NEUROSCIENCE RESEARCH ARTICLE

S. Zelleroth et al. / Neuroscience 397 (2019) 172–183

Toxic Impact of Anabolic Androgenic in Primary Rat Cortical Cell Cultures

Sofia Zelleroth, * Erik Nylander Fred Nyberg, Alfhild Gro¨ nbladh and Mathias Hallberg The Beijer Laboratory, Department of Pharmaceutical Biosciences, Division of Biological Research on Drug Dependence, SE-751 24, Uppsala University, Sweden

Abstract—The use of anabolic androgenic steroids (AASs) among non-athletes is a public health-problem, as abu- sers underestimate the negative effects associated with these drugs. The present study investigated the toxic effects of , , , and trenbolone, and aimed to understand how AAS abuse affects the brain. Mixed cortical cultures from embryonic rats were grown in vitro for 7 days and thereafter treated with increasing concentrations of AASs for 24 h (single-dose) or 3 days (repeated exposure). Cells were co-treated with the -receptor (AR) antagonist flutamide, to determine whether the potential adverse effects observed were mediated by the AR. Cellular toxicity was determined by measuring mitochondrial activity, lactate dehydrogenase (LDH) release, and caspase-3/7 activity. Nandrolone, unlike the other AASs studied, indicated an effect on mitochondrial activity after 24 h. Furthermore, single-dose exposure with testosterone, nandrolone and trenbolone increased LDH release, while no effect was detected with stanozolol. However, all of the four steroids negatively affected mitochondrial function and resulted in LDH release after repeated exposure. Testosterone, nandrolone, and trenbolone caused their toxic effects by induction of apoptosis, unlike stanozolol that seemed to induce necrosis. almost completely prevented AAS-induced toxicity by maintaining mitochondrial function, cellular integrity, and inhibition of apoptosis. Overall, we found that supra-physiological concentrations of AASs induce cell death in mixed primary cortical cultures, but to different extents, and possibly through var- ious mechanisms. The data presented herein suggest that the molecular interactions of the AASs with the AR are primarily responsible for the toxic outcomes observed. Ó 2018 The Author(s). Published by Elsevier Ltd on behalf of IBRO. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Key words: mitochondrial function, membrane integrity, cell death, androgen-receptor, flutamide.

INTRODUCTION as 9.8 % (Althobiti et al., 2018). Many of the misused compounds have not been adequately examined regard- Testosterone and other synthetic anabolic androgenic ing their effect on general health, and even less is known steroids (AASs) are commonly abused in order to about their effects on the brain. Because of this it is of enhance athletic performance and physical appearance importance to gain more knowledge about the acute and (Kindlundh et al., 1998, 1999; Cohen et al., 2007). The long-term adverse effects induced by these substances. average age-of-onset for AAS use is around 20 years It is well known that AASs induce androgenic effects and only 22 % of the users start before age 20 (Pope (development of male sex characteristics) or anabolic et al., 2014). The worldwide prevalence of AAS abuse is effects (muscle building), by actions mediated through poorly understood, but a high frequency of usage occurs the intracellular (AR), which in certain subpopulations, for example in gym communi- regulates transcription in targeted genes (Lee and ties. In these environments the prevalence of usage Chang, 2003). Structural optimization of the testosterone among male young adults has been reported to be as high molecule has subsequently been conducted with the aim of developing AASs with enhanced anabolic effects rela- *Corresponding author at: Box 591, SE-751 24 Uppsala, Sweden. tive to the androgenic effects. However, all AASs so far E-mail addresses: sofi[email protected] (S. Zelleroth), available on the market still cause masculinization when [email protected] (E. Nylander), Fred.nyberg@farmbio. used in high doses and over a long period of time uu.se (F. Nyberg), [email protected] (A. Gro¨ nbladh), (Kicman, 2008). Regimens of AAS misuse often involve [email protected] (M. Hallberg). Abbreviations: 2-way ANOVA, two-way analysis of variance; AAS, doses ranging up to 100 times higher than the therapeutic anabolic androgenic steroids; AR, androgen receptor; CNS, central dose of testosterone that is applied. Doses are often nervous system; DIV, days in vitro; DMSO, dimethyl sulfoxide; LDH, increased gradually during 6–12 weeks, thereafter lactate dehydrogenase; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; PBS, phosphate-buffered saline; SEM, standard decreased and followed by a variable period off the drugs error of mean. (Kanayama et al., 2009). https://doi.org/10.1016/j.neuroscience.2018.11.035 0306-4522/Ó 2018 The Author(s). Published by Elsevier Ltd on behalf of IBRO. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 172 S. Zelleroth et al. / Neuroscience 397 (2019) 172–183 173

In addition to the desired effects, high doses and Testosterone, nandrolone, stanozolol, and trenbolone extended treatment periods with AASs also cause are all commonly abused by AAS-users in their efforts to adverse effects, which have been extensively reviewed gain muscle and improve performance. The chemical elsewhere (Goldman and Basaria, 2018). Furthermore, structures of these AASs are depicted in Fig. 1. In the there are a number of known psychological effects seen present study, supra-physiological concentrations of in AAS users, such as irritability, anxiety, and aggression structurally different AASs caused cellular toxicity in (Kanayama et al., 2008; Lindqvist Bagge et al., 2017), primary cortical cell cultures, but to different extents. which suggest that these steroids affect the brain. These differences between the AASs regarding their Although the impact of AASs on the central nervous sys- effects on mitochondrial function, cell membrane tem (CNS) is not yet fully clarified, several studies report integrity, and activation of specific executioner caspases that these compounds may interact with a variety of neu- are of relevance in order to understand the mechanisms rochemical systems in the rodent brain (Le Greves et al., underlying the AAS-induced effects on the brain that 1997; Hallberg et al., 2000; Kindlundh et al., 2002; Clark result in behavioral and cognitive changes. and Henderson, 2003; Hallberg, 2011; Gronbladh et al., 2014). Supraphysiological doses of AAS may produce a EXPERIMENTAL PROCEDURES variety of neurotoxic effects, which are mechanistically more or less explored. Previous in vitro studies suggest Animals and tissue collection a negative impact on cell survival induced by AASs in All animal experimental procedures were approved by the neuronal-like cells (Estrada et al., 2006; Orlando et al., local animal ethics committee at Uppsala University 2007; Caraci et al., 2011; Basile et al., 2013). These stud- (C15/14) and were conducted in accordance with ies suggest neurodegenerative alterations to be caused Swedish guidelines regarding animal experiments either by direct induction of apoptosis (Estrada et al., (Animal Welfare Act SFS 1998:56) and the European 2006; Basile et al., 2013) or indirectly by increasing neu- Communities Council directive (86/609/EEC). ronal vulnerability to other toxic insults (Orlando et al., Pregnant rats, Sprague–Dawley or Wistar, were 2007; Caraci et al., 2011). This opens up for speculations obtained from Charles River (Sulzfeld, Germany). Mixed regarding the risk of cognitive deficits among AAS- glial and neuronal primary cortical cultures were abusers. Furthermore, a number of studies indicate that prepared according to a previously described method AASs may affect several CNS-related behaviors and the with minor modifications (Diwakarla et al., 2016). Briefly, studies which demonstrate that different AASs may pregnant dams, (n = 10), were anesthetized at embry- induce memory impairment in both rats and humans onic day 17 with increasing concentrations of carbon diox- (Magnusson et al., 2009; Tanehkar et al., 2013; ide. Fetuses were extracted from the mothers and cortical Kanayama et al., 2013; Gronbladh et al., 2013; tissue was dissected in Hank’s balanced salt solution Heffernan et al., 2015) have received particular interest. (Thermo Fisher Scientific, Waltham, MA, USA). The underlying mechanism behind these observations still remains to be clarified, but prolonged exposure to AASs is shown to cause smaller cortical volume and thin- Cell culture ning of the cortex among AAS-users (Bjornebekk et al., Tissue was prepared by digestion with 0.4 mg/mL trypsin 2017), which suggests that AASs may induce cerebral (Sigma-Aldrich, St Louis, MO, USA) at 37 °C for 10 min. atrophy. Furthermore, structural and functional connectiv- The chemical digestion was inhibited with 0.5-mg trypsin ity alterations of amygdala have also been reported in inhibitor (Sigma-Aldrich), and a cell suspension was men after long-term AAS use (Kaufman et al., 2015; obtained by mechanical dissociation with a 1 mL pipette. Westlye et al., 2017; Seitz et al., 2017). The alterations Cells were counted and seeded at appropriate densities seen in amygdala may affect networks between brain into pre-coated Poly-D-lysine (50 mg/mL, Sigma-Aldrich) areas associated with cognitive and emotional control. tissue culture plates (Thermo Fischer Scientific). Cells Taken together, this strengthens the theory about the were grown for 24 h in neurobasal media (Thermo potentially increased risk of persistent alterations in the Fisher Scientific) supplemented with 2 % (v/v) B27 brain, behavioral changes, and development of neurode- (Thermo Fischer Scientific), 10 % (v/v) fetal bovine generative diseases among AAS-users. serum (Thermo Fischer Scientific), 0.5 mM GlutaMAXTM

Fig. 1. Chemical structure of the investigated AASs. Four commonly abused AASs that are administrated either orally or by intramuscular injection as ester prodrugs. A) testosterone, B) nandrolone, C) stanozolol, and D) trenbolone. 174 S. Zelleroth et al. / Neuroscience 397 (2019) 172–183

(Thermo Fischer Scientific), and 100 U/mL penicillin and per well were seeded into a 96-well tissue culture plate streptomycin (Thermo Fisher Scientific). At day one (Thermo Fisher Scientific), and the cells were exposed in vitro (1 DIV), media were replaced with serum-free to either increasing concentrations of AASs alone or media supplemented with 2 % extra B27 (Thermo together with flutamide at 7 DIV for 24 h, (n = 4), and Fischer Scientific). Cell cultures were kept in an 3 days, (n = 4). Following drug treatments, incubator (37 °C/5 % (v/v) CO2) and subsequent partial mitochondrial function as a measure of cell viability was media replacements were performed twice a week. Only determined using the MTT assay (Sigma-Aldrich). The mature cell cultures (7 days in vitro, 7 DIV) were used absorbance was measured using the FLUOstar Omega for the experiments. spectrophotometer (BMG Labtech, Ortenberg, Germany) at 570 nm. Chemicals and drug treatments Cell cytotoxicity assay The AASs (testosterone, nandrolone, stanozolol, trenbolone) and the AR-antagonist (flutamide) were To study whether exposure to the AASs cause membrane purchased from Sigma Aldrich. The AASs and flutamide permeabilization in primary cortical cell cultures, a density were dissolved in 100 % (v/v) dimethyl sulfoxide of 1 105 cells per well were seeded into a 96-well tissue (DMSO, Merck Millipore, Bedford, MA, USA), and culture plate (Thermo Fisher Scientific). At 7 DIV the cells further diluted with water. Final concentrations of AASs were exposed to increasing concentrations of AASs, with administered to the cells were 10 lM, 30 lM and or without flutamide, for 24 h, (n = 4), and 3 days, 100 lM. The final concentration of flutamide was 10 lM (n = 4). After the drug treatments, cell cytotoxicity was and the final concentration of DMSO was 0.1 % (v/v) or studied by measuring the release of lactate less. dehydrogenase (LDH). LDH is an enzyme naturally AASs were added into the cell media for either 24 h occurring in the cell cytoplasm, which upon cell (single-dose) or continually every 24 h for 3 days membrane permeabilization leaks out into the cell media (repeated exposure). To determine the role of ARs, the and acts as a marker for cell cytotoxicity. Culture media AR antagonist flutamide was co-administrated together from cells exposed to AASs for 24 h or 3 days were with the AASs. Controls were treated by replacing the transferred to an empty tissue culture plate. LDH media with media containing 0.1 % (v/v) DMSO. concentration was analyzed using the LDH cytotoxicity detection kit (Roche Life Science, Mannheim, Germany) according to the manufacturer’s instructions. The Immunocytochemistry absorbance was measured at 492 nm using the Immunofluorescent staining was used to investigate the FLUOstar Omega spectrophotometer (BMG Labtech). expression of the AR in primary cortical cell cultures. A Wells containing media-only (background noise) were density of 2.5 104 cells per well was seeded into a 24- subtracted from each well. well tissue culture plate (Thermo Fischer Scientific) containing glass coverslips pre-coated with 50 mg/mL Caspase-3/7 activity poly-D-lysine. At day 7 in vitro (7 DIV) untreated primary cortical cells, (n = 1), were fixed with 4 % (w/v) Apoptosis is a process of programmed cell death, which paraformaldehyde (Sigma-Aldrich) in phosphate- occur in all cells. The onset of apoptosis is regulated by buffered saline (PBS) for 15 min and permeabilized at specific caspases that mediate proteolytic cleavage of room temperature with 0.2 % (v/v) Tween 20 in PBS for certain proteins. Therefore, detection of active caspases 10 min. Non-specific binding was blocked by incubation can be used as a marker of apoptosis. To analyze with PBS containing 10 % (v/v) normal donkey serum whether exposure to AASs induced apoptosis in primary (Sigma Aldrich) and 0.1 % (v/v) Tween 20 for 1 h in cortical cell cultures, a density of 2 104 cells per well room temperature. After blocking, cells were incubated were seeded into a 96-well tissue culture plate (Thermo with the primary antibody rabbit anti-androgen receptor Fisher Scientific). At 7 DIV the cells were either (1:50 dilution; Santa Cruz Biotechnology, Dallas, TX, exposed to increasing concentrations of AASs alone or USA), for 1 h at room temperature. After washing, together with flutamide, for 24 h and 3 days, (n = 3). secondary antibody Alexa 488 donkey anti-rabbit (1:500 Staurosporine (100 nM; Sigma-Aldrich), a well-known dilution; Thermo Fischer Scientific) was added for apoptotic inducer was used as a positive control to 1 h at room temperature. After washing, cover slips ensure detection of caspase activity (data not shown). were mounted with MOWIOL anti-fade mounting After the drug treatments, apoptosis was studied by medium (Sigma-Aldrich) containing 40,6-diamidino-2- analyzing the activity of caspase 3 and caspase 7 using phenylindole, dihydrochloride (DAPI; 2.5:500 dilution; the Caspase 3/7 GLO kit (Promega, Madison, WI, USA) Merck Millipore, Bedford, MA, USA). Images were according to the manufacturer’s instruction with some acquired using 20 objectives on the ImageXpress XL modifications. Briefly, the caspase 3/7-GLO reagent (Molecular Devices, San Jose, CA, USA) microscope. buffer was added (v/v) to the tissue culture plate, containing treated and untreated cells. From this time point on the contents were protected from light. The Cell viability assay tissue culture plate was placed on a shaker at 300 rpm To investigate the effect of AASs on the viability of for 5 min and thereafter incubated for 30 min at room primary cortical cell cultures, a density of 1 105 cells temperature. The contents were then transferred to an S. Zelleroth et al. / Neuroscience 397 (2019) 172–183 175 empty white-walled 96-well plate (Thermo Fisher approximately 60 % neurons as measured by Scientific) and luminescence was analyzed using the microtubule-associated protein 2-positive cells. FLUOstar Omega spectrophotometer (BMG Labtech). Primary cortical cultures expressed the androgen receptor Statistical analysis The immunocytochemistry analysis revealed expression For all measurements, raw data were used for the of the AR in primary cortical cell cultures (Fig. 2) and statistical analyses. Raw data were assumed to follow the cells could therefore be used for further analyses in normal distribution and the statistical choices were the present study. The AR was primarily found to be based on previous report (Lew, 2007). Cells collected expressed in the cell body associated with the nuclei, from fetuses obtained from one individual rat were consid- but also in extranuclear sites such as axons. ered as one individual culture (n = 1). Each individual cul- ture received specific treatment in triplicates and all Single-dose exposure to nandrolone had an effect on analyses were repeated at least three times in individual mitochondrial function in primary cortical cultures cultures (n = 3). The statistical test used was the two- way analysis of variance (2-way ANOVA), with specific Treatment with nandrolone (single-dose exposure for treatment and experiment (culture ID) as factors to 24 h) had an overall effect on mitochondrial activity, account for differences. When an overall effect was found F (3, 9) = 4.323, p 0.05, however, post hoc analysis on treatment, data were further analyzed by Dunnet’s did not reveal any significant differences (Fig. 3). Single- post hoc test, and corrected for multiple comparison using dose exposure to testosterone, stanozolol, and statistical hypothesis testing. The MTT and LDH analyses trenbolone did not alter the overall effect on of treatments with AASs alone and together with flutamide mitochondrial function (Fig. 3). were performed on individual cultures and considered independent experiments. Therefore, data received from Single-dose exposure to testosterone, nandrolone, these experiments were handled and analyzed individu- and trenbolone caused membrane permeabilization ally. In contrast to this, experimental analysis of caspase in primary cortical cultures 3/7 activity in cells treated with AAS alone and together with flutamide were performed on the same individual cul- An overall effect could be detected on LDH release tures. Data from these experiments were therefore han- following 24-h treatment with testosterone, F (3, 9) = dled, analyzed, and presented together. Data are 10.14, p 0.01, nandrolone, F (3, 9) = 7.246, p 0.01, presented as mean values and standard error of mean and trenbolone, F (3, 9) = 4.325, p 0.05, indicating (mean ± SEM). For all statistical analyses a probability membrane permeabilization. Stanozolol did not affect value of p < 0.05 was considered statistically significant. the release of LDH. The post hoc analysis revealed that l Graphpad software (Prism 7.0) was used to assess the exposure to high concentrations (100 M) of statistical analyses. testosterone, nandrolone, and trenbolone significantly increased LDH release when compared with control (Fig. 4). RESULTS Repeated exposure to the AASs caused reduced To investigate the potential toxic effects of supra- mitochondrial function and increased membrane physiological doses of certain structurally diverse AASs, permeabilization in primary cortical cultures an in vitro model of mixed glial and neuronal cortical cell cultures was used. Mature cortical cultures (7 DIV) Repeated administration (every 24 h for 3 days) with showed prominent axonal maintenance and contained testosterone, F (3, 9) = 17.04, p 0.001, nandrolone,

Fig. 2. Expression of the androgen receptor in primary cortical cultures. Immunofluorescent microscopy of fixed 7 DIV primary cortical cultures showed expression of the AR. Primary cortical cultures were stained with antibodies specific for the AR (green) and the nucleus staining DAPI (blue). Images were acquired using 20 objective on the ImageXpress XL microscope. 176 S. Zelleroth et al. / Neuroscience 397 (2019) 172–183

Fig. 3. Single-dose exposure to increasing concentrations of nandrolone, unlike the other AASs, affected mitochondrial function in primary cortical cultures. Mitochondrial function was determined by measuring the MTT after single-dose (24 h) treatment with increasing concentrations (10, 30, 100 lM) of AAS. A) testosterone B) nandrolone, C) stanozolol, and D) trenbolone. Nandrolone was the only AAS that affected mitochondrial function, as indicated by ANOVA. Although, following post hoc test did not show any significant differences. Absorbance measurements are presented as mean values and standard error of mean (mean ± SEM) with three replicates from four individual cultures. Overall differences were calculated with a 2-way ANOVA, followed by Dunnet’s post hoc test in comparison with control when appropriate. Probability values of p < 0.05 were considered statistically significant. Abbreviations: anabolic androgenic steroids (AAS), 3-(4,5-dimethylthiazol-2-yl)-2,5- diphenyltetrazolium bromide (MTT), control (cntrl), two-way analysis of variance (2-way ANOVA).

Fig. 4. Single-dose exposure to increasing concentrations of testosterone, nandrolone and trenbolone, unlike stanozolol, induced LDH release in primary cortical cultures. Cell membrane permeabilization was determined by measuring the LDH released into the cell media after 24 h treatment with increasing concentrations (10, 30, 100 lM) of AAS. A) Testosterone increased LDH release at 100 lM, B) nandrolone increased LDH release at 100 lM, C) stanozolol did not affect membrane permeabilization at any concentration, and D) trenbolone increased LDH release at 100 lM. Absorbance measurements are presented as mean values and standard error of mean (mean ± SEM) with three replicates from four individual cultures. Overall differences were calculated with a 2-way ANOVA, followed by Dunnet’s post hoc test in comparison with control when appropriate. Probability values of p < 0.05 were considered statistically significant, *p < 0.05, **p < 0.01. Abbreviations: lactate dehydrogenase (LDH), anabolic androgenic steroids (AAS), control (cntrl), two-way analysis of variance (2-way ANOVA).

Fig. 5. Repeated exposure to increasing concentrations of the AASs caused mitochondrial dysfunction in primary cortical cultures. Mitochondrial function was determined by measuring the MTT metabolism after repeated (3 days) treatment with increasing concentrations (10, 30, 100 lM) of AAS. A) Testosterone decreased mitochondrial activity at 100 lM, B) nandrolone decreased mitochondrial activity at 30 lM and 100 lM, C) stanozolol decreased mitochondrial activity at 100 lM, and D) trenbolone decreased mitochondrial activity at 100 lM. Absorbance measurements are presented as mean values and standard error of mean (mean ± SEM) with three replicates from four individual cultures. Overall differences were calculated with a 2-way ANOVA, followed by Dunnet’s post hoc test in comparison with control when appropriate. Probability values of p < 0.05 were considered statistically significant, *p < 0.05, **p < 0.01, ****p < 0.0001. Abbreviations: anabolic androgenic steroids (AAS), 3-(4,5- dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), control (cntrl), two-way analysis of variance (2-way ANOVA). S. Zelleroth et al. / Neuroscience 397 (2019) 172–183 177

F (3, 9) = 29.9, p 0.0001, stanozolol, F (3, 9) = 5.781, The toxic effects induced by the AASs are mediated p 0.05, and trenbolone, F (3, 9) = 10.76, p 0.01, by androgen receptor activation resulted in an overall reduction in mitochondrial function, The AR antagonist flutamide prevented the overall indicating reduced viability of the cells. The following reduction in mitochondrial activity seen after single-dose post hoc analysis showed that nandrolone reduced exposure to nandrolone (data are presented in Table 1), mitochondrial function at both 30 lM and 100 lM, while as no reduction in viability was observed following co- the other AASs only caused mitochondrial dysfunction treatment. In addition, flutamide was able to rescue cells at the highest dose (100 lM) when compared with from the mitochondrial dysfunction induced by repeated control (Fig. 5). exposure of testosterone and stanozolol, as no An overall effect on LDH release after repeated reduction in viability was detected following co-treatment administration with testosterone, F (3, 9) = 28.22, (data are presented in Table 1). Furthermore, flutamide p 0.0001, nandrolone, F (3, 9) = 46.79, p 0.0001, was able to reduce the overall negative effect seen on stanozolol, F (3, 9) = 23.84, p 0.0001, and mitochondrial function caused by repeated exposure of trenbolone, F (3, 9) = 8.2, p 0.01, could also be nandrolone F (3, 9) = 14.97, p 0.001, and trenbolone, detected. The post hoc analysis revealed that F (3, 9) = 20.52, p 0.001, but not to control levels. nandrolone and trenbolone increased LDH release at The post hoc tests revealed that the mitochondrial both 30 lM and 100 lM, while a significant increase in dysfunction induced by 30 lM of nandrolone was LDH could only be detected at 100 lM following abolished in the presence of flutamide, while the testosterone and stanozolol treatment when compared reduction detected following exposure to 100 lMof with control (Fig. 6). nandrolone and trenbolone could only be partly inhibited

Fig. 6. Repeated exposure to increasing concentrations of the AASs induced LDH release in primary cortical cultures. Cell membrane permeabilization was determined by measuring the LDH released into the cell media after repeated (3 days) treatment with increasing concentrations (10, 30, 100 lM) of AAS. A) Testosterone increased LDH release at 100 lM, B) nandrolone increased LDH release at 30 lM and 100 lM, C) stanozolol increased LDH release at 100 lM, and D) trenbolone increased LDH release at 30 lM and 100 lM. Absorbance measurements are presented as mean values and standard error of mean (mean ± SEM) with three replicates from four individual cultures. Overall differences were calculated with a 2-way ANOVA, followed by Dunnet’s post hoc test in comparison with control when appropriate. Probability values of p < 0.05 were considered statistically significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Abbreviations: anabolic androgenic steroids (AAS), lactate dehydrogenase (LDH), control (cntrl), two-way analysis of variance (2-way ANOVA).

Table 1. Mitochondrial activity in primary cortical cultures after single dose (24 h) and repeated (3 days) exposure to co-administration of 10 lM flutamide and AAS

Substance Exposuretime Cntrl AAS concentration 2-Way ANOVA

10 lM30lM 100 lM F (Dfn, Dfd) P value

Flutamide + Testosterone 24 h 0.69 ± 0.07 0.69 ± 0.16 0.64 ± 0.13 0.63 ± 0.11 F (3,9) = 0.35 ns 3 days 0.67 ± 0.07 0.77 ± 0,01 0.72 ± 0.04 0.67 ± 0.04 F (3,9) = 1.62 ns Flutamide + Nandrolone 24 h 0.69 ± 0.07 0.71 ± 0.09 0.62 ± 0.09 0.58 ± 0.04 F (3,9) = 2.00 ns 3 days 0.70 ± 0.07 0.80 ± 0.08 0.66 ± 0.02 0.38** ± 0.06 F (3,9) = 15.0 <0.01 Flutamide + Stanozolol 24 h 0.62 ± 0.07 0.65 ± 0.11 0.73 ± 0.16 0.66 ± 0.09 F (3,9) = 0.81 ns 3 days 0.70 ± 0.07 0.84 ± 0.06 0.83 ± 0.07 0.87 ± 0.04 F (3,9) = 1.66 ns Flutamide + Trenbolone 24 h 0.62 ± 0.07 0.72 ± 0.18 0.72 ± 0.15 0.68 ± 0.08 F (3,9) = 0.59 ns 3 days 0.70 ± 0.07 0.95** ± 0.03 0.75 ± 0.04 0.49* ± 0.06 F (3,9) = 20.5 <0.001

Absorbance measurements are presented as mean values and standard error of mean (mean ± SEM) with three replicates from four individual cultures. Overall differences were calculated with 2-way ANOVA, followed by Dunnet’s post hoc test in comparison with control when appropriate. Probability values of p < 0.05 were considered statistically significant, *p < 0.05, **p < 0.01. Abbreviations: anabolic androgenic steroids (AAS), control (cntrl), F ratio (F), degrees of freedom numerator (DFn), degrees of freedom denominator (DFd), and two-way analysis of variance (2-way ANOVA). 178 S. Zelleroth et al. / Neuroscience 397 (2019) 172–183

Table 2. Membrane permeabilization in primary cortical cultures after single dose (24 h) and repeated (3 days) exposure to co-administration of 10 lM flutamide and AAS

Substance Exposuretime Cntrl AAS concentration 2-Way ANOVA

10 lM30lM 100 lM F (Dfn, Dfd) P value

Flutamide + Testosterone 24 h 0.28 ± 0.06 0.24 ± 0.05 0.23 ± 0.06 0.25 ± 0.05 F (3,9) = 1.71 ns 3 days 0.36 ± 0.08 0.35 ± 0.04 0.29 ± 0.02 0.41 ± 0.07 F (3,9) = 1.37 ns Flutamide + Nandrolone 24 h 0.28 ± 0.06 0.26 ± 0.09 0.24 ± 0.06 0.33 ± 0.07 F (3,9) = 1.15 ns 3 days 0.36 ± 0.08 0.34 ± 0.05 0.39 ± 0.06 0.49* ± 0.03 F (3,9) = 7.71 <0.01 Flutamide + Stanozolol 24 h 0.26 ± 0.07 0.23 ± 0.06 0.23 ± 0.07 0.27 ± 0.06 F (3,9) = 1.69 ns 3 days 0.26 ± 0.06 0.23 ± 0.06 0.23 ± 0.04 0.27 ± 0.04 F (3,9) = 0.88 ns Flutamide + Trenbolone 24 h 0.26 ± 0.07 0.22 ± 0.06 0.23 ± 0.06 0.28 ± 0.08 F (3,9) = 1.00 ns 3 days 0.37 ± 0.06 0.30 ± 0.04 0.35 ± 0.03 0.46 ± 0.05 F (3,9) = 6.65 <0.05

Absorbance measurements are presented as mean values and standard error of mean (mean ± SEM) with three replicates from four individual cultures. Overall differences were calculated with a 2-way ANOVA, followed by Dunnet’s post hoc test in comparison with control when appropriate. Probability values of p < 0.05 were considered statistically significant, *p < 0.05. Abbreviations: anabolic androgenic steroids (AAS), control (cntrl), F ratio (F), degrees of freedom numerator (DFn), degrees of freedom denominator (DFd), and two-way analysis of variance (2-way ANOVA).

(data presented in Table 1). In addition, flutamide induced apoptosis in primary cortical cell cultures increased mitochondrial activity when co-administrated (Fig. 7c and 7 g) as activation of caspase 3 and caspase with 10 lM of trenbolone when compared with control 7 could not be detected. Flutamide did not affect (data presented in Table 1). caspase activity of primary cortical cultures when Flutamide inhibited the increased membrane administered alone (data not shown). permeabilization detected after single-dose exposure to testosterone, nandrolone, and trenbolone, as no DISCUSSION significant cytotoxic effect (LDH release) was detected after co-treatment (data presented in Table 2). In The present study demonstrates certain structurally addition, co-administration with flutamide reduced the diverse AASs to cause reduction in cell viability and amount of LDH released after repeated exposure of increase in cell cytotoxicity. This was evidenced by testosterone and stanozolol to similar levels as control. reduced mitochondrial activity and elevated levels of The overall cytotoxic effect induced by nandrolone, LDH, respectively, occurring in a dose-dependent and F (3, 9) = 7.713, p 0.01, and trenbolone, F (3, 9) = time-dependent manner. In addition, the toxic properties 6.646, p 0.05, was reduced by flutamide, but not to of testosterone, nandrolone, stanozolol, and trenbolone control levels. Post hoc analyses demonstrated that the in these assays diverged. effect seen with 30 lM of nandrolone was prevented by The cell viability and cell cytotoxicity were altered at flutamide, but the effect seen at 100 lM was only partly the highest concentrations of AASs (30–100 lM) used in inhibited. In addition, the post hoc test revealed no the present study. These findings support previous cytotoxic effect at any concentration during co-treatment reports suggesting that AASs affect cell survival at with flutamide and trenbolone (data presented in Table 2). micro-molar concentrations causing decreased cell viability in cultured cells (Estrada et al., 2006; Cunningham et al., 2009; Caraci et al., 2011; Basile Testosterone, nandrolone, and trenbolone, unlike et al., 2013; Ma and Liu, 2015). In accordance with the stanozolol, induced apoptosis in primary cortical results presented herein, no detection of any AAS- cultures through activation of the androgen receptor induced toxicity at 10 uM was reported in a previous study Testosterone, nandrolone, and trenbolone caused (Orlando et al., 2007). Interestingly, this concentration, apoptotic cell death. Activation of caspase 3 and and even nano-molar concentrations, were found to caspase 7 were detected after single-dose exposure to enhance excitotoxic cell death. This raises serious con- nandrolone, F (6, 12) = 4.894, p 0.01, and trenbolone, cerns as excitotoxic mechanisms can be triggered by sev- F (6, 12) = 3.293, p 0.05. Caspase 3/7 activity was eral insults, as well as of an underlying neurodegenerative prevented by co-administration with flutamide (Fig. 7b disease (Orlando et al., 2007). Notably, micro-molar con- and 7d). Furthermore, the activation of caspase 3 and 7 centrations of AASs are previously established in the cere- was detected after repeated exposure to testosterone, brospinal fluid of chronic AAS-abusers (Daly et al., 2001). F (6, 12) = 15.58, p 0.0001, and the effect could be The cortical cell death detected herein could be one expla- reduced by flutamide (Fig. 7e). Repeated administration nation for the smaller cortical volume and thinner cortex with nandrolone, F (6, 12) = 8.092, p 0.001, and seen among AAS-using weightlifters (Bjornebekk et al., trenbolone, F (6, 12) = 7.9, p 0.001, further increased 2017), where the extent of the effects are associated with caspase 3/7 activity and the activation was prevented by longer history of AAS exposure and, based on present flutamide (Fig. 7f and 7 h). The post hoc analyses results, possibly by the type of AAS used. With respect revealed that single-dose and repeated administration to the parameters assessed, nandrolone appeared to be with high doses (100 lM) of AASs induced caspase-3/7 the most toxic AAS, while testosterone, stanozolol, and activity when compared with controls. In contrast, neither trenbolone disposed similar effects but to a lesser extent. single-dose nor repeated administration with stanozolol According to the present results, the order of the examined S. Zelleroth et al. / Neuroscience 397 (2019) 172–183 179

Fig. 7. The toxicity caused by testosterone, nandrolone and trenbolone, unlike stanozolol-induced effects, was associated with apoptosis and mediated by androgen receptor activation. Activation of the apoptotic cascade was determined by measuring the activation of caspase 3 and caspase 7 after 24 h (A-D) and 3 days (E-H) treatment with increasing concentrations (10, 30, 100 lM) of testosterone (A, E), nandrolone (B, F), stanozolol (C, G), and trenbolone (D, H). The cells were co-treated with the androgen , flutamide (10 lM), to determine if the effects were mediated by the androgen receptor. Luminescence measurements are presented as mean values and standard error of mean (mean ± SEM) with three replicates from three individual cultures. Overall differences were calculated with 2-way ANOVA, followed by Dunnet’s post hoc test in comparison with control when appropriate. Probability values of p < 0.05 were considered statistically significant, *p < 0.05, ***p < 0.001, ****p < 0.0001. Abbreviations: control (cntrl), two-way analysis of variance (2-way ANOVA).

AASs regarding their toxicity was found to be described to impair social memory (Kouvelas et al., nandrolone > trenbolone > testosterone > stanozolol. 2008), and spatial memory and learning (Magnusson Alterations in the brain following misuse of AASs et al., 2009; Tanehkar et al., 2013). Also, there is a report correlate with behavioral changes including decreased indicating deficits in visuospatial memory of AAS-users, cognitive function (Kanayama et al., 2013). In several where the severity of the deficits is associated with total independent rodent studies is life-time exposure to AASs (Kanayama et al., 2013). 180 S. Zelleroth et al. / Neuroscience 397 (2019) 172–183

AAS-users are also shown to display alterations in brain concentration of nandrolone and trenbolone. However, areas and neurochemical systems involved in cognitive as the caspase activation caused by these AASs was control (Kaufman et al., 2015). However, indications from inhibited by flutamide at this high concentration, the an early study in rodents show that other AASs, like results could instead indicate that some toxic effects , may not affect spatial memory induced by nandrolone and trenbolone, at least in part, and learning (Clark et al., 1995), supporting the hypothe- were mediated through another pathway apart from AR- sis that structurally diverse AASs exert different effects. activation. It is known that AASs induce apoptosis in different The pharmacological differences between the AASs tissues and organs (Wiren et al., 2006; Shokri et al., studied might explain the diverged toxicity detected to 2010; Janjic et al., 2012; Pomara et al., 2015), and there- some extent. Nandrolone has a high ratio regarding fore represents an essential mechanism in AAS-induced anabolic versus androgenic properties and because of toxicity. There are primarily two major signaling pathways this it is one of the most frequently abused steroids. In that activate apoptotic cell death; the mitochondrial path- contrast to testosterone, nandrolone has higher binding way (intrinsic pathway) and the death receptor pathway affinity to the AR. However, while testosterone is (extrinsic pathway). Both signaling pathways result in acti- reduced in vivo by 5a-reductase to the high-affinity vation of caspase proteases, where caspase 3 and cas- ligand (DHT), nandrolone is pase 7 are the executioner caspases (Green and reduced to deliver a metabolite, 5a-dihydro-19-nortestos Llambi, 2015). Dysregulation of apoptosis is associated terone, with poorer AR affinity. In addition, nandrolone is with neurodegenerative disorders (Agostini et al., 2011), a poor substrate to in comparison to hence, long-term in vivo exposure to certain AASs might testosterone, and is not aromatized to to the increase the susceptibility to the onset or progression of same extent (Ryan, 1959). Furthermore, stanozolol has neurodegenerative diseases later in life. The present much weaker binding affinity to the AR compared to both study demonstrated testosterone, nandrolone, and tren- testosterone and nandrolone (Saartok et al., 1984), which bolone to induce apoptosis in cortical cultures in a dose- may be a possible explanation for the lower toxicity dependent and time-dependent manner, as indicated by demonstrated by stanozolol in the present cell assays. activation of the executioner caspases. In contrast, stano- However, even though stanozolol exhibits a low-binding zolol did not induce apoptosis, neither after single nor affinity, it is described as a potent activator of the AR repeated administration. Therefore, it is most likely that (Feldkoren and Andersson, 2005). Trenbolone is a stanozolol caused its toxic effects by initiation of necrosis, nandrolone-derivative and is neither aromatized nor con- i.e. a form of cell damage leading to premature death of verted by 5a-reductase, which makes it more androgenic cells in living tissue by autolysis. The AAS-induced cas- compared to nandrolone (Yarrow et al., 2010). Also, tren- pase 3/7 activation and the decrease in mitochondrial bolone has the highest relative binding affinity to the AR of function detected in the present study, indicate that the the four AASs studied (Scippo et al., 2002). Nevertheless, intrinsic pathway is responsible for the induced apoptosis. trenbolone is shown to have a lower potency to the AR These results are supported by similar findings from some compared to nandrolone (Death et al., 2004). investigators (Basile et al., 2013), whereas others have Based on the literature (Saartok et al., 1984; Scippo suggested testosterone to cause cell death via the extrin- et al., 2002), the order of relative binding affinity of the sic apoptotic pathway (Lopes et al., 2014). Overlap AASs in the present study is trenbolone > nan- between the two signaling pathways has been reported. drolone > testosterone > stanozolol. This ranking is sim- Therefore, specific stimuli may cause a particular type ilar, but not quite comparable, to the AAS toxicity shown in of cell death dependent on the cell type investigated this study, which suggests that the degree of physical and neighboring cells present (Green and Llambi, 2015). binding to the receptor does not fully explain the biological In order to establish whether the AAS-induced toxic activity of the AAS. This suggests that nandrolone may effects were mediated by the AR or not, the cells were induce its effects by an alternative mechanism other than co-treated with the AR-antagonist flutamide. Flutamide that through the classical AR pathway. Specific ligand- prevented the alterations detected in all assays used in induced effects are proposed to depend on how the the present study (MTT metabolism, LDH release and AAS-AR complex interacts with co-regulators and other caspase 3/7 activity) after single-dose treatment with the transcription factors (Holterhus et al., 2002) resulting in AASs. This strongly supports the single-dose effects to enhanced or suppressed transcription of specific genes. be mediated by the AR. Moreover, dependent on type In addition, AASs are suggested to cause non-genomic of AAS studied flutamide was able to either inhibit or effects either by activation of membrane-bound AR decrease the effects detected after repeated treatment. (Foradori et al., 2008) or by modulation of receptors in Flutamide abolished the caspase 3/7 activation after other signaling systems (Bitran et al., 1993; Yang et al., repeated treatment, suggesting the induction of the 2005). apoptotic cascade to be dependent on AR-activation. In the present study stanozolol was the AAS that Effects caused by testosterone or stanozolol on MTT deviated the most from the other steroids investigated. metabolism and LDH release were inhibited by Stanozolol only affected the cells after repeated flutamide, while effects induced by nandrolone and administration with the highest concentration (100 lM) trenbolone were decreased, but not to control levels. A used and was the only AAS that seemed to induce possible explanation could be that the concentration of necrosis. Differences between stanozolol and other flutamide used was too low to compete with the highest types of AASs are previously noted in the literature with S. Zelleroth et al. / Neuroscience 397 (2019) 172–183 181 regard to psychological behavior (Breuer et al., 2001). manuscript. Erik Nylander participated in performing the The present results suggest that stanozolol probably tissue collection, cell culturing, microscopy, and exerts its effects on the brain by a different molecular contributed with statistical advice. Fred Nyberg mechanism compared to the other AASs studied, which contributed with theoretical background, and preparing may explain various behavioral profiles recorded for these the manuscript. Alfhild Gro¨ nbladh contributed with steroids. theoretical background and study design, participated in We believe the present study provides important performing the tissue collection, cell culturing, evaluation information regarding the toxic effects of different AASs, of results, and preparing the manuscript. Mathias doses and administration regimens. Furthermore, we Hallberg contributed with theoretical background, study would like to highlight some potential limitations of the design, evaluation of results, and writing the manuscript. study. The model used (rat embryonical cortical cell cultures) was selected based on its advantages to cell FUNDING lines. Primary cell cultures retain many of the physiological and biochemical features, making them We gratefully acknowledge support from the Kjell och useful to study pathways or isolated mechanisms of Ma¨ rta Beijer Foundation; the Swedish Research Council interest. However, the embryonic cultures do not reflect (grant number 9459); and Swedish Brain Foundation. the intact CNS and therefore it is desirable that complementary in vivo studies are conducted in the DECLARATION OF INTEREST future to elucidate this issue. The AAS concentrations used in the present study (10, 30, 100 lM) were chosen The authors display no conflicts of interest. based on previous literature, which have shown that micro-molar but not nano-molar concentrations cause REFERENCES toxic effects in cultured cells (Estrada et al., 2006; Basile et al., 2013). These concentrations should mimic Agostini M, Tucci P, Melino G (2011) Cell death pathology: the supraphysiological doses used by AAS-users, but perspective for human diseases. Biochem Biophys Res Commun 414:451–455. cannot be interpreted as the actual concentration the Althobiti SD, Alqurashi NM, Alotaibi AS, Alharthi TF, Alswat KA brain is exposed to. The exposure times used in the study (2018) Prevalence, attitude, knowledge, and practice of anabolic (24 h and 3-days) were chosen to investigate an acute androgenic (AAS) use among gym participants. Mater effect in comparison with repeated treatment. The data Sociomed 30:49–52. from this relatively short exposure cannot be directly Basile JR, Binmadi NO, Zhou H, Yang YH, Paoli A, Proia P (2013) transferred to years of AAS abuse, however, our results Supraphysiological doses of performance enhancing anabolic- androgenic steroids exert direct toxic effects on neuron-like cells. give insight on mechanistic differences between treat- Front Cell Neurosci 7:69. ment regimens under controlled circumstances. Bitran D, Kellogg CK, Hilvers RJ (1993) Treatment with an anabolic- In conclusion, it is evident from the present study that androgenic steroid affects anxiety-related behavior and alters the the structurally diverse AASs induced toxic effects in sensitivity of cortical GABAA receptors in the rat. Horm Behav primary cortical cell cultures, but to different extents. 27:568–583. These toxic effects were primarily mediated through the Bjornebekk A, Walhovd KB, Jorstad ML, Due-Tonnessen P, Hullstein IR, Fjell AM (2017) Structural brain imaging of long-term anabolic- AR pathway, although, nandrolone and trenbolone might androgenic steroid users and nonusing weightlifters. Biol induce effects independent from the AR at high Psychiatry 82:294–302. concentrations. These observed alterations may Breuer ME, McGinnis MY, Lumia AR, Possidente BP (2001) contribute to the decline in learning and memory later in Aggression in male rats receiving anabolic androgenic steroids: life that former AAS abusers have reported. The effects of social and environmental provocation. Horm Behav mechanism by which AASs exert their toxic effects 40:409–418. resulting in cognitive deficiency is still a question that Caraci F, Pistara V, Corsaro A, Tomasello F, Giuffrida ML, Sortino MA, Nicoletti F, Copani A (2011) Neurotoxic properties of the needs to be further addressed. Hence, the correlation anabolic androgenic steroids nandrolone and between toxic effects monitored in primary cortical methandrostenolone in primary neuronal cultures. J Neurosci cultures from rat described herein, the underlying Res 89:592–600. molecular mechanisms, and the cognitive impairments Clark AS, Henderson LP (2003) Behavioral and physiological in rodents and humans that appear at supra- responses to anabolic-androgenic steroids. Neurosci Biobehav physiological concentrations of AASs is not clear. Rev 27:413–436. Clark AS, Mitre MC, Brinck-Johnsen T (1995) Anabolic-androgenic steroid and adrenal steroid effects on hippocampal plasticity. ACKNOWLEDGMENT Brain Res 679:64–71. Cohen J, Collins R, Darkes J, Gwartney D (2007) A league of their The authors would like to thank Shanti Diwakarla for proof own: demographics, motivations and patterns of use of 1,955 reading and providing valuable input on the manuscript. male adult non-medical users in the United States. J Int Soc Sports Nutr 4:12. Cunningham RL, Giuffrida A, Roberts JL (2009) induce AUTHOR CONTRIBUTION dopaminergic neurotoxicity via caspase-3-dependent activation of protein kinase Cdelta. Endocrinology 150:5539–5548. Sofia Zelleroth contributed to the theoretical background Daly RC, Su TP, Schmidt PJ, Pickar D, Murphy DL, Rubinow DR and study design, performed the tissue collection, cell (2001) Cerebrospinal fluid and behavioral changes after culturing, cell assays, microscopy, statistical administration: preliminary findings. Arch calculations, evaluation of results, and writing the Gen Psychiatry 58:172–177. 182 S. Zelleroth et al. / Neuroscience 397 (2019) 172–183

Death AK, McGrath KC, Kazlauskas R, Handelsman DJ (2004) Kindlundh AM, Isacson DG, Berglund L, Nyberg F (1999) Factors is a potent androgen and progestin. J Clin associated with adolescent use of doping agents: anabolic- Endocrinol Metab 89:2498–2500. androgenic steroids. Addiction 94:543–553. Diwakarla S, Nylander E, Gronbladh A, Vanga SR, Khan YS, Kouvelas D, Pourzitaki C, Papazisis G, Dagklis T, Dimou K, Kraus Gutierrez-de-Teran H, Ng L, Pham V, Savmarker J, Lundback MM (2008) Nandrolone abuse decreases anxiety and impairs T, Jenmalm-Jensen A, Andersson H, Engen K, Rosenstrom U, memory in rats via central androgenic receptors. Int J Larhed M, Aqvist J, Chai SY, Hallberg M (2016) Binding to and Neuropsychopharmacol 11:925–934. inhibition of insulin-regulated aminopeptidase by macrocyclic Le Greves P, Huang W, Johansson P, Thornwall M, Zhou Q, Nyberg disulfides enhances spine density. Mol Pharmacol 89:413–424. F (1997) Effects of an anabolic-androgenic steroid on the Estrada M, Varshney A, Ehrlich BE (2006) Elevated testosterone regulation of the NMDA receptor NR1, NR2A and NR2B subunit induces apoptosis in neuronal cells. J Biol Chem mRNAs in brain regions of the male rat. Neurosci Lett 226:61–64. 281:25492–25501. Lee DK, Chang C (2003) Molecular communication between Feldkoren BI, Andersson S (2005) Anabolic-androgenic steroid androgen receptor and general transcription machinery. J interaction with rat androgen receptor in vivo and in vitro: a Steroid Biochem Mol Biol 84:41–49. comparative study. J Steroid Biochem Mol Biol 94:481–487. Lew M (2007) Good statistical practice in pharmacology. Problem 2. Foradori CD, Weiser MJ, Handa RJ (2008) Non-genomic actions of Br J Pharmacol 152:299–303. androgens. Front Neuroendocrinol 29:169–181. Lindqvist Bagge AS, Rosen T, Fahlke C, Ehrnborg C, Eriksson BO, Goldman A, Basaria S (2018) Adverse health effects of androgen Moberg T, Thiblin I (2017) Somatic effects of AAS abuse: A 30- use. Mol Cell Endocrinol 464:46–55. years follow-up study of male former power sports athletes. J Sci Green DR, Llambi F (2015) Cell Death Signaling. Cold Spring Harb Med Sport 20:814–818. Perspect Biol 7. Lopes RA, Neves KB, Pestana CR, Queiroz AL, Zanotto CZ, Gronbladh A, Johansson J, Nostl A, Nyberg F, Hallberg M (2013) GH Chignalia AZ, Valim YM, Silveira LR, Curti C, Tostes RC (2014) improves spatial memory and reverses certain anabolic Testosterone induces apoptosis in vascular smooth muscle cells androgenic steroid-induced effects in intact rats. J Endocrinol via extrinsic apoptotic pathway with mitochondria-generated 216:31–41. reactive oxygen species involvement. Am J Physiol Heart Circ Gronbladh A, Johansson J, Nyberg F, Hallberg M (2014) Physiol 306:H1485–H1494. Administration of growth hormone and nandrolone decanoate Ma F, Liu D (2015) 17beta-trenbolone, an anabolic-androgenic alters mRNA expression of the GABAB receptor subunits as well steroid as well as an environmental hormone, contributes to as of the GH receptor, IGF-1, and IGF-2 in rat brain. Growth Horm neurodegeneration. Toxicol Appl Pharmacol 282:68–76. IGF Res 24:60–66. Magnusson K, Hanell A, Bazov I, Clausen F, Zhou Q, Nyberg F (2009) Hallberg M (2011) Impact of anabolic androgenic steroids on Nandrolone decanoate administration elevates hippocampal neuropeptide systems. Mini Rev Med Chem 11:399–408. prodynorphin mRNA expression and impairs Morris water maze Hallberg M, Johansson P, Kindlundh AM, Nyberg F (2000) Anabolic- performance in male rats. Neurosci Lett 467:189–193. androgenic steroids affect the content of substance P and Orlando R, Caruso A, Molinaro G, Motolese M, Matrisciano F, Togna substance P(1–7) in the rat brain. Peptides 21:845–852. G, Melchiorri D, Nicoletti F, Bruno V (2007) Nanomolar T.M. Heffernan L. Battersby P. Bishop T.S., O.N., Everyday Memory concentrations of anabolic-androgenic steroids amplify Deficits Associated with 2015 Anabolic-Androgenic Steroid Use in excitotoxic neuronal death in mixed mouse cortical cultures. Regular Gymnasium Users, The Open Psychiatry Journal, 9: 1-6. Brain Res 1165:21–29. Holterhus PM, Piefke S, Hiort O (2002) Anabolic steroids, Pomara C, Neri M, Bello S, Fiore C, Riezzo I, Turillazzi E (2015) testosterone-precursors and virilizing androgens induce distinct Neurotoxicity by synthetic androgen steroids: oxidative stress, activation profiles of androgen responsive promoter constructs. J apoptosis, and neuropathology: A review. Curr Neuropharmacol Steroid Biochem Mol Biol 82:269–275. 13:132–145. Janjic MM, Stojkov NJ, Andric SA, Kostic TS (2012) Anabolic- Pope Jr HG, Kanayama G, Athey A, Ryan E, Hudson JI, Baggish A androgenic steroids induce apoptosis and NOS2 (nitric-oxide (2014) The lifetime prevalence of anabolic-androgenic steroid use synthase 2) in adult rat Leydig cells following in vivo exposure. and dependence in Americans: current best estimates. Am J Reprod Toxicol 34:686–693. Addict 23:371–377. Kanayama G, Brower KJ, Wood RI, Hudson JI, Pope Jr HG (2009) Ryan KJ (1959) Biological aromatization of steroids. J Biol Chem Anabolic-androgenic steroid dependence: an emerging disorder. 234:268–272. Addiction 104:1966–1978. Saartok T, Dahlberg E, Gustafsson JA (1984) Relative binding affinity Kanayama G, Hudson JI, Pope Jr HG (2008) Long-term psychiatric of anabolic-androgenic steroids: comparison of the binding to the and medical consequences of anabolic-androgenic steroid abuse: androgen receptors in skeletal muscle and in prostate, as well as a looming public health concern? Drug Alcohol Depend 98:1–12. to -binding globulin. Endocrinology 114:2100–2106. Kanayama G, Kean J, Hudson JI, Pope Jr HG (2013) Cognitive Scippo ML, Van de Weerdt C, Willemsen P, Francois JM, Rentier- deficits in long-term anabolic-androgenic steroid users. Drug Delrue F, Muller M, Martial JA, Maghuin-Rogister G (2002) Alcohol Depend 130:208–214. Detection of illegal growth promoters in biological samples using Kaufman MJ, Janes AC, Hudson JI, Brennan BP, Kanayama G, receptor binding assays. Anal Chim Acta 473:135–141. Kerrigan AR, Jensen JE, Pope Jr HG (2015) Brain and cognition Seitz J, Lyall AE, Kanayama G, Makris N, Hudson JI, Kubicki M, Pope abnormalities in long-term anabolic-androgenic steroid users. Jr HG, Kaufman MJ (2017) White matter abnormalities in long- Drug Alcohol Depend 152:47–56. term anabolic-androgenic steroid users: A pilot study. Psychiatry Kicman AT (2008) Pharmacology of anabolic steroids. Br J Res Neuroimaging 260:1–5. Pharmacol 154:502–521. Shokri S, Aitken RJ, Abdolvahhabi M, Abolhasani F, Ghasemi FM, Kindlundh AM, Bergstrom M, Monazzam A, Hallberg M, Blomqvist G, Kashani I, Ejtemaeimehr S, Ahmadian S, Minaei B, Naraghi MA, Langstrom B, Nyberg F (2002) Dopaminergic effects after chronic Barbarestani M (2010) Exercise and supraphysiological dose of treatment with nandrolone visualized in rat brain by positron nandrolone decanoate increase apoptosis in spermatogenic cells. emission tomography. Prog Neuropsychopharmacol Biol Basic Clin Pharmacol Toxicol 106:324–330. Psychiatry 26:1303–1308. Tanehkar F, Rashidy-Pour A, Vafaei AA, Sameni HR, Haghighi S, Kindlundh AM, Isacson DG, Berglund L, Nyberg F (1998) Doping Miladi-Gorji H, Motamedi F, Akhavan MM, Bavarsad K (2013) among high school students in Uppsala, Sweden: a presentation Voluntary exercise does not ameliorate spatial learning and of the attitudes, distribution, side effects, and extent of use. Scand memory deficits induced by chronic administration of nandrolone J Soc Med 26:71–74. decanoate in rats. Horm Behav 63:158–165. S. Zelleroth et al. / Neuroscience 397 (2019) 172–183 183

Westlye LT, Kaufmann T, Alnaes D, Hullstein IR, Bjornebekk A Yang P, Jones BL, Henderson LP (2005) Role of the alpha subunit in (2017) Brain connectivity aberrations in anabolic-androgenic the modulation of GABA(A) receptors by anabolic androgenic steroid users. Neuroimage Clin 13:62–69. steroids. Neuropharmacology 49:300–316. Wiren KM, Toombs AR, Semirale AA, Zhang X (2006) Osteoblast Yarrow JF, McCoy SC, Borst SE (2010) Tissue selectivity and and osteocyte apoptosis associated with androgen action in potential clinical applications of trenbolone (17beta-hydroxyestra- bone: requirement of increased Bax/Bcl-2 ratio. Bone 4,9,11-trien-3-one): a potent anabolic steroid with reduced 38:637–651. androgenic and estrogenic activity. Steroids 75:377–389.

(Received 20 September 2018, Accepted 23 November 2018) (Available online 28 November 2018)