Research ArticleAddiction

µ-Opioid receptor–induced synaptic plasticity in dopamine neurons mediates the rewarding properties of anabolic androgenic steroids

See allHide authors and affiliations

Science Signaling  01 Sep 2020:
Vol. 13, Issue 647, eaba1169
DOI: 10.1126/scisignal.aba1169

Opioid signaling mediates steroid abuse

The dopamine system in the brain mediates feelings of “reward” and is implicated in addiction to various drugs, including anabolic androgenic steroids. Bontempi and Bonci found that the neurological and behavioral effects of androgenic steroids were mediated not by androgen receptors but indirectly by opioid receptors on dopaminergic neurons. A single subcutaneous injection of the steroids nandrolone or testosterone in mice stimulated the release of β-endorphins in a dopaminergic neuron-rich brain region. β-Endorphins activated mu (μ)–opioid receptor (MOR) signaling, which induced excitatory synaptic plasticity in neurons. Blocking MOR activation prevented drug-seeking behavior in steroid-injected mice. The findings may explain the addictive nature of anabolic steroid use.

Abstract

Anabolic androgenic steroids (AAS) have medical utility but are often abused, and the effects of AAS on reward circuits in the brain have been suggested to lead to addiction. We investigated the previously reported correlations between AAS and the endogenous μ-opioid system in the rewarding properties of AAS in mice. We found that a single injection of a supraphysiological dose of natural or synthetic AAS strengthened excitatory synaptic transmission in putative ventral tegmental area (VTA) dopaminergic neurons. This effect was associated with the activation of μ-opioid receptors (MORs) and an increase in β-endorphins released into the VTA and the plasma. Irreversible blockade of MORs in the VTA counteracted two drug-seeking behaviors, locomotor activity and place preference. These data suggest that AAS indirectly stimulate a dopaminergic reward center of the brain through activation of endogenous opioid signaling and that this mechanism mediates the addictive effects of AAS.

INTRODUCTION

Anabolic androgenic steroids (AAS) consist of synthetic derivatives of the primary male sex hormone, testosterone. AAS abuse can lead to physical and psychological dependence and is especially abused by males in first adulthood to enhance athletic performance or for cosmetic purposes (14). The use of AAS by teenagers has also been a primary public health concern because of their potential side effects during a period where brain remodeling and behavioral maturation occur. This, as well as other epidemiological reports, supports studies that link early initiation of AAS misuse with increased risk of psychiatric conditions, indiscriminate and unprovoked aggression, steroid dependence, and the use of other illicit drugs (5, 6). However, whether and why AAS induce physical and psychological dependence is still a matter of debate.

The rewarding properties of AAS have been commonly associated with the improvement of physical appearance, muscular strength, and athletic performance. However, in addition to their influence on somatic features, a growing body of work in adult rodents shows that androgen compounds might have rewarding and reinforcing effects (7). These effects are observed using different behavioral paradigms, such as conditioned place preference (CPP) or self-administration (811). In particular, self-administration of testosterone causes autonomic depression in a similar manner as observed with opioids and is blocked by naltrexone, an opioid antagonist (11). Furthermore, AAS seem to interact with the endogenous opioid system in both humans and rodents, suggesting that some of the AAS effects in the brain might be mediated by opioids (11, 12). In particular, chronic administration of AAS increases the concentration of the endogenous μ-opioid receptor (MOR) agonist β-endorphin in several brain regions including the dopamine (DA) system of the ventral tegmental area [VTA; (13)].

The VTA mediates reward prediction, positive reinforcement, motivation, and reward/drugs-seeking behavior (14, 15). All drugs of abuse interact with DA neurons to increase their activity and the level of DA in brain regions, such as the nucleus accumbens (16). Moreover, drugs of abuse also induce long-term alterations in glutamatergic synaptic transmission on DA neurons, which represents a learning mechanism that is both sufficient and necessary to develop and maintain drug addiction or stress-related behavior (1720). Current hypotheses are that AAS might enhance DAergic activity because testosterone place preference is DA dependent (8, 10). However, there is now insufficient evidence of a direct action by androgens and the mechanisms by which AAS modulate DA neuron activity and synaptic transmission.

Along this line of thought, the first aim of this work was to assess the long-lasting effect of a high dose of the synthetic androgen nandrolone and the endogenous androgen testosterone, two of the most abused AAS, on the physiology of putative DA neurons. The second aim was to understand the mechanism by which AAS interact with the VTA, focusing on the interaction between AAS and the endogenous MOR agonist β-endorphin. Last, we attempted to demonstrate how these mechanisms are involved in the rewarding properties of AAS.

RESULTS

Single administration of AAS induces synaptic plasticity on putative DA neurons

Several lines of evidence show that a single exposure of drugs of abuse is capable of rapidly causing long-lasting neuroadaptive changes onto VTA DA neurons (17, 2123). Synaptic plasticity on DA neurons plays a pivotal role in the development of drug addiction (24). Moreover, it also affects the individual susceptibility to relapse, which is one of the most critical parts in the treatment of drug addiction (24). Thus, studying the long-lasting effect and the mechanisms by which drugs of abuse interact with the VTA are crucial to understand the neurobiology of drug addiction and to discover new therapeutic targets to treat drug dependence. On these bases, we used whole-cell electrophysiological recordings from acute brain slices containing the VTA from male mice treated, 24 hours before the recording, with a single intraperitoneal injection of either vehicle or a supraphysiological concentration of either nandrolone or testosterone. The dose was chosen on the basis of previous studies to simulate a similar amount of AAS abused by people (2528). First, we measured the firing rate of putative DA neurons. Putative DA neurons were recorded medial to the medial nucleus of the optic tract and identified by morphology, tonic spike rate, and presence of a hyperpolarization-induced Ih current. There was no change in the frequency of firing of putative DA neurons at resting (fig. S1, A and B). Moreover, also the membrane capacitance, resistance, and resting membrane potential—after hyperpolarization potential Ih current and neurons firing after different depolarizing steps—were unaltered (fig. S1, C to H). These data suggested that a single injection of AAS did not induce any alteration of the intrinsic properties of putative DA neurons. Next, we studied the long-lasting effect of AAS on glutamatergic synaptic plasticity. First, we pharmacologically isolated spontaneous miniature excitatory postsynaptic currents (mEPSCs; Fig. 1A). Mean amplitude and frequency distribution of mEPSC were significantly increased in nandrolone- and testosterone-treated animals compared to vehicle-treated mice, as indicated by a shift in the cumulative probability distributions (Fig. 1, B and C). The observed increase in amplitude and frequency of mEPSC may be attributable to increased AMPA receptors (AMPARs) function and/or number. To assess synaptic strength, we calculated the ratio of AMPAR-mediated synaptic currents (AMPAR EPSCs) to N-methyl-d-aspartate receptor (NMDAR)–mediated synaptic currents (NMDAR EPSCs). In animals treated with either nandrolone or testosterone, there was a significant increase of AMPAR/NMDAR ratio (Fig. 1D). To understand whether this change in AMPAR/NMDAR was associated to either an increase of AMPAR EPSC or decrease of NMDAR EPSC or both, we performed an input-output experiment pharmacologically isolating either AMPAR or NMDAR EPSC. We reported a significant increase of AMPAR EPSC in mice treated with AAS compared to vehicle-treated mice (fig. S2A). On the contrary, we reported no change in NMDAR EPSC in any experimental condition (fig. S2B), suggesting that the increase of AMPAR/NMDAR ratio might be attributed to an increase of AMPAR function or expression. To verify possible presynaptic alterations, we performed a paired-pulse stimulation experiment to test changes in glutamate release probability. There were no substantial changes in the paired-pulse ratio neither in nandrolone- nor in testosterone-treated mice (Fig. 1E). These data might suggest that there was no presynaptic alteration after treatment with AAS, even though we cannot completely rule out this hypothesis. Drug-evoked potentiation of synaptic transmission is usually, in part, mediated through an exchange of GluR2-containing and GluR2-lacking AMPAR, leading to EPSCs that are sensitive to polyamines and have a rectifying current-voltage relationship (29). Thus, we measured EPSC at −70, 0, and +40 mV to calculate the rectification index (EPSC−70mV/EPSC+40mV). In mice treated with either nandrolone or testosterone, the rectification index was significantly higher than control mice, suggesting the presence of GluR2-lacking AMPAR (Fig. 1F). Last, we tested possible alterations of inhibitory γ-aminobutyric acid (GABA) transmission on putative DA neurons. Neither the treatment with nandrolone nor testosterone showed a long-lasting effect on GABAergic transmission. As a matter of fact, there were no significant differences in neither the frequency nor the amplitude of the miniature inhibitory postsynaptic current (mIPSC; fig. S3, A to C). In summary, these results showed that a single injection of AAS induces long-term alteration of glutamatergic synaptic transmission on putative DA neurons.

Fig. 1 AAS strengthen excitatory synaptic transmission on putative DA neurons.

(A to C) Representative traces (A) and analysis of mEPSC frequency (B) and amplitude (C) in putative VTA DA neurons of mice treated with a single injection of either testosterone (TS; 10 mg/kg), nandrolone (ND; 10 mg/kg), or vehicle (Veh). Cumulative distributions of amplitude and frequency were analyzed by using Kolmogorov-Smirnov test: P < 0.01 versus Veh. (D) AMPAR/NMDAR ratio of mice treated with AAS as described in (A) to (C). (E) Paired-pulse ratio obtained at 50-ms interpulse interval from mice treated as described in (A) to (C). (F) I-V (current-voltage relation) plot of pharmacological isolated AMPAR EPSC (left). All EPSCs were normalized to the EPSC amplitude measured at −70 mV. Rectification index (middle) of AMPAR EPSCs was calculated as the ratio between the amplitude of EPSC obtained at −70 and +40 mV. In (A) to (F), values are means ± SEM of n = 8 neurons recorded from four mice per condition, analyzed by one-way ANOVA with Bonferroni post hoc test: *P < 0.05, **P < 0.01 versus Veh. Representative traces are shown (right).

AAS-induced synaptic plasticity is not mediated by androgen receptor activation

Next, we wanted to describe the mechanism by which AAS modulate glutamatergic synaptic transmission on putative VTA DA neurons. We started our study by investigating the canonical target for androgens: the intracellular androgen receptor. This receptor is expressed in many brain regions, especially in the hypothalamus (30). Androgen receptor expression is reported also in other limbic regions such as the VTA, even though, the physiological role of this receptor in this region remains unknown (31, 32). Thus, we pretreated mice with a subcutaneous injection of the selective androgen receptor antagonist flutamide 60 min before the administration of either AAS or the vehicle. The pretreatment with flutamide did not affect neither the frequency nor the amplitude of mEPSC in testosterone- or nandrolone-treated mice (Fig. 2, A to C). To understand whether treatment with flutamide affected the strength of glutamatergic synaptic transmission, we evaluated the AMPAR/NMDAR ratio. In addition, in this situation, we showed that flutamide did not block the increase of the AMPAR/NMDAR ratio in AAS-treated mice (Fig. 2D). We also reported that pretreatment with flutamide might not affect the presynaptic glutamate release because it did not impair the pair-pulse ratio in any experimental group (Fig. 2E). In the end, we showed that pretreatment with flutamide did not affect also the rectification of AMPAR EPSC in mice treated with AAS (Fig. 2F). These data suggested that AAS induced the alteration of glutamatergic synaptic transmission on putative VTA DA neurons through a noncanonical mechanism without engaging the androgen receptor.

Fig. 2 Androgen receptor antagonist flutamide did not affect AAS-induced synaptic plasticity.

(A to C) Representative traces (A) and analysis of mEPSC frequency (B) and amplitude (C) in putative VTA DA neurons of mice pretreated with flutamide (FLU; 20 mg/kg, sc) or saline (Sal) 1 hour before a single intraperitoneal injection of either testosterone (TS; 10 mg/kg), nandrolone (ND; 10 mg/kg), or vehicle. Cumulative distributions of amplitude and frequency were analyzed by using Kolmogorov-Smirnov test: P < 0.01 for the ND/TS+ conditions versus the Veh conditions. (D) AMPAR/NMDAR ratio in putative DA neuron of mice treated as described in (A) to (C). (E) Paired-pulse ratio obtained at 50-ms interpulse interval in putative DA neurons of mice treated as described in (A) to (C). (F) I-V plot of pharmacological isolated AMPAR EPSC (left). All EPSCs were normalized to the EPSC amplitude measured at −70 mV. Rectification index (middle) of AMPAR EPSCs was calculated as the ratio between the amplitude of EPSC obtained at −70 and +40 mV. Representative traces are shown (right). In (A) to (F), values are means ± SEM of n = 8 neurons recorded from four mice per condition, analyzed by one-way ANOVA with Bonferroni post hoc test: *P < 0.05; #P < 0.05; **P < 0.01; ##P < 0.01 versus Veh + Sal or Veh + FLU, respectively.

AAS increases the plasma and VTA concentration of β-endorphin

Several studies in humans and rodents suggest that AAS may modulate brain activity, enhancing the endogenous opioid system. It has previously been demonstrated that chronic treatment with nandrolone increases the level of β-endorphin in the VTA (13), whereas chronic treatment with testosterone increases the concentration of β-endorphin in the hypothalamus, pituitary gland, and plasma (3335), suggesting that AAS might boost the synthesis or release of endogenous MOR agonists. We decided to investigate whether an acute administration of AAS was sufficient to modulate the plasma and VTA concentration of β-endorphin (Fig. 3A). The results showed that a single injection of either nandrolone or testosterone induced a significant increase of plasma and VTA concentration of β-endorphin (Fig. 3B). The effect was evident 5 to 15 min after the injection with AAS, whereas the level of β-endorphin returned to the basal level after 1 hour. Together, these results showed that AAS were capable of quickly increasing the concentration of the endogenous MOR agonist β-endorphin in both plasma and VTA. β-Endorphin might represent a promising target by which AAS modulate synaptic transmission on DA neurons.

Fig. 3 AAS induce rapid increase of plasma and VTA β-endorphin levels.

(A) Schematic of experiment: Mice were injected with either testosterone (TS; 10 mg/kg), nandrolone (ND; 10 mg/kg), or vehicle (Veh). After different time points (5′, 15′, 30′ and 60′), either VTA or plasma was isolated and processed for β-endorphin quantification through ELISA. (B) Quantification of β-endorphin level in the VTA (left) and plasma (right) of mice treated as described (A). Values are means ± SEM; five mice in each condition were analyzed by one-way ANOVA with Bonferroni post hoc test: *P < 0.05; **P < 0.01 versus Veh. i.p., intraperitoneally.

Blockade of MOR counteracts AAS-induced synaptic plasticity

The rewarding properties of exogenous and endogenous opioids, such as β-endorphin, are well demonstrated, and it is also shown that acute treatment with opioids induces synaptic plasticity in DA neurons (36, 37). Thus, we assumed that AAS might indirectly modulate synaptic transmission in DA neurons, increasing the concentration of β-endorphin and consequently activating MORs. On these bases, we performed intracerebroventricular (ICV) injections of the selective, irreversible MOR antagonist β-funaltrexamine (β-FNA) 24 hours before the injection of either AAS or vehicle (Fig. 4A). Our results showed that pretreatment with β-FNA blocked AAS-induced increase of mEPSC frequency and amplitude. β-FNA did not affect mEPSC amplitude and frequency in vehicle-treated mice (Fig. 4, B to D). We then investigated whether β-FNA was able to block the strength of glutamatergic synapses on putative DA neurons. Our result showed that the irreversible blockade of MOR was capable of dampening the increase of the AMPAR/NMDAR ratio induced by treatment with AAS (Fig. 4E). We also found that pretreatment with β-FNA appeared to have no effect on presynaptic glutamate release because the excitatory prepulse ratio was unaltered in any of the experimental conditions used (Fig. 4F). Rather, pretreatment with β-FNA blocked AAS-induced increases in the AMPAR EPSC rectification index but did not affect the AMPAR rectification index in vehicle-treated mice (Fig. 4G). These data indicate that AAS are capable of modulating synaptic strength on putative DA neurons through a MOR-dependent mechanism that may be attributed to an increase in β-endorphin levels in the VTA.

Fig. 4 AAS-induced synaptic plasticity on putative DA neurons is mediated by MOR activation.

(A) Coronal section depicting cannula implantation in the left lateral ventricle and schematic of experiment: Mice received an ICV injection of the irreversible MOR antagonist β-FNA (0.2 μg in 0.2 μl of saline) or saline (Sal) 24 hours before injecting either testosterone (TS; 10 mg/kg), nandrolone (ND; 10 mg/kg), or vehicle (Veh). Scale bar, 1 mm. Whole-cell patch-clamp recordings were performed 24 hours after the treatment with either AAS or Veh. (B to D) Representative traces (B) and analysis of mEPSC frequency (C) and amplitude (D) recorded in putative VTA DA neurons of mice treated as described (A). Cumulative distributions of amplitude and frequency were analyzed by using Kolmogorov-Smirnov test: P < 0.01 for ND + β-FNA and TS + β-FNA versus ND + Sal, TS + Sal, Veh + Sal, and Veh + β-FNA. (E) AMPAR/NMDAR ratio in putative DA neurons of mice treated as described (A). (F) Paired-pulse ratio obtained at 50-ms interpulse interval. (G) I-V plot and rectification index in putative DA neurons of mice treated as described (A). I-V plot of pharmacological isolated AMPAR EPSC (left). All EPSCs were normalized to the EPSC amplitude measured at −70 mV. Rectification index (middle) of AMPAR EPSCs was calculated as the ratio between the amplitude of EPSC obtained at −70 and +40 mV. Representative traces (right). In (C) to (F), values are means ± SEM of n = 8 neurons recorded from four mice per condition, analyzed by one-way ANOVA with Bonferroni post hoc test: **P < 0.01 versus Veh + Sal; †P < 0.05 and ††P < 0.01 versus Veh + β-FNA; ‡P < 0.05; ‡‡P < 0.01 versus ND + β-FNA; §P < 0.05; §§P < 0.01 versus TS + β-FNA.

We then sought to test this hypothesis. As presented above, AAS rapidly increased the level of β-endorphin, which returned to basal level after almost 1 hour. Thus, to test whether inhibiting MOR activity also blocked AAS-induced synaptic plasticity when β-endorphin returns to basal level, we performed an ICV injection of β-FNA 1 hour after the administration of AAS (fig. S4A). Our results showed that, in this experimental setting, β-FNA did not block AAS-induced increase of mEPSC frequency and amplitude in putative DA neurons (fig. S4, B to D). Moreover, injection of β-FNA 1 hour after the AAS injection did not affect either the AAS-induced increase in the AMPAR/NMDAR ratio (fig. S4E) or the rectification index of AMPAR EPSCs (fig. S4F) on putative DA neurons. These results suggested that AAS-induced increases in β-endorphin levels in the VTA might be necessary to enhance excitatory synaptic transmission on putative DA neurons.

AAS-induced locomotion and place preference requires activation of VTA MOR

The VTA is a brain region centrally involved in the development and expression of a variety of behaviors associated with drug use such as increased locomotor activity and drug-seeking behavior (38). To determine whether AAS modulate locomotor activity, mice were injected with a single administration of either testosterone, nandrolone, or vehicle, and the locomotor activity was monitored for 1 hour. Our results showed that both testosterone and nandrolone significantly increased the locomotor activity 10 to 15 min after the injection (Fig. 5A). The locomotor activity remained high for almost 45 min, which might resemble a similar time course shown with the increase of β-endorphin level. To test whether the increase of the locomotor activity was due to activation of VTA MOR, we pretreated mice with an intra-VTA injection of the irreversible MOR antagonist β-FNA (Fig. 5B and fig. S5) and found that this blocked the increase in locomotor activity in both nandrolone- and testosterone-treated mice (Fig. 5C). These results suggest that the acute increase in locomotor activity induced by AAS was dependent on VTA MOR activation.

Fig. 5 AAS-induced locomotor activity required VTA MOR activation.

(A) Locomotor activity in mice treated with a single intraperitoneal injection with either testosterone (TS; 10 mg/kg), nandrolone (ND; 10 mg/kg), or vehicle (Veh). Values are means ± SEM (nine mice per group), analyzed by one-way ANOVA with Bonferroni post hoc test: P < 0.01 ND versus Veh and P < 0.001 TS versus Veh at 15 min; P < 0.01 ND versus Veh and P < 0.001 TS versus Veh at 30 min. (B) Coronal section depicting intra-VTA cannula implantation. (C) Locomotor activity in mice pretreated with an intra-VTA injection of β-FNA (0.2 μg in 0.2 μl of saline) or saline (0.2 μl; Sal), 24 hours before a single intraperitoneal injection with either TS, ND, or Veh as described in (A). Values are means ± SEM of seven mice per group, analyzed by one-way ANOVA with Bonferroni post hoc test: P < 0.01 for ND + Sal versus TS + β-FNA, TS + Sal versus Veh + Sal, and TS + Sal versus Veh + β-FNA; P < 0.001 for ND + Sal versus Veh + β-FNA and P < 0.0001 for ND + Sal versus Veh + Sal at 15 min; P < 0.05 for TS + Sal versus TS + β-FNA; P < 0.01 for ND + Sal versus Veh + β-FNA; P < 0.001 for ND + Sal versus Veh + Sal, ND + Sal versus ND + β-FNA, and TS + Sal versus Veh + β-FNA; P < 0.0001 for TS + Sal versus Veh + Sal and TS + Sal versus ND + β-FNA at 30 min.

Several papers show the rewarding and reinforcing effect of both testosterone and nandrolone using either CPP or self-administration paradigms (68). To determine whether AAS also induced place preference in our experimental conditions, we performed CPP experiments in mice conditioned with either testosterone or nandrolone (Fig. 6A) and found that both steroids were capable of inducing a strong and significant CPP in mice (Fig. 6B), demonstrating the rewarding effect of AAS. Next, we investigated whether the development of CPP was dependent on VTA MOR activation. We took advantage of the irreversible antagonist β-FNA to induce a chronic pharmacological blockade of VTA MOR. However, to minimize the number of β-FNA injections in the VTA, we assessed the duration of MOR inactivation after a single microinjection of β-FNA in the VTA (fig. S6A). MORs are expressed on presynaptic GABAergic terminals, and their activation reduces GABA release on DA neurons. To verify when MOR activity was restored after β-FNA injection, we analyzed the effect of the MOR agonist [D-Ala2, N-MePhe4, Gly-ol]-enkephalin (DAMGO) on putative DA neurons GABAergic synaptic transmission. DAMGO did not modify the frequency of mIPSC 1 or 2 days after β-FNA injection (fig. S6, B to F) but did significantly by 3 days after (fig. S6, G and H). These data suggested that the intra-VTA injection of β-FNA could block the activity of VTA MOR for 48 hours. Thus, we microinjected β-FNA in the VTA (fig. S7) every 2 days, 20 hours before the AAS conditioning session (Fig. 6, C and D). Our results showed that irreversible chronic blockade of VTA MOR dampened the development of both nandrolone and testosterone place preference (Fig. 6, E and F). To finally confirm that β-FNA injection did not affect the result of the CPP due to its possible aversive effect, we performed another CPP experiment, wherein the AAS and vehicle conditioning sessions were inverted and β-FNA was injected in the VTA (fig. S8) 20 hours before the vehicle conditioning session (fig. S9A). Our result showed that also in this situation, β-FNA was still capable to block the development of AAS-induced place preference (fig. S9, B and C). Together, these data indicate that the rewarding effect of AAS is mediated by VTA MOR activation, likely due to an increase of β-endorphins in the VTA.

Fig. 6 Irreversible blockade of VTA MOR counteracted AAS-induced place preference.

(A) Schematic of experiment: During the pretest, mice were allowed to explore all compartments and were monitored for 15 min to assess any preference for a certain compartment. During conditioning, mice were injected in the morning during days 2, 4, and 6 and in the afternoon during days 3 and 5 with either testosterone (TS; 10 mg/kg) or nandrolone (ND; 10 mg/kg) and were confined to the not preferred compartment. In the afternoon of days 2, 4, and 6 or in the morning of days 3 and 5, mice were treated with vehicle (Veh) and were confined to the preferred side. The test day occurred in the middle of the day, and the mice were allowed to explore all compartments for 15 min. (B) Conditioned place preference (CPP) in mice conditioned as described (A). CPP score was calculated, subtracting the time spent in the drug-paired side to the Veh-paired side. Values are means ± SEM (10 mice per group). Two-tailed paired Student’s t test; ***P < 0.001, ****P < 0.0001. (C) Coronal section depicting intra-VTA cannula implantation (scale bar: 2mm). (D) Schematic of experiment: The experiment was performed as shown previously (A), except for the fact that β-FNA (0.2 μg in 0.2 μl of Sal) was injected in the VTA every 48 hours to maintain VTA MOR blocked for the whole duration of the experiment. The injection occurred on day 1 after the pretest and on days 3 and 5 after the conditioning session to not allow the association between any possible aversive effect of β-FNA with any compartment of the CPP apparatus. (E and F) CPP in mice conditioned as described (D). Values are means ± SEM (10 mice per group); two-tailed paired t test comparing the time spent on the drug-paired side before and after conditioning. ***P < 0.001, ****P < 0.0001.

DISCUSSION

AAS abuse remains a poorly understood and underappreciated social problem despite the risk of addiction and side effects. According to several studies, about 30 to 32% of subjects using AAS will suffer from dependence; this may lead to severe side effects such as increased aggressive behavior and depression, which is associated with a high risk of suicide (39, 40). Thus, studying how AAS modulate the function of brain regions usually involved in drug addiction and the mechanisms underlying the rewarding properties of AAS might be of pivotal importance to find therapeutic targets capable of counteracting AAS abuse and side effects.

A major hypothesis in the pathogenesis of addictive behavior points at drug-induced plasticity within the VTA DA system as one of the major causes of compulsive drug-seeking and relapse behavior (24, 34). All drugs of abuse typically increase the excitatory transmission on DA neurons, and it is demonstrated how these long-term alterations of synaptic transmission are related to the behavioral effects of addictive drugs (that is, CPP or drug sensitization or reinstatement of drug self-administration). Some studies also report that AAS might interact with the DAergic system because AAS place preference is DA dependent, and chronic treatment with AAS induces alteration of DA receptors in many brain regions (810).

In this study, we found that AAS can directly affect the physiology of DA neurons. Our results revealed that a single administration of two of the most common AAS was capable of strengthening the excitatory synaptic transmission on putative VTA DA neurons. We reported an increase of mEPSC amplitude and frequency as well as an increase of the AMPAR/NMDAR ratio due to an increase of AMPAR EPSC and increase of the rectification index of AMPAR EPSC. These data demonstrated that AAS, as well as other drugs of abuse, were capable of strengthening excitatory transmission on putative DA neurons. We also found that this effect was not mediated by androgen receptor activation, which suggests the involvement of noncanonical mechanisms. Administration of AAS in humans induces a feeling of pleasure within 15 to 20 min of administration, which indicates the presence of a fast, nongenomic action (41). Previous studies suggest several receptors as potential targets for the nongenomic effect of androgens. These include GABA receptors, the sigma-1 receptor (σ1R), and a putative membrane receptor for androgens (42, 43). Moreover, several studies report that the nongenomic action of androgens might also be linked to the endogenous opioid system (12). Fifty percent of people abusing AAS meet the diagnostic criteria for opioid use disorder (12). Furthermore, self-administration of testosterone in rodents is blocked by the presence of the opioid antagonist naltrexone. Pretreatment with testosterone induces CPP to subthreshold doses of morphine, and chronic treatment with nandrolone increases the level of β-endorphin in the VTA (1113). In our study, we also showed that a single injection of either testosterone or nandrolone quickly increased the VTA and plasma level of β-endorphin. This increase of β-endorphin supports the assumption that androgens might affect neuronal activity by enhancing the endogenous opioid system activity, even though the mechanisms and the body regions by which androgens promote the release of β-endorphin are not yet known and require further investigation.

The rewarding effect of β-endorphin and its role in natural and drug reward are well described (36, 37). However, the involvement of the endogenous β-endorphin in drug-induced synaptic plasticity is still a matter of debate. Another study shows that a raise in the endogenous level of β-endorphin might be linked to the rewarding properties of ultraviolet light (43). Thus, although it is not completely demonstrated, it would not be unexpected if β-endorphin played a role in the rewarding properties of AAS as well. We showed that preexposure to the selective irreversible MOR antagonist β-FNA dampened the AAS-induced synaptic plasticity in putative VTA DA neurons, demonstrating that activation of MOR was required for AAS action in the VTA. MOR can regulate DA neuron activity and synaptic transmission either through a presynaptic mechanism, reducing GABA release or promoting glutamate release (44), or through a postsynaptic mechanism because they are also expressed in DA neurons (45). Moreover, some authors report that opioids induce synaptic plasticity on DA neurons, regulating the release of other neuromodulators such as orexin (23). Our study does not uncover whether AAS-dependent synaptic plasticity and behavioral effects are mediated by presynaptic or postsynaptic VTA MOR or whether other neuromodulators are involved. This, too, will require further investigation.

Taking advantage of the CPP paradigm to study whether the rewarding effect of AAS was dependent on VTA MOR activation, our results further showed that the irreversible inhibition of MOR in the VTA was sufficient to block AAS-induced place preference and locomotor activity. Another study shows that systemic injection with naloxone blocks testosterone self-administration, suggesting that the endogenous opioids system might be involved in the rewarding properties of this natural hormone. Along this line, our findings provide evidence that AAS-induced enhancement of μ-opioid signaling is involved in the physiological, long-lasting alterations in the VTA commonly seen after exposure to drugs of abuse and that are implicated in the development of drug-seeking behavior. Overall, we did not report any substantial difference between testosterone and nandrolone in any experiment. This suggests that, regarding the aspects studied in this paper, they probably work through the same mechanisms and induce the same effect.

Although our results need to be replicated in other behavioral paradigms, the same mechanism (that of MOR-mediated glutamatergic synaptic transmission on putative VTA DA neurons) may be involved in the rewarding properties of AAS. The interaction between AAS and endogenous opioid system might be of pivotal importance here, given the co-abuse of AAS and opioids. AAS have been previously defined as a gateway drug to opioid dependence. It is also plausible that if AAS addiction is wired through opioid signaling, then people suffering of AAS addiction might be treated with the same approved medication used for opioid abuse, such as buprenorphine. Moreover, further investigations about the plasma concentration of endogenous opioids or of brain opioid receptor binding should also be performed in people abusing AAS. Ultimately, understanding the connection between androgens and opioids could be crucial toward understanding the development of side effects, such as aggressive behavior or depression, commonly seen with chronic abuse of AAS.

MATERIALS AND METHODS

Animals

Male mice (8 to 10 weeks old) were group housed in a colony maintained under a 12-hour light/dark cycle with food and water available ad libitum. All experiments were carried out during the mouse dark cycle. All analyses were performed on mice whose genotype was unknown to the experimenter. All procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health.

Drugs

Testosterone was purchased from Cayman Chemicals (Ann Arbor, MI, USA; catalog no. 15645). Nandrolone was purchased from Sigma-Aldrich (St. Louis, MO, USA; catalog no. N7252). Flutamide, β-FNA, DAMGO, and 6,7-Dinitroquinoxaline-2,3-dione (DNQX) disodium salt were purchased from Tocris (Minneapolis, MN; catalog nos. 4094, 0926, 1171, and 2312). d-aminophosphovalerate (d-APV), tetrodotoxin citrate (TTX), and picrotoxin were purchased from Abcam (Cambridge, MA; catalog nos. ab120003, ab120055, and ab120315). Testosterone and nandrolone were dissolved in 45% (2-hydroxypropyl)-β-cyclodextrin (Sigma-Aldrich, catalog no. H107). Flutamide and β-FNA were dissolved in 0.9% NaCl solution (saline). DAMGO, DNQX, d-APV, TTX, and picrotoxin were dissolved in water.

Electrophysiology

Male mice were treated with a single intraperitoneal injection of either vehicle or testosterone (10 mg/kg) or nandrolone (10 mg/kg). After 24 hours, mice were anesthetized with Euthasol and decapitated. Brains were quickly removed and placed in ice-cold low-sodium artificial cerebrospinal fluid (ACSF). For pharmacological experiments with antagonist, mice were pretreated with flutamide (20 mg/kg) or β-FNA ICV injection (2 μg in 0.2 μl of saline) 1 or 24 hours before the injection of AAS, respectively. Horizontal sections (250 μm) containing the VTA were prepared in ice-cold ACSF using a vibrating blade microtome (Leica VT1200, Buffalo Grove, IL, USA). Right after cutting, slices were recovered for 10 min at 32°C and then transferred to holding ACSF at room temperature. Cutting and recovery were performed with ACSF containing the sodium substitute N-methyl-d-glucamine (NMDG) (45): 92 mM NMDG, 20 mM Hepes (pH 7.35), 25 mM glucose, 30 mM sodium bicarbonate, 1.2 mM sodium phosphate, 2.5 mM potassium chloride, 5 mM sodium ascorbate, 3 mM sodium pyruvate, 2 mM thiourea, 10 mM magnesium, 14 mM sulfate, and 0.5 mM calcium chloride (46). The ACSF used for holding slices before recording was identical but contained 92 mM sodium chloride instead of NMDG and contained 1 mM magnesium chloride and 2 mM calcium chloride. The ACSF used to perfuse slices during recording contained 125 mM sodium chloride, 2.5 mM potassium chloride, 1.25 mM sodium phosphate, 1 mM magnesium chloride, 2.4 mM calcium chloride, 26 mM sodium bicarbonate, and 11 mM glucose. All ACSF solutions were saturated with 95% O2 and 5% CO2. For recording, a single slice was transferred to a heated chamber (32°C) and perfused with normal ACSF (2.5 ml min−1) using a peristaltic pump (World Precision Instruments, Sarasota, FL, USA). Visualization of putative VTA DA neurons was performed with an upright microscope equipped for differential interference contrast microscopy (BX51WI, Olympus, Waltham, MA, USA). Putative DAergic neurons in the VTA were located medial to the medial nucleus of the optic tract and identified by morphology, tonic spike rate, and presence of a hyperpolarization-induced Ih current, which can be a reasonable predictor of DAergic identity in mice (47). Whole-cell patch-clamp recordings were made using a MultiClamp 700B amplifier (1 kHz low-pass Bessel filter and 10 kHz digitization) with pClamp 10.3 software (Molecular Devices, San Jose, CA, USA). Voltage-clamp recordings of excitatory synaptic transmission were made using glass pipets with resistance of 1.5 to 3 megohms, filled with internal solution containing the following: 117 mM cesium methanesulfonate, 20 mM Hepes, 0.4 mM EGTA, 2.8 mM NaCl, 5 mM Tetraethylammonium chloride (TEA-Cl), 2.5 mM Mg-ATP (adenosine triphosphate), and 0.25 mM Na-GTP (guanosine triphosphate); pH 7.2 to 7.3 and 280 to 285 mOsm. Spermine (0.1 mM) was added to the internal solution for measuring the current-voltage relationship of AMPAR currents. For inhibitory transmission, GABA-A mIPSCs were recorded with glass microelectrodes (1.5 to 3 megohms) containing the following: 128 mM KCl, 20 mM NaCl, 10 mM Hepes, 1 mM MgCl2, 1 mM EGTA, 0.3 mM CaCl2, 2 mM Mg-ATP, and 0.25 mM Na-GTP (280 to 285 mOsm) to detect mIPSC as large inward currents at −65 mV. Input and series resistance was continually monitored online; if either parameter changed by more than 20%, data were not included in the analysis. Membrane potentials were not corrected for junction potentials (estimated to be 10 mV). Excitatory afferents were stimulated at 0.1 Hz with a bipolar stimulating electrode placed 100 to 300 μm rostral to the recording electrode. Excitatory paired-pulse ratios were acquired at −70 mV by having a second afferent stimulus of equal intensity at 50 ms after the initial stimulus. The ratio was calculated from the peak amplitude of the second and the first stimuli. The experiments looking at AMPAR/NMDAR ratios were carried out by evoking the dual-component EPSC at +40 mV, and then d-APV (50 μM) was bath applied to isolate AMPAR EPSC. NMDAR responses were calculated by digital subtraction of the average response in the presence of d-APV from the response isolated in its absence. To yield the AMPAR/NMDAR ratio, the peak of the AMPAR EPSC was divided by the peak of the NMDAR EPSC. We measured the rectification of AMPAR-mediated currents by pharmacologically isolating AMPAR EPSC at a holding potential ranging from −70 to +40 mV and then normalizing the peak amplitude of each holding potential to the current at −70 mV. mEPSCs were pharmacologically isolated by having picrotoxin (100 μM) and tetrodotoxin (20 nM) present throughout the experiment and sampled at 1 kHz while clamping the cells at −70 mV. mIPSCs were isolated by having NBQX (2,3-dihydroxy-6-nitro-7-sulfamoylbenzo[f]quinoxaline) (10 μM) and d-APV (50 μM) and tetrodotoxin (20 nM) and were clamped at −65 mV. Two hundred events per cell were acquired and detected using a threshold of 8 pA. Analyses of mEPSC and mIPSC were performed offline and verified by eye using the MiniAnalysis program (v6.0, Synaptosoft). For current-clamp recordings looking at cell excitability, electrodes were filled with the following: 135 mM K-gluconate, 10 mM Hepes, 4 mM KCl, 4 mM Mg-ATP, and 0.3 mM Na-GTP; pH 7.2 to 7.3 and 280 to 285 mOsm. We measured cell excitability with incremental steps of current injections and monitored the number of spikes fired during a fixed current injection (2 s). All values were obtained after cells had reached a stable response, and then averages of three cycles for each cell were taken. No more than one to two cells per animal were used for each experimental setting.

Surgery and cannula implantation

Mice were anesthetized using ketamine (1 ml of 100 mg/ml solution) mixed with xylazine (0.1 ml of 100 mg/ml solution) in injectable saline (8.9 ml) and intraperitoneally administered in a volume of 0.1 ml/10 g of mouse weight. For the assessment of the MOR irreversible blockade by β-FNA, a solution containing β-FNA (0.2 μg) diluted in saline solution and 50% (v/v) of fluorescent microspheres (Thermo Fisher Scientific, Waltham, MA, USA; catalog no. F10720) was injected in the VTA [AP (anterior-posterior), −2.9; ML (medio-lateral), ±1.6; DV (dorso-ventral), −4.6] in a volume of 0.2 μl. Fluorescent beads were used to detect the site of injection during the patch-clamp recording, ensuring that the drug has been delivered to the VTA. Slices without detectable fluorescence in the VTA were excluded. For the pharmacological studies with β-FNA, mice were implanted with guide cannulas above the VTA (AP, −2.9; ML, ±1.6; DV, −3.6; the microneedle protruded 1 mm from the cannula tip) or the left lateral ventricle for ICV injection (AP, −0.6; ML, −1.2; DV, −1.8). For VTA cannula implantation, two holes were drilled through the skull for simultaneous placement of guide cannulas. A third hole was drilled about 2 mm anterior to the guide cannula holes; this hole was enlarged for the placement of an anchor screw, which provided an additional surface area for the skull cap to adhere. Dental cement was applied to the exposed cranium with the aim of securing the guide cannulas and the anchor screw in place. After about 4 hours of recovery, all animals were returned to their reverse light/dark cycle holding room and given free access to food and water.

Behavioral analyses

Locomotor activity was performed in a circular locomotor box with a 40-cm diameter. The day before the experiment, mice were habituated to the apparatus for 30 min. On the next day, mice were injected with either vehicle, testosterone (10 mg/kg), or nandrolone (10 mg/kg) and returned to the same locomotor box. Locomotor activity was monitored for 1 hour. For experiments with intra-VTA cannula microinjection, mice were allowed to recover from the surgery for at least 2 weeks. Mice received an intra-VTA injection of β-FNA (0.2 μg in 0.2 μl of saline) 24 hours before the test. Locomotor activity was monitored using the Noldus EthoVision software.

The CPP apparatus consisted of three chambers. The conditioning compartments consisted of a grid rod-style floor white compartment and a mesh-style floor black compartment connected to a neutral central small compartment (Med Associates Inc., St. Albans, VT, USA). On the first day, mice were allowed to explore the chambers for 15 min. The time spent in each chamber was recorded to determinate any preference for one compartment. During conditioning days 2, 4, and 6, mice received an intraperitoneal injection in the morning of either nandrolone (10 mg/kg) or testosterone (10 mg/kg) and were confined in the less preferred compartment for 30 min. In the afternoon, they received a vehicle injection and were confined in the preferred compartment. On days 3 and 5, the session order was inverted. The test occurred on day 7, where the mice were allowed to explore all compartments for 15 min. During another experimental setting, mice were treated with an intra-VTA injection of either β-FNA (0.2 μg in 0.2 μl of saline) or saline (0.2 μl of saline). The microinjection was performed every 48 hours to irreversibly block the VTA MOR for the entire duration of the experiment (during both AAS and vehicle conditioning sessions). The injection of β-FNA is performed 20 hours before the next conditioning session to avoid any association between β-FNA injection and the chambers of the CCP apparatus. Time spent in the drug-paired chamber minus time spent in the saline-paired chamber was assessed.

ELISA assay

Tissue samples containing the VTA or plasma were obtained from mice injected with either testosterone (10 mg/kg), nandrolone (10 mg/kg), or vehicle at different time points (5, 15, 30, and 60 min). Tissue samples were weighted and homogenized in cold 1× phosphate-buffered saline (PBS) (pH 7.4), and the supernatant was collected after centrifuging for 20 min at 2000 to 3000 rpm. For plasma extraction, whole blood was collected into EDTA-coated Eppendorf (10% 0.5 M sterile EDTA of the expected blood volume into a 2-ml Eppendorf tube). Samples were incubated at room temperature for 10 to 20 min and then centrifuged for 20 min at 2000 to 3000 rpm. Enzyme-linked immunosorbent assay (ELISA) for β-endorphin (BioVision, Milpitas, CA, USA; catalog no. E4458) was performed following the manufacturer’s instruction.

Immunohistochemistry and histological analysis

After behavioral testing, all mice were deeply anesthetized and transcardially perfused with cold 4% paraformaldehyde solution (PFA; Sigma-Aldrich, catalog no. P6148). Coronal sections containing the VTA (100 μm) were directly mounted on slides with Fluoroshield mounting medium with DAPI (4′,6-diamidino-2-phenylindole) (Abcam, catalog no. ab104135). After histological verification, mice with incorrect cannula implantation were excluded from data analysis. For patch-clamp experiments, after recording, horizontal slices were put in cold 4% PFA overnight. The following day, the slices were washed three times (15 min each) in 1× PBS and mounted on slides. Sections were imaged with an upright confocal laser scanning microscope (Olympus FV-1000).

Statistical analysis

Electrophysiology data were analyzed with Clampex and MiniAnalysis. Behavior data were analyzed with EthoVision. Quantitative data are presented as the means ± SEM performed by GraphPad Prism 6 software (InStat, GraphPad Software). All comparisons relate test to control data from littermate animals collected during the same time period. Statistical significance was assessed by t test and one-way analysis of variance (ANOVA). The significant differences were identified by post hoc analysis using the Bonferroni post hoc method for multiple comparisons. Assessments were considered significant with P < 0.05.

SUPPLEMENTARY MATERIALS

stke.sciencemag.org/cgi/content/full/13/647/eaba1169/DC1

Fig. S1. Intrinsic properties of putative DA neurons were not affected by AAS treatment.

Fig. S2. Treatment with AAS increased AMPAR EPSCs, but not NMDAR EPSCs, in putative DA neurons.

Fig. S3. GABAergic synaptic transmission in putative DA neurons was not affected 24 hours after AAS treatment.

Fig. S4. Blockade of VTA MOR 60 min after AAS injection did not affect AAS-induced synaptic plasticity in putative DA neurons.

Fig. S5. Histological verification of VTA cannula placement used for locomotor experiments.

Fig. S6. β-FNA irreversibly blocked VTA MOR activity for 48 hours.

Fig. S7. Histological verification of VTA cannula placement used for CPP experiments.

Fig. S8. Histological verification of VTA cannula placement used for CPP experiments with inverted AAS and vehicle conditioning sessions.

Fig. S9. β-FNA blocked AAS-induced place preference also when the daily AAS and vehicle conditioning sessions were inverted.

REFERENCES AND NOTES

Acknowledgments: We are thankful to M. Pignatelli for assistance with electrophysiology experiments. We thank M. Lalama for assistance in the editing of the manuscript. We thank A. Mateja for expert statistical assistance. Funding: This research was supported by the Intramural Research Program of the National Institute on Drug Abuse (NIDA). Author contributions: L.B. designed and performed experiments. Both authors wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Stay Connected to Science Signaling

Navigate This Article