Dr Michel Barrot, Institut des Neurosciences Cellulaires et Intégratives, 21 rue René Descartes, 67084 Strasbourg Cedex, France. E-mail: firstname.lastname@example.org
BACKGROUND AND PURPOSE The tail of the ventral tegmental area (tVTA), also called the rostromedial tegmental nucleus, is a newly defined brain structure and a potential control centre for dopaminergic activity. It was identified by the induction of DeltaFosB following chronic cocaine exposure. In this work, we screened 20 drugs for their ability to induce FosB/DeltaFosB in the tVTA.
EXPERIMENTAL APPROACH Immunohistochemistry following systemic drug administration was used to study FosB/DeltaFosB induction in the tVTA of adult rats. Double-staining was used to determine whether dopamine or GABA neurones are involved in this induction.
KEY RESULTS The acute injection of the psychostimulant drugs cocaine, D-amphetamine, (+/−)-3,4-methylenedioxymethamphetamine (MDMA), methylphenidate or caffeine, induced the expression of FosB/DeltaFosB in the tVTA GABAergic cells. No induction was observed following exposure to ethanol, diazepam, γ-hydroxybutyric acid (GHB), morphine, ketamine, phencyclidine (PCP), Δ9-tetrahydrocannabinol (THC), sodium valproic acid or gabapentin. To evaluate the role of monoamine transporters in the psychostimulant-induced expression of FosB/DeltaFosB, we tested the antidepressant drugs reboxetine, nortriptyline, fluoxetine and venlafaxine (which target the noradrenaline and/or the 5-hydroxytryptamine transporters), the 5-hydroxytryptamine releasing agent dexfenfluramine, and the dopamine transporter inhibitor GBR12909. Only GBR12909 was able to induce FosB/DeltaFosB expression in the tVTA, showing that this induction is mediated by dopamine.
CONCLUSIONS AND IMPLICATIONS Newly described brain structures may help to increase our knowledge of brain function, pathology and targets for treatments. FosB/DeltaFosB induction in the tVTA is a common feature of drugs sharing psychostimulant properties but not of drugs sharing risk of abuse.
One possible approach to study the function of a particular brain region is to look for stimuli that recruit this region. Hence, in the present work, we screened 20 different drugs for their ability to induce the expression of FosB/ΔFosB in the tVTA. We showed that only psychostimulant drugs were able to recruit these Fos proteins locally. This action affects tVTA GABAergic cells and is probably mediated through the recruitment of the dopaminergic system as it was observed after treatment with the specific dopamine reuptake inhibitor GBR12909.
The nomenclature for drugs and for their molecular targets conforms to the British Journal of Pharmacology Guide to Receptors and Channels (Alexander et al., 2009).
Animal care and procedures were performed in accordance with the European Communities Council Directive 86/6609/EEC. The animal facilities are legally registered for animal housing and experimentation (veterinary Animal House Agreements B67-482-1 and C67-482-1). The scientists in charge of the experiments possess the French certificate authorizing experimentation on living animals, obtained from the government veterinary office. Male Sprague Dawley rats were used in all the experiments (n= 130 rats for all the studies, 280–340 g, Janvier, France), housed under standard conditions (22°C, lights on 07 h 00 min–19 h 00 min) with food and water available ad libitum.
Drug doses were chosen based on behavioural and/or molecular studies in rodents that are relevant to the related human use of these drugs. Drugs were injected i.p. or s.c. at the following doses: cocaine hydrochloride, 2.5, 5, 10, 20 or 40 mg·kg−1 i.p. (Cooper, Melun, France); D-amphetamine sulphate, 1 mg·kg−1 i.p. (Sigma-Aldrich, St Quentin Fallavier, France) (Gruner et al., 2009); (+/−)-3,4-methylenedioxymethamphetamine hydrochloride (MDMA), 5 mg·kg−1 i.p. (Sigma-Aldrich) (Stephenson et al., 1999); methylphenidate hydrochloride, 10 mg·kg−1 i.p. (Sigma-Aldrich) (Gruner et al., 2009); caffeine, 2.5, 10, 25, 60 or 100 mg·kg−1 i.p. (Sigma-Aldrich) (Deurveilher et al., 2006); 1-(2-[bis-(4-fluorophenyl)methoxy]ethyl)-4-(3-phenylpropyl)piperazine dihydrochloride (GBR12909), 15 mg·kg−1 i.p. (Biotrend, Zurich, Switzerland) (Gruner et al., 2009); ethanol (15% solution), 1.5 or 5 g·kg−1 i.p. (Vilpoux et al., 2009); diazepam, 1.5 mg·kg−1 i.p. (Sigma-Aldrich) (Chaouloff et al., 1997); γ-hydroxybutyric acid sodium salt (GHB), 1 g·kg−1 i.p. in a 2 mL·kg−1 volume (Sigma-Aldrich) (Maitre, 1997); morphine sulphate, 10 or 50 mg·kg−1 s.c. (Francopia, Paris, France); ketamine hydrochloride, 50 mg·kg−1 i.p. (Centravet, Taden, France) (Tose et al., 2009); phencyclidine hydrochloride (PCP), 3 or 10 mg·kg−1 i.p. (Sigma-Aldrich) (Castellani and Adams, 1981); Δ9-tetrahydrocannabinol solution (THC) in 30% (2-hydroxypropyl)-β-cyclodextrin, 3 mg·kg−1 i.p. (Sigma-Aldrich) (Lepore et al., 1995); sodium valproic acid, 20 mg·kg−1 i.p. (Sigma-Aldrich); gabapentin, 50 mg·kg−1 i.p. (Teva Santé, Paris la Défense, France) (Pedersen and Blackburn-Munro, 2006); reboxetine mesylate, 0.8 mg·kg−1 i.p. (Edronax®, Pharmacia Gmbh, Karlsruhe, Germany) (Yalcin et al., 2009); nortriptyline hydrochloride, 15 mg·kg−1 i.p. (Sigma-Aldrich) (Beck, 1995); fluoxetine hydrochloride, 10 mg·kg−1 i.p. (Biotrend) (Cryan et al., 2005); venlafaxine hydrochloride, 5 mg·kg−1 i.p. (Effexor®, Wyeth, Paris la Défense, France) (Millan et al., 2001); S-(+)-fenfluramine hydrochloride (dexfenfluramine), 4 mg·kg−1 i.p. (Sigma-Aldrich) (Vickers et al., 1996). Unless otherwise indicated, the drugs were prepared in 0.9% NaCl and injected in a volume of 1 mL·kg−1. For the three highest doses of caffeine, this drug was injected using a 20 mg·mL−1 solution due to limits of solubility. Control animals received an injection of 0.9% NaCl. As the tVTA is a brain region with no or very few FosB/ΔFosB staining under basal conditions (Perrotti et al., 2005; Kaufling et al., 2009; 2010), the induction of these transcription factors is easy to detect locally. It allowed us to use only three animals (triplicate) for most of the drugs tested.
Three hours following the drug injection, the rats were perfused under deep chloral hydrate anaesthesia (800 mg·kg−1; Sigma, St. Louis, MO, USA) with 100 mL phosphate buffer (0.1 M, pH 7.4) followed by 500 mL of a paraformaldehyde solution (4% in phosphate buffer). Brains were removed and post-fixed overnight in the same fixative. Forty µm mesencephalic sections were subsequently cut following the frontal plane on a vibratome (VT 1000S, Leica, Rueil-Malmaison, France). Sections were serially collected in phosphate buffer with 0.9% NaCl [phosphate-buffered saline (PBS)]. For the time course of cocaine response, perfusions were done 30 min and 1, 1.5, 3, 6, 24, 48 and 96 h post-injection.
Sections for immunohistochemistry were washed in PBS (3 × 10 min), incubated 15 min in a 1% H2O2/50% ethanol solution if used for a peroxidase reaction, washed in PBS (3 × 10 min) and incubated in PBS containing Triton X-100 and 5% donkey serum for 45 min. Sections were then incubated overnight at room temperature in PBS with Triton X-100, 1% donkey serum and the primary antibody(ies). Triton X-100 was used at 0.5%, except for glutamic acid decarboxylase (GAD) immunohistochemistry for which it was reduced to 0.2% to limit the background noise and facilitate identification of the cell body.
The rabbit anti-FosB polyclonal antibody (Santa-Cruz Biotechnology, Santa Cruz, CA, USA; catalogue number SC-48; 1:2000 for peroxidase reaction, 1:200 for immunofluorescence) was raised against an internal region of the transcription factor FosB (Santa-Cruz data sheet). It recognizes both FosB and its splice variant ΔFosB, but no other member of the Fos proteins (Perrotti et al., 2004; 2005; Luis-Delgado et al., 2006). The sheep anti-tyrosine hydroxylase (TH) polyclonal antibody (Chemicon, Temecula, CA, USA; catalog number AB1542; 1:1000) was raised against a sodium dodecyl sulphate-denaturated TH from rat phaeochromocytoma (Chemicon data sheet). The mouse anti-GAD 67 kDA monoclonal antibody (Chemicon, catalogue number MAB5406; 1:10 000) was raised against a recombinant fusion protein containing N-terminal regions of GAD 67 kDA not shared by GAD 65 kDA (Chemicon data sheet).
Sections for immunohistofluorescence were washed in PBS (3 × 10 min), incubated with a donkey Cy3 or FITC fluorophore-labelled secondary antibody (Jackson ImmunoResearch, West Grove, PA, USA; 1:400) for 1 h 30 min, and washed in PBS (3 × 10 min) before being mounted in Vectashield (Vector Laboratories, Burlingame, CA, USA). Sections for the peroxidase reaction were washed in PBS (3 × 10 min), incubated with a biotinylated donkey anti-rabbit secondary antibody (Amersham Biosciences, Orsay, France; 1:200 in PBS containing Triton X-100, 1% donkey serum) for 1 h 30 min, washed in PBS (3 × 10 min) and incubated with PBS containing the avidin-biotin-peroxidase-complex (ABC) (ABC Elite, 0.2% A and 0.2% B; Vector Laboratories) for 1 h 30 min. After being washed in Tris-HCl buffer (0.05 M, pH 7.5; 3 × 10 min), bound peroxidase was revealed by incubation in 0.025% 3.3′-diaminobenzidine tetrahydrochloride, 0.0006% H2O2 (Sigma) in Tris-HCl buffer. Sections were incubated for approximately 10 min and washed again. Sections were serially mounted on gelatine-coated slides, air dried, dehydrated in graded alcohols, cleared in Roti-Histol (Carl Roth GmbH & Co., Karlsruhe, Germany) and coverslipped with Eukitt.
Analysis and illustrations
Coded slides were used to analyse the number of FosB/ΔFosB-positive nuclei throughout the tVTA. Blind counting was done using a Nikon Eclipse 80i microscope with the Neurolucida® 8.0 software (MicroBrightField Inc., Williston, VT, USA). We counted the number of FosB/ΔFosB-positive nuclei bilaterally in the tVTA, between −5.80 and −7.30 mm from the bregma, analysing a 40 µm section every 160 µm. Section drawings presenting the FosB/ΔFosB signal were done using a microscope equipped with a camera lucida (Nikon Eclipse E600, Nikon Instruments, Kingston, UK). Data for FosB/ΔFosB analysis are expressed as mean ± SEM of positive nuclei per hemi-tVTA. Statistical analysis was performed with STATISTICA 7.1 (Statsoft, Tulsa, OK, USA), using Student's t-test to compare the drug action with the control. The significance level was set at P < 0.05.
The analysis of double-labelling fluorescence was done on 3 to 6 frontal sections per animal using an epifluorescence microscope (Leica DMRD). Pictures were taken by using a microscope (Leica) with a digital camera (Cool-snap, Princeton, NJ, USA). Adobe Photoshop 7.0 was used to adjust contrast, brightness and sharpness. The colour channels were individually adjusted for the merged pictures. Abbreviations and structure limits are based on the frontal diagrams from the atlas of Paxinos and Watson (1998).
Expression of FosB/ΔFosB in the tVTA by psychostimulant drugs
We previously reported that chronic exposure to cocaine induced the transcription factor ΔFosB in the tVTA (Perrotti et al., 2005; Kaufling et al., 2010) and we recently observed a similar induction with acute cocaine (Kaufling et al., 2009; 2010). We thus used acute cocaine as a positive control for the present study. Few FosB/ΔFosB-positive nuclei were observed in the tVTA following saline administration (Figure 1A), which is in agreement with previous reports (Perrotti et al., 2005; Kaufling et al., 2009). In contrast, acute injections of the psychostimulants cocaine, D-amphetamine, MDMA, methylphenidate or caffeine resulted in a strong expression of the transcription factor FosB/ΔFosB in the tVTA (Figure 1B–F). This induction was observed in each of the animals receiving the psychostimulant injection.
This bilateral induction started in the most posterior part of the VTA and the cluster of nuclear staining extended more caudally, shifting dorsally and slightly laterally to become embedded within the decussation of the superior cerebellar peduncle (Figures 2 and 3A). The overall quantification of this tVTA staining (Figure 2) and the quantification along the anteroposterior axis (Figure 3A) revealed that all the drugs induced FosB/ΔFosB with a similar anteroposterior profile. The induction obtained with caffeine was significant but much lower than that observed with the other psychostimulants (Figures 2E and 3A).
A time-course analysis of the cocaine response (20 mg·kg−1) revealed a rapid and long-lasting induction of FosB/ΔFosB (Figures 3B and S2) (F7,23= 9.2, P < 0.00001). This protein induction could be detected within 30 min post-injection (insert of Figure 3B, P < 0.001) and FosB/ΔFosB was still present within the tVTA 4 days after cocaine exposure. The rapid induction and its peak at 3 h are in agreement with current knowledge regarding psychostimulant-induced FosB expression in other brain regions such as the nucleus accumbens (Nestler et al., 2001; McClung et al., 2004). These data confirmed this timepoint as adequate for our experiments. The long-lasting effect suggests that stable variants of ΔFosB were also locally expressed (Nestler et al., 2001; McClung et al., 2004; Perrotti et al., 2005). No induction of FosB/DeltaFosB was observed over time after saline injection (F2,6= 0.6, P > 0.5).
The expression of FosB/DeltaFosB induced by both cocaine and caffeine was dose dependent (cocaine: F5,12= 43.7, P < 0.00001; Figures 3C and S3) (caffeine: F5,12= 11.4, P < 0.001; Figures 3D and S3). For cocaine, the maximal effect was reached with 20 mg·kg−1. For caffeine, it increased up to the highest tested dose, 100 mg·kg−1. Even at this high dose of caffeine, the induction of FosB/ΔFosB remained much smaller than that observed with the psychostimulant drugs that directly target amine uptake sites (Figures 2 and 3C,D).
The double-labelling by immunofluorescence (Figures 4 and S4) revealed that FosB/ΔFosB induced by the acute injection of these psychostimulants was in each case almost always present in the GABAergic neurones (98–100% depending on the drug, Figure 4G2), with no detectable expression in dopaminergic neurones (Figure 4G1).
Lack of expression of FosB/ΔFosB in the tVTA after exposure to other drugs
Drugs stimulating GABAergic transmission were unable to induce FosB/ΔFosB in the tVTA. This lack of induction was observed with ethanol (Figures 5A and S5A); with the anxiolytic drug diazepam (Figure 5B), an agonist with allosteric activity at the benzodiazepine site on GABAA receptors; and with GHB (Figure 5C), a GABA metabolite acting through both its own receptors and GABA receptors, predominantly GABAB (Carter et al., 2009). The opiate analgesic and drug of abuse morphine (Figures 5D and S5B), an agonist of opioid receptors, was also unable to induce FosB/ΔFosB in the tVTA. Dissociative drugs that primarily act through NMDA antagonism, such as the anaesthetic ketamine (Figure 5E) and PCP (Figures 5F and S5C) did not induce FosB/ΔFosB in the tVTA. The cannabinoid agonist THC (Figure 5G) also had no effect on FosB/ΔFosB expression in the tVTA. All these drugs are however liable to induce drug abuse.
Lastly, the anticonvulsant drug valproic acid (Figure 5H) which inhibits histone deacetylases and also favours GABA transmission through an indirect mechanism, and the anticonvulsant drug gabapentin (Figure 5I), which binds the α2-δ protein subunits of voltage-gated calcium channels (CaVα2-δ), were unable to induce FosB/ΔFosB in the tVTA.
The lack of FosB/ΔFosB induction with all of the said drugs was quantified (Figure 5J).
Expression of FosB/ΔFosB in the tVTA by inhibition of the DAT
The most prominent induction of FosB/ΔFosB in the tVTA was observed with psychostimulant drugs acting on monoamine transporters: cocaine, d-amphetamine, MDMA, methylphenidate. To evaluate the role of the different monoamines in this induction, we tested inhibitors of the various transporters: dopamine transporter (DAT), noradrenaline transporter (NET) and/or 5-hydroxytryptamine transporter (SERT). The antidepressant drug reboxetine (Figure 6A), which is a specific NET inhibitor, the tricylic antidepressant drug nortriptyline (Figure 6B), a NET/SERT inhibitor with a predominant NET action, the antidepressant drug fluoxetine (Figure 6C), a SERT inhibitor, and the antidepressant drug venlafaxine (Figure 6D), a non-tricyclic NET/SERT inhibitor, were unable to induce the expression of FosB/ΔFosB in the tVTA. The anorexigenic drug dexfenfluramine (Figure 6E), which is a 5-hydroxytryptamine releasing agent, was also unable to induce FosB/ΔFosB in the tVTA. In contrast, the DAT inhibitor GBR12909, at a dose that promotes an awake state (Gruner et al., 2009), induced a strong expression of FosB/ΔFosB in the tVTA (Figures 3A and 6F,G). This induction of FosB/ΔFosB following GBR12909 exposure displayed the same anteroposterior profile as the psychostimulant drugs (Figure 3A), and was also mainly present in GABAergic cells (98%, Figure 4F,G2).
Discussion and conclusions
In this study, we tested pharmacological compounds for their ability to recruit a recently discovered brain region, the tVTA. We show that an acute injection of the psychostimulant drugs cocaine, D-amphetamine, MDMA, methylphenidate and caffeine, induced the expression of FosB/ΔFosB in tVTA cells that were identified as GABAergic by double immunostaining. No induction was observed following exposure to ethanol, diazepam, GHB, morphine, ketamine, PCP, THC, sodium valproic acid or gabapentin. By testing various drugs targeting the aminergic systems (reboxetine, nortriptyline, fluoxetine, venlafaxine, dexfenfluramine, GBR12909), we found that the DAT inhibitor GBR12909 was the only one able to induce FosB/ΔFosB expression in the tVTA. This shows that the dopaminergic system has a critical role in the recruitment of tVTA induced by pharmacological compounds. Our results suggest that the tVTA is a common target for drugs sharing psychostimulant properties.
The VTA is key structure for adaptive and goal-directed behaviours, motivation, reward and mood, and it is thus implicated in various psychopathological disorders (Le Moal and Simon, 1991; Nestler and Carlezon, 2006; Grace et al., 2007; Iversen and Iversen, 2007; Schultz, 2007). Its functional heterogeneity along the anteroposterior axis was first observed 30 years ago (Arnt and Scheel-Krüger, 1979). This functional heterogeneity then remained unstudied until the last 10 years. Behavioural studies showed that cocaine (Rodd et al., 2005a), nicotine (Ikemoto et al., 2006), cannabinoids (Zangen et al., 2006), opioid peptides such as the endogenous ligand of µ-opioid receptors endormorphin-1 (Zangen et al., 2002), but also ethanol (Rodd-Henricks et al., 2000; Rodd et al., 2005b) or its metabolite acetaldehyde (Rodd-Henricks et al., 2002), are self-administered if delivered into the posterior VTA; but they are not or poorly self-administered if delivered into the anterior VTA. The functional distinction between anterior and posterior VTA is also supported by studies using local viral-mediated gene transfer to manipulate the AMPA receptor subunit GluR1 (Carlezon et al., 2000), the phospholipase Cγ (Bolanos et al., 2003) or the transcription factor CREB (Olson et al., 2005). This last study identified the functional transition between anterior and posterior VTA around −5.5 mm from the bregma in the rat. However, the anterior versus posterior subdivisions of the VTA only cover the rostral and central tiers of this structure. There is less information available on the caudal-most tier, corresponding to the tVTA which is mainly GABAergic and can be revealed by ΔFosB induction following cocaine exposure (Perrotti et al., 2005; Kaufling et al., 2009). Our results show that various psychostimulant drugs recruit the tVTA after acute administration, as indicated by the induction of the transcription factor FosB/ΔFosB. Behaviourally, future studies are now needed to evaluate whether tVTA also supports drug self-administration and to functionally differentiate the posterior VTA from the tVTA. However, the partial overlap between the posterior VTA and tVTA (Kaufling et al., 2009) might make the latter studies challenging.
Cocaine, D-amphetamine, MDMA and methylphenidate target monoamine transporters. However, the NET inhibitor reboxetine, the NET/SERT inhibitors nortriptyline and venlafaxine, the SERT inhibitor fluoxetine, and the 5-hydroxytryptamine releasing agent dexfenfluramine are unable to induce FosB/ΔFosB in the tVTA. In contrast, FosB/ΔFosB can be induced in the tVTA by the DAT inhibitor GBR12909. A marked increase in dopamine transmission thus appears to be sufficient to induce the local expression of this transcription factor. The dopamine involved may originate from the sparse dopaminergic cells or fibres within the tVTA itself. These fibres may arise from local cell bodies or from dopaminergic cell bodies in VTA and substantia nigra pars compacta, which are tVTA afferents (Kaufling et al., 2009). However, the dopamine may also be of somatodendritic origin from the VTA itself (Kalivas and Duffy, 1988) and diffuse by volume transmission to the nearby tVTA. Another hypothesis that should not be discarded is the possibility of a system-wide polysynaptic recruitment of tVTA. Indeed, tVTA receives inputs from various brain regions recruited by psychostimulants. This includes, but is not restricted to, prefrontal cortical areas, nucleus accumbens or lateral habenula (Geisler et al., 2008; Jhou et al., 2009a; Kaufling et al., 2009; Brinschwitz et al., 2010).
An increase in somatodendritic and axonal dopamine has also been observed after morphine exposure (Kalivas and Duffy, 1988), whereas we found that morphine did not induce the expression of FosB/ΔFosB in the tVTA. Two explanations may be proposed to account for this discrepancy. Firstly, the somatodendritic release of dopamine induced by morphine has a lower capacity for diffusion than that induced by the previously cited psychostimulants. Indeed, DAT are still effective in animals administered with morphine, preventing an important diffusion of the dopamine. Secondly, it has been shown that tVTA neurones express the µ-opioid receptors (Jhou et al., 2009a) and direct stimulation of these opioid receptors by morphine may inhibit the cAMP/PKA pathway and prevent FosB/ΔFosB induction in tVTA neurones. It is also important to remember that the induction of FosB/ΔFosB does not necessarily reflect an electrophysiological activity of the corresponding cell and that a lack of induction does not reflect a lack of electrophysiological effect.
The FosB/ΔFosB induction in the tVTA after caffeine is mild but significant at various doses. Caffeine is not an inhibitor of the monoamine transporter; it is a non-selective adenosine receptor antagonist and a competitive non-selective phosphodiesterase inhibitor (Nehlig et al., 1992). Due to caffeine's low affinity for phophodiesterases, its in vivo action, particularly its psychostimulant effect, is thought to be mediated through its inhibitory effects on adenosine receptors (Ferré, 2008). Studies on the striatal complex have revealed that both pre- and post-synaptic mechanisms, which depend on A1 and A2A receptors and on their interactions with dopamine receptors, are involved in the psychostimulant effect of caffeine (Ferré, 2008). Future work is however needed to evaluate whether similar mechanisms are implicated in the recruitment of tVTA cells by caffeine.
In conclusion, among the 20 drugs tested, only the psychostimulant drugs induced the expression of FosB/ΔFosB in the tVTA. Newly described brain structures may help to increase our knowledge on brain functions, pathology and pharmacological targets. The tVTA, a potential control centre for dopaminergic activity, appears to be a common target for drugs sharing psychostimulant properties rather than for drugs sharing a risk of abuse. Future work is needed to understand the functional implication of tVTA in the stimulant and/or arousing properties of these drugs, and to determine whether this newly defined brain area is a critical neuroanatomical substrate for such properties.
This work was supported by the Centre National de la Recherche Scientifique (contract UPR3212), the University of Strasbourg and the Fondation pour la Recherche Médicale (JK, MB). We thank Stéphane Doridot for animal care.