|Home | About | Journals | Submit | Contact Us | Français|
Chronic cocaine use produces numerous biological changes in brain, but relatively few are functionally associated with cocaine reinforcement. Here we show that daily intravenous cocaine self-administration, but not passive cocaine administration, induces dynamic up-regulation of the AMPA glutamate receptor subunits GluR1 and GluR2 in the ventral tegmental area (VTA) of rats. Increases in GluR1 protein and GluR1S845 phosphorylation are associated with increased GluR1 mRNA in self-administering animals, while increased GluR2 protein levels occurred despite substantial decreases in GluR2 mRNA. We investigated the functional significance of GluR1 up-regulation in the VTA on cocaine self-administration using localized viral-mediated gene transfer. Over-expression of GluR1WT in rat VTA primarily infected dopamine neurons (75%), and increased AMPA receptor-mediated membrane rectification in these neurons with AMPA application. Similar GluR1WT over-expression potentiated locomotor responses to intra-VTA AMPA, but not NMDA, infusions. In cocaine self-administering animals, over-expression of GluR1WT in the VTA markedly increased the motivation for cocaine injections on a progressive ratio schedule of cocaine reinforcement. In contrast, over-expression of protein kinase A-resistant GluR1S845A in the VTA reduced peak rates of cocaine self-administration on a fixed ratio reinforcement schedule. Neither viral vector altered sucrose self-administration, and over-expression of GluR1WT or GluR1S845A in the adjacent substantia nigra had no effect on cocaine self-administration. Taken together, these results suggest that dynamic regulation of AMPA receptors in the VTA during cocaine self-administration contributes to cocaine addiction by acting to facilitate subsequent cocaine use.
Cocaine self-administration produces numerous neurobiological changes, but relatively few have been related to the behavioral context of reinforcement, where cocaine is administered contingent upon the execution of an instrumental response. For example, the release of dopamine, acetylcholine, glutamate and other neurotransmitters is enhanced by response-contingent cocaine administration (Wilson et al., 1994; Zhang et al., 1994; Dworkin et al., 1995; Hemby et al., 1997; Mark et al., 1999; You et al., 2007), and self-administered cocaine produces differential effects on gene expression compared to passive administration of cocaine (Freeman et al., 2002; Jacobs et al., 2002; Thomas et al., 2003; Stefanski et al., 2007; Miguens et al., 2008). There are even fewer instances where such response-contingent regulation has been shown to act reciprocally to directly influence cocaine reinforcement itself. Such reciprocal interaction between reinforcement-related neurobiological change and subsequent cocaine-seeking behavior is critical to understanding the pathological process of cocaine addiction.
Dopamine neurons in the ventral tegmental area (VTA) play a key role in the reinforcing effects of cocaine. Previous studies found that even a single exposure to cocaine can lead to a transient enhancement in the excitability of dopamine neurons resulting from increased sensitivity to amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) glutamate receptor stimulation (White et al., 1995; Zhang et al., 1997) and potentiated synaptic responses resembling long-term potentiation (LTP) in the VTA (Ungless et al., 2001; Saal et al., 2003; Borgland et al., 2004). Repeated cocaine exposure also enhances susceptibility to LTP induction (Liu et al., 2005). This transient enhancement in VTA dopamine neuron excitability may persist for up to 5 days after passive cocaine administration (Zhang et al., 1997; Ungless et al., 2001; Borgland et al., 2004). However, in cocaine self-administering animals, increased AMPA-mediated excitation of VTA dopamine neurons persists for up to 3 months of withdrawal, and does not occur in animals that receive cocaine passively via yoked intravenous injections (Chen et al., 2008). Thus, this long-lasting neuroplasticity is related specifically to the reinforcing context of cocaine self-administration.
Neuroplasticity in VTA dopamine neurons could be driven by changes in the amount of AMPA receptor subunits, or phosphorylation of the GluR1 subunit at serine residues 831 and 845 that are both necessary and sufficient for AMPA receptor trafficking in synaptic membranes (Malinow and Malenka, 2002; Song and Huganir, 2002; Esteban et al., 2003; Derkach et al., 2007). Neurobiological changes that alter the excitability of dopamine neurons could profoundly affect motivational responses to drugs of abuse, and promote cocaine-seeking behavior during abstinence when exposed to drug-associated environmental cues and stress that activate this system (Stewart, 1983; Berridge and Robinson, 1998; Self and Nestler, 1998; Stewart, 2000; Shalev et al., 2002; Phillips et al., 2003; Self and Choi, 2004). Here, we found that cocaine self-administration dynamically regulates AMPA receptor subunits in the VTA in a manner specific to cocaine reinforcement. We then investigated the functional significance of GluR1 up-regulation in the VTA on cocaine self-administration behavior using viral-mediated gene transfer to over-express GluR1 in VTA neurons.
Male Sprague-Dawley rats, weighing 275–300 gm on arrival (Charles River, Kingston, NY, USA), were housed individually in a climate-controlled environment (21°C) on a 12-hr light-dark cycle. All animals were maintained according to the guidelines of the National Institutes of Health and approved by the Institutional Animal Care and Use Committee of the University of Texas Southwestern Medical Center. To facilitate acquisition of cocaine self-administration, animals were temporarily maintained on a restricted diet of lab chow at 85% of their original body weight, and trained to lever-press for 45 mg sucrose pellets on a fixed ratio 1 (FR1) reinforcement schedule until acquisition criteria were achieved (100 pellets self-administered for 3 consecutive days). Animals were then fed ad libitum for at least 1 day before surgery.
Under sodium pentobarbital anesthesia (60 mg/kg, i.p.), a catheter composed of SILASTIC tubing (Green Rubber, Woburn, MA, USA) and treated with tridodecylmethyl ammonium chloride (TDMAC) heparin (Polysciences Inc., Warrington, PA), was surgically placed in the animal’s jugular vein. The catheter was secured with Mersilene surgical mesh (General Medical, New Haven, CT) at the jugular vein, and passed subcutaneously to exit the animals’ back through a 22 gauge cannula (Plastics One, Roanoke, VA) imbedded in dental cement on a Marlex surgical mesh (Bard Inc., Cranston, RI). For experiments with brain infusions, animals also underwent stereotaxic surgery to implant 26 gauge bilateral guide cannulae (Plastics One, Roanoke, VA) in the VTA or the substantia nigra (SN). Stereotaxic coordinates for the VTA and the SN were: −5.6 mm posterior to bregma, ±0.8 mm (VTA) or ±1.5 mm (SN) lateral, and −7.0 mm ventral to dura (Paxinos and Watson, 1998). Dummy cannulae (33 gauge) were left in place throughout the experiment. Animals received a prophylactic injection of penicillin (60,000 IU/0.2 ml, i.m.) and antibiotic ointment to the catheter exit wound daily. Catheters were flushed daily with 0.2 ml of heparinized (20 U/ml), bacteriostatic saline containing gentamycin sulfate (0.33mg/ml).
Operant chambers (Med Associates Inc., St. Albans, VT) for cocaine and sucrose self-administration were contextually different from the animals’ home cage, and located in different rooms. Each chamber was equipped with an infusion pump (Razel Model A pump, Stamford, CT) and 10 ml glass syringe connected to a fluid swivel (Instech, Plymouth Meeting, PA) by Teflon tubing. Tygon® tubing enclosed by a metal spring connected the swivel to the animal’s catheter exit port and was secured to Teflon threads on the catheter assembly. Following 1 week of recovery from surgery, animals were trained to self-administer cocaine in 4 hr sessions for 5–6 days/week. A single lever-press response at the active lever produced a 0.5 mg/kg intravenous (iv) injection of cocaine (NIDA, Research Triangle Park, NC) delivered in 0.05 ml saline over 2.5 sec, concurrent with illumination of a cue light located above the active lever while the house light was extinguished. Each injection was followed by an additional 12.5 sec time-out (TO) period when the house light remained off, and active lever-press responses had no scheduled consequence during this period. Responses on the inactive lever were recorded but had no consequences.
Animals were trained to self-administer cocaine for 3 weeks as described above, and brain tissue was collected at different time points (immediately after the final session, 1 day or 3 weeks withdrawal). Self-administering rats were paired with yoked partners that received an identical amount and temporal pattern of cocaine injections but not contingent upon lever-press behavior. An additional group of animals received yoked cocaine injections for the first time during the final session after receiving yoked saline injections in previous sessions to compare with the chronic cocaine groups with no withdrawal. Another group of animals self-administered saline throughout all sessions to control for potential surgical or other influences relating to the testing procedures.
For determination of AMPA and NMDA receptor subunit levels and their phosphorylation status, brain tissue from the VTA and the SN was dissected following brief (1.6 sec) microwave irradiation (5kW) aimed at the head (Muromachi, Kikai Co. Ltd., Tokyo, Japan) as previously described (Edwards et al., 2007). Following microwave fixation, VTA and SN tissue was obtained from chilled coronal brain slices (~−4.8 to −6.3 mm posterior to bregma) using a 16-gauge punch. Immediately following brain dissection, tissue was homogenized by sonication and boiled for 5 min in lysis buffer (320 nM sucrose, 5 nM HEPES, 50nM NaF, 1 mM EGTA, 1 mM EGTA, 1% SDS containing Protease Inhibitor Cocktail and Phosphatase Inhibitor Cocktails I and II; Sigma, St. Louis, MO), and stored at −80°C until further analysis.
Following protein determination by the Lowry method, 20–40 µg protein aliquots were separated by SDS-PAGE on 7.5–10% acrylamide using Tris/Glycine/SDA buffer (Bio-Rad, Hercules, CA) and electrophoretically transferred to PVDF membranes. Membranes were blocked with 5% non-fat dry milk in phosphate-buffered saline (PBS) containing 0.1% Tween 20 overnight at 4°C, washed and incubated in affinity-purified rabbit polyclonal anti-pGluR1S831 or anti-pGluR1S845 (1:2500; Millipore, Billerica, MA), and stripped and re-probed with antibodies for total GluR1 protein (1:5,000; Millipore). Other blots were probed with antibodies for GluR2, NR2A, NR2B or mouse monoclonal anti-GAD (1:5,000; Millipore,), or anti-pTHS40 (1:1500; Cell Signaling Technology, Inc., Danvers MA) followed by mouse monoclonal anti-TH (1:200,000; Millipore). All blots were stripped and re-probed for (3-tubulin as a protein loading control. Following labeling with primary antibodies, blots were washed and labeled with the appropriate species-specific peroxidase-conjugated secondary antibodies (1:25,000; Vector Laboratories, Burlingame, CA, USA), labeled proteins were detected by enhanced chemiluminescence, and densitized using the NIH image 1.57 as described previously (Edwards et al., 2007). Under these conditions, target protein amounts were linear over a 3–4 fold range. Each blot contained tissue from 5–6 age- and group-matched untreated controls that remained in their home cages but were handled daily to allow normalization of data between blots.
In separate study groups, tissue was dissected without microwave fixation in animals euthanized immediately or 1 day after withdrawal from chronic yoked or self-administered cocaine. Total RNA was extracted from the VTA tissue using Trizol reagent (Invitrogen, Carlsbad, CA), precipitated with isopropanol and treated with DNase to remove genomic DNA (Ambion, Austin, TX). RNA was reverse transcribed to cDNA using a first-strand synthesis kit (Invitrogen, Carlsbad, CA). Cycle thresholds (Ct) were determined in triplicates of a same sample by the ΔΔCt method using the following primer sequences for GluR1: 5'-GTCCGCCCTGAGAAATCCAG-3', 5'-CTCGCCCTTGTCGTACCAC-3', GluR2: 5'-GCCGAGGCGAAACGAATGA-3', 5'-CACTCTCGATGCCATATACGTTG-3', and glyceraldehyde-3-phosphate dehydrogenase (GAPDH): 5'-AACGACCCCTTCATTGAC-3', 5'-TCCACGACATACTCAGCAC-3'.
PC12 cells (Clontech, Palo Alto, CA) were plated at a density of 106 cells per well on 35 mm culture plates maintained at 37°C, 5% CO2 in DMEM supplemented with 10% fetal calf serum and 1% solution of 5000 U/ml penicillin and 5000 µg/ml streptomycin (Invitrogen, Carlsbad, CA) as previously described (Kumar et al., 2005). Cells were infected with 1 µl/well HSV vectors (4.0 × 107 infectious units/ml) after 70–80% confluence in 35-mm wells as described previously (Bachtell et al., 2008). After 24 hr, cells were incubated for 30 min with either 5 µM forskolin, phorbol-12-myristate-13-acetate (PMA) or fresh medium. Cells were harvested using a lysis buffer (1% SDS, 20 mm HEPES, pH 7.4, 100 mm NaCl, 20% glycerol, 1 mm EDTA, 1 mm EGTA, 10 µg/ml leupeptin, 10 µg/ml aprotenin, 10 µg/ml pepstatin, 1 mm phenylmethylsulphonyl fluoride, 50 mm NaF, 0.1 mm sodium orthovanadate, and 10 mm sodium pyrophosphate), briefly sonicated and centrifuged, and levels of GluR1, pGluR1S845 and pGluR1S831 were determined in 20 µg protein aliquots by western blot as described above. Uniformity of protein loading concentration was confirmed with Ponceau S solution staining (Sigma, St. Louis, MO).
For immunocytochemistry and confocal microscopy of HSV-GluR1 over-expression, naïve rats received unilateral stereotaxic infusions of HSV-GluR1WT or HSV-GluR1S845A (1.0 µl/side) in the VTA, while the contralateral side (balanced left and right) received infusions of the PBS solution delivered through 26 gauge Hamilton microsyringes (Hamilton, Reno, NV) at 5.6 mm posterior to bregma, ±0.8 mm lateral, and −8.0 mm ventral to dura. After 2 days, rats were anesthetized with chloral hydrate and killed via intracardiac perfusion of PBS followed by 4% paraformaldehyde (20 min, 12 ml/min). The brains were post-fixed in 4% paraformaldehyde overnight and cryoprotected in 20% glycerol/PBS at 4 °C for 3 days. Coronal brain sections (30 µm) were blocked with 3% normal donkey serum and 0.3% Triton-X in PBS for 60 min prior to incubation with rabbit polyclonal anti-GluR1 (1:1,000; Millipore) and mouse monoclonal anti-TH (1:5000; Millipore) in 3% normal donkey serum and 0.3% Tween-20 for 18–20 h. After washing with PBS, sections were incubated with fluorescent-tagged secondary antibodies for 60 min (Cy2-conjugated donkey anti-mouse for TH; Cy3-conjugated donkey anti-rabbit for GluR1; Jackson Immunoresearch, West Grove, PA). After incubation with secondary antibodies, VTA sections were counterstained with DAPI (1:5000; Roche Applied Science, Mannheim, Germany) for 20 min at room temperature. VTA sections were sequentially dehydrated in 70%, 95% and 100% of ethanol and Citrosolv, and cover-slipped with DPX (Sigma-Aldrich). Negative controls for antibody labeling indicated a lack of specific staining when omitting or diluting the primary antibody.
Confocal microscopy was performed to quantify the relative percent of ectopic HSV-GluR1 expression in TH-positive neurons, and to determine the percentage of infected TH-positive neurons within a 0.5 mm diameter zone of highest HSV-mediated GluR1 expression in the VTA. Simultaneous epifluorescence (Cy2 and Cy3) images were obtained at 20X and 63X magnification with a laser-scanning confocal microscope (Zeiss Axiovert 200 and LSM510-META; Thornwood, New York) using three lasers comprised of argon (458, 477, 488 and 514 nm), HeNe1 (543 nm) and HeNe2 (633 nm). Laser-scanning and optical sectioning in the Z plane were performed using multitrack scanning with a section thickness of 1.45 µm for 20X or 0.45 µm for 63X magnification (Donovan et al., 2008). Co-localization was assessed by analysis of adjacent Z sections, orthogonal sectioning through Z sections, and three-dimensional reconstruction with rotation. Confocal images were imported into Adobe Photoshop (Adobe Systems, Mountain View, CA) for composition of merged images. Single- and double-labeled GluR1WT and TH-positive cells were counted within the infected region across 3 slices per animal and averaged to obtain an individual percent colocalization for each of 6 animals. The mean percentage of colocalization across animals was determined and expressed as both a percentage of TH-positive expression in ectopic GluR1WT-expressing cells and the percentage of total TH-positive cells within the infected region that express GluR1WT. Dendritic labeling of GluR1 in HSV-GluR1WT and HSV-GluR1S845A expressing TH-positive neurons was quantified by measuring the length of GluR1-labeled processes from the soma in 3–5 cells/animal under blinded conditions and expressed as the mean process length (µm)/animal (3–4 animals per group).
VTA slice cultures were prepared from postnatal day 25–35 rats anaesthetized with isoflurane as described previously (Han et al., 2006; Krishnan et al., 2007; Cao et al., 2010b). A tissue block containing midbrain was taken and sliced in ice-cold solution containing (in mM) 254 sucrose, 3 KCl, 1.25 NaH2PO4, 10 d-glucose, 24 NaHCO3,2 CaCl2, and 2 MgSO4. Slices, 300 µm thick were transferred to a holding chamber in 34°C containing artificial cerebrospinal fluid (aCSF, in mM: 128 NaCl, 3 KCl, 1.25 NaH2PO4, 10 d-glucose, 24 NaHCO3,2 CaCl2, and 2 MgSO4, pH 7.35, 295–305 mOsm). After 45–60 min recovery, slices were transferred onto the membrane of Millicell (Millipore) containing culture medium: MEM with 30 mM HEPES, 20 mM D-glucose, 5% B27, 5.0 mM L-glutamine, and 25 U/ml streptomycin/penicillin. After 60 min incubation, GFP-tagged HSV-GluR1WT and HSV-GluR1S845A vectors were pipetted onto the VTA area of the slice surface. Slices were maintained overnight at 34°C, and then put into a recording chamber perfused with standard aCSF at a rate of 2.5 ml/min. All solutions, except for culture medium, were saturated with 95% O2 and 5% CO2. GFP positive cells were visualized with an upright fluorescence microscope using infrared differential interference contrast (IR-DIC) illumination. Whole-cell voltage-clamp recordings were performed under continuous single-electrode voltage clamp mode (AxoClamp 2B, Axon Instruments Inc., Union City, CA, USA). Electrodes (2–4MΩ) were filled with pipette solution containing (in mM) 115 potassium gluconate, 20 KCl, 1.5 MgCl2, 10 phosphocreatine, 10 Hepes, 2 ATP-Mg and 0.5 GTP (pH 7.2, 285 mOsm). Data acquisition was made using DigiData 1322A and pClamp 8 (Axon Instruments).
In these experiments, putative dopamine neurons in the VTA were identified by large hyperpolarization-activated currents (Ih) as described previously (Ungless et al., 2003; Cao et al., 2010a). Ih current was evoked by a family of 10 mV voltage steps (duration 600 ms) from −60 mV to −140 mV holding potentials. In separate experiments, Ih-positive neurons were recorded in 150 µm thick slices, and cells were filled with biocytin (1% internal recording solution) using 20 pA depolarized current injections. Slices were fixed immediately after recording in 4% formaldehyde for 2 hours and then stored at 4°C in PBS. Slices were processed for TH immunoreactivity as discussed above.
Locomotor activity was recorded in the using 1.95 m circular test chambers with a 12 cm wide runway, and equipped with four pairs of photocells located at 90-degree intervals. Drug naïve animals with bilateral guide cannulae in the VTA were habituated to the locomotor apparatus for 2 hr prior to testing, and given 1.0 µl intra-VTA infusions of the HSV vectors through bilateral 33-gauge infusion cannulae extending 1 mm beyond the guide cannulae tip over a 5 min period. Infusion cannulae were left in place for an additional 2 min to allow for diffusion. Each subsequent test session incorporated a 2-hr habituation phase followed by an intra-VTA infusion of either a PBS vehicle, AMPA (10 ng/side, bilateral) or NMDA (500 ng/side, bilateral) in a volume of 0.5 µl/side over 100 sec through bilateral 33-gauge infusion cannulae in counter balanced order over 3 consecutive test days. The injectors were left in place for 30 sec, gently removed, and the animals were placed back into the locomotor apparatus where locomotor activity was recorded for 1 hr.
Animals were trained to self-administer cocaine (0.5 mg/kg/injection) for 3 weeks as described above, then the response requirement was raised over subsequent sessions from a fixed ratio (FR) 1 to FR5 and continued until cocaine intake stabilized to within 10% of the mean of 3 consecutive sessions. Following stabilization, animals were trained in a within-session FR5 dose-response procedure with each injection dose (1.0, 0.3, 0.1, 0.03, and 0 mg/kg) available in descending order in sequential 60 min components following an initial 30 min loading phase (0.5 mg/kg/injection). The unit dose/injection was adjusted by reducing the injection volume (0.2, 0.06, 0.02, 0.006, and 0 mls) to produce unit injection doses of 1, 0.3, 0.1, 0.03, and 0 mg/kg cocaine, respectively. Animals were trained until the injection dose producing peak self-administration rates remained constant for 3 consecutive sessions. Following stabilization of peak rate cocaine doses, animals received intra-VTA or intra-SN infusions of the HSV vectors (1 µl/side) through bilateral guide cannulae as described above. Animals were then tested in within-session dose-response tests during HSV-GluR1 over-expression of (post-infusion days 2–5), and 7–10 days after HSV infusions when HSV-mediated over-expression is diminished to undetectable levels in brain (Carlezon et al., 1997; Sutton et al., 2003; Bachtell et al., 2008).
Following FR5 dose-response testing, rats were re-stabilized on the FR5 schedule (0.5 mg/kg/injection) in daily 4 hr sessions and then trained on a progressive ratio (PR) schedule at either 0.5 or 1.0 mg/kg/injection for 2 weeks. The number of active lever-presses for each successive cocaine injections increased according to the calculation [5e(injection number × 0.2)]−5; i.e., responses/injection increased as 1, 2, 4, 6, 9, 12, 15, 20, 25, 32, 40, 50, etc.). The highest ratio of responses/injection achieved before a 1 hr period when no further injections were earned (break point) was determined in daily tests until they varied <10% for 3 consecutive days. Animals then received the same HSV infusions as in FR5 testing and break points were determined for post-injection days 2–5 (over-expression) and again on days 7–10 (post-expression).
Bilateral guide cannulae were implanted in the VTA of drug-naïve rats as described above. Following a week of recovery, rats were food restricted to 85% body weight and trained to self-administer sucrose pellets on a FR1:TO 15 sec reinforcement schedule for a maximum of 50 sucrose pellets available. Following acquisition, the reinforcement schedule was increased to FR5 for at least 3 weeks and until the latency to consume 50 pellets stabilized to <10% variance from the mean of 3 consecutive sessions. Animals showing stable responses on the FR5 schedule received intra-VTA infusions of HSV- vectors and were tested during HSV-GluR1 over-expression (2–5 days post-infusion). Animals were given 1 week break from food self-administration and then trained on a progressive ratio schedule until break points stabilized on the following schedule progression (1, 2, 5, 9, 13, 19, 25, 33, 41, 51, 61, 76, 91, 111, etc.). Animals then received the same HSV vectors in the VTA and break points were determined during GluR1 over-expression (2–5 days post infusion).
Following completion of behavioral testing, animals were anesthetized with chloral hydrate (300 mg/kg, i.p.) and bilateral infusions of 0.5 µl cresyl violet were delivered in the VTA or SN through the guide cannulae. Five minutes after the cresyl violet infusions, animals were decapitated, brains dissected and infusion sites were identified in 1 mm coronal slices. Only animals with correct bilateral infusion sites in the VTA and the SN were included in the data analysis.
Biological and locomotor data were analyzed by 1-factor ANOVA across study groups, or unpaired T-tests if only 2 groups were compared. Self-administration data were analyzed by 2-factor ANOVA (dose × HSV treatment) with repeated measures on dose (FR only), and by 1-factor ANOVA at each dose, followed by post-hoc tests with Fisher's Least Significant Difference (LSD) test. Statistical significance was preset at p < 0.05.
Animals were allowed to self-administer intravenous cocaine injections for 4 hr/day over 18 days, and brain tissue was collected at different time points following the final self-administration session (no withdrawal, 1 day withdrawal or 3 weeks withdrawal). Each self-administering animal was yoked to an animal that received the same number and temporal pattern of cocaine injections passively and independent of lever-press behavior (Figs. 1a). Cocaine self-administration was stable by the third week of training with average daily intake (days 13–18) ranging from 44.8 to 50.2 mg/kg in self-administering animals and their chronic yoked partners. After 1 day withdrawal from cocaine self-administration, GluR1 protein levels were increased by 82% in the VTA when compared to untreated controls (F2,44 = 10.020, P = 0.001), but GluR1 levels were not altered in animals that received cocaine by passive yoked injection (Fig. 1b,d). These results indicate that GluR1 regulation was related to cocaine delivery in a response-contingent (reinforcing) manner. Increases in GluR1 protein levels in self-administering animals were paralleled by a 74% increase in protein kinase A (PKA)-mediated GluR1S845 phosphorylation (F2,44 = 3.887, P = 0.028), but not GluR1S831 phosphorylation mediated by protein kinase C (PKC) and/or calcium/calmodulin-dependent kinase II (CaMKII). Concomitant increases in phosphorylated and total GluR1 suggest that GluR1S845 phosphorylation is co-regulated with the total amount of GluR1 protein in VTA neurons rather than enhanced PKA activity per se. Since GluR1S845 phosphorylation is necessary and sufficient for extra-synaptic membrane insertion of newly synthesized AMPA receptors (Song and Huganir, 2002; Esteban et al., 2003; Derkach et al., 2007), these findings suggest an increase in GluR1-containing AMPA receptors in extra-synaptic membranes after 1 day withdrawal from cocaine self-administration.
Levels of total GluR1 and GluR1S845 phosphorylation were normalized immediately following a return to 4 hr of cocaine self-administration (Fig. 1c). At the same time, GluR1 mRNA increased by ~1.45 fold in the VTA in cocaine self-administering but not yoked animals, and this increase persisted after 1 day cocaine withdrawal (Fig. 1e,f; F4,20 = 5.212, P = 0.005). Together, these findings indicate that increased GluR1 protein after 1 day cocaine withdrawal likely results from reinforcement-related activation of GluR1 gene expression during and after cocaine self-administration, but GluR1 protein may be degraded during cocaine self-administration, resulting in dynamic daily fluctuations in GluR1 levels during ongoing cocaine use. In contrast to chronic cocaine self-administration, an initial exposure to passive yoked cocaine injections in drug naïve animals increased PKA-mediated GluR1S845 phosphorylation by 40% (F4,77 = 4.495, P = 0.03), but tolerance developed to acute cocaine regulation of GluR1S845 phosphorylation after chronic yoked administration (Fig. 1b,c).
The GluR2 AMPA receptor subunit displayed a similar reinforcement-related 31% up-regulation in the VTA after 1 day withdrawal from self-administered but not yoked cocaine (Fig. 2a–c; F2,44 = 3.95, P = 0.026). GluR2 levels were normalized along with GluR1 immediately after return to 4 hr of cocaine self-administration. However, in contrast to GluR1 up-regulation, increases in GluR2 protein occurred despite a substantial 64–69% reduction in GluR2 mRNA in both chronic yoked and self-administering animals after 1 day withdrawal (Fig. 2d,e; F4,21 = 4.253, P = 0.011). Thus, reinforcement-related up-regulation of GluR2 protein is uncoupled from down-regulation in GluR2 mRNA in the VTA.
There were no changes in the phosphorylation or amount of GluR1 and GluR2 in animals self-administering saline injections throughout the 18 days of training compared to untreated controls (Figs. 1c and and2b),2b), indicating a lack of surgical or procedural influences unrelated to cocaine self-administration. In addition, regulation of GluR1 and GluR2 in cocaine self-administering animals occurred specifically in the mesolimbic dopamine system, since there was no regulation in adjacent SN tissue (Table 1), and cocaine administration (self or passive) failed to regulate the NMDA glutamate receptor subunits NR1, NR2A and NR2B in the VTA (Table 2). Moreover, Table 2 shows there was no regulation of the dopamine and GABA synthesizing enzymes tyrosine hydroxylase (TH) and glutamic acid decarboxylase (GAD), respectively, consistent with our previous TH work in self-administering animals after 1 day withdrawal (Schmidt et al., 2001). Finally, reinforcement-related up-regulation of GluR1 and GluR2 protein in the VTA failed to endure after 3 weeks withdrawal from chronic cocaine self-administration (data not shown).
In order to study the functional role of GluR1 up-regulation on cocaine self-administration behavior, we generated HSV vectors expressing wild-type GluR1, or phosphorylation-resistant GluR1 mutants. Figure 3 illustrates HSV-mediated over-expression of GluR1 in vitro (PC12 cells) and in the VTA of rats after in vivo brain infusions. All 3 viral vectors over-expressed GluR1 in PC12 cells when compared to lacZ-expressing controls (Fig. 3a,b). Treatment with the PKA activator forskolin increased S845 phosphorylation of GluR1WT and the PKC/CaMKII-resistant GluR1S831A, but not the PKA-resistant mutant GluR1S845A as expected. Conversely, treatment with the PKC activator PMA increased S831 phosphorylation of GluR1WT and the PKA-resistant mutant GluR1S845A, but not the PKC/CaMKII-resistant GluR1S831A. Intra-VTA infusions of HSV-GluR1WT led to prominent ectopic expression of GluR1 two days later (Fig. 3c). Quantification of the HSV-GluR1WT-infected cells in the VTA revealed that a majority (75%) of GluR1 over-expressing cells were TH-positive dopamine neurons (N = 6 animals), and that 25% of all dopamine neurons were infected by HSV-GluR1WT within the infected region.
Higher magnification microscopy showed that dopamine neurons infected with HSV-GluR1WT showed prominent GluR1 labeling throughout cell bodies and dendritic processes (Fig 3d). In contrast, dopamine neurons infected with PKA-resistant HSV-GluR1S845A showed GluR1 labeling in cell bodies but with very little dendritic labeling. Quantification of dendritic GluR1 labeling in infected neurons of the VTA found significantly less perisomal labeling with expression of PKA-resistant GluR1S845A (27.1 ± 3.77 µm/dendrite) when compared to GluR1WT (67.4 ± 11.9 µm/dendrite; T1,9 = 2.714, P = 0.024; N = 3–4 animals/group), consistent with a recent report showing that GluR1S845 phosphorylation facilitates dendritic trafficking of GluR1 in hippocampal neurons (Kessels et al., 2009).
The effect of HSV-GluR1 infection on functional membrane AMPA receptors was examined in VTA slice cultures (Fig. 4a–e). In these slice cultures, 90% of biocytin-filled VTA neurons that exhibited a large Ih current were TH positive (Fig. 4a,b). Therefore, putative dopamine neurons in the VTA were identified by the presence of a large Ih current as described in previous studies (Ungless et al., 2003; Cao et al., 2010a). VTA slice cultures were infected with GFP-tagged HSV-GluR1 and GFP-positive neurons exhibiting large Ih currents were recorded by whole cell voltage clamp (Fig. 4c–e). Current responses to bath-applied AMPA (1 µM) were determined at +40 mV and −60 mV and the ratio of these currents was used to measure rectification. Figure 4d illustrates that AMPA-induced currents typical in GluR1WT-GFP expressing VTA neurons show increased rectification at positive potentials compared to non-infected neurons, indicative of an abundance of newly inserted homomeric GluR1-containing AMPA receptors. Over-expression of GluR1WT-GFP increased rectification by ~40%, while over-expression of PKA-resistant GluR1S845A-GFP had no effect compared to uninfected controls (Fig. 4e). These results indicate that putative dopamine neurons infected by HSV-GluR1WT-GFP show physiologically significant membrane incorporation of over-expressed GluR1WT, but not when infected with the PKA-resistant GluR1 mutant. However, peak currents (−60 mV) elicited by bath application of AMPA failed to differ with over-expression of GluR1WT-GFP, averaging 128 ± 23 pA when compared to 136 ± 25 pA (non-infected) and 137 ± 22 pA (GluR1S845A-GFP), and suggesting that homomeric GluR1-containing AMPA receptors may replace existing membrane AMPA receptors.
Over-expression of GluR1WT in VTA neurons in vivo also enhanced the locomotor response to intra-VTA infusions of AMPA when compared to LacZ-expressing controls (Fig. 4f; F(3,27) = 20.320, P = 0.0001). In contrast, over-expression of PKA-resistant GluR1S845A had no effect on AMPA-mediated locomotion, consistent with electrophysiological responses to AMPA described above. In addition, neither GluR1WT nor GluR1S845A over-expression in VTA neurons affected the locomotor response to intra-VTA infusions of NMDA, indicating a specific enhancement of AMPA-mediated behavior.
Given that GluR1 in the VTA is dynamically regulated during daily cycles of cocaine self-administration and withdrawal (Fig. 1), we investigated the functional role of GluR1 up-regulation in the VTA on cocaine self-administration behavior in rats using fixed and progressive ratio schedules of cocaine reinforcement. HSV-mediated over-expression of GluR1WT had no effect on the rate of cocaine intake with low fixed ratio response demands over a wide dose range (Fig. 5a,b), but dramatically increased the level of effort exerted by rats to maintain cocaine self-administration on the more demanding progressive ratio schedule of reinforcement (Fig. 5d,e). This effect is reflected by an almost 3-fold greater ratio of lever-presses/cocaine injection achieved by rats over-expressing GluR1WT than by lacZ-expressing controls when tested at a high cocaine dose (F3,34 = 3.934, P = 0.0164). Conversely, over-expression of the GluR1S845A mutant in VTA neurons reduced peak rates of fixed ratio cocaine self-administration at a lower cocaine dose (F3,40 = 4.544, P = 0.0078), but had no effect on the effort exerted for higher dose cocaine injections on the progressive ratio schedule. These results indicate that VTA expression of GluR1WT enhances cocaine reinforcement at higher doses, while the PKA-resistant GluR1S845A mutant may have a moderate dominant negative effect on cocaine reinforcement that is evident at lower threshold doses for maintaining self-administration behavior. The effects of HSV-GluR1WT and HSV-GluR1S845A on fixed and progressive ratio cocaine self-administration had diminished completely 7–10 days after intra-VTA infusions (Fig. 5c,f), consistent with the transient over-expression profile of these HSV vectors described in previous studies (Carlezon et al., 1997; Sutton et al., 2003; Bachtell et al., 2008). There was no effect of the PKC/CaMKII-resistant GluR1S831A mutant in either fixed or progressive ratio self-administration testing.
In order to determine whether the effects of over-expressed GluR1WT and GluR1S845A in VTA neurons produced generalized effects on reinforcement mechanisms or response capacity, the HSV-GluR1WT and HSV-GluR1S845A vectors were infused into the VTA of rats trained to self-administer sucrose pellets on similar fixed and progressive ratio reinforcement schedules. Neither GluR1WT nor GluR1S845A over-expression in VTA neurons altered the rate of sucrose pellet self-administration on the fixed ratio schedule, or the level of effort exerted for sucrose pellets on the progressive ratio reinforcement schedule (Fig. 6a,c), indicating specific alterations in cocaine reinforcement. In addition, when HSV-GluR1WT and HSV-GluR1S845A were infused into the adjacent SN region, there were no effects on fixed or progressive ratio cocaine self-administration (Fig. 6b,d), indicating that cocaine reinforcement is altered by GluR1WT and GluR1S845A over-expression specifically in the mesolimbic dopamine pathway. Figures 6e and 6f show the anatomical location of HSV infusion sites in the VTA and the SN, respectively, based on histological confirmation for all animals included in this study.
We found that chronic daily cocaine self-administration produces reinforcement-related up-regulation in AMPA, but NMDA, glutamate receptor subunits in the VTA. This up-regulation consists of increases in GluR1 and GluR2 AMPA receptor subunits after 1 day withdrawal from cocaine self-administration, and a corresponding increase in PKA-mediated GluR1S845 phosphorylation, without increasing PKC/CaMKII-mediated GluR1S831 phosphorylation. GluR1S845 phosphorylation facilitates both dendritic and extra-synaptic membrane trafficking of GluR1-containing AMPA receptors, while GluR1S831 phosphorylation facilitates translocation to synaptic membranes in hippocampal neurons (Song and Huganir, 2002; Esteban et al., 2003; Derkach et al., 2007; Kessels et al., 2009). Thus, an abundance of phosphorylated GluR1S845-containing AMPA receptors could promote extra-synaptic membrane incorporation of AMPA receptors that are primed for synaptic translocation with subsequent activity-dependent GluR1S831 phosphorylation in VTA neurons, granted that similar trafficking mechanisms exist in both hippocampal and VTA neurons. In this regard, our results parallel the development of LTP in excitatory synapses on VTA dopamine neurons that occurs specifically in animals that self-administer cocaine, and not in animals receiving cocaine by passive yoked injection (Chen et al., 2008). However, in contrast to the prolonged (up to 3 months withdrawal) enhancement of excitatory responses in VTA dopamine neurons reported by Chen and colleagues, we found that up-regulation of GluR1 and GluR2 is no longer detected after 3 weeks withdrawal from cocaine self-administration. Thus, while reinforcement-related up-regulation in GluR1 and GluR2 could contribute to the initial development of LTP in self-administering animals, other enduring mechanisms may underlie its persistence.
Increases in GluR1 protein are likely mediated by increased gene expression since cocaine self-administration also increased GluR1 mRNA. One potential mechanism could involve an increase in extracellular glutamate levels in the VTA that occurs only in cocaine self-administering animals and not in their yoked partners (You et al., 2007). Thus, a glutamate-NMDA receptor-Ca++ signaling cascade culminating in a reinforcement-specific increase in GluR1 expression could account for these findings. Another possibility is that BDNF-TrkB receptor signaling could contribute to GluR1 up-regulation, since BDNF increases GluR1 expression in cultured hippocampal neurons (Caldeira et al, 2007), and LTP in VTA dopamine neurons is facilitated by BDNF during withdrawal from cocaine administration (Pu et al., 2006). Whether the latter response would be specific to self-administration behavior is unknown. In contrast to GluR1, increases in GluR2 protein were accompanied by decreases in GluR2 mRNA. GluR2 subunits are recycled constitutively and inserted into synaptic membranes independent of activity/phosphorylation changes (Passafaro et al., 2001; Shi et al., 2001). However, since GluR1S845 phosphorylation promotes membrane stabilization and protection from lysosomal degradation (He et al., 2009), it is possible that an up-regulation in phosphorylated GluR1 ultimately would form protective heteromers with GluR2 and account for increases in GluR2 after 1 day withdrawal from cocaine self-administration.
Our results suggest that transient increases in GluR1 and GluR2, along with decreases in GluR2 mRNA, are dynamically modulated by daily cocaine use patterns, as these changes were obliterated after engaging in subsequent cocaine self-administration. Normalization of GluR1 during cocaine self-administration could involve NMDA-mediated dephosphorylation of GluR1S845 leading to internalization and degradation (He et al., 2009). In any event, such dynamic increases in VTA AMPA receptors could contribute to daily cocaine use by facilitating the excitability of VTA dopamine neurons in response to environmental cues or stress that activate these neurons (see Introduction) and trigger craving responses in human cocaine addicts (Robbins et al., 1997; Sinha et al., 1999). We also found that the ability of cocaine to directly increase PKA-mediated GluR1S845 phosphorylation in the VTA is reduced after chronic administration, similar to our previous findings in striatal and amygdalar subregions, but not in cortical or hippocampal sub-regions (Edwards et al., 2007). These results suggest that a generalized tolerance develops to cocaine-induced GluR1S845 phosphorylation throughout the mesolimbic dopamine system with chronic cocaine exposure.
Several studies have reported increases in AMPA or NMDA receptor subunits in the VTA following repeated passive cocaine administration via intraperitoneal injections (Fitzgerald et al., 1996; Churchill et al., 1999), while others have not (Lu et al., 2001). These changes have been related to short-term increases in the excitability of VTA dopamine neurons (White et al., 1995; Zhang et al., 1997; Saal et al., 2003). Given these previous findings, the lack of GluR1 or GluR2 up-regulation with repeated passive administration by yoked cocaine injections is unclear, but could relate to the iv route of administration or the higher cocaine dose received with 4 hr of administration. For example, a single intraperitoneal cocaine injection can induce LTP in VTA dopamine neurons (Ungless et al., 2001), while yoked iv injections fail to induce these changes (Chen et al., 2008). In regards to cocaine self-administration, one study found reductions in the GluR5 and NR1 glutamate receptor subunits in the VTA of rats 15–16 hr after prolonged daily access (8 hr) in rats, but no change in GluR1 and GluR2 protein (Hemby et al., 2005). Another study found that GluR2 levels were up-regulated in the VTA of human cocaine overdose victims (Tang et al., 2003). Yet another study found that NR2A and NR2B protein levels are not altered in the VTA following brief or extended access to cocaine self-administration (Ben-Shahar et al., 2009), in agreement with a lack of NMDA receptor subunit regulation in the present study. Our results demonstrate that GluR1 and GluR2 regulation is highly dynamic during daily cocaine use, oscillating from increases after 1 day withdrawal to normalization immediately following self-administration, which could account for much of this discrepancy.
We used HSV viral vectors to study the role of reinforcement-related GluR1 up-regulation in the VTA on cocaine self-administration behavior. While dopamine neurons were not targeted genetically, the vectors were infused in a VTA region rich in dopamine neurons, resulting in a 3 to 1 ratio of dopamine to non-dopamine neuronal infection. Electrophysiological recordings from putative VTA dopamine neurons indicate that HSV-mediated GluR1 over-expression increased membrane rectification but not peak currents with AMPA receptor stimulation, consistent with membrane insertion of newly synthesized GluR2-lacking AMPA receptors that replace existing AMPA receptors, and similar to the LTP induced by a single cocaine injection (Bellone and Luscher, 2006). The enhancement of AMPA- (but not NMDA-) mediated locomotion confirmed the functional activity of GluR1 over-expression in vivo.
Transient over-expression of GluR1 in VTA neurons also markedly enhanced cocaine reinforcement by almost tripling the degree of lever-press behavior rats would engage to maintain cocaine intake at a high injection dose. These changes occurred in the absence of overt alterations in responses for a natural reward (sucrose), or in the stable regulation of cocaine intake under less demanding self-administration schedules. In addition, the effect was related specifically to GluR1 over-expression in the mesolimbic dopamine cell body region. These results suggest that dynamic up-regulation of GluR1 in the VTA produced specifically by the reinforcing context of cocaine self-administration would act reciprocally to strengthen the reinforcing efficacy of cocaine on subsequent self-administration occasions. The reinforcing efficacy of cocaine returned to control levels after dissipation of GluR1 over-expression, indicating a direct effect of GluR1 in the VTA on cocaine reinforcement. Given that endogenous GluR1 up-regulation also normalized after longer cocaine withdrawal, such transient increase in GluR1 could facilitate cocaine addiction by prolonging daily self-administration binges.
Although the PKC/CaMKII-resistant GluR1S831A did not alter cocaine reinforcement, it is surprising that the PKA-resistant mutant GluR1S845A attenuated low dose cocaine reinforcement when over-expressed in the VTA. Over-expressed GluR1S845A failed to reach dendritic processes in VTA neurons, suggesting a necessary role for GluR1S845 phosphorylation in dendritic trafficking, and complementing recent findings that enhanced GluR1S845 phosphorylation increases dendritic trafficking (Kessels et al., 2009). The lack of dendritic trafficking with the GluR1S845A mutant could promote a dominant negative effect by interacting with endogenous GluR1 and GluR2 subunits and impeding their trafficking through dendritic processes. Although similar dominant negative interactions are prevalent in the literature, further work is needed to elucidate the mechanism for the attenuation of reinforcement by GluR1S845A at lower cocaine injection doses. Nevertheless, our results indicate that PKA-mediated GluR1S845 phosphorylation is necessary for GluR1WT enhancement of cocaine reinforcement.
Vulnerability to cocaine self-administration is associated with enhanced excitability of VTA dopamine neurons in rats (Marinelli and White, 2000), and blockade of ionotropic glutamate receptors in the VTA attenuates relapse to cocaine-seeking behavior (Vorel et al., 2001; Sun et al., 2005). Furthermore, neuroadaptations leading to LTP in VTA dopamine neurons are necessary to induce subsequent neuroplasticity in target regions such as the nucleus accumbens (Mameli et al., 2009), although other behavioral responses to cocaine such as conditioned place preference and locomotor sensitization are unaltered by selective loss of GluR1 in midbrain dopamine neurons (Engblom et al., 2008). Given that repeated electrical stimulation of the lateral hypothalamus reduces both GluR1 levels in the VTA and cocaine- (but not sucrose-) seeking behavior in rats (Levy et al., 2007), the amelioration of reinforcement-related GluR1 up-regulation reported here may have therapeutic efficacy. Treatments that preclude such dynamic regulation of AMPA glutamate receptors during ongoing cocaine use could also prevent the vicious cycle of cocaine-seeking and -taking behavior that leads to cocaine addiction in humans.
This work is supported by United States Public Health Service Grant DA 18743 and the Wesley Gilliland Professorship in Biomedical Research (UTSW). Certain preliminary behavioral data was previously reported in the Ann. N.Y. Acad. Sci., 1003:372-374 (2003).