Assessing Combined Effects of Varenicline and N-acetylcysteine on Reducing Nicotine Seeking in Rats

Abstract

Nicotine addiction is a chronic relapsing brain disorder and cigarette smoking is the leading cause of preventable death in the United States. Currently, the most effective pharmacotherapy for smoking cessation is Varenicline (VRN), which reduces both positive and negative reinforcement by nicotine. Clinically, VRN attenuates withdrawal symptoms and promotes abstinence, but >50% of smokers relapse within 3 months following a quit attempt. This may indicate that VRN fails to ameliorate components of nicotine-induced neuroplasticity that promote relapse vulnerability. Animal models reveal that glutamate dysregulation in the nucleus accumbens is associated with nicotine relapse. N-acetylcysteine (NAC) normalizes glutamate transmission, and prolongs cocaine abstinence. Thus, combining VRN and NAC may promote and maintain, respectively, nicotine abstinence. In rats, we found that VRN effectively reduced nicotine self-administration and seeking in early abstinence, but not seeking later in abstinence. In contrast, NAC reduced seeking only later in abstinence. Because VRN and NAC are sometimes associated with mild adverse effects, we also evaluated a sequential approach combining subthreshold doses of VRN during self-administration and early abstinence with subthreshold doses of NAC during late abstinence. As expected, subthreshold VRN did not reduce nicotine intake. However, subthreshold VRN and NAC reduced seeking in late abstinence, suggesting a combined effect. Overall, our results suggest that combining subthreshold VRN and NAC is a viable and drug-specific approach to promote abstinence and reduce relapse while minimizing adverse effects. Our data also suggest that different components and time points in addiction engage the different neurocircuits targeted by VRN and NAC.

INTRODUCTION

Nicotine addiction is brain disorder characterized by chronic relapse and cigarette smoking is the leading cause of preventable premature death in the United States [1]. Nicotine use is particularly difficult to curtail due to both positive and negative reinforcing properties. On one hand, nicotine is positively reinforcing because it results in increased dopamine release in mesolimbic reward circuitry [2]. On the other hand, nicotine use is negatively reinforced because acute nicotine administration normalizes the impaired cognition and negative affect experienced during periods of nicotine abstinence [3]. These simultaneous positive and negative reinforcing properties make nicotine abstinence particularly difficult to achieve and maintain.

Pharmacotherapeutic treatment for nicotine addiction was markedly advanced by the introduction of Varenicline. Varenicline (VRN), a potent partial agonist of α4β2 receptor, low-efficacy partial agonist of α3β4, α6, α3β2β3, and potent full agonist of α7 containing receptor [4], has become the most effective first-line pharmacotherapy for promoting smoking cessation [5–8]. VRN blunts the positive reinforcing effects of smoking by producing a sustained release of mesolimbic dopamine [2]. Additionally, VRN attenuates negative affect during smoking withdrawal [9] by normalizing amygdala hyperreactivity [10], especially in nicotine users most vulnerable to withdrawal-induced negative affect [11]. Likely for these reasons, VRN increases success of attempts to quit smoking by a factor of 2–3 compared to attempts with no pharmacotherapy [12].

Despite VRN’s clinical efficacy, relapse rates among VRN users still exceed 50–60% after three-months of abstinence [2,8]. One possible explanation for the high relapse rates with VRN treatment is that VRN fails to act directly on the long-term neuroplasticity induced by repeated use of addictive drugs that is known to underly enduring relapse vulnerability [13]. VRN administration to abstinent smokers normalizes resting-state functional connectivity (rsFC) in an amygdala-centric network [10], but has no effect on frontostriatal-mediated reward processing [14]. Importantly, frontal and striatal regions have known implications in drug craving [15], and neural activation in frontostriatal and amygdala regions can predict cue-induced smoking relapse in humans [16]. Further, animal research has demonstrated that VRN fails to reduce relapse of nicotine seeking induced by nicotine cues [17, 18], adding credence to the idea that VRN does not reduce craving induced by drug-paired cues, which may be a contributing factor to relapse.

Animal models of drug use have revealed that relapse across drug classes is occasioned by dysregulation of glutamate in the cortico-accumbens network. Re-exposure to drug-related cues downregulates expression of the glial glutamate transporter, GLT-1, that is expressed in high density by astroglia near synapses, resulting in spillover of synaptically-released glutamate outside of the synaptic cleft [19]. Nicotine is among the drugs that reduce GLT-1 in the nucleus accumbens [20, 21], and restoring GLT-1 in animal models by treatment with drugs like N-acetylcysteine (NAC) or ceftriaxone reduces relapse of nicotine seeking [22–24]. NAC has also been shown to reduce cue-induced relapse of nicotine seeking via indirect stimulation of presynaptic group 2 metabotropic glutamate receptors [25]. Due to a long history of clinical use as an antioxidant, such as in the treatment of acetaminophen overdose [26], NAC has proven particularly tractable for testing in clinical populations with substance abuse disorders and other psychiatric diseases [27–30]. While the effects are sometimes mixed, NAC generally reduces cue reactivity [28] as well as smoking withdrawal symptoms (e.g., craving) and relapse among nicotine-deprived smokers [31] (however, see [32–34]). Because VRN effectively aids in initiating abstinence and NAC reduces craving and cue reactivity during abstinence, an approach combining both pharmacotherapies may increase achievement and maintenance of abstinence from smoking in individuals with tobacco use disorder.

Finally, both VRN and NAC have been associated with moderate adverse side effects. For example, VRN is associated with nausea [35] and mixed reports for increased risk of cardiac symptoms [35–37]. Initial reports on VRN also indicated heightened risk of serious neuropsychiatric symptoms [38], but subsequent investigations have failed to confirm these results [39]. NAC has also been associated with mild side effects, including gastrointestinal events [40], allergic reactions [41, 42] and local irritation with some routes of delivery [40]. Thus, exploring low-dose protocols to reduce adverse side effects while maintaining efficacy for VRN and NAC is warranted.

The present experiments addressed two primary questions in a rat model of nicotine seeking. First, does treatment with VRN or NAC alone reduce self-administration or relapse of nicotine seeking? Second, are subthreshold doses that may reduce adverse side effects of either or both drugs sufficient to reduce nicotine self-administration and relapse? We found that VRN was effective at reducing nicotine self-administration and relapse during early abstinence, but ineffective at reducing relapse later in abstinence. NAC was ineffective at reducing nicotine self-administration or relapse during early abstinence, but was effective at reducing relapse in later abstinence. Based on these results, we sequentially administered low doses of VRN during self-administration and early abstinence followed by NAC during late abstinence. The sequential combination effectively reduced relapse later in abstinence. Because we did not assess directly the nature of the pharmacodynamic interaction between VRN and NAC, we refer to the effect of VRN and NAC on nicotine seeking simply as “combined”. Finally, these effects appear to be drug-specific, as sequentially combined VRN and NAC had no effect on sucrose seeking.

MATERIALS AND METHODS

Animal housing and surgery

One-hundred eighty-seven male Sprague Dawley rats (250 g; Charles River Laboratories) were individually housed using a 12:12 hour dark/light cycle with ad libitum water in a temperature- and humidity-controlled environment. All experimentation occurred during the dark phase, and animals were allowed to acclimate to the vivarium environment for a week prior to surgery. Rats were ~70 days old when implanted with an indwelling jugular catheter (surgical details have been described previously: [43, 44]). Briefly, rats were anesthetized and an indwelling, back-mounted cannula was implanted and attached to a silastic catheter inserted into the right jugular vein. Animals were mildly food restricted following recovery from surgery and prior to beginning self-administration (i.e., fed ~24g/day). A single injection of methohexital (Brevital, ~5mg/kg, i.v.) was used to confirm catheter patency following self-administration. All procedures were in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the Assessment and Accreditation of Laboratory Animal Care.

Drugs Used

(−)-Nicotine hydrogen tartrate salt (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in sterile saline (0.9% weight/vol NaCl) and brought to pH 7.4 with sodium hydroxide (Fisher Scientific, United States of America; 0.784 M). Nicotine solutions and syringes were kept at −4°C while not being used, and new nicotine solutions were made every two days. Nicotine was delivered intravenously contingent on lever presses as described below.

Varenicline tartrate (VRN) (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in sterile saline (0.9% weight/vol NaCl) according to the different doses used (0.3, 1.0 or 3.0 mg/kg) for intraperitoneal (i.p.) injections. VRN was prepared daily and brought to pH 7.4 with sodium hydroxide.

N-acetyl-L-cysteine (NAC) (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in sterile saline (0.9% weight/vol NaCl) according to the different doses used (10 or 30 mg/kg) for injections (i.p.). NAC was prepared daily and brought to pH 7.4 with sodium hydroxide.

General Methods

Self-Administration:

All procedures occurred in standard operant chambers equipped with two retractable levers, a house light, cue light, and 2900 Hz tone generator (Med Associates). Before nicotine self-administration training, animals were food deprived and underwent a single 2-hour food training session in which presses on the active lever resulted in the delivery of a single food pellet (45 mg, Noyes) on a fixed-ratio 1 (FR1) schedule of reinforcement. One day later, animals began 2h sessions of nicotine (Experiments 1–3) or sucrose (Experiment 4) self-administration on an FR1 schedule. Each active lever press resulted in a nicotine infusion (0.02 mg/kg/infusion, Experiments 1–3) or delivery of a single sucrose pellet (Experiment 4) and the simultaneous presentation of a compound light (above the active lever) and tone (2900 Hz) stimulus. Each nicotine/pellet delivery was followed by a 20 second timeout during which responses were recorded but did not produce nicotine or cues. An inactive lever was also provided to control for changes in general activity. Rats underwent self-administration for a minimum of 10 days (Fig 1A).

Figure 1.

Schematic of experimental procedures. A) To test for effects on nicotine intake and early context seeking, rats underwent 7 days of self-administration without treatment. Then, 2 hours prior to days 8–10 of self-administration, days 11–13 of abstinence, and context seeking test (day 14) rats received SAL, VRN, or NAC. B) To test for effects on late cue seeking, rats were exposed to 10 days of self-administration and 7 days of abstinence without treatment. Then, rats received SAL, VRN, or NAC on days 18–21 of abstinence and 2 hours prior to a cued seeking test (day 22). C) To determine whether pre-treatment with VRN or NAC protected against later cue seeking, rats underwent 7 days of self-administration without treatment. Then, 2 hours prior to days 8–10 of self-administration, days 11–13 of abstinence, and context seeking test (day 14) rats received SAL, VRN, or NAC. Then, cue seeking (day 22) was assessed following 7 days of abstinence with no treatment. D) To test for effects of sequential combination of VRN then NAC, rats underwent 7 days of self-administration without treatment. Then, 2 hours prior to days 8–10 of self-administration, days 11–13 of abstinence, and context seeking test (day 14) rats received SAL or VRN. Following 7 days of abstinence with no treatment, rats received NAC on days 18–21 of abstinence and 2 hours prior to a cue seeking test (day 22). E) Active responding, inactive responding, and nicotine infusions during the first 7 sessions of self-administration (i.e., before any treatment) for all animals combined across experiments (mean, standard error). Active responding was significantly greater than inactive responding 2-way repeated-measures ANOVA, response type F(4.757,1394)=7.76, p < .0001). SAL = saline, VRN = varenicline, NAC = N-acetylcysteine. *p<.05

Abstinence:

Following self-administration, animals were left undisturbed in the housing facility (i.e., abstinence) except when receiving i.p. injections as described below. We chose an abstinence model rather than operant extinction because human nicotine users do not commonly encounter situations in which cigarettes or electronic nicotine delivery systems fail to provide nicotine reinforcement. Further, like humans [45], animals often exhibit signs of increased craving over periods of abstinence [46], suggesting that abstinence is an appropriate model with direct translational relevance for studying nicotine relapse.

Context Seeking:

Context seeking was assessed following self-administration and 3 days of abstinence. During context seeking tests (2-hr), lever presses were recorded but no longer produced drug or drug-paired cues (Fig 1B).

Cue Seeking:

All cued seeking tests (2-hr) began with one light/tone cue delivered response-independently, after which each active lever press produced the cue as in self-administration, but no infusion or pellet. Cue seeking was either assessed after self-administration and 10 days of abstinence (Fig 1C) or after context seeking and 7 days of abstinence (Fig 1D).

Experiment 1: Effects of VRN on nicotine self-administration, context seeking, and cued seeking.

To assess the effects of VRN on nicotine intake and seeking, rats received injections of Saline (SAL, N = 38) or VRN (0.3, 1.0, or 3.0 mg/kg, N = 9, 17, 15, respectively) during self-administration and context seeking (Fig 1A). To assess the effects of VRN on cued seeking alone, a separate group of rats received SAL (N = 30) or VRN (0.3, 1.0, or 3.0 mg/kg, N = 7, 14, 7, respectively) 4 days prior to and 2 hours before a cued seeking test (Fig 1B). Finally, to determine if VRN administered during nicotine self-administration and early abstinence protected against later relapse, a third group of rats underwent a cue seeking test following 7 days of abstinence after receiving SAL (N = 10) or VRN (3.0 mg/kg, N = 8) during self-administration and context seeking (Fig 1C).

Experiment 2: Effects of NAC on nicotine self-administration, context seeking, and cued seeking.

To assess the effects of NAC on nicotine intake and seeking, rats received intraperitoneal injections of Saline (SAL, N = 38) or NAC (10 or 30 mg/kg, N = 6 & 14, respectively) during self-administration and context seeking (Fig 1A). To assess the effects of NAC on cued seeking alone, a separate group of rats received SAL (N = 30) or NAC (10 or 30 mg/kg, N = 7 or 8, respectively) 4 days prior to and 2 hours before a cued seeking test (Fig 1B). Finally, to determine if NAC administered during nicotine self-administration and early abstinence protected against later relapse, a third group of rats underwent a cue seeking test following 7 days of abstinence after receiving SAL (N = 10) or NAC (30 mg/kg, N = 7) during context seeking (Fig 1C). Because the procedures in Experiments 1 and two were identical except for the use of VRN or NAC, the same SAL animals were used as controls in both experiments.

Experiment 3: Effects of combined subthreshold doses of VRN + NAC on nicotine self-administration, context seeking, and cued seeking.

To assess the effects of a sequential approach combining subthreshold doses of VRN then NAC on nicotine intake and seeking, rats received intraperitoneal injections of Saline (SAL) or VRN (0.3 or 1.0 mg/kg) during self-administration and context seeking (Fig 1D). To assess the effects of the sequential approach on cued seeking, the same rats received NAC (10 mg/kg) or SAL, 4 days prior to and 2 hours before a cued seeking test 7 days after context seeking (Fig 1D). Thus, Ns for each treatment were SAL + SAL (N=8), SAL + NAC (N=7), VRN 0.3 + NAC (N=9), VRN 1.0 + NAC (N=25).

Experiment 4: Effects of combined subthreshold doses of VRN + NAC on sucrose self-administration, context seeking, and cued seeking.

To determine if the effects of combined subthreshold doses of VRN + NAC were specific to nicotine, rats received intraperitoneal injections of SAL (N = 8) or VRN (1.0mg/kg, N = 8) during sucrose self-administration and context seeking (Fig 1D). To assess the effects of the sequential approach on cued seeking of sucrose, the same rats received NAC (10 mg/kg) or SAL, 4 days prior to and 2 hours before a cued seeking test 7 days after context seeking (Fig 1D). A subset of rats from Experiments 1–3 were also exposed to VRN, NAC, or SAL in a novel open field arena to further assess motoric effects and anxiety-like behavior (see supplement).

Statistics

Data were statistically evaluated using Prism 9.0 (GraphPad, Inc) to conduct analysis of variance (ANOVA) for multiple comparisons. Following a significant interaction score in one or two-way ANOVAs, post hoc tests for multiple comparisons were performed (Dunnett’s or Sidak’s multiple comparisons tests, respectively). For within-subject comparisons repeated measures ANOVAs were conducted.

RESULTS AND DISCUSSION

Experiment 1: VRN reduced nicotine intake and early context seeking, but not late cue seeking.

Figure 1E shows active lever pressing, nicotine infusions, and inactive lever pressing across the first 7 sessions of self-administration (i.e., before treatment with VRN or NAC) for all rats across all conditions in Experiments 1–4. Active lever pressing was greater than inactive lever pressing, indicating that active pressing was motivated by nicotine infusions.

In Experiment 1, VRN 3.0 mg/kg reduced nicotine seeking (active lever presses) during self-administration (as evidenced by differences in nicotine seeking before and after treatment; procedure Fig 1A; results Figs 2A & 2B) and context seeking (procedure Fig 1A; results Fig 2C). There was no effect of VRN on cue seeking either when treatment occurred prior to and during the test (procedure Fig 1B; results Fig 2D) or when treatment ended 7 days prior to testing (procedure Fig 1C; results Fig 2E). Importantly, there was also no effect of VRN .03 mg/kg or VRN 1.0 mg/kg at any point in Experiment 1 (Fig 2A–E).

Figure 2.

VRN 3.0 mg/kg reduces nicotine intake and context seeking in early abstinence. A) Active lever presses during nicotine self-administration. (2-way repeated-measures ANOVA, group F(3,75)=0.92, p=.43, session F(2.69, 201.9)=4.83, p=.004, interaction F(27,675)=2.31, p < .001) VRN 3.0mg/kg differs from SAL (p<.001) Group Ns same as B. B) Active lever presses pre (mean sessions 6–7) and post treatment (mean sessions 9–10; 2-way repeated-measures ANOVA, group F(3,80)=1.57, p=.03, session F(1,80)=4.93, p<.001, interaction (F(3,80)=7.16, p <.001) pre-post differ for VRN 3.0mg/kg (p=.001). C) Active lever presses during early context seeking (One-way ANOVA, F(3)=4.53, p=.006). VRN 3.0mg/kg differs from SAL (p=.002). D) Active lever presses during cued seeking with treatment (One-way ANOVA, F(3)=1.19, p=.32). E) Active lever presses during cued seeking without treatment (paired t-test, t(7)=.89, p=.40). Data shown as mean ± SEM, and N is shown above the bars. *p<.05, **p<0.01 using Sidak’s (2-way ANOVA) or Dunnett’s (One-way ANOVA) post hoc test. SAL = saline, VRN = varenicline, NAC = N-acetylcysteine.

These results are consistent with literature showing that VRN helps smokers achieve abstinence [12], and with findings that relapse is common, even among VRN users [2, 8]. Experiment 1 also replicated prior preclinical work showing that VRN reduces nicotine seeking during self-administration [17, 18]. Some prior work found immediate reductions in nicotine seeking following the first administration of VRN [e.g., 17], rather than the gradual reduction in seeking noted in Experiment 1. Due to procedural differences in reinforcement schedule, session duration and schedule, administration of extinction, and VRN dosing regimen, the reason for the discrepancy in onset of VRN effects is unclear. However, the overall finding that VRN reduces nicotine self-administration is consistent. Further, previous work has demonstrated an increasing effect of repeated low-doses of VRN [17] that is consistent with the present results.

Experiment 1 also indicated that VRN was effective at reducing nicotine seeking in early abstinence. It is possible that this effect could be due to lowered nicotine seeking during self-administration for animals that received VRN. For a full characterization of VRN effects, it may be important for future work to distinguish whether VRN reduced seeking in early abstinence due to acute effects or because of a reduced motivation to consume nicotine carried over from treatment during self-administration. However, for the translational purposes of the present study, the reduction in nicotine seeking during early abstinence is viewed as a clinically-relevant advantage of VRN treatment.

Finally, the finding that VRN fails to reduced cued nicotine seeking is consistent with previous reports [17, 18]. Thus, our data support the use of VRN as a first-line treatment to assist with achieving smoking abstinence, but also highlight the need to combine VRN with other approaches to help maintain abstinence across longer periods.

Experiment 2: NAC reduced late cue seeking, but not nicotine intake or early context seeking.

Regardless of dose, NAC had no effect on nicotine seeking during self-administration or context seeking during early abstinence in Experiment 2 (as evidenced by no differences in nicotine seeking before and after treatment; procedure Fig 1A; results Figs 3A–C). NAC 30 mg/kg reduced cued seeking when treatment occurred prior to and during the test (procedure Fig 1B; results Fig 3D), but not when treatment ended 7 days prior to testing (procedure Fig 1C; results Fig 3E). Importantly, there was also no effect of NAC 10 mg/kg at any point in Experiment 2 (Fig 3A–E).

Figure 3.

NAC 30 mg/kg reduces cue seeking in late abstinence. A) Active lever presses during nicotine self-administration. (2-way repeated-measures ANOVA, group F(2,55)=0.68, p=.51, session F(2.4,132.2)=1.23, p=.30, interaction F(18,495)=.65, p =.86). Group Ns same as B. B) Active lever presses pre (mean sessions 6–7) and post treatment (mean sessions 9–10; 2-way repeated-measures ANOVA, group F(2,55)=2.0, p =.14, session F(1,55)=0.02, p=.89, interaction F(2,55)=1.6, p=.20). C) Active lever presses during early context seeking (One-way ANOVA, F(3)=.14, p=.87). D) Active lever presses during cued seeking with treatment (One-way ANOVA, F(3)=4.29, p=.02). NAC 30mg/kg differs from SAL (p=.01). E) Active lever presses during cued seeking without treatment (paired t-test, t(6)=.12, p=.91). Data shown as mean ± SEM, and N is shown above the bars. *p<.05, **p<0.01 using Sidak’s (2-way ANOVA) or Dunnett’s (One-way ANOVA) post hoc test. SAL = saline, VRN = varenicline, NAC = N-acetylcysteine.

These results indicate that NAC is not an effective first-line therapy for helping achieve and maintain smoking abstinence. Indeed, the literature is mixed with regards to the efficacy of NAC at promoting abstinence [32–34]. While our study and many prior preclinical studies using different rodent treatment protocols have shown that NAC reduces nicotine seeking [23, 24, 47, 48], one preclinical study has reported that NAC can also reduce nicotine self-administration [24]. The discrepancy of this finding and the current data is likely due to differences in treatment parameters. The previous study used relatively higher doses of NAC (30–90 mg/kg) and a more demanding reinforcement schedule (FR5). It is not clear how these parameters influence the effects of NAC, nor is it clear which parameters more appropriately model nicotine use in humans. However, a primary goal of the present study was to identify the lowest doses of NAC capable of reducing nicotine seeking, necessitating our use of low doses. The mixed evidence on the efficacy of NAC for reducing nicotine intake provided by the current study and that of Ramirez-Nino [24] is reflective of the human literature, where some evidence indicates reductions in smoking following NAC treatment [e.g., 31] and other evidence fails to find reductions in smoking [e.g., 49].

Conversely, NAC was effective at promoting abstinence late in treatment. These results are consistent preclinical studies examining the effects of NAC on cued reinstatement [22–25]. However, NAC was ineffective at reducing cued seeking assessed 7 days after treatment, indicating that prior administration was not protective against relapse. Our data and prior preclinical and clinical studies indicate that NAC may not be an effective standalone therapy for reducing nicotine use, but may help to reduce relapse induced by cues after abstinence has been achieved.

Experiment 3: Sequentially combined subthreshold doses of VRN and NAC reduced late cue, but not nicotine intake or early context seeking.

As in Experiment 1, subthreshold VRN (0.3 or 1 mg/kg) alone did not reduce nicotine intake during self-administration and VRN 1mg/kg produced only a modest reduction in context seeking during early abstinence (procedure Fig 1D; results Figs 4A–C). However, subsequent subthreshold NAC (10 mg/kg) administered prior to and during the test reduced cued seeking only in rats that previously received VRN 1mg/kg (procedure Fig 1D; results Fig 4D). Importantly, NAC 10 mg/kg alone did not affect responding at any point in Experiment 2 (Fig 3A–E). As expected, responding did not differ between SAL + SAL and SAL + NAC 10mg/kg groups during self-administration (2-way repeated measures ANOVA, group F[1,13]=3.55, p=.082), context seeking (unpaired t-test t[13]=.89, p=.39) or cued seeking (unpaired t-test t[13]=1.784, p=.10). Thus, we combined these two groups for analysis (Control Group, N=15; Fig 4A–D).

Figure 4.

VRN 0.3 & 1.0 mg/kg + NAC 10 mg/kg reduces cue seeking in late abstinence. A) Active lever presses during nicotine self-administration. (2-way repeated-measures ANOVA, group F(2,46)=2.15, p=.13, session F(2.78, 127.8)=2.62, p=.06, interaction F(18,414)=1.33, p =.17). Group Ns same as B. B) Active lever presses pre (mean sessions 6–7) and post treatment (mean sessions 9–10; 2-way repeated-measures ANOVA, group F(2,46)=1.6, p =.21, session F(1,46)=0.07, p=.80, interaction F(2,46)=2.3, p=.11). C) Active lever presses during early context seeking (One-way ANOVA, F(2,30)=1.8, p=.18). D) Active lever presses during cued seeking with treatment (One-way ANOVA, F(2,23)=5.88, p=.009). VRN .3 + NAC 10mg/kg differs from SAL + NAC 10mg/kg (p=.03). VRN 1 + NAC 10mg/kg differs from SAL + NAC 10mg/kg (p=.006). Data shown as mean ± SEM, and N is shown above the bars. *p<.05, **p<0.01 using Sidak’s (2-way ANOVA) or Dunnett’s (One-way ANOVA) post hoc test. SAL = saline, VRN = varenicline, NAC = N-acetylcysteine.

Because the doses of VRN and NAC in the sequential combinations used in Experiment 3 did not affect behavior when administered alone (in Experiments 1 & 2), these data suggest a combined effect of VRN and NAC on reductions in nicotine seeking. Further, this combined effect may explain the meaningful reduction in late cue seeking, but the lack of effect in self-administration and only modest reduction in context seeking in early abstinence. That is, no meaningful effect of subthreshold VRN was found, but cue seeking was reduced when subthreshold NAC was preceded by subthreshold VRN. Thus, future studies should examine the effects of combined subthreshold VRN and NAC throughout treatment to determine their effects on nicotine use and relapse.

Experiment 4: Effects of sequentially combined subthreshold doses of VRN + NAC on sucrose self-administration, context seeking, and cued seeking:

To ensure that the effects of the sequentially combined subthreshold doses of VRN and NAC were specific to nicotine seeking, we also evaluated their effects on sucrose seeking under the same experimental conditions as Experiment 3. The sequentially-combined subthreshold doses of VRN and NAC had no effect on self-administration, early context seeking, or late cued seeking of sucrose (Fig 5).

Figure 5.

VRN 1.0 mg/kg + NAC 10 mg/kg does not affect sucrose intake or relapse. A) Active lever presses during sucrose self-administration. (2-way repeated-measures ANOVA, group F(1,14)=.02, p=.89, session F(2.59,36.22)=15.86, p<.0001, interaction F(9,126)=2.75, p=.06). Group Ns same as B. B) Active lever presses before (mean sessions 6–7) and after treatment (mean sessions 9–10; 2-way repeated-measures ANOVA, group F(1,14)=.02, p =.90, session F(1,14)=5.7, p=.03, interaction F(1,14)<.001, p=.99). C) Active lever presses during early context seeking (unpaired t-test, t(14)=.32, p=.75). D) Active lever presses during cued seeking with treatment (unpaired t-test, t(14)=.32, p=.76). Data shown as mean ± SEM, and N is shown above the bars. SAL = saline, VRN = varenicline, NAC = N-acetylcysteine.

Translational Utility and Mechanisms:

Using a translational animal model of nicotine seeking, we found that VRN, but not NAC was effective at reducing nicotine intake and promoting abstinence early in treatment. Conversely, NAC, but not VRN was effective at promoting abstinence late in treatment. Neither VRN nor NAC was effective at reducing cued seeking assessed 7 days after treatment, indicating that prior administration of neither drug was protective against relapse. Sequentially administering subthreshold doses of VRN during self-administration and early abstinence followed by NAC during late abstinence effectively reduced cued nicotine seeking later in treatment, suggesting a combined effect of of VRN and NAC. To determine if this effect was selective for nicotine, we evaluated VRN and NAC effects on sucrose seeking, and found no effect on sucrose self-administration, context seeking in early abstinence, or cued seeking late in abstinence. These results indicate that sequentially-combined VRN and NAC are selective for nicotine seeking, suggesting that this therapy may be used to reduce nicotine seeking in humans without influencing motivation for other non-nicotine rewards.

The possibility of combining NAC and VRN to treat nicotine addiction (tobacco use disorder) in humans has advanced only as far as a small open safety trial of non-treatment-seeking smokers that revealed no adverse effects of the drug combination and a modest reduction in cigarette use [32]. Relatedly, a similar clinical approach revealed that administration of another smoking cessation aid, bupropion, combined with NAC reduced several inflammatory and metabolic symptoms in heavy smokers [50]. These clinical results, combined with our preclinical data, suggest that sequentially combining VRN and NAC is a promising approach to reduce smoking and relapse.

VRN functional potency varies by 24-fold for α4β2, α3β4, and α7 receptors, and variation in efficacy ranges from low 15% for α4β2, to high 75–93% for α3β4 and α7, respectively[4]. This drug is a potent partial agonist of α4β2 receptor, less potent high efficacy agonist of α3β4, partial agonist for α6 and α3β2β3, and potent full agonist of α7 containing receptor. A potential mechanism for VRN’s action is that it binds to presynaptic α4- and α6-containing receptors that regulate dopamine release in the striatum [51]. Since the therapeutic use of VRN involves long-term exposure, the higher affinity of VRN for the α4β2 receptors keeps the receptors in a desensitized state, and can even result in channel blockage [4]. However, the potent high affinity and efficacy of activation of α7 receptors may also be important. While it is possible that any of these actions are critical for the effects of VRN, it is also possible that the combination of actions work in concert to achieve the therapeutic effects of assisting with smoking cessation.

A therapeutic approach combining VRN and NAC addresses at least three potential barriers to maintaining abstinence from smoking: positive reinforcement, negative reinforcement, and cue reactivity. Further, VRN attenuates the negative reinforcing properties of smoking by reducing amygdala hyperreactivity, an effect most pronounced in smokers with a high vulnerability to withdrawal-induced negative affect [11]. While VRN affects amygdala-centric circuitry, it does not affect frontostriatal circuitry thought to underlie cue-induced craving and relapse [14]. NAC, however, restores resting-state functional connectivity in this network [52], and restoration of frontostriatal connectivity is associated with reduced craving and smoking relapse compared with placebo-treated smokers [31]. Further, a recent study also found that NAC treatment normalized nicotine-induced downregulation of GLT-1 and glutamate cysteine exchanger expression in the prefrontal cortex of rats [53]. Thus, it appears that the combined VRN + NAC treatment approach proposed here confers effects in these separable circuits to reduce reinforcement and relapse of nicotine seeking.

The possibility of biologically complementary actions by VRN and NAC can explain the behavioral effect observed here; combining subthreshold doses of each drug produced a significant reduction in cued nicotine seeking. Cued drug seeking is well established to involve both amygdala- and frontal cortex to striatum circuits in animal models and human neuroimaging studies [54], and enduring neuroplasticity at glutamatergic synapses in the nucleus accumbens has been widely documented after repeated use of addictive drugs [19, 55]. Thus, the combined effect of NAC and VRN may result from differential effects on distinct aspects of plasticity in glutamatergic transmission in the nucleus accumbens. As mentioned above, the actions of NAC have been strongly tied to restoring glutamate regulation to reduce cued drug seeking, but no studies to date have investigated the effects of VRN on glutamatergic homeostasis in the nucleus accumbens. Regardless of the exact mechanisms, both VRN and NAC have reported, albeit generally mild, adverse effects [56, 57], supporting the potential utility of combining subthreshold doses of each to effectively reduce smoking and relapse while decreasing the incidence of adverse side effects.

Another point of potential translation from the present study is that VRN and NAC appear to be most effective when administered at different stages of nicotine use and withdrawal. Varenicline reduces nicotine intake and seeking in early withdrawal, which is consistent with clinical observations [5, 58]. In contrast, NAC was only effective at reducing nicotine seeking if given during withdrawal. This is consistent with preclinical [22–24, 59] and clinical [28, 31, 33] studies showing that NAC can reduce nicotine craving and smoking; although these effects have not been consistently observed [20, 28, 60]. Thus, regardless of the mechanisms, these data support a potential use of NAC to sustain smoking cessation once VRN treatment has initiated a period of abstinence.

One potential limitation of our study is that it is not entirely clear from our data whether the effects of VRN or NAC depend on time or the method used to induce relapse. While VRN effectively reduced nicotine intake and seeking during early abstinence and NAC reduced seeking late in abstinence, early and late tests also differed in terms of the type of relapse induced. Seeking during early abstinence was induced by a return to the context previously associated with nicotine, while late abstinence was induced by cues previously paired with nicotine. While this distinction is important from a basic research perspective, human smokers frequently encounter both types of stimuli throughout periods of abstinence, diminishing the translational utility of the distinction. Nevertheless, future research is needed to determine if the differential effects of VRN and NAC are dependent on length of the abstinence period or the method of seeking induction.

Conclusions:

Tobacco use is the leading cause of preventable premature death in the United States, highlighting a need for better treatments for those addicted to nicotine. Our translational preclinical data suggest that VRN and NAC have separable but complementary effects on nicotine seeking, and that sequentially combining both drugs is a promising therapeutic approach for maintaining nicotine abstinence while avoiding mild adverse effects sometimes encountered during treatment with VRN and NAC. It is likely that these effects are due to a more complete amelioration of smoking-associated brain circuit impairments, with VRN normalizing amygdala-striatal circuitry contributing to negative affect and NAC normalizing frontostriatal circuitry contributing to cue reactivity.