INTRODUCTION

Despite declines over the past 25 years, more than a fifth of the US population continues to smoke, and an increasing number of tobacco users are intermittent smokers in social settings (‘chippers’). It is likely that environmental cues contribute to social and habitual tobacco use. Nicotine, the primary addictive substance in tobacco smoke, stimulates neuronal activity in the ventral tegmental area and downstream dopamine (DA) projection regions that support addictive behavior (Due et al, 2002; Brody et al, 2004). Mesolimbic DA release is believed to underlie the primary reinforcing effects of drugs of abuse including nicotine (Roberts et al, 1980; Di Chiara and Imperato, 1988; Corrigall et al, 1992; Picciotto et al, 1998). In particular, DA release in the nucleus accumbens (NAc) shell is thought to regulate motivation for drug and natural rewards (Di Chiara et al, 2004). Rodent and nonhuman primate studies show that after repeated pairings of a primary reward with a conditioned stimulus (CS) cue, the CS functionally replaces the primary reinforcer in regulation of DA neuron firing (Schultz et al, 1997; Day et al, 2007). In addicted individuals, the rewarding properties of cues associated with a primary reinforcer such as nicotine are thought to have a more predominant function than primary reinforcement in compulsive drug use (Robinson and Berridge, 1993). Associative learning between drug and conditioned rewards involves the amygdala, cingulate cortex, hippocampus and NAc core and shell (Robbins and Everitt, 2002; Robinson and Berridge, 2003), DA projection areas activated by nicotine and cigarette-associated cues in smokers (Due et al, 2002; Brody et al, 2004) and the striatum, including the NAc, appears to regulate cue-induced craving (David et al, 2007; Franklin et al, 2007; McClernon et al, 2008; Volkow et al, 2008). Rodent studies show that a nicotine-associated context stimulates immediate early gene activation in several mesolimblic DA projection areas (Schroeder et al, 2001; Schiltz et al, 2005), suggesting that neuroplasticity associated with nicotine addiction occurs at a molecular level in the DA system.

A candidate gene regulated in the NAc by in vivo nicotine effects on DA is the transcription factor cyclic AMP-response element binding protein (CREB) (Pandey et al, 2001; Brunzell et al, 2003; Pluzarev and Pandey, 2004; Walters et al, 2005). D1- and D2-type DA receptors are distinguished according to their respective stimulatory and inhibitory effects on adenylyl cyclase and subsequent activation of CREB by phosphorylation of the protein at Ser-133 (Gomperts et al, 2002). In a nicotine place conditioning paradigm where animals are given a choice between two chambers previously paired with either nicotine or saline, mice show elevated expression of the phosphorylated form of CREB (pCREB) in the NAc that corresponds with increases in preference for the nicotine-paired chamber (Walters et al, 2005). This nicotine conditioned place preference (CPP) is also associated with elevated levels of pCREB in the hippocampus (Walters et al, 2005), a brain area that is important for context conditioning (Kim and Fanselow, 1992; Phillips and LeDoux, 1992) and sensitive to nicotine exposure (Levin et al, 2006; Kenney and Gould, 2008). It is not clear, however, whether phosphorylation of CREB precipitates nicotine chamber choice during CPP or whether the animals’ nicotine-seeking behavior results in phosphorylation of CREB. Changes in signaling that precede CPP may be involved in behavioral responses to cue reward. It is interesting that chronic nicotine exposure results in decreases in pCREB in the NAc (Brunzell et al, 2003) and that the rewarding effects of elicit drugs of abuse, such as morphine and cocaine, are enhanced by disruption of CREB activation in the NAc shell (Carlezon et al, 1998; Barrot et al, 2002). Constitutive, global knockout of CREB α/δ isoforms prevents nicotine CPP (Walters et al, 2005), suggesting that CREB activation is critical for this reward-associated behavior; however the function of NAc CREB activation in nicotine CPP remains to be determined.

To address the function of CREB in nicotine place conditioning, we first tested whether nicotine or a nicotine-paired chamber would regulate total CREB or pCREB levels in the hippocampus and NAc of animals isolated to the training chamber during training or 24 h following 3 days of nicotine context conditioning in the absence of nicotine. To expand on previous research (Brunzell et al, 2003; Pluzarev and Pandey, 2004; Walters et al, 2005), we subdivided the NAc into core and shell regions. A nicotine-paired chamber increased levels of pCREB in the NAc shell but not the core or the hippocampus. We hypothesized that changes in CREB signaling that occur while the animal is isolated in a nicotine-paired chamber are CS dependent and necessary for nicotine CPP. To test this hypothesis, we used a herpes simplex viral vector to express a dominant-negative form of CREB with a mutation at the Ser-133 phosphorylation site (mCREB) (Carlezon et al, 1998) or control vector (LacZ) into the NAc shell. Mice showed normal nicotine CPP when infused with LacZ control vector, but shell infusions of mCREB blocked nicotine CPP at a range of doses. Together these data suggest that context-associated changes in CREB activation in the NAc shell are necessary for nicotine CPP.

MATERIALS AND METHODS

Animals

Male C57BL/6J (8- to 12-week old) mice were purchased from Jackson Laboratories or bred in our laboratory. Mice were maintained in a temperature-controlled vivarium (21±2°C) under a 12 : 12 h light–dark cycle and housed four per cage. Food and fluid were available ad libitum. Mice were habituated to handling for at least 3 days before behavioral testing or biochemical studies. All studies were approved by the Yale University Animal Care and Use Committee and followed the NIH Guide for the Care and Use of Laboratory Animals.

Drug Dosing and Administration

Mice undergoing place conditioning received i.p. injections of nicotine in 0.9% saline (0.04, 0.065, 0.09, 0.175, 1.75 mg/kg (by weight of freebase)) or vehicle immediately before training.

CREB and pCREB Induction Studies During Pavlovian Context Conditioning

For studies of CREB signaling under Pavlovian conditions, separate groups of mice received nicotine (NIC; 0.09 mg/kg, i.p.) or saline injection (SAL) once daily for 3 days immediately before placement in a (48 × 22 × 18 cm) training chamber for 30 min. Brains of all SAL animals and a portion of NIC mice were harvested 15 min into the third training session. Approximately 22 h following the third NIC training session, the remaining mice (NIC Chamber) had brains harvested 15 min following saline injection and placement into the nicotine-paired chamber. A Med Associates locomotor apparatus measured beam breaks to calculate distance traveled during exposure to the training chambers for subgroups of animals in each of the three conditions.

Western Blot Analysis

Mouse brains were harvested by rapid decapitation, placed in ice-cold PBS, sectioned in a chilled matrix, and placed in fresh, chilled PBS. The dorsal hippocampus was dissected and 16-gauge punches of NAc core and shell were harvested from 1 mm brain sections and fast-frozen in dry ice. Tissues and blots were processed as described previously (Brunzell et al, 2003). Protein (10 μg) was loaded on a polyacrylamide gradient gel (4–20%) and transferred to nitrocellulose for immunodetection using polyclonal antisera (Cell Signaling, USA) to CREB (1 : 1000) and pCREB (Ser-133; 1 : 500). Blots were washed and then incubated in 1 : 1000 peroxidase-labeled anti-rabbit secondary antiserum (Vector, USA). For NAc core and shell, ECL detection reagents were applied for 1 min, blots were exposed to X-Ray film and developed. Equal loading was assured by Ponceau staining, and blots were subsequently incubated in 1 : 10 000 GAPDH monoclonal antiserum (Upstate Biotechnology) and 1 : 5000 anti-mouse secondary antiserum as a loading control. In the hippocampus, monoclonal pCREB (1 : 500 in mouse) and CREB (1 : 1000 in rabbit) (Cell Signaling) antisera were incubated on the same blot. Fluorescence-tagged anti-mouse and -rabbit secondary antibodies of different wavelengths were used (Licor) and were scanned using the Odyssey scanning system and software.

Unbiased Place Conditioning

Two conditioning chambers with retractable doors were separated by a neutral chamber. The walls and the floors of the chambers varied but lighting was equilibrated so that there was no baseline preference for either chamber. Time spent in each chamber was recorded by beam breaks and calculated using Med-PC IV software.

On day 0, baseline, mice received a vehicle injection, were placed in the neutral chamber, and allowed to explore the apparatus for 15 min. Mice with a chamber bias greater than 75% were dropped from studies. On days 1–3, mice were placed in one conditioning chamber for 30 min following vehicle injection in the morning, and in the afternoon were placed in the opposite conditioning chamber following an injection of drug; vehicle-injected control subjects received saline injections before placement in either chamber. Mice were counterbalanced for drug-paired chamber according to treatment dose, genotype, and baseline preference. On day 4, testing, mice were injected with vehicle, placed in the center compartment, and allowed access to both conditioning chambers for 15 min. Baseline and testing sessions took place at an intermediate time between the AM and PM training sessions. Total time spent in each chamber was measured and changes from baseline preference were calculated.

HSV-Mediated Gene Transfer

HSV-mCREB or HSV-LacZ was infused into the NAc shell 1 day after measurement of baseline preference and 2 days before the initiation of place conditioning. Mice were anesthetized with a mixture of ketamine (130 mg/kg, i.p.) and xylazine (13 mg/kg). Bilateral infusion canulae were 1.5 mm apart and offset by 0.5 mm in length to target the NAc shell when inserted unilaterally at a 20° angle, sparing the overlying cingulate cortex and contralateral structures. Stereotaxic placement of a 10 mm guide canula (o.d./i.d. 0.6/0.45 mm) into the NAc shell minimized backflow into overlying structures (1.5 mm A/P, 2.5 mm L, and −4.0 mm D/V with respect to bregma). Anatomical control animals received an optimal dose of nicotine following a more dorsal infusion of HSV-mCREB (1.5 mm A/P, 2.25 mm L, and −3.25 mm D/V from dura) to further insure that behavioral effects of mCREB were due to infusion into the NAc shell and not into an overlying structure. HSV vector (0.7 μl over 14 min) was infused by microinfusion pump. Canulae were kept in place for 5 min after infusion. Following surgery, mice were housed four per cage. A nonsteroidal anti-inflammatory drug (5 mg/kg, Rimadyl) was used for preemptive analgesia in some mice; all mice received 64 mg of children's acetaminophen applied to wet chow ad libitum during recovery. Most mice recovered from surgery within 48 h without signs of discomfort, infection, or bleeding. Following 2 days recovery, mice underwent CPP training on days when HSV-mediated mCREB expression was at its peak (days 2–4 post-infusion), and underwent subsequent testing on day 5 (Carlezon et al, 1998; Barrot et al, 2002). Following behavioral procedures, mice received lethal doses of chloral hydrate, were perfused intracardially with chilled 4% paraformaldehyde (PFA), and brains were harvested and sectioned for verification of canula placement. A separate group of HSV-LacZ mice brains were harvested 3 days post-infusion for the X-gal enzymatic assay.

X-Gal Immunohistochemistry

Mice were perfused transcardially with ice-cold PBS followed by 4% PFA in 0.1 mol/l phosphate buffer. Brains were removed, postfixed in 4% PFA, and subsequently placed in 30% sucrose in PBS for dehydration. Sections (40 μm) were collected on a freezing microtome and stored in cryoprotectant at −20°C. For X-gal histochemistry, sections were washed 3 × in PBS, incubated for 4 h at 37°C in PBS solution containing 3.1 mmol/l potassium ferrocyanide, 1 mmol/l MgCl, 0.01% sodium deoxycholate, 0.02% Nonidet P-40, and 0.2 mg/ml X-gal. β-Galactosidase enzymatic activity was terminated with 5 min washes in PBS. Sections were mounted, stained with neutral red, dehydrated with increasing concentrations of ethanol, and coverslipped with Permount.

Statistical Analyses

For place conditioning, significant effects of treatment and genotype were determined analyzing conditioning chamber as a within-subject measure (drug paired vs saline paired), with between-subjects measures for viral-mediated transfection (HSV-mCREB or HSV-LacZ) and drug (eg NIC/VEH). Post hoc two-tailed t-tests were performed for all statistics except where previous studies predicted a directional finding to justify use of one-tailed t-tests. Western blot optical densities were normalized to Vehicle controls to enable comparisons across blots. Planned comparisons were made between vehicle-injected (SAL) and nicotine-injected (NIC) animals and between SAL subjects and animals exposed to a nicotine-paired chamber in the absence of nicotine (NIC Chamber). Just before brain harvest, locomotor activity was assessed between treatment groups (SAL, NIC, and NIC Chamber) using time bin as a repeated measure.

RESULTS

A Nicotine-Paired Context Leads to Increases in pCREB in the NAc Shell in the Absence of Changes in Locomotor Activity

Exposure to a nicotine-paired context was sufficient to activate NAc CREB activity in the absence of overt nicotine-seeking (Figure 1a). In comparison to saline-injected control animals that received no nicotine exposure, saline-injected animals exposed to a chamber where they had previously received nicotine showed a significant increase in levels of pCREB (p=0.05) in the NAc shell. Slight elevations in NAc shell pCREB levels during nicotine Pavlovian training failed to reach significance (p>0.1), suggesting that the CS properties of the context rather than the primary rewarding properties of the nicotine regulate this effect following 3 days of training. Neither nicotine nor a nicotine-associated context resulted in changes in pCREB levels in the hippocampus or NAc core (Figure 1b and c) and there was no effect of treatment condition on total CREB levels in the NAc shell, core, or hippocampus (p's>0.1).

Figure 1
figure 1

Nicotine context conditioning increases phosphorylated cyclic AMP-response element binding protein (pCREB) levels in the nucleus accumbens (NAc) shell but not the NAc core or dorsal hippocampus. (a) In comparison to animals injected with saline in the training chamber (SAL, S; n=12), mice that received nicotine (0.09 mg/kg, i.p. by weight of freebase) showed no effect of treatment on levels of pCREB during the third day of Pavlovian training (NIC, N; n=13), but did show NAc shell-specific elevations in pCREB following exposure to the chamber previously paired with nicotine (NIC Chamber, NC; n=4; p<0.05). Total CREB levels in the NAc shell were not regulated by NIC or exposure to the NIC-paired Chamber. (b, c) There was no effect of NIC, or NIC Chamber exposure on either NAc core or hippocampal levels of CREB or pCREB (p's>0.1). (d) There was no difference in locomotor activity across groups of SAL, NIC, and NIC Chamber mice during the 15 min post-injection immediately before brain harvest (p>0.1). Activity data are presented as distance traveled in cm across 5 min time bins. All brains were harvested 15 min following injection and placement into the chamber. Data from western blotting are presented as percent of SAL control data. Error bars represent SEM; *p⩽0.05 compared to SAL animals.

Full size image

Figure 1d shows nicotine-associated locomotor activity on the third day of training and during exposure to a nicotine-paired chamber in the absence of nicotine. There was no difference in locomotor activity across experimental conditions on the day of brain harvest (F2,11=0.448, p=0.65), and no interaction of Pavlovian condition with time bin on distance traveled (F2,11=1.662, p=0.23). These data further suggest that the nicotine-associated context and not the overt behavior of the animal led to the observed induction of pCREB in the NAc shell.

Unbiased Nicotine CPP is Dose Dependent

Nicotine CPP in C57BL/6J mice was tightly regulated by dose (F3,39=2.87, p=0.05). As previously reported in rats (Le Foll and Goldberg, 2005) and mice (McGeehan and Olive, 2003), C57BL/6J mice showed nicotine CPP at the 0.09 mg/kg, i.p. dose of nicotine as measured by a preference for the drug-paired chamber in comparison to saline controls (p=0.05; Figure 2a). The highest dose of nicotine (0.175 mg/kg) did not result in a preference for the nicotine-paired chamber, presumably due to potential aversive effects of nicotine. Unlike s.c. administration under similar conditions (Walters et al, 2006), however, i.p. nicotine did not result in conditioned place aversion to the nicotine-paired chamber even at a dose of nicotine one log higher than the highest dose (1.75 mg/kg, i.p.; data not shown), perhaps due to faster kinetics of the i.p. route of administration.

Figure 2
figure 2

Nonbiased nicotine conditioned place preference (CPP) is tightly regulated by dose. (a) C57BL/6J mice showed a trend for nicotine CPP at the 0.04 mg/kg dose (n=12), a significant increase in preference for the nicotine-paired chamber at the 0.09 mg/kg dose (n=8; p=0.05), and no preference at the highest dose of i.p. nicotine, 0.175 mg/kg (n=8) as compared to saline-injected animals (n=9). All nicotine doses are calculated by weight of the freebase. Data are presented in seconds as mean percent change from baseline preference; error bars reflect SEM; *p<0.05 compared to vehicle control animals. (b) There was no difference in time spent in the conditioning chamber at baseline indicating that animals showed no overall bias for either chamber before training (p's>0.1), suggesting that change in preference was due to nicotine reward rather than anxiolytic effects of nicotine treatment.

Full size image

Decreases in CREB Activity in the NAc Shell Oppose Nicotine Reward Learning

We used a modified herpes simplex viral vector (HSV) to express mCREB, a dominant-negative form of CREB with a mutation of Ser-133 to alanine, in the NAc shell to test directly whether place-conditioning-associated changes in CREB activity are necessary for nicotine CPP. The serine to alanine mutation in this HSV-mCREB construct reduces cyclic AMP-response element (CRE)-mediated transcription in CRE-LacZ reporter mice (Barrot et al, 2002); hence, mCREB expression disrupts NAc CREB activity and will oppose increases in NAc CREB activity that may occur during place conditioning and following nicotine CPP (Supplementary Figure S1). There was a significant interaction of HSV infusion with dose of nicotine (F3,148=2.836, p=0.04) on change in chamber preference. Mice infused with HSV-LacZ control vector, like untreated animals, showed dose-dependent nicotine CPP (Figure 3a), but pretraining NAc shell infusion of HSV-mCREB blocked preference for the nicotine-paired chamber across a range of doses (Figure 3a and b). HSV-mCREB infusion into the NAc shell also resulted in a significant preference for the saline chamber at the optimal dose for nicotine CPP in untreated mice (0.09 mg/kg; p=0.036), which resulted from a shift in time spent in the center compartment and did not reflect a significant decrease in preference for the drug-paired chamber. A subgroup of animals tested for locomotor activity on the day of testing showed no effect of HSV infusion or prior nicotine treatment on locomotor activity (Supplementary Figure S2). Reconstructions of infusion canula placement and an HSV-LacZ X-gal enzymatic activity assay demonstrate that HSV vector expression was local to the site of infusion (Figure 4). Control animals that received infusions 0.75 mm dorsal to the NAc shell showed similar nicotine CPP to HSV-LaZ-infused animals following administration of 0.09 mg/kg, i.p. nicotine, as measured by an increase in preference for the drug-paired chamber over the saline-paired chamber (p=0.05). Preference for the drug-paired chamber was significantly greater in saline-injected controls (t54=1.675, p=0.05), suggesting that blockade of nicotine CPP in these experiments chiefly involved inhibition of CREB in the NAc shell and not some overlying structure (Figure 5).

Figure 3
figure 3

Blockade of cyclic AMP-response element binding protein (CREB) activation in nucleus accumbens (Nac) shell with HSV-mCREB abolishes nicotine conditioned place preference (CPP) across a range of doses. (a) HSV-LacZ-treated mice showed a preference for the nicotine-paired chamber at the 0.065 (n=18) and 0.09 (n=20) mg/kg, i.p. doses of nicotine but not at the 0.04 mg/kg dose (n=22) as compared to saline-treated controls (n=18). (b) HSV-mCREB infusion into the NAc shell blocked nicotine CPP at all i.p. nicotine doses as demonstrated by the lack of preference for the drug-paired chamber (All p's>0.2; n=15–23). HSV-mCREB-infused mice showed preference for the saline-paired chamber at the 0.09 mg/kg, i.p. dose of nicotine. (c, d) Mice showed no preference for the drug- or saline-paired chamber at baseline. Data are presented as changes from baseline preference ±SEM; *p⩽0.05 compared to vehicle control mice.

Full size image
Figure 4
figure 4

HSV infusions into the nucleus accumbens (NAc) shell. (a) Reconstruction of canulae placement for HSV-mCREB- (closed circles) and HSV-LacZ-infused animals (open circles). (b) A representative mouse NAc section harvested 3 days post-HSV-LacZ infusion. X-gal enzymatic reaction demonstrates that HSV infusions were local to the NAc shell infusion site. The use of guide canulae helped to insure that viral vector that may have traveled up the infusion canulae did not infect overlying tissues. Coronal section of striatum used with permission: Rosen GD, Williams AG, Capra JA, Connolly MT, Cruz B, Lu L et al. (2000). The Mouse Brain Library@ www.mbl.org. Int Mouse Genome Conference 14: 166.

Full size image
Figure 5
figure 5

HSV-mCREB infusions into structures above the nucleus accumbens (NAc) shell did not disrupt nicotine conditioned place preference. (a) Surgery control animals (n=18) showed a conditioned place preference to 0.09 mg/kg, i.p. nicotine as measured by a significant increase in preference for the nicotine-paired chamber compared to HSV saline-injected animals (n=38). (b) Mice showed no preference for either chamber at baseline. (c) Reconstructions of canulae placement for HSV-mCREB (closed circles) in animals that received infusions 0.75 mm dorsal to the NAc shell target structure. Data are presented as mean changes from baseline preference and mean time spent in either chamber at baseline ±SEM; *p<0.05 compared to saline-injected controls.

Full size image

DISCUSSION

These studies demonstrate that activation of CREB in the NAc shell is critical for nicotine CPP, a form of appetitive contextual conditioning. We further show that a nicotine-paired context increases levels of pCREB in the NAc shell in the absence of overt nicotine-seeking or locomotor activation, suggesting that exposure to an environment associated with nicotine can result in activation of the transcription factor CREB. Context-associated elevations of pCREB were specific to the NAc shell and not observed in the NAc core or the hippocampus using these parameters. Under place conditioning training conditions where animals were given a choice of chamber, disruption of CREB activation in the NAc shell blocked nicotine CPP, suggesting that induction of pCREB by exposure to contextual cues may support nicotine reward learning.

These data expand on previous studies of CREB αδ knockout mice (Walters et al, 2005) suggesting that regulation of CREB is critical for nicotine CPP. We show that in animals that develop with normal CREB function, inactivation of CREB in the NAc shell in adulthood is sufficient to disrupt nicotine CPP. Previous reports have shown that HSV-mCREB infusion into the NAc shell enhances cocaine and morphine CPP (Carlezon et al, 1998; Walters and Blendy, 2001; Barrot et al, 2002). In contrast, the current study shows that nicotine CPP cannot occur without CREB activity in the NAc shell. This suggests that although various drugs of abuse converge on the NAc in terms of DA release (Di Chiara and Imperato, 1988), the circuit-level adaptations that underlie nicotine reward differ from those of other drugs of abuse (Nestler, 2005). Interestingly, chronic nicotine results in decreases in pCREB in the NAc (Brunzell et al, 2003), and thus may contribute to the ability of cigarette smoking to promote the transition to use of other drugs of abuse (Lai et al, 2000). Other studies, however, show that elevated pCREB is associated with cocaine CPP in the NAc (Miller and Marshall, 2005; Walters et al, 2005) and antisense-induced reductions of NAc CREB in shell or core attenuate the primary reinforcing efficacy of cocaine as well as its ability to act as primer during reinstatement (Choi et al, 2006), perhaps due to more chronic dosing regimens of cocaine self-administration (McClung and Nestler, 2003). Results obtained with CPP do not always translate onto intravenous self-administration data, but unfortunately, reliable nicotine self-administration is difficult to establish in mice.

Place conditioning provides a valuable tool for assessing Pavlovian contributions to changes in signaling and consequent reward learning. The present studies expand on earlier data, which did not discriminate between subdivisions of the NAc (Pandey et al, 2001; Brunzell et al, 2003; Walters et al, 2005), to show a dichotomy between the NAc core and shell in nicotine-associated regulation of pCREB. Previous data show that CPP testing, but not a similar regimen of nicotine exposure and withdrawal in the homecage, results in elevations of NAc pCREB (Walters et al, 2005). In the current study mice showed an increase in NAc shell levels of pCREB while isolated to a nicotine-paired chamber 22 h after the last of 3 daily pairings of nicotine with that chamber. Hence, Pavlovian nicotine place conditioning resulted in elevations of pCREB in the absence of overt nicotine-seeking. Microdialysis studies have shown that elevated DA release, which ought to lead to D1-associated activation of CREB, occurs in the NAc shell but not the core during early acquisition sessions (fewer than seven), indicating that CREB in the NAc shell may contribute to initiation of nicotine use (Spina et al, 2006). These Pavlovian conditioning data support the hypothesis that the CS properties of the context are sufficient to activate CREB in the NAc shell.

At the molecular level, activation of pCREB in the NAc follows a pattern that is similar to that seen for DA neuron activity during stimulus reward learning. Studies in primates and rodents show that after repeated pairing of a rewarding stimulus with a cue, DA neuron firing shifts from responding to the primary rewarding stimulus, becoming dependent instead on the conditioned cue (Schultz et al, 1997; Day et al, 2007). Similarly, elevations of pCREB were observed in the NAc in response to acute nicotine exposure, but not following 4 days of exposure to nicotine in the homecage (Walters et al, 2005) or during Pavlovian training on the third day of nicotine exposure in these studies. Instead, NAc pCREB levels were elevated in response to the nicotine-paired chamber, suggesting that the CS properties of the context and not the rewarding properties of nicotine regulate pCREB in the NAc shell after training that is sufficient to lead to nicotine CPP.

CREB signaling is necessary for neural plasticity associated with learning and addiction (Silva et al, 1998; Carlezon et al, 2005). Genes such as fos and BDNF that are regulated by nicotine and thought to be important for tobacco addiction (Schroeder et al, 2001; Li et al, 2008) are upregulated as a result of CREB overexpression and are inhibited by expression of mCREB in the NAc of transgenic mice (McClung and Nestler, 2003). Disruption of CREB activity could interfere with expression of many genes that have CRE sequences in their promoters. For example, CREB activity is necessary for the regulation of μ-opiate receptors in response to nicotine administration, and this upregulation contributes to nicotine CPP (Walters et al, 2005). Thus, CRE-mediated neuroadaptations that normally occur in response to nicotine may lead to lasting changes in systems that regulate incentive salience and drug reward.

There was no effect of nicotine or a nicotine-paired chamber on regulation of pCREB in the hippocampus in the current study, as was observed using a 4-day regimen of CPP training in previous studies (Walters et al, 2005). One potential reason for this difference may be that animals were not actively seeking the nicotine-paired chamber in the current study, but were rather confined to that chamber by the experimenter. Some evidence suggests that regulation of CREB in the dorsal hippocampus is more critical for spatial navigation than context conditioning (Pittenger et al, 2002). Although CREB was not regulated in the NAc core or hippocampus during exposure to a nicotine-paired chamber on the third day of training, we cannot conclude that CREB in these brain areas is not critical for learning during earlier trials or for consolidation processes related to nicotine CPP (Josselyn and Nguyen, 2005). Human studies show that the dorsal striatum and orbitofrontal cortex may also be involved in cue reward for drugs of abuse (David et al, 2007; Franklin et al, 2007; McClernon et al, 2008; Volkow et al, 2008). Whereas guide canulae prevented expression of mCREB along the canulae tracks, damage to these overlying structures could have impaired cue reward. LacZ control animals showed levels of nicotine CPP similar to untreated controls, however, suggesting that potential unilateral damage during infusion did not affect nicotine CPP.

In summary, these studies show that activation of CREB in the NAc shell is critical for nicotine CPP. The phosphorylation state of CREB may therefore regulate the motivational valence for nicotine or the conditioned rewarding effects of nicotine-associated cues, such as a context associated with nicotine. A nicotine-paired environment can elicit CREB phosphorylation in the NAc shell, suggesting that contextual cues may drive changes in CREB that promotes nicotine reward learning. We conclude from the current study that pharmacological agents that decrease CREB phosphorylation might be novel targets for smoking cessation.