Department of Physiology, Brain Korea 21 Project for Medical Science, Brain Research Institute, Yonsei University College of Medicine, Seoul, South Korea
Address Correspondence and reprint requests to Dr. Jeong-Hoon Kim, Department of Physiology, Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752, South Korea. E-mail: firstname.lastname@example.org
Glycogen synthase kinase 3β (GSK3β), which is abundantly present in the brain, is known to contribute to psychomotor stimulant-induced locomotor behaviors. However, most studies have been focused in showing that GSK3β is able to attenuate psychomotor stimulants-induced hyperactivity by increasing its phosphorylation levels in the nucleus accumbens (NAcc). So, here we examined in the opposite direction about the effects of decreased phosphorylation of GSK3β in the NAcc core on both basal and cocaine-induced locomotor activity by a bilateral microinjection into this site of an artificially synthesized peptide, S9 (0.5 or 5.0 μg/μL), which contains sequences around N-terminal serine 9 residue of GSK3β. We found that decreased levels of GSK3β phosphorylation in the NAcc core enhance cocaine-induced hyper-locomotor activity, while leaving basal locomotor activity unchanged. This is the first demonstration, to our knowledge, that the selective decrease of GSK3β phosphorylation levels in the NAcc core may contribute positively to cocaine-induced locomotor activity, while this is not sufficient for the generation of locomotor behavior by itself without cocaine. Taken together, these findings importantly suggest that GSK3β may need other molecular targets which are co-activated (or deactivated) by psychomotor stimulants like cocaine to contribute to generation of locomotor behaviors.
It has been well known that psychomotor stimulants like cocaine produce the increase in locomotor activity by activation of dopaminergic neurotransmission in the nucleus accumbens (NAcc) (Kalivas and Stewart 1991; Hyman 1996), which is a neuronal substrate mediating the rewarding effects of drugs of abuse (Robbins et al. 1989; Koob and Le Moal 2001; Goto and Grace 2008). Interestingly, it has been suggested that, in addition to a classic cAMP-mediated pathway, dopamine has a distinct signaling pathway involving glycogen synthase kinase 3β (GSK3β), which is a serine/threonine kinase abundantly present in the brain, and accordingly it contributes to dopamine and psychomotor stimulant-induced characteristic behaviors (Beaulieu et al. 2007, 2009).
Recently, it has been shown that inhibition of GSK3β by systemic injection of valproate or a more specific inhibitor, SB216763, attenuates hyperactivity produced by acute injection of psychomotor stimulants like cocaine or amphetamine (AMPH) (Miller et al. 2009; Enman and Unterwald 2012). Further, direct microinjection of SB216763 into the NAcc core, but not in the shell, has been shown to block the expression of locomotor sensitization induced by repeated injection of cocaine or meth-AMPH (Xu et al. 2009, 2011). These results suggest that locomotor activity produced by psychomotor stimulants is dependent on GSK3β activity in the NAcc, especially in the core. However, these studies only used inhibitors for GSK3β yet, which result in the increase of phosphorylation levels at its serine 9 residue and supposedly a subsequent reduction in its activity (Dajani et al. 2001; Frame et al. 2001). As a result, it still remains unexplored what effects the supposed increase of GSK3β activity in the NAcc in the opposite way by reducing the phosphorylation levels at the same residue will actually bring to locomotor activity. Thus, we examined the effects of direct manipulation leading to reduction of GSK3β phosphorylation levels in the NAcc core on both basal and cocaine-induced locomotor activity.
Materials and methods
Male Sprague-Dawley rats weighing 200–230 g on arrival were obtained from Orient Bio Inc. (Seongnam-si, Korea). They were housed three per cage in a 12-h light/dark cycle room (lights out at 8:00 pm) and all experiments were conducted during the day time. Rats had access to water and food ad libitum at all times. All animal use procedures were conducted according to an approved Institutional Animal Care and Use Committee protocol.
Drug and peptide
Cocaine hydrochloride (Belgopia, Louvain-la-Neuve, Belgium) was dissolved in sterile 0.9% saline. S9 peptide, which consists of 21 amino acids (a.a.) including a small peptide (11 a.a.), commonly referred to as protein transduction domain (Choi et al. 2006), and a portion (10 a.a.) of the N-terminus sequence of GSK3β (GRPRTTSFAE) known as the substrate site for Akt and thereby competes with GSK3β against its phosphorylation (Dajani et al. 2001; Frame et al. 2001), was artificially synthesized and kindly provided by Professor Soo Young Lee at Center for Cell Signaling and Drug Discovery Research, Ewha Womans University (Seoul, South Korea). It was dissolved to final working concentrations of 1.0 or 10.0 μg/μL in 0.9% saline.
Brain tissue preparation for molecule analysis
Animals were decapitated at different time points (15 and 60 min) (for experiment 1) or 60 min only (for experiment 3), after saline or acute cocaine intra-peritoneal (IP) injections. For experiment 4, animals were decapitated 2 weeks after the last saline or cocaine injections. Brains were rapidly removed and coronal sections (1.0 mm thick extending 1.60–2.60 mm from bregma) were obtained with an ice-cold brain slicer. The NAcc core tissues were obtained in the circular punch with 1.0 mm diameter and the shell in the rectangular shape cut excluding the core and surrounding regions (for experiment 1 and 4) (see Fig. 1a), or the core region only (for experiment 3) on an ice-cold plate, immediately frozen on dry ice and stored at −80°C. They were prepared bilaterally and pooled for each individual animal's protein isolation.
Tissues were homogenized in lysis buffer containing 0.32 M sucrose, 2 mM EDTA, 1% sodium dodecyl sulfate, 10 μg/mL aprotinin, 10 μg/mL leupeptin, 1 mM phenylmethylsulfonyl fluoride, 10 mM sodium fluoride, and 1 mM sodium orthovanadate. The concentration of protein was determined by using Pierce BCA protein assay kit (Pierce, Rockford, IL, USA). Samples were then boiled for 10 min and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Proteins were transferred electrophoretically to nitrocellulose membranes (Bio-Rad, Hercules, CA, USA), which were then blocked with 5% skim milk in PBS-T buffer (10 mM phosphate-buffered saline plus 0.05% Tween-20). Specific antibodies against total GSK3β (1 : 10 000 dilution in PBS-T with 5% skim milk; Cell Signaling, Beverly, MA, USA), phospho-GSK3β (specific to detect phosphorylated GSK3 β at serine 9; 1 : 1000 dilution in PBS-T with 5% BSA; Cell Signaling) and anti-β-actin (1 : 10 000 dilution in PBS-T with 5% skim milk; Abcam, Cambridge, UK) were used to probe the blots. Primary antibodies were detected with peroxidase-conjugated secondary antibodies, anti-rabbit IgG (1 : 2000 dilution in PBS-T with 5% skim milk; KOMA Biotech, Seoul, Korea) or anti-mouse IgG (1 : 5000 dilution in PBS-T with 5% skim milk; Cell Signaling) followed by enhanced chemiluminescence (ECL) reagents (Amersham Biosciences, Arlington Heights, IL, USA) and exposure to X-ray film. Band intensities were quantified based on densitometric values using Fujifilm Science Lab 97 Image Gauge software (version 2.54) (Fujifilm, Tokyo, Japan).
Rats were anesthetized with intraperitoneal (IP) ketamine (100 mg/kg) and xylazine (6 mg/kg), placed in a stereotaxic instrument with the incisor bar at 5.0 mm above the interaural line and implanted with chronic bilateral guide cannulas (22 gauge tubing size with 0.71 mm in outer and 0.39 mm in internal diameters; Plastics One, Roanoke, VA, USA) aimed at the NAcc core (A/P, +3.4; L, ±1.5; D/V, −7.5 mm from bregma and skull) (Pellegrino et al. 1979). Cannulas were angled at 10° to the vertical, positioned 1 mm above the final injection site, and secured with dental acrylic cement anchored to stainless steel screws fixed to the skull. After surgery, 28 gauge (tubing size with 0.36 mm in outer and 0.18 mm in internal diameters) obturators were placed in the guide cannulas, and rats were returned to their home cages for a 7-day recovery period.
Bilateral intracranial microinjections into the NAcc core were made in the freely moving rat. Injection cannulas (28 gauge) connected to 1 μL syringes (Hamilton, Reno, NV, USA) via PE-20 tubing were inserted to a depth 1 mm below the guide cannula tips. Injections were made in a volume of 0.5 μL per side over 30 s. After 1 min, the injection cannulas were withdrawn and the obturators were replaced.
Locomotor activity was measured with a bank of six activity boxes (35 × 25 × 40 cm) (IWOO Scientific Corporation, Seoul, Korea) made of translucent Plexiglas. Each box was individually housed in a PVC plastic sound attenuating cubicle. The floor of each box consisted of 21 stainless steel rods (5 mm diameter) spaced 1.2 cm apart center-to-center. Two infrared light photo beams (Med Associates, St. Albans, VT, USA) positioned 4.5 cm above the floor and spaced evenly along the longitudinal axis of the box estimated horizontal locomotion. It was counted as a single locomotor activity occurred only when rats interrupted two beams consecutively, avoiding any possible confounding measures like rearing or grooming in a spot covering just a single beam.
Design and procedure
One week after arrival, rats were administered once with either saline or cocaine (15 mg/kg, IP) at two different time points (15 and 60 min). Then, the brain tissues (both core and shell) were punched out and prepared for western blot analysis.
Once they were recovered from surgery, rats were randomly assigned to six groups and allowed to stay in the locomotor activity boxes for 60 min to adjust to the new environment. Then, they were bilaterally microinjected into the NAcc core with either saline or S9 (0.5 or 5.0 μg/0.5 μL/side) immediately followed by a single IP injection of either saline or cocaine (15 mg/kg), and their locomotor activity was measured for 60 min.
About 1 week after cannular instalment surgery, six different groups of rats were bilaterally microinjected into the NAcc core with either saline or S9 peptide (0.5 or 5.0 μg/0.5 μL/side) immediately followed by a single IP injection of saline or cocaine (15 mg/kg). They were back to cage and placed in a quiet place for 60 min, then, decapitated and the brain tissues (core only) were prepared for western blot analysis.
Rats were administered repeatedly with either saline or cocaine (15 mg/kg, IP), once a day for seven consecutive days. Two weeks after drug-free withdrawal period, they were decapitated and the brain tissues (both core and shell) were prepared for western blot analysis.
After completion of the Experiment 2, rats were anesthetized and perfused via intra-cardiac infusion of saline and 10% formalin. Brains were removed and further post-fixed in 10% formalin. Coronal sections (40 μm) were subsequently stained with cresyl violet for verification of cannula tip placements. Only rats with injection cannula tips located bilaterally in the NAcc core were included in the data analyses. Of the 49 rats surgically prepared for testing, seven (two for saline and five for S9 microinjections) were dropped for failing to meet this criterion. Any evident neurotoxicity other than mechanical damage resulting from the cannula implantation was also examined under the microscopy with cresyl violet stained sections.
For Experiment 3, rats were decapitated and brain slices on an ice-cold plate were observed to locate cannula tracks. Only rats with injection cannula tips located bilaterally in the NAcc core were included in the data analyses. Of the 35 rats surgically prepared for testing, five (one for saline and four for S9 microinjections) were dropped for failing to meet this criterion.
The data were analyzed with either t-test or two-way anova followed by post hoc Bonferroni comparisons. Differences between experimental conditions were considered statistically significant when p <0.05.
Cocaine decreases GSK3β phosphorylation levels in the NAcc core, but not in the shell
The expression levels of GSK3β in the NAcc core and shell were examined in the tissues obtained at two different time points (15 and 60 min) after either saline or cocaine (15 mg/kg, IP) injections (Fig. 1). The two-way anova with time and drug as two different factors conducted on the ratio of phosphorylated to total GSK3β levels in the core revealed a significant effect of drug (F1,17 = 6.84, p <0.05). Post hoc Bonferroni comparisons conducted on these data revealed that the ratio of the phosphorylated to total GSK3β in the NAcc core was significantly decreased (p <0.05) in cocaine compared to saline injected rats at 60 min time points measured. These effects, however, were not present in the shell (Fig. 1).
Inhibition of GSK3β inactivation in the NAcc core enhances cocaine-induced hyper-locomotor activity, while leaving basal activity unchanged
GSK3β is supposedly inactivated when its phosphorylation levels at serine 9 residue are increased (Dajani et al. 2001; Frame et al. 2001). To examine what effects inhibition of such GSK3β inactivation in the ΝΑcc core might produce on locomotor activity, we made a direct microinjection into this site of S9 peptide, which is a synthetic peptide consisting of 21 amino acids that include serine 9 residue of the N-terminus sequences of GSK3β known as the substrate site for Akt (also known as protein kinase B). As a result of its sequence homology, S9 peptide supposedly competes with endogenous GSK3β against its phosphorylation by Akt, and thereby makes GSK3β less phosphorylated and supposedly consequent inhibition of its inactivation.
Rats were first habituated to the activity boxes for 1 h and both basal and cocaine-induced locomotor activities were sequentially measured for another 1 h either with or without S9 microinjection. The two-way anova with microinjection and IP as two different factors conducted on the 1 h total locomotor activity counts revealed multiple significant effects of microinjection (F2,36 = 4.66, p <0.02), IP (F1,36 = 100.59, p <0.001), and microinjection X IP interactions (F2,36 = 4.48, p <0.02). As expected, cocaine produced higher locomotor activity than saline when no S9 was present in the NAcc core (p <0.001 by post hoc Bonferroni comparisons). Further, interestingly, microinjection of S9 into this site dose-dependently enhanced these effects only in cocaine IP rats (p <0.001 by post hoc Bonferroni comparisons for higher dose of S9 compared to saline) (Fig. 2a). However, S9 alone without cocaine produced no significant effects on basal locomotor activity. Time-course analyses of these findings showed that the ability of S9 in the NAcc core to enhance the increase in locomotor activity observed in cocaine administered rats was apparent throughout the 1 h time course measured (Fig. 2b). The location for the injection cannula tips used in these experiments is depicted in Fig. 2c.
Microinjection of S9 in the NAcc core decreases GSK3β phosphorylation levels
To confirm whether S9 in the NAcc core actually reduced the phosphorylation levels of GSK3β in this site, we made a bilateral microinjection of S9 to a separate group of rats and measured the ratio of phosphorylation to total levels of GSK3β in the NAcc core. The two-way anova with microinjection and IP as two different factors conducted on the ratio of phosphorylated to total GSK3β levels in this site revealed multiple significant effects of microinjection (F2,24 = 10.05, p <0.001) and IP (F1,24 = 22.86, p <0.001). As expected, we found that S9 alone without cocaine dose-dependently decreased the ratio of phosphorylated to total GSK3β levels in the NAcc core (p <0.05 by post hoc Bonferroni comparisons for higher dose of S9 compared to saline) (Fig. 3), while interestingly we also found that cocaine in the presence of S9 further decreased this ratio of GSK3β levels in this site (p <0.05 by post hoc Bonferroni comparisons for a higher dose of S9 with compared to without cocaine).
Chronic cocaine decreases basal levels of GSK3β phosphorylation in the NAcc core, but not in the shell
As our data shows that the decrease of GSK3β phosphorylation levels in the NAcc core enhances cocaine-induced hyper-locomotor activity similar to sensitized locomotor activity usually obtained by chronic cocaine (Fig. 2), we examined whether GSK3β phosphorylation levels in the NAcc are different in chronic cocaine compared from matched saline animals. When measured at 2 weeks of drug-free withdrawal period after 7 days of daily cocaine injections, t-test revealed a significant effect of chronic cocaine [t(5) = 2.62, p <0.05] on the expression levels of GSK3β in the NAcc core, but not in the shell (Fig. 4).
The present results revealed that inhibition of GSK3β inactivation in the NAcc core by temporarily reducing its phosphorylation levels at serine 9 residue enhances cocaine-induced hyper-locomotor activity, while leaving basal activity unchanged. This is the first direct demonstration, to our knowledge, that the selective decrease of GSK3β phosphorylation levels in the NAcc core may actually contribute positively to psychomotor stimulant-induced locomotor activity.
In the past few years, people have shown that GSK3β activity in some brain areas is regulated by psychomotor stimulants treatment. For example, phosphorylation levels of GSK3β are reduced not only in the dorsal striatum by acute administration of AMPH or cocaine (Beaulieu et al. 2007; Miller et al. 2009) and by chronic AMPH (Mines and Jope 2012) but also in the NAcc core by chronic cocaine and meth-AMPH (Xu et al. 2009, 2011). Consistent with these results, we also observed acute cocaine decreases the ratio of phosphorylation to total levels of GSK3β distinctively in the NAcc core, but not in the shell (Fig. 1). Further, we found that these effects are more significantly expressed at later time point (60 min) than earlier (15 min), which may reflect a delayed kinetic property with a peak around between 60 and 90 min that appears in the cAMP-independent dopamine pathway through an Akt-GSK3β signaling cascade (Beaulieu et al. 2007).
Literature has affluently shown that the core and the shell are distinguished in both anatomical structures and behavioral output functions (Jongen-Relo et al. 1994; Di Chiara 2002; Zahm 2002; Meredith et al. 2008); for example, the core mediates the motivational impact of Pavlovian conditioned stimuli, while the shell does the impact of primary reinforcers (Cardinal et al. 2002). Further, psychostimulant sensitization has shown not only to increase c-fos mRNA expression or dopamine transmission but also to significantly induce structural plasticity in the core, but not in the shell (Cadoni et al. 2000; Li et al. 2004; Nordquist et al. 2008). These results suggest that our present findings, in which GSK3β differentially responds to acute cocaine in the core, may also reflect intrinsic differences in signaling pathways set wired in the core in terms of mediating psychostimulant-induced behaviors. Interestingly, it has recently been shown that specific inhibition of GSK3β activity in the NAcc core actually blocks cocaine- or meth-AMPH-induced locomotor sensitization (Xu et al. 2009, 2011), indicating that GSK3β in the core is more specifically involved in psychostimulant-induced behaviors.
In the literature, it has been consistently shown that GSK3β inhibitors such as lithium chloride or valproic acid increase phosphorylation levels for GSK3β at serine 9 residue and accordingly attenuate psychomotor stimulants-induced increase in locomotor activity (Beaulieu et al. 2008; Xu et al. 2009, 2011; Enman and Unterwald 2012). However, these studies have not investigated in the opposite way yet about the possible effects of direct inhibition of GSK3β inactivation in the NAcc on locomotor activity. Although it has previously been shown that either transgenic mice over-expressing GSK3β or more selectively GSK3β knockin mice enhance locomotor activity in response to either a novel environment or psychomotor stimulants like AMPH or methylphenidate (Prickaerts et al. 2006; Polter et al. 2010; Mines et al. 2013), these studies not only manipulated GSK3β gene affecting the whole brain but also permanently throughout animal's life-span. Thus, we have found a new way for a local and temporal inhibition of GSK3β inactivation and used in this study an artificially synthetic S9 peptide, which competes with endogenous GSK3β for the N-terminal phosphorylation site at serine 9 residue, resulting in less phosphorylation and consequent inhibition of GSK3β inactivation (Dajani et al. 2001; Frame et al. 2001). As a consequence, a bilateral microinjection into the NAcc core of this artificial peptide interestingly enhances cocaine-induced hyper-locomotor activity in a dose-dependent manner (Fig. 2) with an accompanied decrease of GSK3β phosphorylation levels (Fig. 3), which is exactly opposite result to the ones obtained when GSK3β inhibitors were used, but simultaneously consistent with the ones obtained from transgenic or knockin mice as shown in literature. Thus, our findings favorably support the notion that GSK3β activity in the NAcc core is importantly involved in cocaine-induced locomotor activity. Further, our present results also show that the decrease of GSK3β phosphorylation alone in the NAcc core without the help of cocaine is not sufficient by itself for the generation of locomotor behavior (see Figs 2 and 3). Similarly, GSK3β knockin mice data showed that a constitutive GSK3β activation enhanced locomotor activity only when a certain type of stimulus (e.g., novel environment or psychomotor stimulants) was present (Polter et al. 2010; Mines et al. 2013). In our case, we habituated rats before locomotor measurement, so that rats with compared from the ones without S9 showed no different locomotor activity unless there was cocaine present. These results importantly suggest that activation of GSK3β may become somehow workable only when other molecular targets are co-activated (or deactivated) by psychomotor stimulants like cocaine or even a novel environment (as in knockin data). It remains in the future to find out what they might be and how, though.
Rather surprisingly, our present results indicate that the ratio of phosphorylated to total GSK3β in the NAcc core is not directly proportional to the amount of locomotor activity produced; for example, rats injected with cocaine in the presence of low dose of S9 did not show less phosphorylation levels compared with rats with cocaine alone, although locomotor activity in the former was higher than the latter (see Fig. 3 and compare with Fig. 2). These results may suggest that the prior decrease of GSK3β phosphorylation by S9 in the NAcc core may just help set the stage for following cocaine to more easily produce enhanced hyper-locomotor activity similar to sensitized locomotion normally observed in rats with chronic cocaine treatment (Robinson and Berridge 1993). To confirm this hypothesis, we measured GSK3β phosphorylation levels in the NAcc at 2 weeks of drug-free withdrawal period after 7 days of daily cocaine injections and found that basal levels of GSK3β phosphorylation in the NAcc core, but not in the shell, are significantly reduced in rats with cocaine compared with saline pre-treated (Fig. 4). These results imply that reduced basal levels of GSK3β phosphorylation in the NAcc core may be related with the production of behavioral sensitization by chronic cocaine treatment, and in our case, although it is speculative yet, temporal reduction of GSK3β phosphorylation in the NAcc core by S9 may have pulled forward neuronal state to help set the stage similar to the one obtained by chronic cocaine, so that seemingly sensitized locomotor activity is produced when cocaine comes in. Taken together, our present findings indicate that GSK3β phosphorylation levels in the NAcc core may be an important regulator of cocaine-induced locomotor activity and suggest that it is worth closer examination at this molecule in terms of locomotor sensitization in the future.
This study was supported by the Korea Healthcare Technology R&D Project funded by the Ministry for Health, Welfare & Family Affairs (A084692) (J.-H. K.). The authors have no conflict of interest to declare.