Abstract
The identification of the molecular mechanisms involved in nicotine addiction and its cognitive consequences is a worldwide priority for public health. Novel in vivo paradigms were developed to match this aim. Although the β2 subunit of the neuronal nicotinic acetylcholine receptor (nAChR) has been shown to play a crucial role in mediating the reinforcement properties of nicotine, little is known about the contribution of the different α subunit partners of β2 (i.e., α4 and α6), the homo-pentameric α7, and the brain areas other than the ventral tegmental area (VTA) involved in nicotine reinforcement. In this study, nicotine (8.7–52.6 μg free base/kg/inf) self-administration was investigated with drug-naive mice deleted (KO) for the β2, α4, α6 and α7 subunit genes, their wild-type (WT) controls, and KO mice in which the corresponding nAChR subunit was selectively re-expressed using a lentiviral vector (VEC mice). We show that WT mice, β2-VEC mice with the β2 subunit re-expressed exclusively in the VTA, α4-VEC mice with selective α4 re-expression in the VTA, α6-VEC mice with selective α6 re-expression in the VTA, and α7-KO mice promptly self-administer nicotine intravenously, whereas β2-KO, β2-VEC in the substantia nigra, α4-KO and α6-KO mice do not respond to nicotine. We thus define the necessary and sufficient role of α4β2- and α6β2-subunit containing nicotinic receptors (α4β2*- and α6β2*-nAChRs), but not α7*-nAChRs, present in cell bodies of the VTA, and their axons, for systemic nicotine reinforcement in drug-naive mice.
- nicotinic acetylcholine receptor (nAChR)
- ventral tegmental area (VTA)
- lentiviral vector
- nicotine
- intravenous self-administration
- drug-naive mice
Introduction
Nicotine is the principal substance responsible for tobacco addiction (Peto et al., 1996; Balfour, 2002; Silagy et al., 2004), but it also enhances attention and cognitive performance (Newhouse et al., 2004). Nicotine binds to neuronal nicotinic acetylcholine receptors (nAChRs), a heterogeneous family of pentameric ligand-gated ion channels (Corringer et al., 2000; Changeux and Edelstein, 2005). In the brain, six α (α2-α7) and/or three β (β2-β4) subunits potentially assemble in multiple combinations with a broad diversity of pharmacological and electrophysiological properties (McGehee and Role, 1995; Le Novère et al., 2002). In genetically engineered mice lacking the β2 subunit (β2-KO mice), β2 has been shown to contribute to cognitive functions and nicotine reinforcement (Picciotto et al., 1995, 1998; Granon et al., 2003). α4 subunit containing nicotinic receptors (α4*-nAChRs) have been implicated in conditioned place preference for nicotine (Tapper et al., 2004), whereas the role of α6*-nAChRs remains insufficiently characterized, although both of them are quantitatively expressed in dopaminergic (DAergic) neurons (Le Novère et al., 1996; Champtiaux et al., 2002, 2003). The ventral tegmental area (VTA) is considered the principal brain region mediating the reinforcing properties of multiple drugs of abuse, including nicotine, but its precise contribution is still debated (Laviolette and van der Kooy, 2004; Nashmi et al., 2007).
In the present study we address two important issues: the first concerns the mode of administration of nicotine. We attempted to identify the brain regions implicated when nicotine is self-administered systemically rather than directly injected into the VTA (Maskos et al., 2005; Besson et al., 2006; David et al., 2006). We adopted intravenous nicotine self-administration (SA) because it is considered the animal paradigm most closely resembling smoking behavior in humans (Rose and Corrigall, 1997; Corrigall, 1999). In the past, drug SA studies have been conducted using chronic procedures in rodents, in which animals were used as their own controls (Koob and Weiss, 1990). Recently, studies have been published by several groups using an acute mouse model in which animals are tested in pairs using a contingent and a yoked control mouse, which enables rapid assessment of the reinforcing properties of a compound. Although this acute model is thought to assess the initiation rather than the chronic maintenance of drug-taking behavior, it was demonstrated that drug-naive mice acutely self-administer the same drugs that humans abuse, such as cocaine (Kuzmin et al., 1992, 1996c,d, 2000; Kuzmin and Johansson, 2000; Rasmussen et al., 2000; Blokhina et al., 2005; Lesscher et al., 2005), morphine (Kuzmin et al., 1996a, 1997), amphetamine (Cossu et al., 2001), cannabinoids (Martellotta et al., 1998a), γ-hydroxybutyric acid (Martellotta et al., 1998c; Fattore et al., 2000, 2001), scopolamine (Rasmussen and Fink-Jensen, 2000), and nicotine (Martellotta et al., 1995; Rasmussen and Swedberg, 1998; Fattore et al., 2002; Paterson et al., 2003). These studies demonstrated that drug-naive mice exhibit SA behavior similar to rats chronically trained under the same schedule of reinforcement (FR1) and using the same operant response, i.e., nose-poking (Fattore et al., 1999, 2002; Solinas et al., 2003; Spano et al., 2004; Crombag et al., 2005).
Second, we analyzed the respective roles played by the non-β2 subunits expressed in the DAergic system, namely the α4, α6, and α7 subunits, in the modulation of acute nicotine reinforcement, and evaluated, using the lentiviral vector technology, to what extent their presence in the mesolimbic system is required for nicotine reinforcement. Acute nicotine intravenous SA was performed in drug-naive wild-type (WT) (C57BL/6J) mice, β2-KO, α4-KO, α6-KO, and α7-KO mice, as well as in KO mice with the corresponding subunit selectively re-expressed in the VTA, and β2-VEC mice with selective re-expression in the substantia nigra (SN).
Materials and Methods
Subjects.
Male C57BL/6J (Charles River), β2-, α4-, α6-, α7-nAChR KO mice and their corresponding WT controls were used, weighing 24–28 g at the time of experiments. The animals were housed eight per cage with food and water available ad libitum, kept under standard conditions (temperature 21 ± 1°C, 60–65% relative humidity) on reversed 12 h light/dark cycle (light on 7:00 P.M.) and left undisturbed for at least 10 days before starting the experimental procedure. SA sessions took place during the dark phase of the cycle, between 9:00 and 12:00 A.M. All experiments were performed in strict accordance with both the Guide for the Care and Use of Laboratory Animals (National Institutes of Health) and the E.C. regulations for animal use in research (CEE n° 86/609).
Lentiviruses.
The lentiviral expression vectors are derived from the pHR' expression vectors first described by Naldini et al. (1996), with several subsequent modifications (Maskos et al., (2005). In the lentiviruses used in this study, the bicistronic expression of mouse wild-type β2, α4 or α6 nAChR subunit cDNAs and the eGFP cDNA are under the control of the mouse phosphoglycerate kinase (PGK) promoter (see Figure 6A). The β2 subunit expressing lentivirus is referred to as pTRIPΔU3[PGK_β2_IRES2_eGFP_WPRE]. Further details can be found in Maskos et al. 2005).
To create the pTRIPΔU3[PGK_α4_IRES2_eGFP] construct, a three step strategy was used. First, on the expression vector containing the wild-type α4 cDNA, obtained from Jerry Stitzel, the NheI site located within the α4 cDNA was removed by mutagenesis (QuickChange, Stratagene), without affecting amino acid sequence. Second, the modified α4 cDNA was cloned into the pIRES2-eGFP expression plasmid (Clontech) between XhoI and BamHI of the multiple cloning site using a linker oligonucleotide NotI /BamHI, creating a NheI site after the α4 stop codon. Third, the wild-type mouse α4 subunit was finally ligated between the XhoI-NheI sites of the pTRIPΔU3[PGK_β2_IRES2_eGFP_WPRE] vector previously described to obtain the complete vector sequence of 13599 bp (see Fig. 6A).
To construct the pTRIPΔU3[PGK_α6_IRES2_eGFP] construct, a three-step strategy was used. First, in the expression vector containing the wild-type α6 cDNA plus FLAG sequence, obtained from Ines Ibanez-Tallon, the ClaI restriction site located at the 5′ extremity of the α4 cDNA and the XhoI site in the 3′ extremity were replaced by mutagenesis (QuickChange, Stratagene) with XhoI and NheI, respectively. Second, another mutagenesis was performed on this modified vector to introduce a stop codon between the end of the α6 cDNA and the FLAG sequence. Third, the wild-type mouse α6 subunit cDNA was finally ligated between the XhoI-NheI sites of the pTRIPΔU3[PGK_IRES2_eGFP_WPRE] vector to replace the β2 cDNA, and to obtain the complete vector sequence of 13090 bp (see Fig. 6A). All modified regions were verified by DNA sequencing.
Lentivirus stereotaxic injections.
Mice aged 11–13 weeks were anesthetized using 250 μl of ketamine 1.5% (Merial)/xylazine 0.05% (Bayer Healthcare) in PBS. The mouse was introduced into a stereotaxic frame adapted for use with mice (Cunningham and McKay, 1993). Lentivirus (2 μl at 75 ng of p24 protein per μl) was injected bilaterally at: antero-posterior −3.4 mm, lateral ± 0.5 mm from bregma and −4.4 mm from the surface for VTA injection. To target the SN, injections were at: antero-posterior −3.0 mm (from bregma), lateral ± 1.3 mm and dorso-ventral −4.3 mm from the skull, as described by Avale et al. (2008). All procedures were performed in accordance with European Commission directives 219/1990 and 220/1990, and approved by Animalerie centrale and Médecine du travail, Institut Pasteur. The mice were tested after 5–6 weeks of viral expression.
Drugs.
For SA experiments, (−)-nicotine bitartrate (Sigma) was freshly dissolved in 0.9% saline, and the pH adjusted to 7.2 ± 0.1 with NaOH (0.1N). The drug doses used in this study (8.7, 17.5, 26.3, 35, and 52.6 μg/kg/inj) are referred to the free base, and correspond to nicotine concentrations of 0.025, 0.05, 0.075, 0.1 and 0.15 mg/kg/inj of salt.
Self-administration procedure.
Mice were tested in pairs of identical test cages (8 × 8 × 8 cm inner size), each presenting a central frontal hole (diameter 25 mm) 1 cm above the box floor fitted with an infrared sensor interfaced to an operating computer that controlled an automatic syringe pump (PHM-100A, Med Associates) (Fig. 1A). A rear vertical chink (5 mm wide) was made on the opposite wall through which the tail was extended outside the box and taped to a horizontal surface allowing access to the lateral tail veins with a 27G winged needle (external diameter 0.4 mm), connected to the syringe through Teflon tubing.
A, Set-up and simplified scheme for intravenous self-administration experiments. Each nose-poke of the Active (A) mouse activates a computer-operated syringe pump which delivers a nicotine injection into the tail vein of both the A and the yoked Passive (P) mouse. B, Concentration-dependent nicotine intravenous SA in C57BL/6J drug-naive mice. Each bar (y-axis at left) represents the mean ± SEM of the cumulative nose pokes (NPs) of the active mice (A, black bars) and passive yoked controls (P, white bars) over the 30 min session. Each gray bar (y-axis at right) represents the R value for mice self-administering saline (Sal) or different nicotine concentrations over the 30 min session. Drug doses are expressed as μg/kg/infusion. **p < 0.01 vs yoked passive and saline groups. ANOVA followed by Dunnett's test (n = 12–18 pairs of animals). ***p < 0.001 vs saline control group. ANOVA followed by Dunnett's test (n = 12–18 pairs of animals).
Because mice were housed on a reversed light/dark cycle, they were kept in the dark during transportation from the housing to the experimental room. This room was always the same, exclusively dedicated to SA experiments in mice, and had the same environmental conditions as the housing room, i.e., T = 21 ± 1°C and 60–65% humidity.
Self-administration sessions were conducted over several seasons, but care was taken in maintaining constant room temperature and humidity, and in ensuring that no sounds entered the experimental room. Experiments were conducted in the dark: a dim light located above the apparatus was switched on for incannulation of the tail vein only, whereas the box containing the animal was covered with black cloth so that mice were never exposed to the light. No infrared lighting or other expedients were used to provide vasodilation of the tail vein or to facilitate needle insertion. No detergents, soap or alcohol were used between sessions, distilled water only being used to clean boxes at the end of each session.
Animals were first placed in the test cage for 10 min of habituation (pre test) with their tails taped but no needle inserted. Pairs of animals were selected on the basis of an approximate equal level of nose-poking activity during pre test. Thereafter, the matched pairs were placed into the experimental boxes, one mouse defined as active (A) and the other one passive (P), and needles inserted in the lateral tail vein. Animals were randomly allocated to the different experimental groups and allowed access to the drug under a continuous reinforcement (FR-1) schedule. A nose-poking modus operandi was used for nicotine SA, because it reflects a more “natural” behavior for mice and requires less motor and motivational output than lever-pressing.
Each nose-poke (NP) of the active (A) mouse activated the computer-operated syringe pump delivering either nicotine or saline infusion (volume of injection: 1.0 μl, infusion time: 1 s) both to the A and the passive (P) mouse, so that animals received the same amount of the drug simultaneously. A white cue light was activated by each NP of the A mouse for 1 s, i.e., during the delivery of the drug, thus serving as a drug-associated cue along with the noise of the activated syringe pump. No cue light was activated by NPs of the yoked P mouse.
NPs of the P mouse were recorded but had no scheduled consequences. A short time-out period (i.e., inactivation of NP responses) of 2 s was imposed after each drug infusion. Each treatment included not <5 pairs of animals. SA sessions lasted 30 min and each mouse was used only once. The number of NPs for both animals (A and P mice) in each treatment group was analyzed with two-way ANOVA to evaluate effects of drug delivery mode (“contingently” versus “noncontingently”), unit dose (including vehicle), and interactions between group and drug dose. For post hoc comparisons, Dunnett's test was used to compare single groups of mice, and respective vehicle controls. Statistical significance was set at p < 0.05. The whole study was designed as a between-subjects (independent groups) experiment, because each treatment we describe was performed on a single set of animals.
To obtain gradual measurements of the reinforcing effects of nicotine, the mean ratio (or Reinforcement Index) R (Kuzmin et al., 1996a,b; Martellotta et al., 1998a,b) between the number of responses (NPs) of the active (A) and passive (P) mice during the 30 min session was calculated, nicotine effect being considered rewarding, neutral or aversive when R was greater than, equal to, or <1, respectively.
Receptor autoradiography.
Brains from WT, β2-KO, α4-KO, α6-KO, and re-injected mice were dissected, frozen in dry ice, and stored at −80°C until use. Twenty-micrometer-thick coronal sections were obtained by on a cryostat at −20°C, and thaw mounted on Menzel Gläser SuperFrost Plus microscope slides. Sections were kept at −80°C until the binding assay. Sections used for [125I]epibatidine binding were incubated at room temperature with 200 pm [125I]epibatidine (NEN Perkin-Elmer; specific activity 2200 Ci/mmol) in Tris 50 mm pH 7.4 for 1 h. Then, sections were rinsed twice in the same buffer for 5 min, once briefly in distilled water, and exposed to film. Details of the procedure were described previously (Zoli et al., 1998).
The [125I]conotoxin MII binding procedure was adapted from Whiteaker et al. (2000). Sections were preincubated in binding buffer (144 mm NaCl, 1.5 mm KCl, 2 mm CaCl2, 1 mm MgSO4, 20 mm HEPES, 0.1% BSA, pH 6.8) + 1 mm PMSF at 25°C for 15 min. Binding reaction was performed in binding buffer supplemented with 0.1% BSA, 5 mm EDTA, 5 mm EGTA, and 10 μg/ml aprotinin, leupeptin and pepstatin. The sections were incubated at 25°C for 2 h with 0.5 nm [125I]conotoxin MII. Then, sections were rinsed for 30 s once in binding buffer at 25°C, twice in binding buffer at 4°C, twice in binding buffer diluted 1/10 at 4°C, and finally once in 5 mm HEPES at 4°C. The slides were dried on a slide drying bench before being exposed for 24 h to Kodak Biomax MS films with appropriate 125I standards.
Results
Nicotine dose–response curve in C57BL/6J drug-naive mice
The mouse model for acute intravenous SA in drug-naive animals is schematically presented in Figure 1A. Both the “active” (A) and the “passive” (P) mouse are in neighboring cages, cannulated in the tail vein, and exposed to the same amount of nicotine at the same time, determined by the nose-poke behavior of the A mouse. The procedure was first validated using saline and nicotine in C57BL/6J mice, which represents the background strain of WT and KO mice used in this study. The data are presented both in terms of the number of nose pokes (NP) of the A and P mouse, and their ratio, the Reinforcement Index, R, see Materials and Methods.
As shown in Figure 1B, an R value close to 1 implied no statistically significant difference in the mean number of responses of A and P C57BL/6J mice when saline (Sal) was made available (red bar). Conversely, a nicotine concentration of 26.3 μg per kg and per infusion (μg/kg/inf) significantly increased responding in the active mice (18.5 ± 1.3), although not affecting nose-poking activity of yoked passive animals (4.2 ± 2.6), so that resulting R values were >1 (red bar). By increasing the R value above unity, the nicotine concentration of 26.3 μg/kg/inf is considered to possess reinforcing properties. ANOVA analysis confirmed highly significant differences in nose-poking activity among experimental groups (F(7,107) = 26.31, p < 0.01). Conversely, no differences were obtained between active and passive mice at nicotine concentrations of 8.7, 17.5, 35 and 52.6 μg/kg/inf, with R values being close to 1.
This experiment demonstrated that nicotine exerts a positive reinforcing effect in C57BL/6J drug-naive mice following a dose-dependent U-shaped curve, as already reported for nicotine in DBA mice (Fattore et al., 2002), and for other drugs of abuse (Martellotta et al., 1998a,b; Ledent et al., 1999; Fattore et al., 2000; Cossu et al., 2001), under the same experimental conditions.
Absence of nicotine intravenous SA in β2-nAChR KO drug-naive mice
Figure 2A illustrates nicotine intravenous SA in mice lacking the β2 subunit of the nAChR (β2-KO mice). It can be noted that when saline infusions (Sal) were made contingent on nose-poking activity, the R value was close to 1, indicating no statistically significant difference in the mean number of NPs between active and passive β2-KO mice (Fig. 2A, right). Yet, unlike C57BL/6J mice (Fig. 1B) and WT mice (Fig. 2A, left), active β2-KO animals to which free access to nicotine was given at the dose of 26.3 μg/kg/inf did not increase their rate of responding with respect to either passive animals receiving noncontingent nicotine infusions or control animals receiving contingent vehicle (saline) infusions (F(3,29) = 1.39, ns). All of the doses tested were ineffective in inducing nicotine SA behavior in β2-KO mice, so that the reinforcement index (R) was equivalent to 1 (Fig. 2A, right).
A, Nicotine intravenous SA in β2-nAChR knock-out (KO) and wild-type (WT) mice. Each bar (y-axis at left) represents the mean ± SEM of the cumulative nose pokes (NPs) of the β2-WT (left) and β2-KO (right) active mice (A, black bars) and passive yoked controls (P, white bars) over the 30 min session. Each gray bar (y-axis at right) represents the R value for β2-WT (left) and β2-KO (right) mice self-administering nicotine over the 30 min session. Doses are expressed as μg/kg/inf. #p < 0.05 vs corresponding saline group and **p < 0.01 vs corresponding yoked passive group. ANOVA followed by Dunnett's test (n = 7–13 pairs of animals). ***p < 0.001 vs corresponding saline group. ANOVA followed by Dunnett's test (n = 7–13 pairs of animals). B, Nicotine intravenous SA in ventral tegmental area (VTA) re-expressed β2-KO mice (β2-VEC-VTA mice). Each bar (y-axis at left) represents the mean ± SEM of the cumulative nose pokes (NPs) of the β2-VEC-VTA active mice (A, black bars) and passive yoked controls (P, white bars) over the 30 min session. Each gray bar (y-axis at right) represents the R value for β2-VEC-VTA mice self-administering saline (Sal) or nicotine over the 30 min session. Doses are expressed as μg/kg/inf. **p < 0.01 vs yoked passive mice and saline controls. ANOVA followed by Dunnett's test (n = 6–8 pairs of animals). *p < 0.05 vs saline control group. ANOVA followed by Dunnett's test (n = 6–8 pairs of animals). C, Nicotine intravenous SA in substantia nigra (SN) re-expressed β2-KO mice (β2-VEC-SN mice). Each bar (y-axis at left) represents the mean ± SEM of the cumulative nose pokes (NPs) of the β2-VEC-SN active mice (A, black bars) and passive yoked controls (P, white bars) over the 30 min session (n = 6–14 pairs of animals). Each gray bar (y-axis at right) represents the R value for β2-VEC-SN mice self-administering saline (Sal) or nicotine over the 30 min session. Doses are expressed as μg/kg/inf (n = 6–14 pairs of animals).
Recovery of nicotine intravenous SA after re-expression of β2 in the VTA but not the SN of β2-KO mice
As shown in Figure 2B, when saline was made contingent on NP responses of VTA re-expressed β2-KO mice (β2-VEC-VTA), active and passive animals displayed a similar mean number of responses, as indicated by the R value (gray bar). However, when mice were allowed to intravenously self-administer nicotine at 26.3 μg/kg/inf (black bars), active mice significantly increased their rate of nose-poking activity with respect to both yoked passive mice (white bars) receiving nicotine noncontingently, and control mice receiving saline (Sal). Accordingly, the nicotine concentration of 26.3 μg/kg/inf increased R above unity (gray bar), and an overall significant effect of drug was revealed by post hoc comparisons (F(3,21) = 10.74, p < 0.01). Nicotine thus exerts positive reinforcing effects in VTA re-expressed β2-KO mice, where it is able to sustain acute SA behavior.
In another set of mice, the β2 subunit was selectively re-expressed in the substantia nigra (β2-VEC-SN) and the mice tested under the same conditions. As shown in Figure 2C, nicotine at 26.3 μg/kg/inf did not modify the R index (gray bar) with respect to saline (Sal), thus showing no intravenous nicotine SA at this unit nicotine dose. Lower and higher concentrations of 17.5 and 35 μg/kg/inf, respectively, did not result in nicotine SA as well, thus proving absence of any shift in the dose–response curve. Nicotine thus exerts positive reinforcing effects only when the β2-subunit is re-expressed in the VTA, but not in the SN.
Identifying the role played by homomeric α7*-nAChRs in the VTA
The α7*-nAChRs in the VTA have been described to modulate the discharge pattern of DA neurons (Mameli-Engvall et al., 2006) and to mediate long-term effects of chronic exposure to nicotine (Besson et al., 2007). McGehee and Mansvelder had also established their importance for nicotine induced LTP in the VTA (Mansvelder and McGehee, 2000; Mansvelder et al., 2002).
Their role was thus tested in acute nicotine SA. As shown in Figure 3, we found that, as in the β2-WT and VTA-VEC mice, the only effective dose of nicotine in sustaining SA behavior was 26.3 μg/kg/inf, the only nicotine concentration significantly increasing R >1 (gray bar). Under these conditions, both α7-WT (left) and α7-KO (right) active mice (black bars) displayed a robust increase in NP numbers over saline (Sal) controls and over the corresponding passive mice (white bars), thus showing that the lack of α7-subunit does not interfere with the acute rewarding effects of nicotine.
Nicotine intravenous SA in α7-nAChR KO and WT mice. Each bar (y-axis at left) represents the mean ± SEM of the cumulative nose pokes (NPs) of the α7-WT (left) and α7-KO (right) active mice (A, black bars) and passive yoked controls (P, white bars) over the 30 min session. Each gray bar (y-axis at right) represents the R value for α7-WT (left) and α7-KO (right) mice self-administering saline (Sal) or nicotine over the 30 min session. Doses are expressed as μg/kg/inf. #p < 0.05 vs corresponding saline group and **p < 0.01 vs corresponding yoked passive group. ANOVA followed by Dunnett's test (n = 6–8 pairs of animals). **p < 0.01 and ***p < 0.001 vs corresponding saline group. ANOVA followed by Dunnett's test (n = 6–8 pairs of animals).
The role played by the α4 subunit partner of β2 in the VTA
The α4 subunit is among the main partners of the β2 subunit, coexpressed in DAergic and GABAergic neurons of the VTA (Klink et al., 2001; Champtiaux et al., 2002, 2003). Mice deleted for the α4 subunit and their corresponding WT littermates were therefore tested for nicotine reinforcement at the same doses as for the β2-KO mice. As shown in Figure 4A, α4-WT mice (left) exhibited a U-shaped dose–response curve for nicotine. Yet, as for β2-KO animals, α4-KO animals (right) did not reveal R values significantly different from 1 (gray bars), showing the absence of nicotine reinforcement at all of the doses tested.
A, Nicotine intravenous SA in α4-nAChR KO and WT mice. Each bar (y-axis at left) represents the mean ± SEM of the cumulative nose pokes (NPs) of the α4-WT (left) and α4-KO (right) active mice (A, black bars) and passive yoked controls (P, white bars) over the 30 min session. Each gray bar represents R value for α4-WT (left) and α4-KO (right) mice self-administering saline (Sal) nicotine over the 30 min session. Doses are expressed as μg/kg/inf. **p < 0.01 vs corresponding yoked passive and/or saline groups. ANOVA followed by Dunnett's test (n = 5–14 pairs of animals). **p < 0.01 vs corresponding saline group. ANOVA followed by Dunnett's test (n = 5–14 pairs of animals). B, Nicotine intravenous SA in VTA re-expressed α4-KO mice (α4-VEC-VTA mice). Each bar (y-axis at left) represents the mean ± SEM of the cumulative nose pokes (NPs) of the α4-VEC-VTA active mice (A, black bars) and passive yoked controls (P, white bars) over the 30 min session. Each gray bar (y-axis at right) represents the R value for α4-VEC-VTA mice self-administering saline (white bar) or nicotine (black bars) over the 30 min session. Doses are expressed as μg/kg/inf. **p < 0.01 vs corresponding saline group. ANOVA followed by Dunnett's test (n = 6–10 pairs of animals). **p < 0.01 vs saline control group. ANOVA followed by Dunnett's test (n = 6–10 pairs of animals).
A lentivirus was then constructed expressing the α4-subunit (see Fig. 6A). As shown in Figure 4B, nicotine reinforcement was promptly restored by selective re-expression of the α4 subunit in the VTA, a nicotine concentration of 26.3 μg/kg/inf being able to increase the R value >1 (gray bar). Along with previous results from experiments with β2-VEC-VTA mice, these data highlight the critical role played by the α4β2* containing subtype of nAChRs in mediating the acute reinforcing properties of nicotine.
The role played by the α6 subunit in the VTA
This subunit has generated substantial interest since its original description because of its differential expression in the DAergic system, and its role in locomotion (Le Novère et al., 1996, 1999). Yet, α6-KO animals did not show any overt developmental, neurological or behavioral deficits (Champtiaux et al., 2002, 2003). However, when tested in our SA paradigm, although α6-WT animals (left) did self-administer nicotine at the unit dose of 26.3 μg/kg/inf, α6-KO drug-naive mice (right) did not do so, as illustrated in Figure 5A. The α6-KO did not self-administer nicotine even within an extensive range of lower (8.7–17.5 μg/kg/inf) or higher (35–52.6 μg/kg/inf) doses, showing R values close to 1 (gray bars). This finding provides the first behavioral phenotypes identified so far in the α6-KO line.
A, Nicotine intravenous SA in α6-nAChR KO and WT mice. Each bar (y-axis at left) represents the mean ± SEM of the cumulative nose pokes (NPs) of the α6-WT (left) and α6-KO (right) active mice (A, black bars) and passive yoked controls (P, white bars) over the 30 min session. Each gray bar (y-axis at right) represents the R value for α6-WT (left) and α6-KO (right) mice self-administering saline (Sal) or nicotine over the 30 min session. Doses are expressed as μg/kg/inf. #p < 0.05 vs corresponding saline group and **p < 0.01 vs corresponding yoked passive or saline control groups. ANOVA followed by Dunnett's test (n = 6–8 pairs of animals). **p < 0.01 vs corresponding saline group. ANOVA followed by Dunnett's test (n = 6–8 pairs of animals). B, Nicotine intravenous SA in VTA re-expressed α6-KO mice (α6-VEC-VTA mice). Each bar (y-axis at left) represents the mean ± SEM of the cumulative nose pokes (NPs) of the active mice (A, black bars) and passive yoked controls (P, white bars) over the 30 min session. Each gray bar (y-axis at right) represents the R value for α6-VEC-VTA mice self-administering saline (Sal) or nicotine over the 30 min session. Doses are expressed as μg/kg/inf. ##p < 0.001 vs corresponding saline group and **p < 0.01 vs corresponding yoked passive or saline control groups. ANOVA followed by Dunnett's test (n = 9–11 pairs of animals). **p < 0.01 vs saline controls. ANOVA followed by Dunnett's test (n = 9–11 pairs of animals).
The α6 subunit was then selectively re-expressed using a lentiviral vector (Fig. 6A) in the VTA of α6-KO mice (Fig. 6B) and the mice tested under the same conditions. Active mice (black bars) significantly (p < 0.01) increased their rate of nose-poking for nicotine at 26.3 μg/kg/inf with respect to their yoked passive mice (white bars), and the R value was therefore >1, as indicated in Figure 5B (gray bars). Conversely, active and passive α6-VEC-VTA mice displayed a similar mean number of responses when either saline (Sal) or nicotine at 17.5 and 35 μg/kg/inf were made contingent on NP responses. Nicotine thus exerts positive reinforcing effects in VTA re-expressed α6-KO mice, where it is able to sustain acute SA behavior. This experiment demonstrates that α6β2* nAChRs in the VTA and/or in its projections are necessary and sufficient to establish nicotine SA behavior.
A, Lentiviral vectors used in the re-expression experiments Maps of lentiviral expression vectors. Diagrams of the three lentiviral vectors used in this study, between and including the LTR regions. LTR, long terminal repeat; RNA pack, genomic RNA packaging signal; RRE, rev response element; cPPT, central polypurine tract; CTS, central termination sequence; PGK, promoter of the mouse phosphoglycerate kinase gene; IRES2, internal ribosome entry sequence; eGFP, enhanced green fluorescent protein; WPRE, woodchuck hepatitis B virus post-transcriptional regulatory element; 3′-PPT, 3′-polypurine tract; ΔU3, deletion of the U3 portion of 3′-LTR. B, 125I-Epibatidine and 125I α-conotoxin MII autoradiography. Top: 125I-Epibatidine autoradiography, binding sites are shown in coronal sections from two different levels, VTA (Bregma −3.52 mm) and SN (Bregma −3.08 mm), in WT, β2-KO, α4-KO and vectorised mice. From left to right, wild-type (WT) signal in the VTA (arrow), in SN (arrow head), absence in the β2-KO group, restoration in β2-VEC-VTA mice, absence of epibatidine binding in the α4-KO group, restoration in α4-VEC-VTA mice. Asterisk marks α6β2* and α3β2*-nAChRs in the superior colliculus. Bottom: 125I-conotoxin MII autoradiography, from left to right: Presence of α-conotoxin MII binding sites in WT mice, absence in α6-KO mice and unilateral restoration in α6-VEC-VTA mice. The restoration of α6* binding sites in vectorised α6 KO mice is illustrated with a brain injected unilaterally, the other hemisphere being used as control. Asterisk marks α3β2*-nAChRs in the superior colliculus. Coronal drawings are modified from Paxinos and Franklin (2001).
Verifying the injection sites in VEC mice
Figure 6B illustrates representative samples of specific ligand binding experiments performed on WT, KO and VEC mice, covering the different genotypes and injection paradigms. In the upper part of the figure, binding of iodinated epibatidine reveals the lack of high affinity binding sites in β2 and α4 KO mice, whereas these are restored in the corresponding vectorised mice. As shown before for β2 injection into the SN (Avale et al., 2008), β2 re-expression in the SN yields a strong signal in the area of the SN pars compacta, sparing the VTA.
In the lower part, binding of iodinated conotoxin MII was performed to identify α6*-nAChRs. In α6-KO mice no specific signal was obtained in the VTA, whereas a strong signal was visible in WT mice, and in VEC mice.
Discussion
We have developed a robust, reproducible method for intravenous nicotine self-administration in drug-naive mice and applied it to a detailed study of the role of homo- and heteropentameric nAChRs in this acute paradigm. The method is robust because it was performed in different seasons, over a couple of years, and yielded reproducible results as demonstrated by the comparable data obtained with WT mice of the different KO strains. In the Materials and Methods section we have identified key issues required, like a reversed light/dark cycle, carrying out experiments in the dark, and “nose-poking” rather than “lever-pressing” as modus operandi. Because lever-pressing activity requires more motor and motivational output than nose-poking, it likely may prove to be a less “natural” behavior for mice, and therefore not be the most suitable to unmask behavioral differences in drug-taking behavior.
Our work demonstrates first of all that, in line with previous studies, drug-naive C57BL/6J mice are sensitive to nicotine's reinforcing effects (Robinson et al., 1996; Stolerman et al., 1999). They display a typical inverted U-shaped dose–response curve to intravenous nicotine, as observed in two previous studies (Fattore et al., 2002; Paterson et al., 2003). Moreover, in agreement with the data of Paterson et al. (2003), a sharp dose–response curve was obtained, and a dose of 26.3 μg/kg/inf was confirmed as being the only reinforcing one in sustaining acute SA behavior (Rasmussen and Swedberg, 1998). These findings, that only one specific dose of nicotine is able to sustain SA behavior, are reminiscent of human studies in which adult smokers clearly titrate their nicotine intake carefully to experience the positive effects of the drug (Benowitz and Jacob, 1985; Benowitz, 2001). Similarly, adolescent smokers do titrate their nicotine intake in response to smoking low yield cigarettes by taking more puffs per cigarette (Kassel et al., 2007).
Also in line with previous studies, β2-KO mice no longer self-administer nicotine intravenously (Picciotto et al., 1998), or into the VTA (Maskos et al., 2005). In addition, the data obtained with β2-VEC-VTA mice are consistent with the observation that nicotine increases the burst firing of VTA DAergic neurons (Grenhoff et al., 1986; Mameli-Engvall et al., 2006), decreases electrical self-stimulation in the VTA (Ivanova and Greenshaw, 1997), and dose-dependently enhances DA release when perfused into the VTA (Tizabi et al., 2002; Rahman et al., 2004). Accordingly, infusion of nicotinic antagonist into the VTA significantly decreases nicotine SA in the rat (Corrigall et al., 1994).
Moreover, re-expression of the β2 subunit in the SN pars compacta (Avale et al., 2008) instead of the VTA of β2-KO mice did not yield any nicotine SA. The present study thus unequivocally identifies the presence of the nAChR β2 subunit in the VTA as the critical condition for nicotine intravenous SA in drug-naive animals. Consistent with this conclusion, nicotine SA is altered by lesions of the posterior portion of the pedunculopontine tegmental nucleus, which innervates VTA DAergic neurons, but not by lesions of the anterior region projecting to SN pars compacta (Alderson et al., 2006). Yet, additional brain structures such as the central linear nucleus raphé, located posterodorsally to the VTA, and the supramammilary nucleus of the posterior hypothalamus, located anterior to the VTA, may also contribute to the primary reinforcing effects of nicotine, at least in the rat (Ikemoto et al., 2006). Findings from the present study also demonstrate that drug-naive α7-KO mice do self-administer nicotine in a manner indistinguishable from their WT counterparts. This finding is consistent with the pharmacological evidence that the α7 subunit is not involved in chronic nicotine SA in trained rats (Grottick et al., 2000), nor in nicotine-induced conditioned place preference (Walters et al., 2006). Recent work (Besson et al., 2007; Salas et al., 2007) suggests that α7*-nAChRs play a key role in long-term adaptations to passive chronic nicotine, rather than in the short-term response to nicotine SA.
These in vivo data also invite comparison with the work of McGehee and Mansvelder performed in slices (Mansvelder and McGehee, 2000; Mansvelder et al., 2002). They could show that short application of nicotine leads to the desensitization of heteromeric receptors on both the GABAergic and DAergic neurons in the VTA. Because α7*-nAChRs on glutamatergic axons in the VTA do not desensitize, the net effect is an increased glutamatergic stimulation of the DA neurons in the VTA, explaining the increase in DA cell firing. This increase in elicited DA release is then considered the key step in nicotine reinforcement. It is difficult here to extrapolate from these in vitro studies to our acute reinforcement paradigm. In slices, important afferents, like the cholinergic input from the mesopontine tegmentum, as discussed above (Alderson et al., 2006), get severed. Our in vivo work suggests that this α7*-nAChR mediated effect is not crucial for nicotine SA in an acute paradigm, and we also been able to establish recently that α7-KO mice exhibit only a slightly altered electrophysiological response to nicotine injection when DAergic neurons are recorded in vivo (Mameli-Engvall et al., 2006).
The α4 subunit partner of the β2 subunit has also been implicated in reward: selective activation of α4*-nAChRs by low doses of nicotine that do not activate other nAChR subtypes has been reported to be sufficient for nicotine-induced reward, as tested by conditioned place preference (Tapper et al., 2004). Consistent with this view, α4-KO mice did not show significant nicotine SA and this behavior was promptly restored only after the targeted re-expression of the α4 subunit in the VTA. α4β2*-nAChRs are thus major players in acute nicotine SA in drug-naive mice.
A rather surprising result of our study is the loss of nicotine SA in α6-KO mice. The α6 subunit is to a large extent coexpressed with the α4 subunit in DAergic neurons of the VTA. α4β2*-, α6β2*-, and α4α6β2*-nAChRs have been found enriched in the terminal fields of these neurons in the Nucleus accumbens (NuAcc) (Champtiaux et al., 2003; Salminen et al., 2007). We thus demonstrate that α6*-nAChRs cannot compensate for the absence of the α4 subunit, and vice versa, although both are widely expressed in VTA neurons. Importantly, targeted re-expression of the α6 subunit in the VTA of α6-KO mice promptly restored nicotine SA behavior, confirming the VTA as the key area in mediating nicotine reinforcement.
Our in vivo findings can be seen as a behavioral correlate of recent in vitro work using fast-scan cyclic voltammetry in slices of mouse NuAcc (Exley et al., 2008). This study indeed showed that α-conotoxin MII, a specific antagonist of α6*-nAChRs, selectively suppressed DA release evoked by single and low-frequency action potentials, whereas DA release was potently enhanced by high-frequency bursts. This filtering effect, previously described for nicotine (Rice and Cragg, 2004; Zhang and Sulzer, 2004), is considered the key event in the establishment of saliency of nicotine-associated cues. In this paradigm, α6*-nAChRs therefore dominate the control of DA neurotransmission, when nicotine is applied to the slice. Our in vivo results on α4-KO and α6-KO mice together with the in vitro study of Exley et al. (2008) thus highlight the crucial role of α4* and α6*-nAChRs in nicotine reinforcement.
Our re-expression technique targets both DA and GABAergic neurons in the VTA (Maskos et al., 2005), such that both transmitter systems have their nAChRs restored in the VEC mice. Thus the questions remains as to the relative importance of nAChRs in GABA vs DA neurons for the SA behavior measured here. McGehee and Mansvelder have shown that heteromeric nicotinic receptors play a role in both neurotransmitters systems of the VTA (Mansvelder and McGehee, 2000; Mansvelder et al., 2002) for the increase in DA cell firing, and consequently DA release, considered the key step in the reinforcing effects of a drug. It would thus be interesting to further dissect the action of nicotine in the VTA with respect to neurotransmitter subtype. This issue can be addressed by the selective targeting of nAChRs to GABAergic and DA cells using lentiviral vectors with specific promoters. This work is currently under way (Tolu et al., 2007).
In conclusion, we have demonstrated that the presence of the β2, α4, and α6 subunits in cell bodies of the VTA and their corresponding mesolimbic terminals is necessary and sufficient for acute intravenous SA of a pharmacological dose of nicotine. It is thus tempting to speculate that a specific α4α6β2*-nAChR is the key mediator of intravenous SA. This hypothesis can be tested once FRET experiments using tagged nicotinic subunits will become possible in vivo (Nashmi et al., 2003; Drenan et al., 2008), to prove the presence and identify the functional properties of a given subunit combination.
On the contrary, the α7 subunit is not involved in mediating the acute reinforcing properties of nicotine. These findings illustrate the validity of this new powerful experimental paradigm for a detailed in-depth genetic analysis of nicotine reinforcement in the mouse.
Footnotes
-
This work was supported by grants from the Agence Nationale pour la Recherche (ANR Neuroscience, Neurologie et Psychiatrie 2005), Institut Pasteur, the Collège de France, the Centre National de la Recherche Scientifique CNRS URA 2182, the Association de Recherche sur le Cancer (ARC), the Institut National du Cancer (INCa), Mission Interministérielle de Lutte contre la Drogue et la Toxicomanie (MILDT), the Fondation pour la Recherche Médicale (FRM), and European Commission Contracts ‘Nicotine Dependence’ (NIDE) and ‘Nicotine and Ageing’ to U.M. and J.-P.C, and National Institutes of Health Grants MH 53631, GM 48677, and DA12242 to J.M.M. We thank Jerry Stitzel for his generous gift of the mouse α4 subunit, Ines Ibanez-Tallon for the mouse α6-FLAG clone, Philippe Faure, Morgane Besson, Brian Molles, Sylvie Granon, and Arnaud Cressant for comments on this manuscript, and Barbara Tuveri for animal care.
- Correspondence should be addressed to either of the following: U. Maskos, uwe.maskos{at}pasteur.fr, or W. Fratta, E-mail: wfratta{at}unica.it, at the above addresses.
