Nicotine enhances contextual fear conditioning (Davis & Gould, 2006; Davis, Porter, & Gould, 2005; Feiro & Gould, 2005; Gould & Higgins, 2003; Gould & Wehner, 1999) but not cued fear conditioning (Gould, Feiro, & Moore, 2004). Recent studies suggest that this effect is mediated by α4β2 nicotinic acetylcholine receptors (nAChRs; Davis & Gould, 2006; Wehner et al, 2004). Little is known of the cellular mechanisms that underlie the effect of nicotine on contextual fear conditioning. In contrast, the cellular mechanisms that underlie contextual fear conditioning are well characterized. For example, extracellular signal-regulated kinase 1/2 (ERK 1/2) is among several intracellular signaling proteins that are implicated in contextual fear conditioning (Atkins, Selcher, Petraitis, Traskos, & Sweatt, 1998). Research indicating that nicotine administration, like conditioning, increases ERK 1/2 activity (Valjent, Pages, Herve, Girault, & Caboche, 2004) suggests that this second messenger may be involved in the enhancement of contextual fear conditioning by nicotine.

Numerous studies have shown that ERK 1/2 activation is necessary for the formation of long-term memories of contextual fear conditioning (for review see Sweatt, 2001). ERK 1/2 is activated after training (Ahi, Radulovic, & Spiess, 2004; Atkins et al., 1998; Sananbenesi, Fischer, Schrick, Spiess, & Radulovic, 2002, 2003; Selcher, Atkins, Trzaskos, Paylor, & Sweatt, 1999), and inhibition of mitogen-activated protein kinase kinase (MEK), the immediate upstream activator of ERK 1/2, produces deficits in contextual fear conditioning (Ahi et al., 2004; Atkins et al., 1998; Sananbenesi et al., 2002; Selcher et al., 1999). Genetic knockout studies have also implicated ERK 1/2 in fear conditioning; MEK dominant negative mice, which have decreased ERK 1/2 activity, show deficits in contextual fear conditioning (Shalin, Zirrgiebel, Honsa, Julien, Miller, Kaplan, & Sweatt, 2004).

Although the involvement of ERK 1/2 in the effects of nicotine on contextual fear conditioning has not been investigated until now, nicotine has been shown to activate ERK 1/2 in multiple brain areas (Brunzell, Russell, & Picciotto, 2003), including the hippocampus (Valjent et al., 2004), an area shown to be integral to contextual fear conditioning (Logue, Paylor, & Wehner, 1997; Philips & Ledoux, 1992). Studies of synaptic plasticity further suggest that ERK 1/2 may be involved in the effects of nicotine. Hippocampal long-term potentiation (LTP), which has long been regarded as a representative model of the cellular and molecular processes underlying learning (Bliss & Collingridge, 1993; Izquierdo & McGaugh, 2000), activates ERK 1/2 (English & Sweatt, 1996), and inhibition of ERK 1/2 activation impairs LTP induction (Atkins et al., 1998; Coogan, O’Leary, & O’Connor, 1999; English & Sweatt, 1997; Impey, Obrietan, et al., 1998; Winder et al., 1999). Just as nicotine enhances contextual fear conditioning (Gould & Wehner, 1999), nicotine facilitates the induction of LTP (Fujii, Ji, Morita, & Sumikawa, 1999) and can even induce hippocampal LTP (Matsuyama, Matsumoto, Enomoto, & Nishizaki, 2000). Additionally, inhibition of ERK 1/2 disrupts nicotine-facilitated LTP (Wang, Chen, Zhu, & Chen, 2001). These results suggest that nicotine-enhanced activation of ERK 1/2 may facilitate synaptic plasticity and could mediate the effects of nicotine on contextual fear conditioning.

The present experiments investigate the role of ERK 1/2 in nicotine enhancement of contextual fear conditioning. If nicotine enhances contextual fear conditioning through increases in ERK 1/2 activation, then this enhancement should be susceptible to interference by a dose of an ERK 1/2 inhibitor that is subthreshold for inhibiting contextual fear conditioning. To test this, SL327, a selective, systemic inhibitor of ERK 1/2 activation, was coadministered with nicotine at training, testing, or both training and testing.

Eight- to 12-week-old male C57BL/6J mice, obtained from the Jackson Laboratory (Bar Harbor, ME), were maintained on a 12-hr light–dark cycle with lights on at 7:00 a.m. Mice were housed in groups of 4 with ad libitum access to food and water. All procedures were approved by the Temple University Institutional Animal Care and Use Committee.

Mice were trained and tested in four identical Med Associates (St. Albans, VT) conditioning chambers housed in sound-attenuating cubicles. Ventilation fans provided a 69-dB background noise. The conditioning chambers consisted of stainless steel sides; Plexiglas front, back, and ceiling panels; and a grid floor composed of 18 stainless steel bars connected to a wiring harness that delivered a 0.57-mA unconditioned stimulus (US) foot shock. The conditioned stimulus (CS) was an 85-dB white noise delivered by speakers mounted on the left side of each chamber. The conditioning program was run with Med Associates software.

SL327—a selective systemic MEK 1 inhibitor (Atkins et al., 1998) structurally analogous to U0126 (Favata et al., 1998) that does not have significant effects on PKA, PKC, or CaMKII (Selcher et al., 1999)—and (-)-nicotine hydrogen tartrate salt were purchased from Sigma-Aldrich (St. Louis, MO). Injections were counterbalanced across training and testing such that each mouse received two injections prior to training and two injections prior to testing: SL327 or vehicle, and nicotine or saline. SL327 at 5 and 10 mg/kg (10 ml/kg injection volume), dissolved in a 10% Cremaphor (Sigma-Aldrich) and 90% saline solution, was administered by intraperitoneal injection 20 min prior to training and testing in order to establish a subthreshold dose for pairing with nicotine. Prior research demonstrated that 10 mg/kg of SL327 significantly inhibits hippocampal ERK activity (Selcher et al., 1999). The timing of SL327 administration was based on peak ERK 1/2 inhibition time (Selcher et al., 1999), ERK 1/2 activation times subsequent to contextual fear conditioning (Sananbenesi et al., 2002), and preliminary behavioral pharmacology experiments in our lab. Nicotine tartrate (0.25 mg/kg equivalent to 0.09 mg/kg freebase) was dissolved in saline and administered by intraperitoneal injection (10 ml/kg injection volume) 5 min prior to training and testing. The dose of nicotine, administration route, and injection time were based on prior studies (Gould & Higgins, 2003; Petersen, Norris, & Thompson, 1984).

All procedures were conducted during the mice’s light cycle. Only contextual fear was tested because nicotine enhances contextual but not cued fear conditioning (Gould et al., 2004; Gould & Higgins, 2003). In addition, contextual fear conditioning shows greater disruption by ERK 1/2 inhibition than cued fear conditioning (Selcher et al., 1999). For conditioning, baseline freezing was assessed during the first 120 s, and then mice received two paired presentations, separated by 120 s of a CS (30-s, 85-dB white noise) paired with a coterminating US (2-s, 0.57-mA footshock). The training session was terminated 30 s following the second CS–US presentation. Testing occurred 24 hr later, during which mice were placed in the conditioning chambers and contextual fear conditioning was assessed over 5 min. For a detailed description of the training and testing procedures, see Gould & Lommock (2003).

Fear was quantified as percent freezing, calculated by dividing observed freezing by the total observations during that period (Gould & Wehner, 1999). Percent freezing data were analyzed with SPSS version 13. Data sets were tested for significance with analyses of variance (ANOVAs). Homogeneity of variance was tested with Levene’s statistic (Cohen, 2001). Tukey’s post hoc tests of statistical significance (p < .05; Cohen, 2001) were conducted when data sets had equal variance. Data sets with unequal variance were tested with Games-Howell post hoc tests (Maxwell & Delaney, 2003). In addition to contextual freezing, baseline freezing was analyzed to assess general drug effects.

In Experiment 1, a dose of SL327 subthreshold for disrupting contextual fear conditioning was determined. Mice received SL327 (0 [vehicle], 5 or 10 mg/kg) 20 min prior to contextual fear conditioning, and saline 5 min prior to conditioning to match later experiments. An ANOVA indicated a significant difference among groups for contextual freezing, F(2, 35) = 5.72, p < .05 (see Figure 1A), and no difference for baseline freezing, F(2, 35) = 0.00, p > .05. Games-Howell post hoc tests revealed that mice that received 10 mg/kg of SL327 show significantly less contextual fear conditioning than mice in the vehicle group. Mice receiving 5 mg/kg SL327 did not significantly differ from vehicle. Thus, the 10-mg/kg dose of SL327 disrupted contextual fear conditioning, whereas the 5-mg/kg dose was subthreshold for disrupting contextual fear conditioning.

Open in a separate window

Figure 1

Effects of SL327 on nicotine-enhanced contextual fear conditioning

(A) Administration of SL327, a selective inhibitor of extracellular signal-regulated kinase (ERK) 1/2, blocks contextual fear conditioning at a dose of 10 mg/kg but not at a dose of 5 mg/kg. Therefore, SL327 is subthreshold for the disruption of contextual fear conditioning at a dose of 5 mg/kg. (B) Subthreshold inhibition of ERK 1/2 at training of contextual fear conditioning blocks the enhancement of contextual fear conditioning by nicotine (0.09 mg/kg). SL327 (5 mg/kg) was administered only prior to training; prior to testing, mice received a matched vehicle injection. (C) Subthreshold inhibition of ERK 1/2 at training and testing of contextual fear conditioning blocks the enhancement of contextual fear conditioning by nicotine (0.09 mg/kg). (D) Subthreshold inhibition of ERK 1/2 at testing has no significant effect on nicotine-enhanced contextual fear conditioning. SL327 (5 mg/kg) administered only prior to testing; prior to training, mice received a matched vehicle injection. Asterisk (*) indicates significant difference from vehicle control groups. Data are mean (± SEM).

In Experiment 2, the role of ERK 1/2 in the enhancement of contextual fear conditioning by nicotine was investigated. The subthreshold dose of SL327 (5 mg/kg) established in Experiment 1 was administered 20 min prior to training; vehicle was administered 20 min prior to testing. Nicotine (0.09 mg/kg) or saline was administered 5 min prior to both training and testing. Thus, mice received vehicle/saline, vehicle/nicotine, SL327/nicotine, or SL327/saline at training and vehicle/saline or vehicle/nicotine at testing. An ANOVA showed a significant effect of drug treatment on contextual fear conditioning, F(3, 60) = 6.71, p < .05 (see Figure 1B), but not baseline freezing, F(3, 60) = 1.00, p > .05. Tukey’s post hoc comparisons revealed that the vehicle/nicotine group demonstrated significantly higher levels of contextual fear conditioning than all other drug treatments. Notably, the SL327/ nicotine group demonstrated significantly lower levels of contextual fear conditioning than the vehicle/nicotine group but levels of contextual fear conditioning that were not different from the vehicle/saline group. These results demonstrate that a subthreshold dose of SL327 administered prior to training of contextual fear conditioning attenuates the enhancement of contextual fear conditioning by nicotine.

The effects of SL327 both at training and testing and at testing alone on nicotinic enhancement of contextual fear conditioning were also examined. Coadministration of SL327 and nicotine at training and testing yielded results identical to coadministration prior to training alone; ANOVAs indicated a significant effect of drug condition on contextual fear conditioning, F(3, 76) = 4.05, p < .05 (see Figure 1C), but not baseline freezing, F(3, 76) = 1.45, p > .05. Tukey’s post hoc comparisons indicated that the vehicle/nicotine group demonstrated greater levels of contextual fear conditioning than all other groups. In addition the SL327/nicotine group displayed no enhancement compared with the vehicle/saline group. These results indicate that attenuation of nicotine-enhanced contextual fear conditioning by SL327 at training only was not due to a state-dependent effect, because state-dependent learning theory would predict that learning occurring in one state (SL327 and nicotine) would not be retrieved in another state (vehicle and nicotine), but learning and retrieval occurring during the same state (SL327 and nicotine) would not be disrupted. Because the effects of SL327 on enhancement of contextual fear conditioning by nicotine are consistent regardless of SL327 treatment at testing, it is unlikely that the effects of SL327 on nicotine-enhanced memories are due to state-dependent effects.

In contrast to the effects of coadministration of SL327 and nicotine at training only, coadministration of SL327 with nicotine at testing did not significantly alter nicotine enhancement of contextual fear conditioning. ANOVAs showed a significant effect of drug condition on contextual fear, F(3, 75) = 7.16, p < .05 (see Figure 1D). Tukey’s post hoc comparisons revealed that the vehicle/nicotine and SL327/ nicotine groups demonstrated higher levels of contextual fear than did the vehicle/saline group. Additionally, there was no significant effect of SL327 on retrieval at testing, because the SL327/saline group was not statistically different from the vehicle/saline group. There was no significant effect of drug condition on baseline freezing, F(3, 75) = 0.32, p > .05. These results indicate that there is no significant effect of a subthreshold dose of SL327 on retrieval of nicotine-enhanced contextual fear conditioning.

The present experiment extends previous reports (Atkins et al., 1998; Selcher et al., 1999) indicating that ERK 1/2 is involved in contextual fear conditioning by demonstrating that the enhancement of contextual fear conditioning by nicotine is blocked by a dose of the ERK 1/2 inhibitor SL327 that is subthreshold for disrupting contextual fear conditioning. This effect was seen when SL327 was administered at training but not testing. Thus, these results suggest that the enhancement of contextual fear conditioning by nicotine is mediated by ERK 1/2 activation at training. However, the mechanisms through which nicotine may increase ERK 1/2 activation during learning are still unknown.

Nicotine has been shown to activate ERK 1/2 in cultures of PC12h cells (Nakayama, Numakawa, Ikeuchi, & Hatanaka, 2001), SH-SY5Y cells (Dajas-Bailador, Soliakov, & Wonnacott, 2002), cultured hippocampal neurons (Dajas-Bailador et al., 2002), and multiple brain regions (Brunzell et al., 2003; Valjent et al., 2004). Furthermore, Hu, Liu, Chang, and Berg (2002) demonstrated that nicotine induces activation of CREB and upregulation of cFOS in hippocampal neurons, and that activation of CREB by nicotine was dependent upon ERK 1/2 phosphorylation. Recent evidence indicates that hippocampal CREB activity is essential for the formation of long-term memory (Florian, Mons, & Roullet, 2006; Graves, Dalvi, Lucki, Blendy, & Abel, 2002; Impey, Smith, et al., 1998; Wood, Kaplan, Park, Blanchard, Oliveira, Lombardi, & Abel, 2005). Thus, nicotine can increase the activity of signaling cascades underlying learning-related plasticity, but how nicotine is coupled to these cascades is unclear.

There are at least three routes by which nicotine could act to enhance learning-related ERK 1/2 activity: (a) direct activation of the ERK 1/2 signaling cascade, (b) facilitation of ERK activity mediated by N-methyl-d-aspartate receptors (NMDARs), and (c) stimulation of the release of neurotransmitters that can activate ERK 1/2. Nicotinic receptors can gate Ca++ influx (Barrantes, Westwick, & Wonnacott, 1994; Seguela, Wadiche, Dineley-Miller, Dani, & Patrick, 1993), an event that can activate intracellular signaling molecules such as ERK 1/2 (Rosen, Ginty, Weber, & Greenberg, 1994). Thus, during contextual fear conditioning, nicotine-mediated calcium influx through nAChRs could boost ERK 1/2 activation and enhance learning.

Nicotine-mediated effects at nAChRs may also interact with NMDAR-mediated effects to enhance ERK 1/2 signaling and contextual fear conditioning. Multiple lines of evidence support this possibility. First, nAChRs and NMDARs co-localize on the same neurons, suggesting that processes mediated by each receptor could interact during learning (Risso et al., 2004). Second, nAChRs and NMDARs have been shown to critically interact during acquisition of contextual fear conditioning (Gould & Lewis, 2005); similar results have been found for other types of learning (Ciamei, Averson, Cestari, & Castellano, 2002; Levin, Bettegowda, Weaver, & Christopher, 1998). Finally, nAChRs and NMDARs have been shown to interact during synaptic plasticity. Specifically, nicotine enhances NMDAR-dependent LTP (Sawada, Ohno-Shosaku, & Yamamoto, 1994; Yamazaki, Jia, Niu, & Sumikawa, 2006). The mechanisms through which nicotine could interact with NMDARs are unknown, but at least three possibilities exist: nAChR activation could contribute to NMDAR activation, nAChR activation could boost NMDAR signaling, and nicotine could stimulate glutamate release.

NMDARs are critically involved in contextual fear conditioning (Bardgett et al., 2003; Csernansky et al., 2005; Gould & Lewis, 2005; Gould, McCarthy, & Keith, 2002); activation of NMDARs requires conjunctive release of glutamate with postsynaptic depolarization (Nowak, Bregestovski, Ascher, Herbet, & Prochiantz, 1984). Nicotine-mediated nAChR gating of cations could contribute to the depolarization necessary for NMDAR activation. In addition, NMDARs, like nAChRs, gate calcium (Riedel, Platt, & Micheau, 2003); thus, as discussed earlier, nicotine-induced activation of nAChRs could contribute to activation of cell-signaling cascades involved in contextual fear conditioning that are activated by NMDARs. Finally, nicotine can stimulate the release of glutamate (Fedele, Varnier, Ansaldo, & Raiteri, 1998; McGehee, Heath, Gelber, Devay, & Role, 1995; Radcliffe, Fisher, Gray, & Dani, 1999). Increased levels of glutamate could facilitate NMDAR activation, which could lead to ERK 1/2 activation during contextual fear conditioning.

Additionally, nicotine has been shown to modulate the release of other neurotransmitters. Acetylcholine, norepinephrine, dopamine, and 5-hydroxytryptamine are all released in response to nicotine administration (Grady et al., 2002; Grady, Marks & Collins, 1994; Rao, Correa, Adams, Santori, & Sacaan, 2003; Singer et al., 2004). These neurotransmitters have also been implicated in contextual fear conditioning (Ji, Wang, & Li, 2003; Melik, Babar-Melik, Ozgunen, & Binokay, 2000; Pezze, Heidbreder, Feldon, & Murphy, 2001; Roberts, Krucker, Levy, Slanina, Sutcliffe, & Hedlund, 2004; Soares, Fornari, & Oliveira, 2006) and are associated with activation of ERK 1/2 (Berkeley & Levey, 2003; Chen, Nguyen, Pike, & Russo-Neustadt, 2007; Errico, Crozier, Plummer, & Cowen, 2001; Runyan & Dash, 2004). Ergo, nicotine-stimulated neurotransmitter release could enhance contextual fear conditioning through activation or enhancement of ERK 1/2 activity.

In summary, the present findings are the first to suggest that nicotine enhances contextual fear conditioning by increasing ERK 1/2 activity at training. The mechanisms by which nicotine could be coupled to ERK 1/2 activity are as yet unknown. However, recent evidence indicates that nicotine could enhance ERK 1/2 activation through at least three mechanisms: direct activation of the ERK 1/2 signaling cascade via calcium influx, interaction with NMDARs, and enhancement of the release of neurotransmitters that effect learning-related processes and increase ERK 1/2 activation. In future studies we will examine the mechanisms through which nicotine enhances ERK 1/2 activation during contextual fear conditioning.