective steroids or with the vehicle ethanol alone. Three independent triplicate experiments were performed at different times. Firefly and Renilla luciferase activity were determined using the Dual Luciferase reporter gene kit (Promega Corp.). 2.5. Data analysis For means of comparability of experiments performed at different times and to enable later cluster analysis (see below), transactivation data were normalized. Fold induction of each of the three different reporter genes in a single experiment (ratio of luciferase activity (steroid-treated)/(ETOH-treated)) was defined to be 1 at 10 nM R1881. Induction values of the same reporter gene

P.M. Holterhus et al. / Journal of Steroid Biochemistry & Molecular Biology 82 (2002) 269–275

due to R1881 at other concentrations than 10 nM or due to any concentration of the other hormones assayed at the same time were expressed relatively to this value. We calculated the mean of the three independent triplicate experiments. Statistical significance of differences between reporter gene induction due to different steroids was calculated with Student’s t-test. To further analyze the transactivation data and to categorize them into steroid-dependent gene expression profiles, normalized reporter gene induction due to the eight different hormones were log2 -transformed and centered by the means for each of the six different hormone concentrations used. This has been done for each of the three different promoters. Subsequently, hierarchical clustering for the used steroids (centered average linkage clustering) was performed and clustered promoter activation data has been visualized by Tree View software [20].

3. Results In CHO-cells, transiently transfected with the AR expression plasmid pSVAR0 the expected 110/112 kDa band in Western blots indicating the expression of the AR was demonstrated. In contrast, in mock-transfected cells an AR-specific signal was absent (data not shown). None of the eight different steroid hormones used induced any reporter gene activation in the absence of pSVAR0 in CHO-cells (data not shown), indicating the specificity of the transactivation data as a measure of AR-dependent signaling. All the eight steroids showed a concentration dependent induction of the MMTV-luc reporter gene due to the AR (Fig. 1a). Activity due to the synthetic reference steroid R1881 reached a plateau at 1 nM with no further significant changes at higher concentrations. Overall activation due to dihydrotestosterone was lower than R1881 with a maximum at 100 nM. Androstenedione and dehydroepiandrosterone showed relatively low activation of the MMTV-promoter at 0.001 and 0.01 nM with higher activity induced by androstenedione than dehydroepiandrosterone. However, both of them gained considerable activity with increasing hormone concentrations reaching levels comparable with R1881 at 10 nM. Moreover, the anabolic steroids oxandrolone, nandrolone, and stanozolol were also potent activators of the AR at the MMTV-promoter. At concentrations ≥10 nM, oxandrolone and nandrolone even induced a similar activation level to R1881. Remarkably, the anabolic stanozolol exceeded R1881-induced activity at concentrations higher than 10 nM. The general dose-activation relationship of stanozolol was almost identical with that due to testosterone at the MMTV-promoter (Fig. 1a). While there was a similar concentration dependent induction of the GRE-OCT-luc reporter gene, the most striking feature of this promoter was the well-reproducible, extraordinary high activation due to testosterone at all hormone concentrations (Fig. 1b). Furthermore, dihydrotestosterone,

271

R1881, and the anabolic steroids nandrolone and stanozolol shared approximately the same activation profile across the whole range of hormone concentrations except for a lower activation due to dihydrotestosterone at 0.001 nM. Dehydroepiandrosterone and androstenedione as well as oxandrolone showed considerably lower activity than the other hormones at low concentrations of 0.001–0.1 nM. However, with increasing hormone concentrations up to pharmacological levels they gained substantial activity. At concentrations ≥10 nM, androstenedione, dehydroepiandrosterone, oxandrolone as well as dihydrotestosterone, R1881, nandrolone, and stanozolol all demonstrated about the same activation level (Fig. 1b). With (ARE)2 TATA, the “virilizing” androgens testosterone, dihydrotestosterone and their synthetic analogon R1881 showed the highest activities among all other androgens in a concentration dependent manner (Fig. 1c). Maximum activation was due to dihydrotestosterone at 10 nM. Interestingly, the anabolic steroid nandrolone shared its induction levels with these three androgens. In contrast, androstenedione and dehydroepiandrosterone, together with the anabolics oxandrolone and stanozolol showed a striking and statistically significant deficiency of AR-induced (ARE)2 TATA-luc activation compared with GRE-OCT-luc and MMTV-luc activation, predominantly at lower concentrations up to 0.1 nM (P < 0.001–0.05). At ligand concentrations of 10 nM and higher, they gained substantial activity. In case of the most selective anabolic steroid stanozolol, even under pharmacological concentrations up to 100 nM the activation of the (ARE)2 TATA-luc reporter gene remained low (Fig. 1c). 3.1. Analysis of promoter activation profiles The design of our study represents an artificial model imitating different natural androgen target genes in a hypothetical androgen target tissue. We presented a set of transactivation data comprising three androgen regulated promoters activated by eight hormones at six different concentrations measured in 1296 single transfections resulting in 144 final, normalized data points. Classical arrangement of the data in the form of histograms is neither sufficient alone nor intuitive to detect or present the underlying patterns of activation. Therefore, we added a strategy originally developed for the analysis of large scale gene expression data [20,21]. Fig. 2 shows an “Eisen-plot” with a hierarchical dendrogram of the used androgens based on the analysis of the similarity of reporter gene activation profiles across the promoters and across the concentrations. Our essential finding is the division of the steroids into two main subgroups by computed clustering: (1) virilizing androgens including nandrolone, (2) testosterone-precursor hormones together with the anabolic steroids oxandrolone and stanozolol. In the group of predominantly “virilizing” androgens, testosterone clusters at a separate branch which is due to the extraordinary high GRE-OCT-luc activation

272

P.M. Holterhus et al. / Journal of Steroid Biochemistry & Molecular Biology 82 (2002) 269–275

Fig. 1. Transactivation studies of androgen responsive luciferase reporter gene constructs: (a) relative fold induction of MMTV-luciferase reporter gene. Columns represent the means of three independent triplicate experiments. Error bars represent+1S.D.; (b) relative fold induction of GRE-OCT-luciferase reporter gene. Columns represent the means of three independent triplicate experiments. Error bars represent +1S.D.; (c) relative fold induction of (ARE)2 TATA-luciferase reporter gene. Columns represent the means of three independent triplicate experiments. Error bars represent+1S.D.

P.M. Holterhus et al. / Journal of Steroid Biochemistry & Molecular Biology 82 (2002) 269–275

273

Fig. 1. (Continued ).

and the relatively high MMTV-activity. The leading pattern of the data set, however, is the considerably low activation of the (ARE)2 TATA-promoter due to androstenedione, dehydroepiandrosterone, oxandrolone, and stanozolol. Stanozolol clusters at a separate branch reflecting its obvious capability as a more potent activator of MMTVand GRE-OCT-promoters than androstenedione, dehydroepiandrosterone, and oxandrolone.

4. Discussion We generated a model with three different artificial androgen responsive target genes in a virtual androgen responsive tissue. We clearly demonstrated that structurally different androgens with discrete biological actions are characterized by distinct promoter activation profiles upon activating the AR. Therefore, our experimental data suggest that steroid-specific modulation of target gene expression programs due to differential activation of the AR could be a relevant molecular mechanism contributing to the diversity of actions of structurally different androgens in vivo. The separation of the steroids by cluster analysis into two distinct subgroups based on their reporter gene activation profiles is of additional interest. Taking into account the limited value of artificial-target instead of natural-target

genes, our experimental data imply that similarity in reporter gene expression profiles could reflect to some extent a similarity in biological actions. Remarkably, the two virilizing androgens dihydrotestosterone and testosterone, and the high affinity ligand R1881 group together in the left main branch in the dendrogram indicating relatively similar reporter gene activation patterns (Fig. 2). Interestingly, testosterone often induced higher reporter gene induction than dihydrotestosterone which is clearly visible in the histograms (Fig. 1a and b) but also in the cluster diagram where it clusters separately (Fig. 2). This was most striking in the experiments with the GRE-OCTpromoter (Fig. 1b) but to a less extent also when we used the MMTV-promoter (Fig. 1a). In contrast, Deslypere et al. reported a generally 10 times higher activity of dihydrotestosterone compared with testosterone in AR-dependent induction of the MMTV-promoter in CHO-cells at concentrations <1 nM [22]. This is well in accordance with the higher stability of the dihydrotestosterone-AR complex [23] and is probably an important mechanism leading to different actions of these two androgens in vivo [22]. However, Kemppainen et al. presented equal or even lower activity of dihydrotestosterone-induced MMTV-activation in CV1-cells compared with testosterone at 0.01 and 0.1 nM, thus confirming our data [13]. We found dihydrotestosterone to be the most active ligand in AR-induced activation

274

P.M. Holterhus et al. / Journal of Steroid Biochemistry & Molecular Biology 82 (2002) 269–275

Fig. 2. Cluster analysis of promoter activation profiles. The dendrogram represents the similarity of expression profiles due to the different hormones. The colors represent the distance of relative fold induction relatively to the mean in each row in log2 space. Increasing red intensity means increasingly higher promoter activation than the mean, increasing green intensity means decreasing promoter activation compared with the mean. Black color means no deviation from the mean.

of the TATA-promoter (Fig. 1c). These considerations indicate that the relative biological activity comparing different androgens as measured in vitro in the form of AR-induced reporter gene activation can be variable and does not necessarily reflect the physical binding affinity of an androgen to the AR alone. We assume that this variability in the transactivation capability is influenced by both, variations in the biological environment (experimental conditions, e.g. degree of over-expression, but also tissue-specific factors in vivo) as well as the molecular structure of target gene promoters resulting in their preferential activation due to distinct androgen-AR-complexes. Hence, our observations provide an additional explanation for the different in vivo actions of testosterone and dihydrotestosterone. In contrast to oxandrolone and stanozolol, nandrolone has a higher androgenic/anabolic ratio [8,24]. This may be the

reason why the reporter gene activation profile due to nandrolone being more similar to the virilizing androgens than to the other two anabolics (Fig. 1 a–c, Fig. 2). The relative similarity in reporter gene activation of dehydroepiandrosterone and androstenedione with oxandrolone and stanozolol (Fig. 2) is less obvious since the anabolic activity of androstenedione is seemingly low [25]. It seems evident that other functional characteristics of the steroids than only those associated with apparently preferential anabolic or androgenic activity in vivo also contribute to the observed differences in the promoter activation profiles. However, based on the fact that dehydroepiandrosterone and androstenedione were potent activating ligands of the AR leading to differential target gene activation patterns, our data support a relevant contribution of testosterone-precursor hormones to mechanisms of in vivo androgen action, e.g. the differential morphogenesis in normal and abnormal male sexual differentiation. It has to be assumed that a differential metabolism of the androgens in target tissues [9] may add to differential target gene activation profiles and thus to steroid-specific biological actions as well. In our study, we have not particularly investigated the metabolism of the used steroids. As we have focused on only one cell type (CHO-cells), cell-type-specific differences in the metabolism of the steroids are not reflected by the activation profiles in the presented data. Also, possible interactions of the steroids with other intracellular binding sites than the AR which likely contribute to differential actions in vivo [10–12] are not covered by the used CHO-model because neither of the three used reporter genes showed steroid-induced activation in the absence of a transfected AR expression plasmid. The interaction of a steroid with the ligand-binding pocket of the AR is a highly specific molecular event. There are probably 18 [26] or 19 [27] amino acids in the AR that interact with the bound ligand. Alterations of the steroid molecule in agonists and antagonists cause a different spatial interaction with the residues of the ligand-binding pocket of the AR [27]. The reporter genes that we selected as models for androgen responsive target genes in this study are characterized by considerable differences in the composition of their promoters. They are not only different in the sequence, spacing and composition of the consensus hormone responsive elements, but they also either allow or exclude the interaction with other transcription factors than the activated AR (e.g. SP1 in GRE-OCT-luc; NF1 and Oct1 in MMTV-luc; none in (ARE)2 TATA-luc). We have to assume that the specific ligand-induced receptor conformation determines how the hormone receptor complex can specifically interact with co-regulators and neighboring transcription factors. This predicts for preferential mechanisms of target gene activation depending on the structure of the DNA response element. Differential nucleosome rearrangement may be part of this process [28]. In conclusion, our model provides first experimental evidence for the existence of distinct target gene expression profiles due to AR-activation by structurally different

P.M. Holterhus et al. / Journal of Steroid Biochemistry & Molecular Biology 82 (2002) 269–275

androgenic steroid hormones. In fact, we showed that the AR transduces a ligand-specific signal down to the target gene level. High throughput applications like cDNA-microarrays and proteomics will help to uncover the natural steroid-specific target genes and their expression programs in natural androgen responsive target tissues. This will add significantly to a comprehensive understanding of the diversity of biological actions of androgens and consequently to new concepts on their role in physiology and pathophysiology and their potential value as medication in the treatment of clinical conditions. Acknowledgements The study was supported by DFG grants Hi 497/4-3 to OH as well as KFO 111/1 to OH and to PMH. The authors thank N. Homburg for excellent technical assistance. They are indebted to G. Jenster and A.O. Brinkmann for providing androgen receptor and luciferase reporter gene constructs. References [1] O. Hiort, P.M. Holterhus, E.M. Nitsche, Physiology and pathophysiology of androgen action, Baillieres Clin. Endocrinol. Metab. 12 (1) (1998) 115–132. [2] G.H. Sinnecker, O. Hiort, L. Dibbelt, N. Albers, H.G. Dörr, H. Hauss, U. Heinrich, M. Hemminghaus, W. Hoepffner, M. Holder, D. Schnabel, K. Kruse, Phenotypic classification of male pseudohermaphroditism due to steroid 5 alpha-reductase 2 deficiency, Am. J. Med. Genet. 63 (1) (1996) 223–230. [3] O. Hiort, H. Willenbring, N. Albers, W. Hecker, J. Engert, L. Dibbelt, G.H. Sinnecker, Molecular genetic analysis and human chorionic gonadotropin stimulation tests in the diagnosis of prepubertal patients with partial 5 alpha-reductase deficiency, Eur. J. Pediatr. 155 (6) (1996) 445–451. [4] A.L. Boehmer, A.O. Brinkmann, L.A. Sandkuijl, D.J. Halley, M.F. Niermeijer, S. Andersson, F.H. de Jong, H. Kayserili, M.A. de Vroede, B.J. Otten, C.W. Rouwe, B.B. Mendonca, C. Rodrigues, H.H. Bode, P.E. de Ruiter, H.A. Delemarre-van de Waal, S.L. Drop, 17Beta-hydroxysteroid dehydrogenase-3 deficiency: diagnosis, phenotypic variability, population genetics, and worldwide distribution of ancient and de novo mutations, J. Clin. Endocrinol. Metab. 84 (12) (1999) 4713–4721. [5] W. Twesten, P.M. Holterhus, W.G. Sippell, M. Morlot, H. Schumacher, B. Schenk, O. Hiort, Clinical, endocrine, and molecular genetic findings in patients with 17beta-hydroxysteroid dehydrogenase deficiency, Horm. Res. 53 (1) (2002) 26–31. [6] S. Andersson, N. Moghrabi, Physiology and molecular genetics of 17beta-hydroxysteroid dehydrogenases, Steroids 62 (1) (1997) 143– 147. [7] H.R. Gribbin, S.G. Flavell Matts, Mode of action and use of anabolic steroids, Br. J. Clin. Pract. 30 (1) (1976) 3–9. [8] C.D. Kochakian, Definition of androgens and protein anabolic steroids, Pharmacol. Ther. B 1 (2) (1975) 149–177. [9] M. Tóth, T. Zakár, Relative binding activities of testosterone, 19-nortestosterone and their 5␣-reduced derivates to the androgen receptor and other androgen-binding proteins: a suggested role of 5␣-reductive steroid metabolism in the dissociation of “myotropic” and “androgenic” activities of 19-nortestosterone, J. Steroid Biochem. 17 (6) (1982) 653–660. [10] M. Mayer, F. Rosen, Interaction of anabolic steroids with glucocorticoid receptor sites in rat muscle cytosol, Am. J. Physiol. 229 (5) (1975) 1381–1386.

275

[11] L. Fernandez, R. Chirino, L.D. Boada, D. Navarro, N. Cabrera, I. del Rio, B.N. Diaz-Chico, Stanozolol and danazol, unlike natural androgens, interact with the low affinity glucocorticoid-binding sites from male rat liver microsomes, Endocrinology 134 (3) (1994) 1401– 1408. [12] A.J. Ellis, T.E. Cawston, E.J. Mackie, The differential effects of stanozolol on human skin and synovial fibroblasts in vitro: DNA synthesis and receptor binding, Agents Actions 41 (1–2) (1994) 37–43. [13] J.A. Kemppainen, E. Langley, C.I. Wong, K. Bobseine, W.R. Kelce, E.M. Wilson, Distinguishing androgen receptor agonists and antagonists: distinct mechanisms of activation by medroxyprogesterone acetate and dihydrotestosterone, Mol. Endocrinol. 13 (3) (1999) 440–454. [14] P.M. Holterhus, O. Hiort, G.H.G. Sinnecker, Stanozolol acts as activating ligand of the androgen receptor in reporter gene induction, Exp. Clin. Endocrinol. Diabetes 105 (Suppl. 1) (1997) 11. [15] A.O. Brinkmann, P.W. Faber, H.C. van Rooij, G.G. Kuiper, C. Ris, P. Klaassen, J.A. van der Korput, M.M. Voorhorst, J.H. van Laar, E. Mulder, J. Trapman, The human androgen receptor: domain structure, genomic organization and regulation of expression, J. Steroid Biochem. 34 (1–6) (1989) 307–310. [16] G. Jenster, T.E. Spencer, M.M. Burcin, S.Y. Tsai, M.J. Tsai, B.W. O’Malley, Steroid receptor induction of gene transcription: a two-step model, Proc. Natl. Acad. Sci. U.S.A. 94 (15) (1997) 7879–7884. [17] M. Truss, M. Beato, Steroid hormone receptors: interaction with deoxyribonucleic acid and transcription factors, Endocr. Rev. 14 (4) (1993) 459–479. [18] C. Chen, H. Okayama, High-efficiency transformation of mammalian cells by plasmid DNA, Mol. Cell. Biol. 7 (8) (1987) 2745–2752. [19] P.M. Holterhus, H.T. Bruggenwirth, O. Hiort, A. Kleinkauf-Houcken, K. Kruse, G.H. Sinnecker, A.O. Brinkmann, Mosaicism due to a somatic mutation of the androgen receptor gene determines phenotype in androgen insensitivity syndrome, J. Clin. Endocrinol. Metab. 82 (11) (1997) 3584–3589. [20] M.B. Eisen, P.T. Spellmann, P.O. Brown, D. Botstein, Cluster analysis and display of genome-wide expression patterns, Proc. Natl. Acad. Sci. U.S.A. 95 (25) (1998) 14863–14868. [21] V.R. Iyer, M.B. Eisen, D.T. Ross, G. Schuler, T. Moore, J.C. Lee, J.M. Trent, L.M. Staudt, J. Hudson Jr., M.S. Boguski, D. Lashkari, D. Shalon, D. Botstein, P.O. Brown, The transcriptional program in the response of human fibroblasts to serum, Science 283 (5398) (1999) 83–87. [22] J.P. Deslypere, M. Young, J.D. Wilson, M.J. McPhaul, Testosterone and 5␣-dihydrotestosterone interact differently with the androgen receptor to enhance transcription of the MMTV-CAT reporter gene, Mol. Cell. Endocrinol. 88 (1–3) (1992) 15–22. [23] M. Kaufmann, L. Pinsky, The dissociation of testosterone and 5alpha-dihydrotestosterone-receptor complexes formed within cultured human genital skin fibroblasts, J. Steroid Biochem. 18 (2) (1983) 121–125. [24] F. Neumann, Pharmacological and endocrinological studies on anabolic agents, Environ. Qual. Saf. Suppl. 5 (1976) 253–264. [25] B.B. Rasmussen, E. Volpi, D.C. Gore, R.R. Wolfe, Androstenedione does not stimulate muscle protein anabolism in young healthy men, J. Clin. Endocrinol. Metab. 85 (1) (2000) 55–59. [26] P.M. Matias, P. Donner, R. Coelho, M. Thomaz, C. Peixoto, S. Macedo, N. Otto, S. Joschko, P. Scholz, A. Wegg, S. Basler, M. Schafer, U. Egner, M.A. Carrondo, Structural evidence for ligand specificity in the binding domain of the human androgen receptor. Implications for pathogenic gene mutations, J. Biol. Chem. 275 (34) (2000) 26164–26171. [27] N. Poujol, J.M. Wurtz, B. Tahiri, S. Lumbroso, J.C. Nicolas, D. Moras, C. Sultan, Specific recognition of androgens by their nuclear receptor. A structure-function study, J. Biol. Chem. 275 (31) (2000) 24022–24031. [28] G. Jenster, Coactivators and corepressors as mediators of nuclear receptor function: an update, Mol. Cell. Endocrinol. 143 (1–2) (1998) 1–7.