Influence of the Anabolic-Androgenic Steroid Nandrolone on Cannabinoid

Neuropharmacology 50 (2006) 788e806 www.elsevier.com/locate/neuropharm

Inﬂuence of the anabolic-androgenic steroid nandrolone on cannabinoid dependence

Evelyne Ce´le´rier a, Therese Ahdepil b, Helena Wikander b, Fernando Berrendero a, Fred Nyberg b, Rafael Maldonado a,*

a Laboratori of Neurofarmacologia, Facultat de Cie´ncies de la Salut i de la Vida, Universitat Pompeu Fabra, C/Doctor Aiguader 80, 08003 Barcelona, Spain b Department of Pharmaceutical Biosciences, University of Uppsala, Box 591 Biomedicum, S751 24 Uppsala, Sweden Received 8 April 2005; received in revised form 29 November 2005; accepted 29 November 2005

Abstract

The identiﬁcation of the possible factors that might enhance the risk of developing drug addiction and related motivational disorders is crucial to reduce the prevalence of these problems. Here, we examined in mice whether the exposure to the anabolic-androgenic steroid nandrolone would affect the pharmacological and motivational effects induced by D9-tetrahydrocannabinol (THC), the principal psychoactive component of Cannabis sativa. Mice received nandrolone using pre-exposure (during 14 days before THC treatment) or co-administration (1 h before each THC injection) procedures. Both nandrolone treatments did not modify the acute antinociceptive, hypothermic and hypolocomotor effects of THC or the development of tolerance after chronic THC administration. Nandrolone pre-exposure blocked THC- and food-induced conditioned place preference and increased the somatic manifestations of THC withdrawal precipitated by the CB1 cannabinoid antagonist rimonabant (SR141617A). The aversive effects of THC were not changed by nandrolone. Furthermore, nandrolone pre-exposure attenuated the anxiolytic-like effects of a low dose of THC without altering the anxiogenic-like effects of a high dose in the lit/dark box, open ﬁeld and elevated plus-maze. Biochemical experiments showed that chronic nandrolone treatment did not modify CB1 receptor binding and GTP-binding protein activation in the caudate-putamen and cerebellum. Taken together, our results suggest that chronic nandrolone treatment alters behavioural responses related to cannabinoid addictive properties. Ó 2005 Elsevier Ltd. All rights reserved.

Keywords: Anabolic-androgenic steroid; Nandrolone; THC; Dependence; Anxiety; Mice

1. Introduction (AAS), used as doping substances, as a risk factor for the consumption of drugs of abuse such as cannabis. Considerable efforts are now devoted to the study of risk Derivates from Cannabis sativa, whose main psychoactive factors that may increase individual vulnerability to drugs of constituent is the D9-tetrahydrocannabinol (THC) (Mechoulam abuse. An important factor that could enhance the risk for et al., 1970), are today the most consumed illicit drugs world- an addictive process and motivational related disorders is the wide (Smart and Ogborne, 2000). Cannabinoid effects are exposure to pharmacological compounds able to modify the mediated by the activation of two receptors, the CB1 cannabi- physiological equilibrium of the rewarding system (Koob noid receptor highly abundant in the central nervous system, and Le Moal, 2001). The purpose of the present study was and the CB2 receptor, mainly located in the cells of the to investigate the effects of anabolic-androgenic steroids immune system (Ameri, 1999). The cannabinoid system is closely related to several neurobiological pathways involved in motivation, mood and addictive behaviours, particularly the dopaminergic and the opioid system (for reviews, see * Corresponding author. Tel.: þ34 93 542 28 45; fax: þ34 93 542 28 02. Manzanares et al., 1999; Maldonado and Rodrı´guez de E-mail address: [email protected] (R. Maldonado). Fonseca). Accordingly, cannabinoids have been reported to

0028-3908/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuropharm.2005.11.017 E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 789 produce behavioural and biochemical changes in animals treatment with THC were evaluated. In order to mimic human similar to other drugs of abuse and leading to dependence abuse regimes, supraphysiologic doses of nandrolone were ad- (Maldonado and Rodrı´guez de Fonseca, 2002). However, ministered using two different protocols, i.e. co-administration one important aspect of marijuana activity is the complexity during THC treatment and pre-exposure before THC of evoked emotional/affective responses and the possibility treatment. of dual euphoric-dysphoric effects (Halikas et al., 1985). In- deed, reinforcing and aversive as well as anxiolytic and anxio- 2. Methods genic effects have been reported after the administration of THC and other cannabinoid agonists (Chaperon and Thie´bot, 2.1. Animals 1999; Ghozland et al., 2002; Berrendero and Maldonado, e 2002). Albino male CD-1 mice (CRIFFA, France) weighing 20 22 g were housed five per cages and maintained at a controlled temperature AAS are synthetic derivates of testosterone widely used in (21 1 C) and humidity (55 10%) environment. The mice were given ac- the clinic as androgen replacement therapy and as chemother- cess to food and water ad libitum. Lighting was maintained at 12-h cycles (on apy for certain types of cancer (Wilson and Griffin, 1980). at 7 a.m. and off at 7 p.m.). All the experiments were performed during the During the last five decades, AAS has been used at doses light phase of the dark/light cycle. The animals were habituated to the exper- 10e100 times the therapeutic range by many athletes and imental room and handled for 1 week before the start of the experiments. All animal procedures met the guidelines of the National Institute of Health de- bodybuilders to enhance their physical performance, increase tailed in the ‘‘Guide for the Care and use of Laboratory Animals’’, the Euro- muscle mass and intensify training regimens (Wilson, 1988; pean Communities directive 86/609/EEC regulating animal research and were Lukas, 1993; Yesalis and Bahrke, 1995). The chronic use of approved by the Local Ethical Committees. All experiments were performed high doses of AAS has been reported to cause several physical with the investigators being blind to the treatment conditions. and psychological side-effects such as liver dysfunction, coro- nary heart disease, reproductive dysfunction, acne, depression, 2.2. Drugs personality changes and aggressive behaviour (Williamson and Young, 1992; Pope and Katz, 1994; Bahrke et al., THC and the AAS nandrolone decanoate were purchased from Sigma (Poole, UK). The selective CB1 cannabinoid receptor antagonist 1996). Interestingly, a concurrent abuse of AAS has also SR141617A (rimonabant) was generously provided by Sanofi Research been reported among addicts and others not connected to (France). Nandrolone was dissolved in vehicle (10% ethanol/10% cremophor sports (DuRant et al., 1993; Lukas, 1993; Yesalis and Bahrke, EL/80% distilled water) and injected intramuscularly (i.m.) in a volume of 1995; Kindlundh et al., 1999). Several studies have suggested 2 ml/g body weight. THC was dissolved in vehicle (5% ethanol/5% cremophor the association between use of AAS and consumption of alco- EL/90% distilled water) and injected intraperitoneally (i.p.) in a volume of 10 ml/g body weight. Rimonabant was dissolved in vehicle (10% ethanol/ hol, tobacco and illicit drugs, such as cannabis, opiates, am- 10% cremophor EL/80% distilled water) and injected i.p. in a volume of phetamine and ecstasy (DuRant et al., 1993; Yesalis and 20 ml/g body weight. Control mice received equivolumic vehicle injections. Bahrke, 1995; Kindlundh et al., 1999; Kanayama et al., 2003). Based on these clinical and epidemiological studies, 2.3. Behavioural experiments AAS exposure has been proposed to serve as a ‘‘gateway’’ for the misuse of other drugs of abuse (Arvary and Pope, The effects of nandrolone treatment on acute and chronic THC effects 2000). In line with this hypothesis, recent animal studies were evaluated. Two different protocols (pre-exposure or co-administration) have shown that AAS can be self-administered by laboratory were used for nandrolone administration in an attempt to mimic conditions similar to those used during human abuse regimes: drug intake concomitant animals (Ballard and Wood, 2005). AAS have also been re- to nandrolone use or drug intake with a past of long-term consumption of nan- ported to evoke neurobiochemical and behavioural alterations drolone. In the first protocol (pre-exposure), nandrolone was chronically ad- related to dependence, mood and motivation in rodents espe- ministered once daily during 14 days before starting THC treatment. In the cially by affecting the endogenous opioid and dopamine sys- second protocol (co-administration), nandrolone was administered 1 h before tem (Lukas, 1993; Menard et al., 1995; Clark et al., 1996; each THC injection. The i.m. injections of nandrolone were given alternatively ` in the left and the right hind leg when repeated administration was required. A Le Greves et al., 1997; Johansson et al., 1997, 2000a,b; supra-therapeutic dose of nandrolone (15 mg/kg) was chosen (i) because it Thiblin et al, 1999; Hallberg et al., 2000; Schlussman et al., mimics the dose self-administered by heavy nandrolone abusers (Williamson 2000; Ce´le´rier et al., 2003; Kindlundh et al., 2003 and for re- and Young, 1992) and (ii) because it has been previously shown to induce bio- view, see Clark and Henderson, 2003). chemical changes in the endogenous opioid and dopamine system in rodents Interactions between sex-steroids and the endocannabinoid (Johansson et al., 1997). system on reproduction and endocrine responses have been well documented (Gonzalez et al., 2000; MacCarrone et al., 2.3.1. Acute effects of THC The acute effects induced by different doses of THC (5, 10 or 20 mg/kg, 2000; Corchero et al., 2001). However, the possible influence i.p.) on nociception, locomotion, and rectal temperature were evaluated in of these interactions on the behavioural responses mediated by each experimental group. the central nervous system remains to be elucidated. In the Two nociceptive tests were used, the tail-immersion and the hot-plate. In present study, we use different behavioural models in mice the tail-immersion test, mice were gently placed in a restrainer cylinder. The to evaluate the effects of the AAS nandrolone on several nociceptive threshold was assessed as described previously (Janssen et al., 1963), by measuring the time to withdraw the tail immersed in a thermostated THC behavioural responses related to its addictive properties. water bath (50 0.1 C) (Clifton-Scientific Instruments, England), with a cut For this purpose, the effects of nandrolone exposure on the be- off latency of 15 s to prevent tissue damage. Nociceptive responses were also havioural and somatic responses induced by acute and chronic measured using a hot-plate analgesia meter (Colombus, Ohio, USA). A glass 790 E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 cylinder (19 cm high, 19 cm diameter) was used to keep the mice on the heat- 523 s; mean time spent in dotted compartment: 553 s). After this session, ed surface of the plate, which was maintained at a temperature of 52 0.1 C. mice were randomized for pairing to drug or vehicle administration and for The latency for jumping was evaluated as the nociceptive threshold, with a cut- assignment to a compartment. Care was taken to ensure that treatments off time of 240 s. were counterbalanced as closely as possible between compartments to ensure Locomotor responses were evaluated using locomotor activity boxes an unbiased procedure. To evaluate THC rewarding effects, mice were exposed (9 20 11 cm; Imetronic, Lyon, France). The boxes contained a line of to a priming injection of THC (1 mg/kg) 24 h before starting the conditioning photocells 2 cm above the floor to measure horizontal movements and another phase, as previously reported (Valjent and Maldonado, 2000). During the con- line located 6 cm above the floor to measure vertical activity (rearing). On the ditioning phase, mice were alternatively treated for 10 consecutive days with experimental day, the mice were individually placed in the boxes and the am- THC (1 mg/kg) or vehicle. To evaluate THC aversive effects, mice were not bulatory, horizontal (ambulatory movements plus small movements) and ver- exposed to any priming injection and were alternatively treated for 10 consec- tical activities were recorded during 10 min in a low luminosity environment utive days with THC (5 mg/kg) or vehicle during the conditioning phase as (5e15 lx). previously reported (Valjent and Maldonado, 2000). Mice in the drug group Rectal temperature was measured in each mouse using an electronic ther- received THC on days 1, 3, 5, 7 and 9 in their assigned compartment, and ve- mocouple flexible probe (Panlab, Barcelona, Spain). The probe was lubricated hicle on days 2, 4, 6, 8 and 10 in the opposite compartment. Control mice re- and placed 3 cm into the rectum of the mice for 20 s before the temperature ceived vehicle in both compartments alternatively. Doors matching the walls was recorded. of the compartment allowed the confinement of the mice for 45 min immedi- In order to habituate the animals to the test environment and to obtain a sta- ately after THC or vehicle injection. Finally, the test phase was conducted 24 h ble baseline, tail-immersion response was measured for 2 days before the ex- after the last conditioning session exactly as was the preconditioning phase, periment. On the experimental day, basal tail-immersion latency and rectal i.e. with free access to both compartments for 20 min. temperature were evaluated before starting the treatment. Locomotor activity Food reward was evaluated in animals with limited access to food and ha- was evaluated 20 min after THC injection during 10 min. The tail-immersion bituated to sucrose consumption for 5 days before the test. A fixed amount of test was performed 30 min after THC injection and was immediately followed food equivalent to 15% of mouse body weight was applied per day, as previ- by the hot-plate test. Rectal temperature was measured 45 min after THC ously reported (Elmer et al., 1995). Animals were maintained on the same diet injection. throughout the experimental period. During the preconditioning phase, mice were placed in the middle of the neutral area and their location was recorded for 18 min. Mice conditioned to food had free access to it (normal mouse food 2.3.2. Tolerance to THC effects plus sucrose) in the confined compartment on days 1, 3 and 5, and had no ac- As in the previous experiment, the basal tail-immersion response was eval- cess to food in the other compartment on days 2, 4 and 6. Control animals were uated for 2 days before the beginning of THC injections. THC (20 mg/kg, i.p.) alternatively exposed to both compartments during this phase and had no ac- was administered twice a day (10 a.m. and 7 p.m.) during 5 days. Every morn- cess to food in the assigned compartment. Mice were confined in the condi- ing before any THC injection, body weight, tail-immersion latency and rectal tioning compartments during a period of 20 min. The testing phase was temperature were measured. Locomotor activity was recorded 20 min after conducted 24 h after the final conditioning session exactly as the precondition- each morning THC injection for 10 min. Tail-immersion and rectal tempera- ing phase, i.e. free access to each compartment for 18 min. ture were measured 30 and 45 min, respectively, after each morning THC in- In both THC and food experiments, a score was calculated for each mouse jection. The hot-plate test was only performed the first and last day of THC as the difference between the post-conditioning and the pre-conditioning time treatment (day 1 and 5), immediately after the tail-immersion exposure. spent in the drug-paired compartment.

2.3.3. Rimonabant-precipitated THC withdrawal 2.3.5. THC effects on anxiety-like responses Mice chronically treated with THC (20 mg/kg, i.p., twice daily during The anxiolytic- and anxiogenic-like effects of THC were evaluated by 5 days) for the evaluation of tolerance to its pharmacological effects received using three experimental procedures: lit/dark box, open-field and elevated an additional THC morning injection on day 6 (20 mg/kg, i.p.). Four hours lat- plus-maze. Behavioural tests were performed 30 min after THC administration. er, cannabinoid withdrawal was precipitated by injection of the selective CB1 Anxiolytic-like effects of THC were induced by using a low dose (0.2 mg/kg), receptor antagonist rimonabant (10 mg/kg, i.p.). The mice were placed individ- whereas a higher dose of THC (7.5 mg/kg) induced the anxiogenic-like effects. ually into test chambers to evaluate the behavioural signs of withdrawal. The In the lit/dark box procedure (Filliol et al., 2000), mice were individually chambers consisted of transparent round plastic boxes (30 cm in diameter and exposed for 5 min to a box consisting of a small dark compartment 50 cm in height) with a white floor. Somatic signs of withdrawal were evalu- (15 20 25 cm) with black floor and walls, dimly lit (5e10 lx) connected ated 15 min before and 45 min after rimonabant challenge. The number of wet by a 4 cm long tunnel to a larger lit compartment (30 20 25) with white dog shakes, front paw tremors and sniffing were counted. Ptosis was scored, 1 floor and walls, exposed to intense illumination (500 lx). Lines were drawn on for appearance and 0 for non-appearance, within each 5 min time period. Tak- the floor of both compartments to allow measurements of locomotor activity ing into account all the individual signs, a global withdrawal score was calcu- by counting the numbers of squares (5 5 cm) crossed. These lines also di- lated for each mouse by giving each individual sign a relative weight as vide the lit compartment into three equal zones, from the tunnel to the opposite previously reported (Hutcheson et al., 1998). wall, designated as proximal, central and distal zones. To start the experiment, each mouse was placed in the dark compartment facing the lit area. The pa- 2.3.4. Rewarding effects of THC and food rameters recorded were latency for first entrance into the lit area, time spent The rewarding effects of THC and food were evaluated by using the con- in each compartment, number of squares crossed in each compartment, loco- ditioned place preference paradigm. The place preference apparatus consisted motor activity (number total of squares crossed) and number of entries in each of two different cubic compartments (15 15 15 cm) separated by a trian- compartment and into each zone of the lit compartment. gular neutral area (15 cm per side). The two compartments had visual and tac- The open-field consisted of a rectangular area (70 cm wide, 90 cm long and tile differences, where one compartment had dotted walls and rough floor and 60 cm high) brightly illuminated from the top (500 lx). A total of 63 squares the other compartment had striped walls and smooth floor. The movement and (10 10 cm) were drawn with black lines on the white floor dividing the field location of the mice were recorded by computerized monitoring software into central and peripheral areas. To start the experiment, each mouse was (SmartÒ Videotrack, Panlab., Barcelona, Spain) with images relayed from placed in the central area of the field. The following events were recorded dur- a camera placed above the apparatus. The place preference conditioning ing an observation period of 5 min: latency to go out from the central area, schedule consisted of three phases. During the preconditioning phase, the numbers of squares crossed in each area, number of rearing, number of entries mouse was placed in the middle of the neutral area and allowed to explore into the central area and locomotor activity (total number of squares crossed). both compartments, and the time spent in each compartment was measured The elevated plus-maze consisted of two arms (length 30 cm; width 7 cm) during 20 min. No initial preference or aversion was revealed for any compart- forming the shape of a plus situated on a platform 80 cm above ground. One of ment in any of the experiments (mean time spent in striped compartment: the arms was open and the other closed (14 cm). Number of entries and time E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 791 spent into the open and closed arm were recorded during an observation period (nandrolone/vehicle and THC/vehicle treatments). When required, the three- of 5 min. way ANOVAwas followed by corresponding two-way and one-way ANOVAs. Statistical analyses of data from CB1 receptor binding and activation of 2.4. Biochemical experiments GTP-binding protein were assessed by using the unpaired two-tailed Student’s t-test. Statistical significance criterion was P < 0.05 in all cases. CB1 receptor binding and activation of GTP-binding protein was investigated in two brain areas related to cannabinoid behavioural responses and dependence in animals: the caudate-putamen and the cerebellum (Navarro et al., 3. Results 2001; Castan˜e´ et al., 2004). The studies were performed in mice chronically treated with nandrolone (15 mg/kg) or vehicle once daily during 14 days. On day 15, mice were decapitated and their brains were quickly removed 3.1. Effects of nandrolone exposure on acute THC and frozen by immersion in 2-methyl-butane surrounded by dry ice. All sam- responses ples were stored at 80 C during a similar short period of time until pro- cessed for analysis. The influence of nandrolone exposure on the acute pharmacological effects of THC (5, 10, and 20 mg/kg, i.p.) was eval- 2.4.1. Nandrolone effects on cannabinoid CB1 receptor uated by measuring the responses induced in nociception, binding The protocol used was basically the method described by Herkenham et al. body temperature and locomotor activity. Nandrolone (1991). Briefly, brain structures were incubated for 2.5 h at 37 C in a buffer (15 mg/kg, i.m.) or vehicle was administered using the two containing 50 mM Tris with 5% bovine serum albumin (fatty acid-free), pH different protocols: co-administration and pre-exposure. Con- 3 7.4, and 10 nM [ H]CP-55,940 (Perkin Elmer Life Sciences) prepared in the trol experiments showed that nandrolone had no intrinsic same buffer, in the absence or presence of 10 mM non-labelled CP-55,940 effect in any of the parameters studied when injected in (Tocris) to determine total and non-specific binding, respectively. Following this incubation, slides were washed in 50 mM Tris buffer with 1% bovine combination with vehicle. serum albumin (fatty acid-free), pH 7.4, for 4 h (2 2h)at0 C, dipped in ice-cold distilled water, then dried under a stream of cool dried air. Autoradio- 3.1.1. THC antinociceptive effects grams were generated by apposing the labelled tissues, together with autoradiographic standards ([3H]Microscales, Amersham), to tritium-sensitive film Chronic nandrolone pre-exposure for 14 days did not sig- (Hyperfilm-[3H], Amersham) for a period of 10 days and developed for nificantly modify THC acute antinociceptive effects in the 4 min at 20 C. Developed films were analysed and quantified in a computer- tail-immersion and the hot-plate test (Table 1). In the tail- ized image analysis system (MCID, St. Catharines, Ontario, Canada) using the 3 immersion test, significant antinociceptive effects of 10 and standard curve generated from H standards. 20 mg/kg of THC were observed in both vehicle and nandrolone pre-exposed mice. Two-way ANOVA revealed significant 2.4.2. Nandrolone effects on cannabinoid CB1 receptor activation of GTP-binding protein THC effect (F(3,87) ¼ 23.60, P < 0.001), but no nandrolone The protocol used was basically the method described by Sim et al. (1995). effect (F(1,87) ¼ 1.03, NS) and no THC/nandrolone interac- Briefly, the brain structure were rinsed in assay buffer (50 mM Tris, 3 mM tion (F(3,87) ¼ 1.44, NS). In the hot-plate test, significant MgCl2, 0.2 mM EGTA, 100 mM NaCl, and 0.5% bovine serum albumin fatty antinociceptive effects of 10 and 20 mg/kg were found in ve- acid-free, pH 7.4) at 25 C for 10 min, then pretreated for 15 min with an ex- hicle, but not in nandrolone pre-exposed mice. Two-way cess concentration (2 mM) of GDP (Sigma Chemical Co., Madrid, Spain) in assay buffer containing 0.04 nM [35S]GTPgS (Amersham), 2 mM GDP, and ANOVA showed significant THC effect (F(3,87) ¼ 8.22, 5 mM WIN 55 212-2 (Sigma Chemical Co., Madrid, Spain). Basal activity P < 0.001), no nandrolone effect (F(1,87) ¼ 0.07, NS) and was assessed in the absence of agonist, whereas non-specific binding was mea- no THC/nandrolone interaction (F(3,87) ¼ 1.18, NS). sured in the presence of 10 mM unlabelled GTPgS. Slices were rinsed twice in Similarly, nandrolone co-administration did not modify 50 mM Tris buffer, pH 7.4, at 4 C and deionized once in water, then dried un- THC acute antinociceptive effects as evaluated in the tail- der a stream of cool dry air. Autoradiograms were generated by apposing the labelled tissues to film (Kodak Biomax MR) for a period of 2 days and devel- immersion and the hot-plate test (Table 1). In the tail-immer- oped for 4 min at 20 C. Developed films were analysed and quantified in sion test, significant antinociceptive effects of 5, 10 and a computerized image analysis system (MCID, St. Catharines, Ontario, 20 mg/kg of THC were observed in both vehicle and nandro- Canada). lone co-treated mice. Two-way ANOVA revealed significant THC effect (F(3,105) ¼ 26.11, P < 0.001), but no nandrolone 2.5. Statistical analysis effect (F(1,105) ¼ 0.03, NS) nor THC/nandrolone interaction (F(3,105) ¼ 0.40, NS). In the hot-plate test, significant antino- The data obtained for acute THC pharmacological responses, food and THC effects in the place conditioning paradigm, rimonabant-precipitated ciceptive effects of 10 and 20 mg/kg of THC were observed in THC withdrawal and THC effects on anxiety-like response were analysed by both vehicle and nandrolone co-treated mice. Two-way using a two-way ANOVA with nandrolone/vehicle and THC/vehicle treatment ANOVA revealed significant THC effect (F(3,104) ¼ 11.25, as factors of variation. One-way ANOVA was used to reveal the main THC or P < 0.001), but no nandrolone effect (F(1,104) ¼ 3.81, NS) nandrolone effects, followed when appropriate by corresponding post hoc com- or THC/nandrolone interaction (F(3,104) ¼ 0.22, NS). parisons (Dunnett’s test) to indicate groups significantly different from control. In addition, individual comparisons of time spent in the drug-paired compartment during preconditioning and test phases were made in place conditioning 3.1.2. THC hypothermic effects experiments in each experimental group by using the paired two-tailed Chronic nandrolone pre-exposure during 14 days and nan- Student’s t-test. Data from the study of tolerance to THC antinociception, hypothermia and drolone co-administration did not modify THC acute hypother- hypolocomotor effects were analysed by using a three-way ANOVA with time mic effects (Table 1). Similar significant hypothermic effects of as the within subjects factor of variation and two between subjects factors THC were observed in vehicle and nandrolone exposed mice 792 E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806

Table 1 Effects of nandrolone on the acute pharmacological effects of D9-tetrahydrocannabinol (THC) THC doses Tail-withdrawal Hot-plate Rectal Total horizontal latency (s) jumping (s) temperature locomotor activity ( C) (counts) Pre-exposure 0 mg/kg VEH 2.51 0.12 47 4 1.13 0.59 265 21 ND 2.50 0.17 53 12 1.81 0.65 230 23 5 mg/kg VEH 2.70 0.11 57 5 1.46 0.86 258 25 ND 2.55 0.16 66 8 1.56 0.74* 204 32 10 mg/kg VEH 4.84 0.36** 108 15** 5.56 1.08** 96 12** ND 3.73 0.46* 92 9 3.32 1.04** 149 23* 20 mg/kg VEH 5.38 0.37** 99 14** 7.52 0.92** 104 12** ND 5.52 0.36** 100 16 7.87 0.84** 94 9**

Co-administration 0 mg/kg VEH 2.22 0.06 48 3 1.24 0.34 206 13 ND 2.48 0.13 58 5 0.78 0.26 213 13 5 mg/kg VEH 3.76 0.50** 72 11 3.14 1.29** 135 35* ND 3.96 0.39** 107 17 4.83 0.93** 106 23** 10 mg/kg VEH 4.89 0.39** 107 12** 5.38 0.86** 66 11** ND 4.39 0.41** 124 16** 5.53 0.90** 83 17** 20 mg/kg VEH 5.44 0.44** 113 19** 6.71 1.33** 97 18** ND 5.45 0.60** 131 20** 7.87 1.42** 86 16** VEH, vehicle; ND, nandrlone. *P < 0.05, **P < 0.05; Dunnet’s test, comparison with control group receiving vehicle. under both protocols. In the pre-exposure experiment, two-way was administered using the two different protocols, i.e., ANOVA revealed significant THC effect (F(3,87) ¼ 38.40, co-administration and pre-exposure. P < 0.001), but no nandrolone effect (F(1,87) ¼ 0.92, NS) or THC/nandrolone interaction (F(3,87) ¼ 1.18, NS). In the 3.2.1. Tolerance to THC antinociceptive effects co-administration experiment, two-way ANOVA revealed sig- Nandrolone pre-exposure and co-administration did not nificant THC effect (F(3,105) ¼ 32.39, P < 0.001), but no nan- produce any change in the development of tolerance to antino- drolone effect (F(1,105) ¼ 1.15, NS) or THC/nandrolone ciception induced by chronic THC in the tail immersion interaction (F(3,105) ¼ 0.13, NS). (Fig. 1) and the hot-plate (data not shown) tests. Both protocols of nandrolone treatment did not modify the nociceptive 3.1.3. THC hypolocomotor effects responses in these two tests in vehicle-treated mice. In THC Chronic nandrolone pre-exposure during 14 days and nan- chronically treated mice, tolerance to the antinociceptive drolone co-administration did not influence THC acute effects effects of THC in the tail-immersion and hot-plate tests devel- on horizontal locomotor activity (Table 1). Significant hypolo- oped similarly in both vehicle and nandrolone pre-exposed comotor effects of THC were observed in vehicle-pre-exposed mice. Similar results were obtained in THC mice co-treated mice and nandrolone pre-exposed mice under both protocols. with nandrolone. (see three-way ANOVA in Table 2). In the pre-exposure experiment, two-way ANOVA revealed significant THC effect (F(3,92) ¼ 26.67, P < 0.001), but no 3.2.2. Tolerance to THC hypothermic effects significant effect of nandrolone (F(1,92) ¼ 0.66, NS) or Nandrolone pre-exposure and co-administration did not THC/nandrolone interaction (F(3,92) ¼ 0.07, NS). In the co- produce any change in the development of tolerance to the administration experiment, two-way ANOVA also revealed hypothermic effects of THC (Fig. 1). Both protocols of nan- significant THC effect (F(3,78) ¼ 26.06, P < 0.001), but no drolone treatment did not modify body temperature in vehi- significant effect of nandrolone (F(1,78) ¼ 0.09, NS) or cle-treated mice. In THC chronically treated mice, tolerance THC/nandrolone interaction (F(3,78) ¼ 0.52, NS). to the hypothermic effects of THC developed similarly in both vehicle and nandrolone pre-exposed mice. Similar results 3.2. Effects of nandrolone treatments on the development were obtained in THC mice co-treated with nandrolone (see of tolerance to THC effects three-way ANOVA in Table 2).

Changes in nociceptive thresholds, rectal temperature and 3.2.3. Tolerance to THC hypolocomotor effects locomotor activity were evaluated every day during chronic Nandrolone pre-exposure and co-administration did not administration of THC (20 mg/kg, i.p., twice a day for produce any change in the development of tolerance to the hy- 5 days) or vehicle. Nandrolone (15 mg/kg, i.m.) or vehicle plocomotor effects of THC (Fig. 1). Both protocols of E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 793

A: Nandrolone pre-exposure B: Nandrolone co-administration

VEH-VEH 7 7 ND-VEH VEH-THC 6 6 ** ** ND-THC 5 ** 5 ** ** ** ** 4 ** ** 4 ** ** ** ** * ** 3 ** ** 3 ** 2 2 1 1 0 0 012345 012345 Days Days

4 4 2 2 0 0 -2 ** ** -2 ** ** * -4 * -4 ** ** -6 ** ** -6 ** -8 -8 (% of basal value) Tail withdrawal (s) Rectal temperature -10 -10 ** -12 ** 12345 -12 12345 Days Days

300 ** 300 250 250 ** 200 200 ** ** 150 * ** 150 ** ** * ** (counts) 100 ** 100 50 50 * ** Locomotor activity ** 0 0 0 1 2345 012345 Days Days

Fig. 1. Effects of nandrolone treatments on the development of tolerance to THC-induced antinociceptive, hypothermic and hypolocomotor effects. THC (20 mg/ kg, i.p.) was chronically administered twice a day during 5 days. Tolerance development was evaluated in mice pre-exposed (A) to or co-treated (B) with nandrolone (ND) or vehicle (VEH). Locomotor activity was recorded 20 min after each morning THC injection during 10 min. Antinociceptive effects were evaluated 30 min after each morning THC injection by using the tail-immersion test. The rectal temperature was measured just before and 45 min after each morning THC injection. The number of mice per group in the nandrolone pre-exposure experiments was 10e13. The number of mice per group in the nandrolone co-treatment experiments was 11e15. Data are expressed as mean SEM. *P < 0.05, **P < 0.01 (one-way ANOVA, comparison with the respective basal value). nandrolone treatment did not modify locomotor activity in ve- injection of rimonabant in vehicle control mice co-treated or hicle-treated mice. In THC chronically treated mice, tolerance pre-exposed to nandrolone. Rimonabant injection in chronic to the hypolocomotor effects of THC developed similarly in THC treated mice with vehicle pre-exposure precipitated both vehicle and nandrolone pre-exposed mice. Similar results a withdrawal syndrome manifested by the presence of wet were obtained in THC mice co-treated with nandrolone (see dog shake (one-way ANOVA, F(1,24) ¼ 15.07, P < 0.01), three-way ANOVA in Table 2). front paw tremor (F(1,24) ¼ 33.09, P < 0.001), ptosis (F(1,24) ¼ 29.25, P < 0.001) and a significant global with- 3.3. Effects of nandrolone treatments on drawal score (F(1,24) ¼ 62.53, P < 0.01), as revealed by rimonabant-precipitated THC withdrawal syndrome one-way ANOVA (Fig. 2). The intensity of THC withdrawal was increased in nandrolone pre-exposed mice (Fig. 2) (see THC withdrawal was precipitated by the administration of two-way ANOVA in Table 3). One-way ANOVA revealed the selective CB1 receptor antagonist, rimonabant (10 mg/kg a significant increase of sniffing (F(1,21) ¼ 4,61, P < 0.05) i.p.), in mice receiving chronic administration of THC and global withdrawal score (F(1,21) ¼ 5.12, P < 0.05) in (20 mg/kg, i.p., twice daily during 6 days) or vehicle. No sig- mice pre-exposed to nandrolone in comparison with the group nificant signs of withdrawal were observed in any group of pre-exposed to vehicle. mice during the 15 min behavioural observation performed In contrast, the intensity of THC withdrawal was not modified before the administration of rimonabant. Moreover, no behav- by the co-administration of nandrolone (Fig. 2). Rimonabant- ioural manifestations of withdrawal were observed after the precipitated THC withdrawal was manifested in vehicle 794 E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806

Table 2 3.4.1. THC rewarding properties Nandrolone effects on tolerance to D9-tetrahydrocannabinol (THC) pharmaco- In the nandrolone pre-exposure experiment, two-way AN- logical effects OVA revealed no significant THC (F(1,56) ¼ 1.48, NS) and Nandrolone pre-exposure Nandrolone co-administration nandrolone (F(1,56) ¼ 0.80, NS) effects, but significant FPFPTHC/nandrolone interaction (F(1,56) ¼ 4.64, P < 0.05). Re- THC antinociceptive effects (tail-flick test) warding effects of THC (1 mg/kg) were observed in mice pre- THC (1,43) ¼ 63.36 <0.001 (1,48) ¼ 101.25 <0.001 viously receiving intramuscular vehicle for 14 days (Fig. 3A) Time (4,172) ¼ 15.11 <0.001 (4,192) ¼ 15.39 <0.001 as revealed by the score values (one-way ANOVA; ND (1,43) 0.03 NS (1,48) 0.36 NS ¼ ¼ F(1,29) ¼ 5.37, P < 0.05) and the increase in the time spent THC time (4,172) ¼ 14.41 <0.001 (4,192) ¼ 16.29 <0.001 THC ND (1,43) ¼ 0.14 NS (1,48) ¼ 0.22 NS in the drug-paired compartment during the test vs. the precon- ND time (4,172) ¼ 0.08 NS (4,192) ¼ 0.09 NS ditioning phase (Student’s t-test, t(1,14) ¼3.27, P < 0.01). THC ND (4,172) ¼ 1.02 NS (4,192) ¼ 0.13 NS On the contrary, THC did not induce rewarding effects in time mice pre-exposed to nandrolone, as revealed by the similar time spent in the drug-paired compartment during the precon- THC antinociceptive effects (hot-plate test) THC (1.43) ¼ 7.10 <0.05 (1,47) ¼ 13.48 <0.01 ditioning and test phase (Student’s t-test, t(1,14) ¼ 0.193, NS) Time (4,172) ¼ 42.53 <0.001 (4,188) ¼ 163.17 <0.001 and the score values (one-way ANOVA, F(1,29) ¼ 0.47, NS). ND (1,43) ¼ 0.36 NS (1,47) ¼ 2.48 NS Significant differences were also observed in THC-treated THC time (4,172) ¼ 7.58 <0.01 (4,188) ¼ 3.17 NS mice when compared with vehicle and nandrolone pre- THC ND (1,43) 0.19 NS (1,47) 0.35 NS ¼ ¼ exposed mice (one-way ANOVA, F(1,29) ¼ 4.87, P < 0.05). ND time (4,172) ¼ 0.73 NS (4,188) ¼ 0.01 NS THC ND (4,172) ¼ 1.04 NS (4,188) ¼ 0.02 NS The influence of chronic nandrolone co-administration on time THC rewarding effects could not be assessed due to the interference of the previous intramuscular injection. Two-way THC hypothermic effects ANOVA revealed no significant THC (F(1,53) ¼ 1.81, NS) THC (1.43) 74.85 <0.001 (1,47) 87.73 <0.001 ¼ ¼ and nandrolone (F(1,53) ¼ 1.20, NS) effects, and no signifi- Time (4,172) ¼ 25.10 <0.001 (4,188) ¼ 21.74 <0.001 ND (1,43) ¼ 0.04 NS (1,47) ¼ 6.31 <0.05 cant THC/nandrolone interaction (F(1,53) ¼ 0.01, NS). THC THC time (4,172) ¼ 36.79 <0.01 (4,188) ¼ 21.48 <0.001 (1 mg/kg, s.c.) did not induce conditioned place preference THC ND (1,43) ¼ 2.61 NS (1,47) ¼ 1.07 NS when mice received an intramuscular injection of vehicle or ND time (4,172) ¼ 0.76 NS (4,188) ¼ 0.44 NS nandrolone 1 h before the conditioning session (Fig. 3B), as THC ND (4,172) ¼ 1.55 NS (4,188) ¼ 1.06 NS revealed by the similar time spent in the drug-paired compart- time ment on the preconditioning and test phase in the THC-treated THC hypothermic effects mice (Student’s t-test, t(1,11) ¼ 0.42, NS, in the vehicle group THC (1.43) ¼ 46.84 <0.001 (1,47) ¼ 28.36 <0.001 and t(1,11) ¼ 0.68, NS in the nandrolone group) and score val- Time (4,172) ¼ 41.88 <0.001 (4,188) ¼ 36.01 <0.001 ues (one-way ANOVA, F(1,25) ¼ 0.37, NS, in the vehicle ND (1,43) 1.13 NS (1,47) 0.56 NS ¼ ¼ group and one-way ANOVA, F(1,25) ¼ 0.19, NS, in the nan- THC time (4,172) ¼ 7.8 <0.001 (4,188) ¼ 5.60 <0.001 THC ND (1,43) ¼ 3.62 NS (1,47) ¼ 0.21 NS drolone group). On the contrary, a significant conditioned ND time (4,172) ¼ 0.51 NS (4,188) ¼ 1.75 NS place preference was observed in a positive control group of THC ND (4,172) ¼ 1.63 NS (4,188) ¼ 0.31 NS mice receiving THC alone under the same experimental con- time ditions without any intramuscular administration, as revealed Three-way ANOVA with THC and nandrolone (ND) (between subjects) and by the increase in the time spent in the drug-paired compart- time (within subjects) as factors of variations. NS, not significant. ment during the test phase (Student’s t-test, t(1,9) ¼2.39, P < 0.05), and the score values (one-way ANOVA, co-administrated mice by the presence of wet dog shake (one- F(1,25) ¼ 5.06, P < 0.05). way ANOVA, F(1,25) ¼ 30.09, P < 0.001), front paw tremor (F(1,25) ¼ 25.76, P < 0.001), sniffing (F(1,25) ¼ 4.72, P < 3.4.2. THC aversive properties 0.05), ptosis (F(1,25) ¼ 40.76, P < 0.001) and a significantly In the nandrolone pre-exposure experiment, two-way global withdrawal score (F(1,25) ¼ 40.94, P < 0.01), as revealed ANOVA revealed significant THC effect (F(1,51) ¼ 13.68, by one-way ANOVA. Two-way ANOVA (Table 3) revealed P < 0.01), but no effect of nandrolone (F(1,51) ¼ 0.76, NS) similar manifestations of all the withdrawal signs in chronic or THC/nandrolone interaction (F(1,51) ¼ 0.05, NS). Signifi- THC treated mice co-administered with vehicle or nandrolone. cant aversive effects of THC (5 mg/kg) were observed in mice previously receiving intramuscular vehicle for 14 days (Fig. 3A), as revealed by the score values (one-way ANOVA, 3.4. Effects of nandrolone treatments on THC rewarding F(1,27) ¼ 8.87, P < 0.01) and the decrease in the time spent and aversive properties in the drug-paired compartment during the test vs. the preconditioning phase (Student’s t-test, t(1,12) ¼3.25, P < 0.01). The influence of nandrolone exposure on THC rewarding Nandrolone pre-exposure did not modify THC aversive ef- and aversive effects was investigated by using the place condi- fects, as revealed by the score values (one-way ANOVA, tioned paradigm (Fig. 3). F(1,27) ¼ 0.41, NS) and the decrease in the time spent in E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 795

A: Nandrolone pre-exposure

60 ** 200 40 ** VEH ** ** 140 ND 20 100 # 120 0 0 ** Wet dog shakes VEH THC Front paw tremor VEH THC 100 80 # ** 12 ** 6 **** 60 8 4 40 20

Sniffing 4 2 Ptosis Global withdrawal score 0 0 0 VEH THC VEH THC VEH THC

B: Nandrolone co-administration

60 ** **** ** 200 VEH 40 ND 140 20 100 120 ** ** 0 0

Wet dog shakes 100

VEH THC Front paw tremor VEH THC 80 12 6 **** 60 8 * 4 40

Ptosis 20 Sniffing 4 2

** Global withdrawal score 0 0 0 VEH THC VEH THC VEH THC

Fig. 2. Effects of nandrolone treatments on the rimonabant-precipitated THC withdrawal syndrome. Abstinence is precipitated by the administration of the CB1 receptor antagonist rimonabant (10 mg/kg, i.p.) in mice receiving a chronic treatment of THC (20 mg/kg, i.p.), twice a day during 6 days. Mice were pre-exposed (A) to or co-treated (B) with nandrolone (ND) or vehicle (VEH). Somatic signs of withdrawal were observed for 45 min immediately after the rimonabant administration. The global withdrawal score was calculated for each animal by giving each individual sign a relative weight. The number of mice per groupin the nandrolone pre-exposure experiments was 10e13. The number of mice per group in the nandrolone co-treatment experiments was 10e15. Data are expressed as mean SEM. *P < 0.05, **P < 0.01 (one-way ANOVA, comparison with the respective basal value) and #P < 0.05 (one-way ANOVA, VEH- vs ND-treated mice). the drug-paired compartment during the test vs. precondition- could not be assessed due to the interference of the previous in- ing phase (Student’s t-test, t(1,12) ¼ 4.55, P < 0.01). No dif- tramuscular injection. Two-way ANOVA revealed no signifi- ferences were observed in the aversive effects of THC when cant THC (F(1,65) ¼ 2.19, NS) and nandrolone (F(1,65) ¼ compared vehicle and nandrolone pre-exposed mice (one- 0.27, NS) effects, and no significant THC/nandrolone interac- way ANOVA, F(1,25) ¼ 0.31, NS). tion (F(1,65) ¼ 1.17, NS). THC (5 mg/kg, s.c.) did not induce Similar to the previous experiment, the influence of chronic conditioned place aversion when mice received an intramuscu- nandrolone co-administration on THC aversive properties lar injection of vehicle or nandrolone 1 h before the

Table 3 Nandrolone effects on rimonabant-precipitated D9-tetrahydrocannabinol (THC) withdrawal THC P ND P Interaction P Nandrolone pre-exposure Wet dog shakes (1,43) ¼ 43.24 <0.001 (1,43) ¼ 1.55 NS (1,43) ¼ 2.99 NS Front paw tremor (1,43) ¼ 68.20 <0.001 (1,43) ¼ 4.91 <0.05 (1,43) ¼ 4.58 <0.05 Snifﬁng (1,43) ¼ 12.31 <0.01 (1,43) ¼ 5.28 <0.05 (1,43) ¼ 5.28 <0.05 Ptosis (1,43) ¼ 66.32 <0.001 (1,43) ¼ 3.94 NS (1,43) ¼ 0.01 NS Global withdrawal score (1,43) ¼ 109.13 <0.001 (1,43) ¼ 6.31 <0.05 (1,43) ¼ 5.49 <0.05

Nandrolone co-administration Wet dog shakes (1,46) ¼ 47.82 <0.001 (1,46) ¼ 0.04 NS (1,46) ¼ 0.07 NS Front paw tremor (1,46) ¼ 41.15 <0.001 (1,46) ¼ 0.07 NS (1,46) ¼ 0.10 NS Snifﬁng (1,46) ¼ 5.81 <0.05 (1,46) ¼ 3.01 NS (1,46) ¼ 3.01 NS Ptosis (1,46) ¼ 46.17 <0.001 (1,46) ¼ 0.16 NS (1,46) ¼ 1.08 NS Global withdrawal score (1,46) ¼ 60.45 <0.001 (1,46) ¼ 0.09 NS (1,46) ¼ 0.12 NS Two-way ANOVA with THC and nandrolone (ND) (between subjects) as factors of variations. NS, not signiﬁcant. 796 E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806

A: Nandrolone pre-exposure B: Nandrolone co-administration

THC (1 mg/kg) * # 120 * 120

80 80

40 40

0 0 Score

-40 -40

-80 -80 VEH ND VEH ND VEH ND VEH ND _ -120 VEH VEH THC THC -120 VEH VEH THC THC THC

THC (5mg/kg) 120 120

80 80

40 40

0 0 Score

-40 -40

-80 -80

-120 * -120 **

VEH ND VEH ND VEH ND VEH ND *_ VEH VEH THC THC VEH VEH THC THC THC

Food # 160* 160 * 120 120

80 80

40 40 Score

0 0 _ VEH ND VEH ND VEH ND VEH ND -40 _ _ FD FD -40 __FD FD FD

-80 -80

Fig. 3. Effects of nandrolone treatments on rewarding and aversive properties of THC and food. The doses of 1 and 5 mg/kg (i.p.) of THC were used to induce conditioned place preference and aversion, respectively, in mice pre-exposed to (A) or co-treated (B) with nandrolone (ND) or vehicle (VEH). Food (FD)-induced conditional place preference was assessed on lightly food deprived mice. A score was calculated for each mouse as the difference between the post-conditioning and the pre-conditioning time spent in the drug- or food-paired compartment. The number of mice per group was 10e15. *P < 0.05, **P < 0.01 (one-way ANOVA, comparison with the respective basal value) and #P < 0.05 (one-way ANOVA, VEH- vs ND-treated mice). conditioning session (Fig. 3B), as revealed by the similar time aversion was observed in a positive control group receiving spent in the drug-paired compartment during the precondition- THC alone under the same experimental conditions without ing and test phase (Student’s t-test, t(1,14) ¼0.60, NS, in the any intramuscular administration, as revealed by the increase vehicle group and t(1,14) ¼ 0.88, NS in the nandrolone group) in the time spent in the drug-paired compartment during and score values (one-way ANOVA, F(1,29) ¼ 0.15, NS, in the the test phase (Student’s t-test, t(1,9) ¼2.81, P < 0.05) vehicle group and F(1,29) ¼ 0.21, NS, in the nandrolone and the score values (one-way ANOVA, F(1,29) ¼ 4.85, group). On the contrary, a signiﬁcant conditioned place P < 0.05). E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 797

3.5. Effects of nandrolone treatments on the rewarding ANOVA. Anxiolytic-like effects of THC (0.2 mg/kg) were properties of food also observed in the lit/dark box (Fig. 4A and see Table 6 for two-way ANOVAs). THC increased the percentage of en- In mice previously receiving intramuscular vehicle for tries in the lit compartment (F(1,35) ¼ 5.10, P < 0.05) and the 14 days (Fig. 3A), two-way ANOVA revealed no significant entries in the distal area of this compartment (F(1,35) ¼ 21.63, food (F(1,40) ¼ 2.87, NS) and nandrolone (F(1,40) ¼ 2.90, P < 0.001), and decreased the entries in its proximal area NS) effects, but a significant food/nandrolone interaction (F(1,35) ¼ 21.03, P < 0.001), as revealed by one-way (F(1,40) ¼ 4.35, P < 0.05). Significant conditioned place ANOVA. THC also increased the percentage of squares preference to food was observed in these mice, as revealed crossed in the lit compartment (F(1,35) ¼ 42.22, P < 0.001). by the score values (one-way ANOVA, F(1,21) ¼ 0.29, NS) The anxiolytic-like effects of THC (0.2 mg/kg) were not re- and the increase in the time spent in the food-paired compart- vealed in the elevated plus-maze in our experimental condi- ment during the test vs. preconditioning phase (Student’s tions (Fig. 6A and see Table 7 for two-way ANOVA). In t-test, t(1,9) ¼ 3.18, P < 0.05). On the contrary, rewarding mice pre-treated with nandrolone, THC (0.2 mg/kg) did not effects of food were abolished in mice pre-exposed to nandro- produce any anxiolytic-like effect in the open-field lone, as revealed by the similar time spent in the food-paired (Fig. 5A), lit/dark box (Fig. 4A) and elevated plus-maze compartment during the test vs. preconditioning phase (Stu- (Fig. 6A). One-way ANOVA revealed significant differences dent’s t-test, t(1,10) ¼ 0.26, NS) and the score values (one- in THC effects between vehicle and nandrolone pre-treated way ANOVA, F(1,22) ¼ 0.09, NS). Significant differences mice in the open-field (number of rearing: F(1,24) ¼ 44.61, were also observed between vehicle and nandrolone pre- P < 0.001; percentage of squares crossed in the central area: exposed mice conditioned with food (F(1,20) ¼ 7.31, P < 0.05). F(1,24) ¼ 25.94, P < 0.001) and the lit/dark box (entries Similar to the previous experiment, food did not induce into the proximal area of the lit compartment: conditioned place preference in mice receiving an intramuscu- F(1,38) ¼ 40.36, P < 0.001; entries into the distal area of lar injection of vehicle or nandrolone 1 h before the condition- the lit compartment: F(1,38) ¼ 32.71, P < 0.001, and squares ing session (Fig. 3B), as revealed by the similar time spent in crossed into the lit compartment: F(1,38) ¼ 30.41, P < 0.001). the drug-paired compartment on the preconditioning and test The influence of chronic nandrolone co-administration on phase (Student’s t-test, t(1,11) ¼0.51, NS, in the vehicle THC anxiolytic-like effects could not be assessed due to the group and t(1,12) ¼0.17, NS in the nandrolone group), interference of the previous intramuscular injection (Figs. and the score values (one-way ANOVA, F(1,25) ¼ 0.12, 4Be6B and Tables 5e7 for two-way ANOVA). Indeed, NS). Two-way ANOVA revealed no significant food THC (0.2 mg/kg) did not produce any anxiolytic-like effects (F(2,54) ¼ 1.30, NS) and nandrolone (F(1,54) ¼ 0.11, NS) ef- in any of the behavioural tests when mice received a previous fects, and no food/nandrolone interaction (F(1,54) ¼ 0.03, intramuscular injection of vehicle. NS). On the contrary, a significant place preference was observed in a positive control group of mice receiving food under 3.6.2. Influence of nandrolone on the anxiogenic-like effects the same experimental conditions without any intramuscular of THC (7.5 mg/kg) administration, as revealed by the significant increase in the In mice pre-treated with vehicle, the dose of 7.5 mg/kg of time spent in the drug-paired compartment during the test THC produced anxiogenic-like effects in the open-field, lit/ phase (Student’s t-test, t(1,10) ¼2.88, P < 0.05) and the dark box and elevated plus-maze (Figs. 4Ae6A and see Tables score values (one-way ANOVA, F(1,25) ¼ 4.92, P < 0.05). 5e7 for two-way ANOVA). In the open-field, THC decreased the number of rearing (F(1,24) ¼ 19.00, P < 0.001) and the 3.6. Effects of nandrolone treatments on the THC percentage of squares crossed into the central area anxiolytic- and anxiogenic-like responses (F(1,24) ¼ 22.64, P < 0.001), as revealed by one-way ANOVA. In the elevated plus-maze, THC decreased the per- The influence of nandrolone on THC-induced changes in centage of time spent in the open arms (one-way ANOVA, anxiety-like responses was evaluated in the lit/dark box, F(1,29) ¼ 4.95, P < 0.05). In the lit/dark box test, THC in- open-field and elevated plus-maze (Figs. 4e6) and the global creased the percentage of entries in the proximal area of the results obtained are summarized in Table 4. Both protocols of lit compartment (F(1,36) ¼ 9.80, P < 0.01), and decreased nandrolone treatments had no intrinsic effect when injected in the entries in its distal area (F(1,36) ¼ 9.57, P < 0.01) and combination with vehicle in any of the parameters studied. the squares crossed in the lit compartment (F(1,36) ¼ 8.49, P < 0.01), as revealed by one-way ANOVA. In mice pre- 3.6.1. Influence of nandrolone on the anxiolytic-like effects treated with nandrolone, THC (7.5 mg/kg) produced similar of THC (0.2 mg/kg) anxiogenic-like effects in the open-field, the lit/dark box and In mice pre-treated with vehicle, the dose of 0.2 mg/kg of the elevated plus-maze. In the open-field, one-way ANOVA re- THC produced anxiolytic-like effects in the open-field vealed a decrease in the number of rearing (F(1,22) ¼ 17.17, (Fig. 5A and see Table 5 for two-way ANOVA). THC in- P < 0.001) and the squares crossed into the central area creased the number of rearing (F(1,24) ¼ 44.10, P < 0.001) (F(1,22) ¼ 29.94, P < 0.001). In the elevated plus-maze, and the percentage of squares crossed in the central area one-way ANOVA showed a decrease in the time spent in (F(1,24) ¼ 20.07, P < 0.001), as revealed by one-way the open arms (F(1,29) ¼ 14.10, P < 0.01) and the total 798 E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806

A: Nandrolone pre-exposure

400 * # 60 60 * # 300 * * * * * 40 * 40 200 20

20 % Time lit

100 % Squares lit Total squares

0 0 0 VEH THC THC VEH THC THC VEH THC THC 0.2 7.5 0.2 7.5 0.2 7.5

# * VEH 70 # * ND 60 * 50 * # 40 # * * 30 *

% Entries 20

0 Prox Cent Dist Prox Cent Dist Prox Cent Dist VEH THC 0.2 THC 7.5 B: Nandrolone co-administration 400 60 60 300 * * * 40 * * 40 * * 200

20 %Time lit 20

100 % Squares lit Total squares

0 0 0 VEH THC THC VEH THC THC VEH THC THC 0.2 7.5 0.2 7.5 0.2 7.5

70 # VEH

60 ND 50

30 % Entries 20 # * 10 * 0 Prox Cent Dist Prox Cent Dist Prox Cent Dist VEH THC 0.2 THC 7.5

Fig. 4. Effects of nandrolone treatments on THC anxiolytic- and anxiogenic-like responses in the lit/dark box paradigm. THC (i.p.) was acutely administered at the dose of 0.2 mg/kg and 7.5 mg/kg to induce anxiolytic- and anxiogenic-like responses in mice pre-exposed to (A) or co-treated with (B) nandrolone (ND) or vehicle (VEH). The test was performed 30 min after THC or vehicle treatment. The number of mice per group in the nandrolone pre-exposure experiments was 11e20. The number of mice per group in the nandrolone co-treatment experiments was 11e16. Data are expressed as mean SEM. *P < 0.05, **P < 0.01 (one-way ANOVA, comparison with the respective basal value) and #P < 0.05, ##P < 0.01 (one-way ANOVA, VEH- vs ND-treated mice). number of entries (F(1,29) ¼ 5.59, P < 0.05). In the lit/dark (F(1,36) ¼ 5.06, P < 0.05), whereas the entries in its proxi- box, one-way ANOVA revealed a decrease in the squares mal area was increased (F(1,35) ¼ 7.04, P < 0.05). No crossed (F(1,35) ¼ 17.33, P < 0.001) and the time spent signiﬁcant differences were found between vehicle-and nan- (one-way ANOVA, F(1,35) ¼ 4.46, P < 0.05) in the lit com- drolone pre-treated mice receiving the administration of partment, and the percentage of entries in its distal area THC. E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 799

A: Nandrolone pre-exposure B: Nandrolone co-administration 400 400 * *

300 300

200 200 Total squares

100 Total squares 100

0 0 VEH THC THC VEH THC THC 0.2 7.5 0.2 7.5

60 60 * 50 50

40 40 # # 30 30 * Rearing 20 * Rearing 20 * * * * 10 10

0 0 VEH THC THC VEH THC THC 0.2 7.5 0.2 7.5

25 25 * 20 * 20

15 * # 15 * # 10 * * 10 in central area in central area

% square crossed 5 5 % square crossed

0 0 VEH THC THC VEH THC THC 0.2 7.5 0.2 7.5 VEH ND

Fig. 5. Effects of nandrolone treatments on THC anxiolytic- and anxiogenic-like responses in the open-ﬁeld paradigm. THC (i.p.) was acutely administered at the dose of 0.2 mg/kg and 7.5 mg/kg to induce anxiolytic- and anxiogenic-like responses in mice pre-exposed to (A) or co-treated with (B) nandrolone (ND) or vehicle (VEH). The test was performed 30 min after THC or vehicle treatment. The number of mice per group in the nandrolone pre-exposure experiments was 11e20. The number of mice per group in the nandrolone co-treatment experiments was 11e16. Data are expressed as mean SEM. *P < 0.05, **P < 0.01 (one-way ANOVA, comparison with the respective basal value) and ##P < 0.05 (one-way ANOVA, VEH- vs ND-treated mice).

In mice co-treated with vehicle, the dose of 7.5 mg/kg of elevated plus-maze, one-way ANOVA revealed a decrease in THC produced anxiogenic-like effects in the lit/dark box the time spent (F(1,22) ¼ 7.36, P < 0.05) and entries and elevated plus-maze (Figs. 4B and 6B and see Tables 6 (F(1,22) ¼ 7.57, P < 0.05) in the open arms. In the open-field, and 7 for two-way ANOVA). In the lit/dark box, one-way one-way ANOVA revealed a decrease in the number of rearing ANOVA revealed a decrease in the time spent (F(1,31) ¼ (F(1,35) ¼ 6.80, P < 0.05) and an increase in the total number 10.52, P < 0.01) and squares crossed (F(1,31) ¼ 13.48, of squares crossed (F(1,35) ¼ 4.80, P < 0.05). No significant P < 0.01) in the lit compartment. In the elevated plus-maze, differences were found between vehicle and nandrolone co- a decrease was observed in the time spent in the open arms treated mice receiving THC. (F(1,23) ¼ 4.5, P < 0.05). In the open-field, one-way ANOVA revealed a decrease in the number of rearing (F(1,18) ¼ 13.32, 3.7. Effects of nandrolone pre-treatment on CB1 receptor P < 0.01) and an increase in the total number of squares binding and GTPgS-binding protein activation crossed (F(1,18) ¼ 9.95, P < 0.01). In mice co-treated with nandrolone, THC produced similar anxiogenic-like effects in Autoradiography binding studies did not reveal any differ- the lit/dark box and elevated plus-maze. In the lit/dark box, ence between the density of CB1 receptors in the caudate- one-way ANOVA showed a decrease in the time spent putamen (Student’s t-test, t(1,10) ¼ 1.33, NS) and cerebellum (F(1,30) ¼ 10.52, P < 0.01) and squares crossed Student’s t-test, t(1,6) ¼ 1.49, NS) in mice receiving a daily (F(1,30) ¼ 9.15, P < 0.01) in the lit compartment. In the administration of nandrolone (15 mg/kg) during 14 days and 800 E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806

A: Nandrolone pre-exposure B: Nandrolone co-administration 60 60

40 * 40

20 20 Total entries Total entries

0 0 VEH THC THC VEH THC THC 0.2 7.5 0.2 7.5

40 40

30 30 * * * * * 20 * 20 % Time open % Time open 10 10

0 0 VEH THC THC VEH THC THC 0.2 7.5 0.2 7.5

50 50

40 40 *

30 30

20 20 % Entries open % Entries open 10 10

0 0 VEH THC THC VEH THC THC 0.2 7.5 0.2 7.5 VEH ND

Fig. 6. Effects of nandrolone treatments on THC anxiolytic- and anxiogenic-like responses in the elevated plus-maze paradigm. THC (i.p.) was acutely administered at the dose of 0.2 mg/kg and 7.5 mg/kg to induce anxiolytic- and anxiogenic-like responses in mice pre-exposed to (A) or co-treated with (B) nandrolone (ND) or vehicle (VEH). The test was performed 30 min after THC or vehicle treatment. The number of mice per group in the nandrolone pre-exposure experiments was 11e20. The number of mice per group in the nandrolone co-treatment experiments was 11e16. Data are expressed as mean SEM. *P < 0.05, **P < 0.01 (one-way ANOVA, comparison with the respective basal value). control animals receiving vehicle (Fig. 7A and B). Repeated hypolocomotor effects, as previously reported (Martin et al., nandrolone administration did not modify the efficacy of can- 1991). Similar to former results obtained with morphine nabinoid receptor activation. Indeed, the magnitude of the (Ce´le´rier et al., 2003), nandrolone treatment did not produce [35S]GTPgS binding induced by the receptor agonist relevant changes in acute THC antinociceptive and locomotor WIN55,212-2 was similar in the caudate-putamen (Student’s t-test, t(1,10) ¼0.34, NS) and the cerebellum (Student’s Table 4 Summary of nandrolone effects on the anxiolytic- and anxiogenic-like effects t-test, t(1,7) ¼ 0.72, NS) (Fig. 7C and D) of mice chronically of D9-tetrahydrocannabinol (THC) treated with nandrolone (15 mg/kg, i.m., during 14 days) or THC doses Test Nandrolone Nandrolone vehicle. pre-exposure co-administration Lit/ Open- Plus Lit/ Open- Plus 4. Discussion dark field maze dark field maze 0.2 mg/kg Vehicle A A 000 0 The aim of our study was to evaluate the influence of nan- Nandrolone 0 0 0 0 0 0 drolone exposure on several pharmacological and behavioural responses induced by acute and chronic THC treatment and re- 7.5 mg/kg Vehicle Aþ Aþ Aþ Aþ 0Aþ lated to its addictive properties. Acute THC administration in- Nandrolone Aþ Aþ Aþ Aþ 0Aþ duced dose-dependent antinociceptive, hypothermic and A, anxiolytic-like effect; Aþ, anxiogenic-like effect. E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 801

effects. However, nandrolone modiﬁed morphine- but not cannabinoid-induced hypothermia. Both cannabinoids and opioids induce hypothermia by acting on the hypothalamus, but differ- 0.08 NS 0.56 NS 0.06 NS 1.26 NS 0.46 NS 0.12 NS ent neurochemical mechanisms are involved in their re- ¼ ¼ ¼ ¼ ¼ ¼ sponses. Hypothalamic NMDA-NO pathways (Ulugol et al., 2000), and changes in the plasma levels of antipyretic substances such as ACTH and corticosterone (Nikolarakis et al., 1989) play an important role for morphine hypothermic effects. In contrast, changes in the serotoninergic activity in 0.01 (1,39)

< the hypothalamus (Davies and Graham, 1980; Malone and Taylor, 2001) with a permissive role of dopamine via the D2 receptors (Nava et al., 2000) participate in cannabinoid- induced hypothermia. In agreement, nandrolone administra- 4.95 NS (1,45) 5.09 NS (1,45) 7.98 1.78 NS (1,39) 1.16 NS (1,39) 14.15 NS (1,45) ¼ ¼ ¼ ¼ ¼ ¼ tion has been reported to modify endogenous opioid activity (Johansson et al., 2000a), the hypothalamic expression of the NMDA receptor NR1 subunit (Le Greve`s et al., 1997) and the circulating levels of corticosterone and ACTH (Schlussman et al., 2000), without any effect on dopaminergic and serotoninergic transmission in the hypothalamus (Thiblin et al., 1999). These biochemical changes could support a pref- 0.19 NS (1,45) 0.01 NS (1,45) 0.49 NS (1,45) 0.02 NS (1,39) 0.90 NS (1,39) 0.39 NS (1,39) erential modulatory action of nandrolone on morphine ¼ ¼ ¼ ¼ ¼ ¼ hypothermia. Tolerance is an adaptive response to prolonged exposure of Co-administration the organism to different drugs that provide a partial correlate of their addictive properties. Chronic administration of THC develops tolerance to the antinociceptive, hypothermic and hy- 0.001 (1,39) 0.001 (1,39) polocomotor effects, as previously reported (Chaperon and < < Thie´bot, 1999; Ghozland et al., 2002). This tolerance occurs quickly as shown by the complete loss of hypothermic effect on the second day of THC administration. Similar to previous 0.27 NS (1,45) 0.08 NS (1,45) 0.04 NS (1,45) 20.18 28.41 0.14 NS (1,39) results obtained with morphine (Ce´le´rier et al., 2003), the de- ¼ ¼ ¼ ¼ ¼ ¼ velopment of tolerance to THC hypothermia, hypolocomotion and antinociception in the tail-immersion test was not modified by nandrolone. However, nandrolone attenuated tolerance to morphine (Ce´le´rier et al., 2003), but not THC (present 0.001 (1,44) 0.001 (1,44) 0.01 (1,44) study) antinociceptive effects in the hot-plate test. The tail- 9-tetrahydrocannabinol (THC) in the open-field paradigm < < < D withdrawal response is mainly mediated by spinal mechanisms whereas hot-plate responses require the involvement of supraspinal structures (Grossman et al., 1982; Morgan et al., 1989). 36.06 14.89 2.43 NS (1,44) 11.47 2.20 NS (1,44) 0.96 NS (1,44) Nandrolone action on morphine tolerance would involve ¼ ¼ ¼ ¼ ¼ ¼ supraspinal mechanisms that would not be required for the development of tolerance to THC antinociceptive effects, such as changes in endogenous opioid activity in the periacqueductal grey (Johansson et al., 2000a). Thus, cannabinoid actions in the periacqueductal grey and the rostral ventromedial medulla, 0.001 (1,44) two important structures for the control of pain, have been < reported to involve opioid-independent mechanisms (Meng et al., 1998; Hohmann et al., 2005). The rewarding properties of drugs of abuse play a crucial role 0.06 NS (1,44) 0.13 NS (1,44) 1.31 NS (1,44) 18.02 9.56 NS (1,44) 2.80 NS (1,44) for the initiation of the addictive processes and are related to the ¼ ¼ ¼ ¼ ¼ ¼ biochemical changes mainly occurring in the mesolimbic dopa- (1,44) (1,44) Pre-exposure Pre-treatmentFPFPFPFPFPFP Treatment Interactionminergic Co-treatment system (Koob, Treatment 1992). Cannabinoid Interaction effects on the rewarding system are complex (Maldonado and Rodrı´guez de Fonseca, 2002). In the place conditioning paradigm, cannabinoid agonists currently produce aversive-like responses (Hutcheson et al., 1998; Ghozland et al., 2002), although canna- central central Rearing (1,44) % Squares Anxiolytic-like effects of THCTotal (7.5 square mg/kg) (1,44) Rearing (1,44) % Squares Anxiolytic-like effects of THCTotal (0.2 square mg/kg) (1,44) Table 5 Influence of nandrolone on the anxiolytic- and anxiogenic-like effects of Two-way ANOVA with THC and nandrolone (between subject) as factors of variation; NS, not significant. binoid conditioned place preference can be observed under 802 E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806

particular experimental conditions. Herein, we used a priming injection of THC previous to conditioning to reveal the cannabinoid place preference (Valjent and Maldonado, 2000). Acti- 0.16 NS 0.18 NS 0.67 NS 0.67 NS 0.16 NS 0.17 NS 0.30 NS 0.01 NS 0.05 NS 0.14 NS 0.33 NS 2.84 NS vation of mesolimbic dopaminergic system participates in the ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ common substrate for the rewarding properties of major drugs of abuse, including opioids and cannabinoids (Koob, 1992; Manzanares et al., 1999). A mutual interaction between cannabinoid and opioid systems has been reported to obtain rewarding effects (Chaperon and Thie´bot, 1999; Manzanares et al., 0.001 (1,59) 0.001 (1,59) 0.050.05 (1,59) (1,59) < < < < 1999; Maldonado and Rodrı´guez de Fonseca, 2002). Cannabi- noids can also modulate glutamate and GABA transmission in reward circuits, and interact with neuropeptides relevant for 0.01 NS (1,60) 0.06 NS (1,60) 3.80 NS (1,60) 0.01 NS (1,59) 0.50 NS (1,60) 0.25 NS (1,60) 22.05 1.03 NS (1,59) 13.94 0.03 NS (1,60) 5.04 8.42 processing motivation, such as the corticotropin-releasing fac- ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ tor (CRF) (Maldonado and Rodrı´guez de Fonseca, 2002). We showed here that nandrolone pre-exposure is associated with a decrease in THC-induced place preference. A similar decrease of morphine-induced place preference was induced by 0.05 (1,60) 0.05 (1,60) 0.01 (1,60) 0.05 (1,60) nandrolone in our previous study (Ce´le´rier et al., 2003). These < < < < results are somewhat unexpected in regard to the literature compiling the AAS actions on the rewarding system. Thus, an increase in dopaminergic activity has been proposed to ac- 4.57 1.78 NS (1,60) 2.65 NS (1,60) 4.48 10.00 1.32 NS (1,59) 0.35 NS (1,59) 0.91 NS (1,59) 5.49 0.66 NS (1,59) 1.40 NS (1,59) 0.26 NS (1,59) ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ count for some of the positive effects that frequently appear following AAS administration in man (Kashkin and Kleber,

Co-administration 1989). Nandrolone has been found to elevate dopamine levels in the rat nucleus accumbens (Kindlundh et al., 2001) by inter- acting with the endogenous opioid system (Lukas, 1993; Me- 0.001 (1,60) 0.001 (1,60) 0.001 (1,60) nard et al., 1995; Hallberg et al., 2000; Johansson et al., < < < 1997, 2000a,b). Moreover, AAS have been ascribed a sensitiz- ing action on the brain reward system similar to other various psychoactive substances, such as amphetamine (Clark et al., 0.27 NS (1,60) 1.16 NS (1,60) 28.06 18.97 0.58 NS (1,60) 14.03 1.15 NS (1,59) 0.79 NS (1,59) 0.25 NS (1,59) 0.35 NS (1,59) 0.12 NS (1,59) 0.32 NS (1,59) ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ 1996). The unexpected decrease in THC and morphine-induced conditioned place preference after nandrolone pre-exposure might then be explained by AAS-induced imbalances in the different endogenous opioid peptides in structures related to reinforcing/dysphoric properties of the drugs (Johansson 0.05 (1,70) 0.001 (1,69) 0.01 (1,70) 0.01 (1,69) 0.001 (1,69) 0.001 (1,69) et al., 2000a). Indeed, nandrolone enhanced levels of the en- 9-tetrahydrocannabinol (THC) in the lit/dark box < < < < < <

D dogenous k-agonist dynorphin in the striatum, which could account for a dysphoric effect (Johansson et al., 2000a) through an inhibition of the dopaminergic activity in this area (Steiner 0.65 NS (1,70) 6.19 2.42 NS (1,70) 3.62 NS (1,70) 0.60 NS (1,70) 25.05 9.24 7.53 1.52 NS (1,69) 16.77 1.4414.02 NS (1,69) ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ and Gerfen, 1998). This may explain why the aversive properties of THC were not modified. The impairment produced by nandrolone on the rewarding effects induced by a natural stimulus such as food on this place conditioning paradigm further reinforces this hypothesis. Susceptibility to develop addiction 0.001 (1,70) 0.001 (1,70) 0.001 (1,70) is influenced by sources of reinforcement, variable neuroadap- < < < tive mechanisms and neurochemical changes that together lead to altered homeostasis of the brain reward system (Koob and Le Moal, 2001). The changes induced by nandrolone may reflect 32.74 14.44 3.44 NS (1,70) 0.52 NS (1,70) 21.97 2.09 NS (1,70) 0.08 NS (1,69) 0.22 NS (1,69) 3.91 NS (1,69) 0.32 NS (1,69) 0.090.24 NS NS (1,69) (1,69) ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ long-term modifications in the brain reward circuits leading to a progressive decrease in the basal hedonic level, which (1,70) Pre-exposure Pre-treatmentFPFPFPFPFPFP Treatment Interaction Co-treatment Treatment Interaction (1,69) would result in an unpleasant state facilitating the development of an addictive process (Koob and Le Moal, 2001). The effects of nandrolone co-administration on the rewarding properties of THC and food could not be studied in our experimental conditions since conditioned place preference was not observed in the control mice receiving intramus- proximal proximal % Entries % Squares lit (1,70) Anxiolytic-like effects of THCTotal (0.2 squares mg/kg) (1,70) % Time lit (1,70) Table 6 Influence of nandrolone on the anxiolytic- and anxiogenic-like effects of Two-way ANOVA with THC and nandrolone (between subject) as factors of variation; NS, not significant. % Entries distal (1,70) % Entries central (1,70) % Squares lit (1,69) % Time lit (1,69) Anxiolytic-like effects of THCTotal (7.5 squares mg/kg) (1,69) % Entries % Entries central% Entries distal (1,69) (1,69) cular injection of vehicle. In a previous study, morphine E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 803

Table 7 Inﬂuence of nandrolone on the anxiolytic- and anxiogenic-like effects of D9-tetrahydrocannabinol (THC) in the elevated-plus maze paradigm Pre-exposure Co-administration Pre-treatment Treatment Interaction Co-treatment Treatment Interaction FPFPFPFPFPFP Anxiolytic-like effects of THC (0.2 mg/kg) Total square (1.56) ¼ 0.19 NS (1.56) ¼ 0.19 NS (1.56) ¼ 2.73 NS (1.48) ¼ 0.88 NS (1.48) ¼ 0.15 NS (1.48) ¼ 0.97 NS % Squares (1.56) ¼ 1.29 NS (1.56) ¼ 0.14 NS (1.56) ¼ 0.48 NS (1.48) ¼ 1.32 NS (1.48) ¼ 2.93 NS (1.48) ¼ 2.93 NS central Rearing (1.56) ¼ 1.05 NS (1.56) ¼ 0.04 NS (1.56) ¼ 0.63 NS (1.48) ¼ 0.10 NS (1.48) ¼ 1.67 NS (1.48) ¼ 1.67 NS

Anxiolytic-like effects of THC (7.5 mg/kg) Total square (1.56) ¼ 0.03 NS (1.56) ¼ 6.38 < 0.05 (1.56) ¼ 1.33 NS (1.43) ¼ 0.01 NS (1.43) ¼ 4.86 < 0.05 (1.43) ¼ 0.02 NS % Squares (1.56) ¼ 0.11 NS (1.56) ¼ 3.01 NS (1.56) ¼ 0.02 NS (1.43) ¼ 0.14 NS (1.43) ¼ 7.87 < 0.01 (1.43) ¼ 0.94 NS central Rearing (1.56) ¼ 0.73 NS (1.56) ¼ 15.93 < 0.001 (1.56) ¼ 0.41 NS (1.43) ¼ 0.01 NS (1.43) ¼ 11.35 < 0.01 (1.43) ¼ 0.06 NS Two-way ANOVA with THC and nandrolone (between subject) as factors of variation; NS, not signiﬁcant. conditioned place preference was not observed in mice receiv- dependence in humans (Pope and Katz, 1994; Brower et al., ing similar intramuscular vehicle administration (Ce´le´rier 1991). Chronic treatment with nandrolone in mice did not et al., 2003). In contrast, the negative affective properties of produce any sign of withdrawal in vehicle-treated mice after a high dose THC were preserved in the vehicle co-treated injection of rimonabant (this study) or naloxone (Ce´le´rier mice. A hypothesis to explain these results could be that intra- et al., 2003). These results suggest that AAS withdrawal syn- muscular injection just before the rewarding stimulus (food or drome does not directly involve the endogenous cannabinoid drug) would lead to negative emotional states in which the or opioid systems, or cannot be revealed under the standard mice are less sensitive to the reinforcing effects of the drugs, experimental conditions in rodents. Our results also showed without affecting their aversive-related properties. that chronic nandrolone pre-exposure signiﬁcantly enhances The induction of physical dependence and consequent rimonabant-precipitated THC withdrawal, which is in accor- avoidance of an aversive withdrawal state represents an impor- dance with biochemical studies indicating changes in the en- tant aspect of the motivational impetus maintaining drug ad- dogenous opioid system after chronic AAS exposure diction (Koob and Le Moal, 2001). Clinical studies suggest (Menard et al., 1995; Johansson et al., 1997; Harlan et al., that prolonged use of high-doses of AAS may induce physical 2000). Indeed, a bidirectional interaction between opioid and

A B 600 VEH 500 VEH ND ND 400 300 200

fmol/mg tissue 100 0 Caudate Cerebellum Putamen C D 100 VEH VEH ND 80 ND 60 40 20

% of basal binding 0 Caudate Cerebellum Putamen

Fig. 7. Effects of nandrolone chronic treatment on CB1 receptor binding and activation of GTP-binding proteins. CB1 receptor levels and functional activity were analysed in the caudate-putamen and the cerebellum of mice receiving a daily administration of nandrolone (ND) (15 mg/kg) during 14 days and control animals receiving vehicle (VEH). (A) Representative autoradiograms for cannabinoid receptor binding of coronal brain sections from VEH and ND treated mice in the caudate-putamen. (B) Cannabinoid receptor binding (fmol/mg tissue) in the caudate-putamen and the cerebellum of VEH and ND treated mice. (C) Representative autoradiograms for cannabinoid receptor agonist-stimulated [35S]GTPgS binding of coronal brain sections from VEH and ND treated mice in the caudate-putamen. (D) Cannabinoid receptor agonist-stimulated [35S]GTPgS binding (% of stimulation of basal binding) in the caudate-putamen and the cerebellum of VEH and ND treated mice. Data are expressed as mean SEM of 4e6 animals per group. No significant differences between genotypes were observed. 804 E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 cannabinoid systems has been reported for the development These behavioural and biochemical results suggest that nan- and expression of physical dependence manifestations drolone chronic treatment may alter the rewarding system (Maldonado and Rodrı´guez de Fonseca, 2002). and act, not only as a risk factor for the consumption of drugs States of generalized anxiety have been reported to predict with addictive properties such as cannabis, but also inducing a higher frequency in the diagnostics of addiction (Compton a general decrease in the basal hedonic level. Although et al., 2003). Previous studies have reported complex results many aspects of steroid actions still require further research, concerning the role of CB1 cannabinoid receptors in anxiety. our results provide new support to suggestions that the grow- Both anxiolytic and anxiogenic effects have been reported ing use of AAS by both athletes and non-athletes could lead to after THC administration in animals and humans depending relevant public health problems. on the experimental conditions and the cannabinoid dose (Fabre and McLendon, 1981; Sethi et al., 1986; Tunving, 1987; Chaperon and Thie´bot, 1999; Berrendero and Maldonado, Acknowledgements 2002). THC also induced dose-dependent anxiolytic- and anxiogenic-like effects in our study. The anxiolytic-like effects of a This study was supported by grants from Spanish MEC low dose of THC were abolished by nandrolone pre-exposure. (BFU2004-00920/BFI and GEN2003-20651-C06-04 to RM), In contrast, the anxiogenic-like effects of a high dose of THC Generalitat de Catalunya (2002SGR00193 to RM), Redes were not modified. Different neurobiological mechanisms del Instituto de Salud Carlos III (C03/06 and G03/005 to have been reported to be involved in the anxiogenic- and anxi- RM), the European Commission (GENADDICT#OJ 2003/ olytic-like effects of cannabinoids (Rodrı´guez de Fonseca C164, No.005166 to RM), Marato de TV3 (to RM), the Fon- et al., 1996; Manzanares et al., 1999; Are´valo et al., 2001). dation Fyssen (Grant to EC) and the Swedish Medical Re- The m-opioid antagonist b-funaltrexamine and the d-opioid an- search Council (Grant 9459 to FN). tagonist naltrindole have been shown to block THC-induced anxiolytic-like effects, but did not modify the anxiogenic-like References responses induced by this cannabinoid (Berrendero and Maldo- nado, 2002). The changes induced by nandrolone on opioid Are´valo, C., de Miguel, R., Hernańdez-Tristań, R., 2001. Cannabinoid effects systems may then explain such a differential effect on the on anxiety-related behaviours and hypothalamic neurotransmitters. anxiolytic- and anxiogenic-like effects of THC. Pharmacol. Biochem. Behav. 70, 123e131. In order to evaluate the possible mechanisms involved in the Ameri, A., 1999. The effects of cannabinoids on the brain. Prog. Neurobiol. e effects produced by chronic nandrolone treatment on THC re- 58, 315 348. Arvary, D., Pope, H.G.J., 2000. Anabolic-androgenic steroids as a gateway to sponses, we have investigated the effects produced by such opioid dependence. N. Engl. J. Med. 342, 1532. a chronic treatment, one CB1 receptor binding and GTP binding Bahrke, M.S., Yesalis, C.E., Wright, J.E., 1996. Psychological and behavioural protein activation. Indeed, previous studies have revealed that effects of endogenous testosterone and anabolic-androgenic steroids. androgen steroids such as testosterone, induced changes in can- An update. Sport Med. 22, 367e390. nabinoid binding sites in the rat brain (Rodriguez de Fonseca Ballard, C.L., Wood, R.I., 2005. Intracerebroventricular self-administration of commonly abused anabolic-androgenic steroids in male hamsters (Meso- et al., 1994), suggesting that the effects observed in this study cricetus auratus): nandrolone, drostanolone, oxymetholone, and stanozolol. may derive from the effects of nandrolone on the expression Behav. Neurosci. 119, 752e758. and/or functional activity of cannabinoid receptors. Two brain Berrendero, F., Maldonado, R., 2002. Involvement of the opioid system in the areas related to cannabinoid behavioural responses and depen- anxiolytic-like effects induced by D9-tetrahydrocannabinol. Psychophar- e dence, the caudate putamen and the cerebellum, were selected macology 163, 111 117. Brower, K.J., Blow, F.C., Young, J.P., Hill, E.M., 1991. Symptoms and corre- for this study. Chronic nandrolone administration did not pro- lates of anabolic-androgenic steroid dependence. Br. J. Addict. 86, 759e duce any change in CB1 receptor binding or in GTPgS incorpo- 768. ration induced by the cannabinoid agonist WIN 55 212-2 in the Castan˜e´, A., Maldonado, R., Valverde, O., 2004. Role of different brain struc- caudate-putamen and cerebellum. Therefore, the effects of nan- tures in the behavioural expression of WIN 55,212-2 withdrawal in mice. e drolone on THC behavioural effects were not due to a selective Br. J. Pharmacol. 142, 1309 1317. Ce´le´rier, E., Yazdi, M.T., Castane, A., Ghozland, S., Nyberg, F., alteration on the density or functional activity of CB1 receptors, Maldonado, R., 2003. Effects of nandrolone on acute morphine responses, and were rather due to a more general modification on behaviou- tolerance and dependence in mice. Eur. J. Pharmacol. 465, 69e81. ral and hedonic responses. Chaperon, F., Thie´bot, M.H., 1999. Behavioral effects of cannabinoid agents in In summary, differential effects of nandrolone were found animals. Crit. Rev. Neurobiol. 13, 243e281. on THC-induced withdrawal, place conditioning and anxi- Clark, A.S., Henderson, L.P., 2003. Behavioral and physiological responses to anabolic-androgenic steroids. Neurosci. Biobehav. Rev. 27, 413e436. ety-like responses. Indeed, nandrolone decreased both anxio- Clark, A.S., Lindenfeld, R.C., Gibbons, C.H., 1996. Anabolic-androgenic ste- lytic-like and rewarding effects of THC, but did not change roids and brain reward. Pharmacol. Biochem. Behav. 53, 741e745. or even enhance the aversive negative states associated with Compton, W.M., Cottler, L.B., Jacobs, J.L., Ben-Abdallah, A., drug intake, including conditioned aversion, anxiogenic-like Spitznagel, E.L., 2003. The role of psychiatric disorder in predicting e effects and withdrawal syndrome. No changes in CB1 receptor drug dependence treatment outcome. Am. J. Psychiatry 160, 890 895. Corchero, J., Manzanares, J., Fuentes, J.A., 2001. Role of gonadal steroids in binding and GTP binding protein activation were observed the corticotrophin-releasing hormone and proopiomelanocortin gene ex- after nandrolone treatment. Interestingly, nandrolone also depression response to Delta(9)-tetrahydrocannabinol in the hypothalamus creased the rewarding effect of food as a natural stimulus. of the rat. Neuroendocrinol. 74, 185e192. E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806 805

Davies, J., Graham, J., 1980. The mechanism of action of THC on body tem- Kindlundh, A., Isacson, D.G.L., Berglund, L., Nyberg, F., 1999. Factors asso- perature in mice. Psychopharmacology 69, 299e305. ciated with adolescent use of doping agents: anabolic-androgenic steroids. DuRant, R.H., Vaughn, R.I., Ashworth, C.S., Newman, C., Slavens, G., 1993. Addiction 94, 543e553. Use of multiple drugs among adolescents who use anabolic steroids. Kindlundh, A.S., Lindblom, J., Bergstro¨m, L., Wikberg, J.E.S., Nyberg, F., N. Engl. J. Med. 32, 922e926. 2001. The anabolic-androgenic steroid nandrolone decanoate affects the Elmer, G.I., Pieper, J.O., Goldberg, S.R., George, F.R., 1995. Opioid operant density of dopamine receptors in the male rat brain. Eur. J. Neurosci. self-administration, analgesia, stimulation and respiratory depression in 13, 291e296. m-deficient mice. Psychopharmacology 117, 23e31. Kindlundh, A.S., Lindblom, J., Nyberg, F., 2003. Chronic administration Fabre, L.F., McLendon, D., 1981. The efficacy and safety of nabilone (a syn- with nandrolone decanoate induces alterations in the gene-transcript con- thetic cannabinoid) in the treatment of anxiety. J. Clin. Pharmacol. 2, tent of dopamine D1- and D2-receptors in rat brain. Brain Res. 979, 377Se382S. 37e42. Filliol, D., Ghozland, S., Chluba, J., Martin, M., Matthes, H.W., Simonin, F., Koob, G.F., 1992. Neural mechanisms of drug reinforcement. Ann. N.Y. Acad. Befort, K., Gaveriaux-Ruff, C., Dierich, A., LeMeur, M., Valverde, O., Sci. 654, 171e191. Maldonado, R., Kieffer, B.L., 2000. Mice deficient for delta- and mu- Koob, G.F., Le Moal, M., 2001. Drug addiction, dysregulation of reward, and opioid receptors exhibit opposing alterations of emotional responses. allostasis. Neuropsychopharmacology 24, 97e129. Nat. Genet. 25, 195e200. Le Greve`s, P., Huang, W., Johansson, P., Tho¨rnwall, M., Zhou, Q., Nyberg, F., Ghozland, S., Matthes, W.D.H., Simonin, F., Filliol, D., Kieffer, B.L., 1997. Effects of an anabolic androgenic steroid on the regulation of the Maldonado, R., 2002. Motivational effects of cannabinoids are mediated NMDA receptor NR1, NR2A and NR2B subunit mRNAs in brain regions by m-opioid and k-opioid receptors. J. Neurosci. 22, 1146e1154. of the male rat. Neurosci. Lett. 226, 61e64. Gonzalez, S., Bisogno, T., Wenger, T., Manzanares, J., Milone, A., Lukas, S.E., 1993. Current perspectives on anabolic e androgenic steroids. Berrendero, F., Di Marzo, V., Ramos, J.A., Fernandez-Ruiz, J.J., 2000. Trends Pharmacol. Sci. 14, 61e68. Sex steroid influence on cannabinoid CB(1) receptor mRNA and endocan- MacCarrone, M., De Felici, M., Bari, M., Klinger, F., Siracusa, G., Finazzi- nabinoid levels in the anterior pituitary gland. Biochem. Biophys. Res. Agro, A., 2000. Down-regulation of anandamine hydrolase in mouse Commun. 270, 260e266. uterus by sex hormones. Eur. J. Biochem. 267, 2991e2997. Grossman, M.L., Bausbaum, A.I., Fields, H.L., 1982. Afferent and efferent Maldonado, R., Rodrı´guez de Fonseca, F., 2002. Cannabinoid addiction: be- connections of the rat tail flick reflex (a model used to analyse pain control havioral models and neural correlates. J. Neurosci. 22, 3326e3331. mechanisms). J. Comp. Neurol. 206, 9e16. Malone, D.T., Taylor, D.A., 2001. Involvement of somatodendritic 5-HT1A re- Halikas, J.A., Weller, R.A., Morse, C.L., Hoffmann, R.G., 1985. A longitudi- ceptors in THC-induced hypothermia in the rat. Pharmacol. Biochem. nal study of marijuana effects. Int. J. Addict. 20, 701e711. Behav. 69, 595e601. Hallberg, M., Johansson, P., Kindlundh, A.M., Nyberg, F., 2000. Anabolic an- Manzanares, J., Corchero, J., Romero, J., Fernańdez-Ruiz, J.J., Ramos, J.A., drogenic steroids affect the content of substance P and substance P (1e7) Fuentes, J.A., 1999. Pharmacological and biochemical interactions be- in the rat brain. Peptides 21, 845e852. tween opioids and cannabinoids. TiPS 20, 287e294. Harlan, R.E., Brown, H.E., Lynch, C.S., D’Souza, D., Gracia, M.M., 2000. Martin, B.R., Compton, D.R., Thomas, B.F., Prescott, W.R., Little, P.J., The effect on opioid peptides in the rat brain after chronic treatment with Razdan, R.K., Johnson, M.R., Melvin, L.S., Mechoulam, R., Ward, S.J., the anabolic androgenic steroid, nandrolone decanoate. Brain Res. 853, 1991. Behavioural, biochemical, and molecular modeling evaluation of 99e104. cannabinoid analogs. Pharmacol. Biochem. Behav. 40, 471e478. Herkenham, M., Lynn, A.B., Johnson, M.R., Melvin, L.S., de Costa, B.R., Mechoulam, R., Shani, A., Edery, H., Grundfeld, Y., 1970. Chemical basis of Rice, K.C., 1991. Characterization and localization of cannabinoid recep- hashish activity. Science 169, 611e612. tors in rat brain: a quantitative in vitro autoradiographic study. J. Neurosci. Menard, C.S., Hebert, T.J., Dohanich, G.P., Harlan, R.E., 1995. Androgenic- 11, 563e583. anabolic steroids modify b-endorphin immuno-reactivity in the rat brain. Hohmann, A.G., Suplita, R.L., Bolton, N.M., Neely, M.H., Fegley, D., Brain. Res. 669, 255e256. Mangieri, R., Krey, J.F., Walker, J.M., Holmes, P.V., Crystal, J.D., Meng, I.D., Manning, B.H., Martin, W.I., Fields, H.L., 1998. An analgesia Duranti, A., Tontini, A., Mor, M., Tarzia, G., Piomelli, D., 2005. An circuit activated by cannabinoids. Nature 395, 381e383. endocannabinoid mechanism for stress-induced analgesia. Nature 435, Morgan, M., Sohn, J.H., Liebeskind, J., 1989. Stimulation of the periaqueduc- 1108e1112. tal gray matter inhibits nociception at the supraspinal as well as spinal Hutcheson, D.M., Tzavara, E.T., Smadja, C., Valjent, E., Roques, B.P., level. Brain Res. 502, 61e66. Hanoune, J., Maldonado, R., 1998. Behavioural and biochemical evidence Nava, F., Carta, G., Gessa, G.L., 2000. Permissive role of dopamine D2 recep- for signs of abstinence in mice chronically treated with D9-tetrahydrocan- tors in the hypothermia induced by THC in rats. Pharmacol. Biochem. nabinol. Br. J. Pharmacol. 125, 1567e1577. Behav. 66, 183e187. Janssen, P., Neimegeers, C., Dony, J., 1963. The inhibitory effects of fentanyl Navarro, M., Carrera, M.R., Fratta, W., Valverde, O., Cossu, G., Fattore, L., (R4263) and other morphine like analgesics, on the warm water induced Chowen, J.A., Gomez, R., del Arco, I., Villanua, M.A., Maldonado, R., tail reflex in rats. Arzneimittelforsch 13, 502e507. Koob, G.F., Trodriguez de Fonseca, F., 2001. Functional interaction Johansson, P., Ray, A., Zhou, Q., Huang, W., Karlsson, K., Nyberg, F., 1997. between opioid and cannabinoid receptors in drug self-administration. Anabolic androgenic steroids increase b-endorphin levels in the ventral J. Neurosci. 21, 5344e5350. tegmental area in the male rat brain. Neurosci. Res. 27, 185e189. Nikolarakis, K.E., Pfeiffer, A., Stalla, G.K., Herz, A., 1989. Facilitation of Johansson, P., Hallberg, M., Kindlundh, A., Nyberg, F., 2000a. The effect on ACTH secretion by morphine is mediated by activation of CRF releas- opioid peptides in the rat brain, after chronic treatment with the ing neurons and sympathetic neuronal pathways. Brain Res. 498, anabolic androgenic steroid, nandrolone decanoate. Brain Res. Bull. 51, 385e388. 413e418. Pope Jr., H.G., Katz, D.L., 1994. Psychiatric and medical effects of anabolic- Johansson, P., Lindquist, A.-S., Nyberg, F., Fahlke, C., 2000b. The anabolic androgenic steroid use. A controlled study of 160 athletes. Arch. Gen. androgenic steroid nandrolone decanoate affects alcohol intake, defensive Psychiatry 51, 375e382. behaviors and brain opioid peptides in the rat. Pharmacol. Biochem. Rodriguez de Fonseca, F., Cebeira, M., Ramos, J.A., Martin, M., Fernandez- Behav. 97, 1e9. Ruiz, J.J., 1994. Cannabinoid receptors in rat brain areas: sexual differen- Kanayama, G., Cohane, G.H., Weiss, R.D., Pope, H.G., 2003. Past anabolic- ces, fluctuations during estrous cycle and changes after gonadectomy and androgenic steroid use among men admitted for substance abuse treatment: sex steroid replacement. Life Sci. 54, 159e170. an underrecognized problem? J. Clin. Psychiatry 64, 156e160. Rodrı´guez de Fonseca, F., Rubio, P., Menzaghi, F., Merlo-Pich, E., Rivier, J., Kashkin, K.B., Kleber, H.D., 1989. Hooked on hormones? JAMA 262, 3166e Koob, G.F., Navarro, M., 1996. Corticotropin-releasing factor (CRF) 3170. antagonist [D-Phe12, Nle21, 38,CaMeLeu37]CRF attenuates the acute 806 E. Ce´le´rier et al. / Neuropharmacology 50 (2006) 788e806

actions of the highly potent cannabinoid receptor HU-210 on defensive- Thiblin, I., Finn, A., Ross, S.B., Stenfors, C., 1999. Increased dopaminergic withdrawal behavior in rats. J. Pharmacol. Exp. Ther. 276, 56e64. and 5-hydroxytryptaminergic activities in male rat brain following long Schlussman, S.D., Zhou, Y., Johansson, P., Kiuru, A., Ho, A., Nyberg, F., term treatment with anabolic androgenic steroids. Br. J. Pharmacol. 126, Kreek, M.J., 2000. Effects of the androgenic anabolic steroid, nandrolone 1301e1306. decanoate, on adrenocorticotropin hormone, corticosterone and proopio- Tunving, K., 1987. Psychiatric aspects of cannabis use in adolescents and melanocortin, corticotropin releasing factor (CRF) and CRF receptor1 young adults. Pediatrician 14, 83e91. mRNA levels in the hypothalamus, pituitary and amygdala of the rat. Ulugol, A., Dost, T., Dokmeci, D., Akpolat, M., Karadag, C.H., Dokmeci, I., Neurosci. Lett. 284, 190e194. 2000. Involvement of NMDA receptors and nitric oxide in the thermoreg- Sethi, B.B., Trivedi, J.K., Kumar, P., Gulati, A., Agarwal, A.K., Sethi, N., ulatory effect of morphine in mice. J. Neural Transm. 107, 515e521. 1986. Antianxiety effect of cannabis: involvement of central benzodiaze- Valjent, E., Maldonado, R., 2000. A behavioural model to reveal place preference pine receptors. Biol. Psychiatry 21, 3e10. to D9-tetrahydrocannabinoid in mice. Psychopharmacology 147, 436e438. Sim, L.J., Selley, D.E., Childers, S.R., 1995. In vitro autoradiography of recep- Williamson, D.J., Young, A.H., 1992. Psychiatric effects of androgenic and tor-activated G-proteins in rat brain by agonist-stimulated guanylyl 5#- anabolic-androgenic steroid abuse in men: a brief review of the literature. [g[35S]thio]triphosphate binding. Proc. Natl. Acad. Sci. U.S.A. 92, J. Psychopharmacol. 6, 20e26. 7242e7246. Wilson, J.D., 1988. Androgen abuse by athletes. Endocr. Rev. 9, 181e199. Smart, R.G., Ogborne, A.C., 2000. Drug use and drinking among students in Wilson, J.D., Grifﬁn, J.E., 1980. The use and misuse of androgens. 36 countries. Addict Behav. 25, 455e460. Metabolism 29, 1278e1295. Steiner, H., Gerfen, C.R., 1998. Role of dynorphin and enkephalin in the regu- Yesalis, C.E., Bahrke, M.S., 1995. Anabolic-androgenic steroids, current is- lation of striatal output pathways and behavior. Exp. Brain Res. 123, 60e76. sues. Sports Med. 19, 326e340.