Exercise (Wheel-Running) Reduces Chronic Alcohol-Induced Escalation of Drinking in Mice via Elevated BDNF-TrkB Receptor Activity

TL;DR

Exercise reduces excessive alcohol drinking in mice by boosting brain chemicals linked to motivation and reward.

Contribution

This study shows that exercise mitigates chronic alcohol-induced drinking escalation via increased BDNF-TrkB signaling in mice.

Findings

Exercise increased BDNF expression in the medial prefrontal cortex of mice.

Exercise reversed chronic alcohol-induced BDNF deficits and reduced excessive drinking.

Blocking TrkB receptors reversed the protective effects of exercise on alcohol consumption.

Abstract

Alcohol use disorder (AUD) is a significant medical problem and there is great need for developing effective treatment strategies. Brain-Derived Neurotrophic Factor (BDNF) has been shown to play a role in regulating numerous pharmacological and motivational effects of alcohol. We have shown that chronic alcohol-induced escalation of drinking is accompanied by a deficit in BDNF levels in medial prefrontal cortex (mPFC). This study examined whether exercise (wheel-running) attenuates excessive alcohol drinking via increased BDNF expression, thereby mitigating the deficit in mPFC. Adult male C57BL/6J mice were given scheduled (2-hr/day) access to a running wheel in the home-cage 1-hr following opportunity to drink alcohol for 2-hr/day. After six weeks, mice were further separated into groups that received chronic alcohol vapor or control (air) inhalation exposure. Results indicated that alcohol consumption did not alter wheel-running and exercise did not alter alcohol intake during the 6-week baseline. Exercise increased BDNF mRNA and protein expression in mPFC, reversed chronic alcohol-induced reduction in BDNF levels, and attenuated escalated alcohol drinking. Systemic administration of a TrkB receptor antagonist (ANA-12) reversed the beneficial effects of wheel-running in the model. Together, these data provide support for exercise as a potentially effective intervention strategy for treating AUD.

Figures

A](#F1)). Two-way ANOVA indicated a significant main effect of Week [F(5,110) = 14.20, p < 0.0001, partial η2 = 0.3923] due to less running recorded during the first week compared to weeks 2–6. Compared to BDNF levels in the NWH group, BDNF protein expression in mPFC was significantly elevated 24-hr following the 6-week exercise regimen, with a trend of elevated levels at the 1-wk time point ([Fig. 1B](#F1)). In the DG, BDNF protein expression was increased at 24-hr, and this persisted at least 1-week following wheel-running ([Fig. 1C](#F1)). Analysis of mPFC BDNF protein levels indicated a ma

A](#F2)) and DG ([Fig. 2B](#F2)). Compared to controls (NWH-H2O group), both WH-H2O and WH-EtOH groups evidenced elevated *Bdnf* mRNA levels in mPFC for exon I [F(1,32) = 16.09, p < 0.001, partial η2 = 0.3346], exon IV [F(1,32) = 11.68, p = 0.0017, partial η2 = 0.2675], and exon IX [F(1,32) = 10.72, p = 0.0025, partial η2 = 0.2510]. Similarly, in DG ANOVAs revealed main effects of Wheel access for *Bdnf* exon I [F(1,30) = 18.71, p = 0.0001, partial η2 = 0.3841], exon IV [F(1,30) = 6.19, p = 0.0186, partial η2 = 0.1710], and exon IX [F(1,34) = 18.01, p = 0.0002, partial η2 = 0.3463]. Moderate l

B](#F4)). Analysis of these data indicated a significant main effect of Group [F(3,76) = 28.35, p < 0.0001, partial η2 = 0.5281] as well as the Group × Test interaction [F(9,228) = 2.53, p = 0.0087, partial η2 = 0.0909]. NWH-CIE mice evidenced an escalation of drinking, with alcohol intake increasing from 66% during Test 1 to more than doubling (105%) during Test 4 (ps < 0.05). This escalated drinking was significantly attenuated in CIE-exposed mice that had scheduled daily access to a running wheel in their home-cage during baseline and each test cycle (NWH-CIE > WH-CIE, ps < 0.05). Alcohol i

Full text

INTRODUCTION

Alcohol use disorder (AUD) constitutes a major public health concern in the United States and world-wide. Excessive alcohol use has serious medical consequences, and alcohol misuse results in tremendous economic, social, and personal costs to society [1–3]. AUD is a chronic relapsing disorder, and relapse represents a major challenge to treatment efforts. Despite significant advances in our understanding of biological mechanisms and environmental factors that influence motivation to engage in harmful levels of alcohol drinking, there remains a critical need for developing new and more effective therapeutic interventions for AUD [4–7].

Preclinical studies have played a key role in discovery of new targets with promising therapeutic potential [8]. Using our mouse model of alcohol dependence involving repeated cycles of chronic intermittent ethanol (CIE) vapor exposure, we conducted an unbiased genomic analysis of brain tissue and showed significant down-regulation in expression of Bdnf mRNA in medial prefrontal cortex (mPFC) of CIE-exposed mice compared to nondependent controls [9]. Brain-Derived Neurotrophic Factor (BDNF), a member of the neurotrophin family of growth factors, plays an important role in the development and maintenance of the nervous system. BDNF is widely distributed and highly expressed in mammalian brain, it is secreted from neurons in an activity-dependent manner, and BDNF interacts with tropomyosin receptor kinase B (TrkB) receptors and low-affinity p75 neurotrophin (p75NT) receptors to regulate a multitude of cellular processes [10]. BDNF binding at TrkB receptors results in phosphorylation of the receptor and activation of MAPK/ERK1/2, PLCg, and PI3K intracellular signaling pathways. BDNF signaling in brain has been implicated in a number of neuropsychiatric disorders, including AUD [11].

Humans with AUD are reported to have lower circulating levels of BDNF [12], although mixed results have been obtained in studies examining serum BDNF levels in abstinent individuals with a history of AUD diagnosis [13]. The TrkB (Ntrk2) gene locus on chromosome 9 was shown to be associated with alcohol dependence [14], and a functional polymorphism in the Bdnf gene (Val66Met) has been implicated in AUD, with evidence that the Val66 allelic variant confers higher risk for relapse [15] while the 66Met variant is associated with earlier onset and greater severity of alcohol dependence [16]. Mice carrying the homolog of the human 66Met allele (68Met) showed excessive alcohol consumption compared to wildtype Val68 mice [17]. Additional evidence for a connection between BDNF and alcohol-related behaviors comes from studies showing increased drinking in several genetic mouse and rat models with known deficiencies in BDNF expression [11]. Moderate alcohol consumption in mice was reported to increase Bdnf mRNA in the hippocampus [18] while high levels of intake achieved in a binge drinking rat model were associated with suppressed BDNF expression in hippocampus, an effect reversed by administration of a TrkB agonist [19]. Moderate alcohol intake increased Bdnf mRNA levels in dorsolateral (but not dorsomedial) striatum [20]. Also, while viral-mediated decreased BDNF expression in dorsal striatum increased alcohol drinking [20], increasing BDNF expression in this brain region reduced alcohol consumption, effects shown to be mediated by BDNF-TrkB receptor interaction [21]. A similar negative relationship between BDNF activity and alcohol consumption was shown for the central and medial (but not basolateral) nuclei of the amygdala [22]. We and others have shown reduced BDNF expression in mPFC following chronic alcohol self-administration [23, 24] and CIE vapor exposure [9, 24, 25]. We also showed that replenishing BDNF levels in the mPFC either by direct injection of the neurotrophin or viral-mediated overexpression of BDNF reduced or prevented, respectively, escalation of dependence-related alcohol drinking [26]. While elevating BDNF activity in brain appears to have therapeutic potential in treating AUD, from a clinical standpoint, systemically administering the neurotrophic factor is not practical since it does not readily penetrate the blood brain barrier [27]. Consequently, we turned our attention to an alternative approach that increases central BDNF activity – physical exercise.

Exercise has received increased attention as an effective, non-pharmacological treatment approach for various neuropsychiatric illnesses, including PTSD and alcohol/drug addiction. Physical activity (wheel-running, treadmill) has been shown to generally reduce self-administration of alcohol and other drugs such as cocaine, methamphetamine, heroin, and nicotine [28, 29]. Free (24-hr) access to wheel-running reduced home-cage alcohol consumption in several studies [30–35], although others reported different outcomes [36, 37]. Numerous factors likely explain differences in study outcomes, including species/genotype, duration of exercise, timing of exercise during different stages of addiction, and temporal relationship between access to exercise and alcohol/drug availability [28, 38, 39]. Sex is another factor that significantly influences the effects of exercise on alcohol/drug addiction in animals and humans [38, 40]. Clinical studies have reported exercise to reduce alcohol drinking [41–43] and reduce craving and relapse vulnerability in detoxified alcoholics [43–45]. The role of exercise in the treatment of AUD is likely multifaceted, as it improves physical and mental health [46–48], reduces stress [49] and, owing to its rewarding effects, it provides an alternative (rewarding-substitute) to alcohol/drug use [28, 31]. Indeed, the stress-alleviating effects of exercise have led to its consideration as an adjunct treatment for veterans seeking treatment for alcohol/drug addiction [50]. From a mechanistic perspective, exercise (aerobic and resistance) is known to elevate BDNF activity in brain [28, 40, 51, 52].

The present series of studies were conducted to test the general hypothesis that physical activity (wheel-running) reduces escalated alcohol consumption in our mouse model of alcohol dependence. It was hypothesized that exercise-induced increased BDNF production in brain would mitigate chronic alcohol-induced deficits in BDNF expression, thereby, attenuating escalated alcohol drinking. To more directly test whether this effect is related to increased BDNF activity at TrkB receptors, we examined whether administering a TrkB receptor antagonist would reverse the beneficial effects of exercise in our model of excessive alcohol drinking. As such, these studies aim to provide insight into the mechanisms and conditions in which exercise reduces alcohol consumption, generating valuable information on the clinical utility of employing exercise as a promising therapeutic intervention for those suffering with AUD.

MATERIALS and METHODS

Subjects

Male C57BL/6J mice (Jackson Laboratories, Bar Harbor, ME) were obtained at 10 weeks of age (25–27g) and were individually housed under a 12-hr light/dark cycle (lights off at 0800-hr). Mice were provided with food (Diet #2918; Harlan Teklad, Madison, WI) and water ad libitum in a temperature and humidity-controlled AAALAC-accredited facility at the Medical University of South Carolina.

Exercise: Scheduled Wheel-Running

Mice were given home-cage access to an activity wheel (Bioserve cat #S3174 and S3175) 2-hr/day beginning 6 hours in the dark cycle. The amount of exercise (distance ran) was tracked for each mouse and calculated using bike computers (Enkeeo Model BKV-1527) attached to the running wheels. The No-Wheel groups were left undisturbed in their home-cages.

Home Cage Limited Access Alcohol Drinking

Mice were given daily (Mon-Fri) free access to alcohol (15% v/v ethanol) in their home-cage, with water as the alternative fluid for 2-hr, starting 3-hr into the dark cycle. Alcohol intake (g/kg) was calculated based on the amount (ml) of the 15% ethanol solution consumed and the body weight recorded in the middle of each week. Alcohol was prepared fresh daily, and bottles were weighed before and after the 2-hour access period.

Chronic Intermittent Ethanol (CIE) Exposure

Mice were exposed to chronic intermittent alcohol (ethanol) vapor or air in inhalation chambers, as previously described [26, 53–56]. Prior to each day of alcohol vapor exposure, mice were administered 1.6 g/kg alcohol with 1 mmol/kg pyrazole, to stabilize blood alcohol concentrations (BAC). Air exposed mice received only pyrazole. Mice were placed in alcohol vapor or air inhalation chambers 16-hr/day for 4 days during each weekly exposure cycle. The concentration of alcohol in the chambers was monitored daily, and air and alcohol flow rates were adjusted to maintain BACs in the 200–250 mg/dl range throughout the exposure cycles. BACs were determined once a week immediately upon removal from the alcohol vapor chamber as previously described [26, 53, 56].

Bdnf mRNA and BDNF Protein Assays

As detailed in Supplemental Materials, medial prefrontal cortex (mPFC) and the dentate gyrus region of the hippocampus (HPC) were sectioned (2-mm) and dissected on ice, with the mouse brain atlas serving as a guide [57]. For Bdnf mRNA measurement, total RNA was extracted from samples and quantified, and quantitative real-time polymerase chain reaction (qRT-PCR) assays were performed using similar procedures to those previously published by our group [9]. Custom designed TaqMan qRT-PCR primers were used to specifically target and amplify Bdnf exon I, IV, and IX mRNA, as detailed in Supplemental Materials. Cyclophilin (Ppia) was used as the reference gene, and the 2^-DDCT method [58] was used to calculated fold change in expression of the target gene relative to the reference gene, with the NWH-CTL group as the reference condition. BDNF protein in separate mPFC and DG samples was assayed using BDNF Duo ELISA (R&D Systems) as per manufacturer’s instructions. Samples were run in duplicate in 96-well plates, coated with BDNF-capture antibody, and blocked with 1% bovine serum albumin (BSA). Data are expressed as pg BDNF/mg protein. Details of assay procedures are provided in Supplemental Materials.

Drugs

Alcohol and pyrazole were administered intraperitoneally (IP) in a volume of 0.02 mL/g body weight. The TrkB antagonist, ANA-12 (Selleckchem Cat no S7745), was resuspended in 1% (DMSO) in phosphate-buffered saline (PBS) (Sigma Aldrich) at 0.1 mg/mL for 1 mg/kg and 0.05 mg/mL for 0.5 mg/kg doses. All drugs (alcohol, pyrazole, ANA-12) and vehicle were administered intraperitoneally (IP) in a volume of 0.01 mL/g body weight.

Data Analysis

Experimental procedures were performed with experimenters unaware of subject group assignment to minimize any bias in collection of data. Data were analyzed using one-way or two-way ANOVAs followed by Newman-Keuls post-hoc comparisons, as detailed in each of the experiments. Alpha was set at 0.05 to assess significant effects. Analyses were conducted using Statistica (Tibco Software Inc.; Palo Alto, CA).

RESULTS

While some studies have shown exercise reduces home-cage alcohol preference and consumption in mice, in many cases access to wheel-running and alcohol was continuous (24-hr) and simultaneous [31–34]. Since these are essentially mutually exclusive behaviors (wheel-running vs. drinking alcohol (or water) from a sipper tube), we chose to examine an exercise regimen that involved limited access (2-hr/day) to running wheels in the home-cage. Further, to ensure that exercise did not merely alter alcohol intake due to elevated exercise-induced thirst, alcohol was presented for 2-hr/day in the home-cage in a free choice (2-bottle) situation (15% ethanol vs. water) 1-hr prior to the 2-hr access to running wheels.

Limited Daily Access to Wheel-Running Increases BDNF Protein Expression in Medial Prefrontal Cortex and Dentate Gyrus

To determine whether daily scheduled access to an activity wheel in the home-cage increased BDNF expression in brain, mice were given access to activity wheels (WH groups) or not (NWH group) in the home cage for 2-hr/day over 6 weeks. Mice were sacrificed at 24-hr (WH-24-hr) or 1 week (WH-1-wk) following the last session (half the NWH group was sacrificed at each of these time points and then combined for analysis) (N = 12 mice/group). Bilateral samples of medial prefrontal cortex (mPFC) and dentate gyrus (DG) were processed for protein analysis by ELISA. mPFC was targeted because we had previously shown that chronic alcohol exposure reduced BDNF levels in this brain region [9, 25]. The DG was included in our analysis since this brain region has been shown to be highly responsive to exercise effects on BDNF activity [59, 60].

Wheel-running stabilized over the 6-week period (2.5–3 km/2-hr session) and did not differ between groups that were sacrificed at 24-hr or 1-wk following the last exercise session (Fig. 1A). Two-way ANOVA indicated a significant main effect of Week [F(5,110) = 14.20, p < 0.0001, partial η2 = 0.3923] due to less running recorded during the first week compared to weeks 2–6. Compared to BDNF levels in the NWH group, BDNF protein expression in mPFC was significantly elevated 24-hr following the 6-week exercise regimen, with a trend of elevated levels at the 1-wk time point (Fig. 1B). In the DG, BDNF protein expression was increased at 24-hr, and this persisted at least 1-week following wheel-running (Fig. 1C). Analysis of mPFC BDNF protein levels indicated a main effect of Group [F(2,22) = 4.91, p = 0.01729, partial η2 = 0.3085] and post-hoc analyses indicated that BDNF levels were significantly higher in the WH-24hr group compared to the NWH group (p = 0.0139). Similarly, analysis of DG samples indicated a significant main effect of Group [F(2,26) = 6.36, p = 0.0057, partial η2 = 0.3284], with post-hoc analyses indicating elevated levels of BDNF protein at both the 24-hr (p = 0.0201) and 1-wk (p = 0.0062) time points following exercise.

Effects of Alcohol Drinking on Wheel-Running and Bdnf mRNA Levels in mPFC

Next we determined whether limited access to alcohol drinking influenced wheel-running and its ability to elevate Bdnf mRNA expression in the mPFC. Mice were divided into 4 groups based on scheduled home-cage wheel and/or alcohol access (N = 9–12/group). One group was given access to alcohol (15% v/v) for 1-hr starting at 3-hr into the dark cycle with no opportunity to exercise (NWH-EtOH group); a second group was treated the same but was given access to activity wheels for 2-hr/day for 6 weeks, starting 1-hr after the alcohol bottle was removed (WH-EtOH group); another group was given access to the same regimen of wheel-running without alcohol access (WH-H2O group); and the final group served as controls, receiving neither access to running wheels nor alcohol (NWH-H2O group).

Wheel-running (2-hr/day/5-days for 6 weeks) was not significantly altered by alcohol consumption that occurred 1-hr before access to the activity wheels. That is, wheel-running was stable and did not differ between WH-EtOH and WH-H2O groups over the 6-week study (Supplemental Fig. 1). Scheduled access to wheel-running did not significantly influence alcohol intake – i.e., alcohol consumption was similar for WH-EtOH and NWH-EtOH groups (Supplemental Fig. 2). Analysis of body weight registered each week indicated a steady increase that did not differ among groups (Supplemental Fig. 3).

At 24-hr following the final session, all mice were sacrificed, and mPFC and DG samples were dissected and frozen until qRT-PCR analysis was performed targeting Bdnf exon I, exon IV, and exon IX. Bdnf exon I and exon IV transcripts were selected because they have been shown to be especially responsive to exercise and chronic alcohol, respectively [24, 25, 61], and Bdnf exon IX encodes the pre-proBDNF protein [62, 63]. Wheel-running with or without home-cage access to alcohol over 6 weeks significantly increased Bdnf mRNA expression (exons I, IV, and IX) in mPFC (Fig. 2A) and DG (Fig. 2B). Compared to controls (NWH-H2O group), both WH-H2O and WH-EtOH groups evidenced elevated Bdnf mRNA levels in mPFC for exon I [F(1,32) = 16.09, p < 0.001, partial η2 = 0.3346], exon IV [F(1,32) = 11.68, p = 0.0017, partial η2 = 0.2675], and exon IX [F(1,32) = 10.72, p = 0.0025, partial η2 = 0.2510]. Similarly, in DG ANOVAs revealed main effects of Wheel access for Bdnf exon I [F(1,30) = 18.71, p = 0.0001, partial η2 = 0.3841], exon IV [F(1,30) = 6.19, p = 0.0186, partial η2 = 0.1710], and exon IX [F(1,34) = 18.01, p = 0.0002, partial η2 = 0.3463]. Moderate level of alcohol consumption alone (NWH-EtOH group) did not significantly alter BDNF transcriptional activity in either brain region and the EtOH X Wheel access interaction did not achieve significance in mPFC or DG. Thus, limited daily access (2-hr/day) to activity wheels over 6 weeks increased Bdnf mRNA expression in mPFC and DG, and limited home-cage access to alcohol (2-hr/day) did not modify this effect.

Wheel-Running Attenuates Chronic Alcohol (CIE)-Induced Escalated Alcohol Drinking

In a separate cohort of mice, we examined the effects of exercise in our CIE Drinking model. Mice were given free access to alcohol (15% v/v) in the home-cage 2-hr/day 5-days/week during a 6-week baseline period, with half the mice presented with a running wheel for 2-hr in the home-cage 1-hr after alcohol was removed. The remaining mice were left undisturbed. WH and NWH groups were then divided into two groups, receiving either weekly cycles of CIE exposure (NWH-CIE and WH-CIE groups) or control (air) exposure (NWH-CLT and WH-CTL groups) alternating with weekly test drinking days (N = 20/group) (Fig. 3). Wheel-running access and alcohol drinking were suspended during inhalation exposure cycles.

Analysis of wheel-running during the 6-week baseline period revealed a significant main effect of Week [F(5,170) = 11.56, p < 0.0001, partial η2 = 0.2537] due to greater average distance recorded during weeks 4, 5, and 6 compared to the first 3 weeks (p < 0.05) (Supplemental Fig. 4). Analysis of wheel-running during test cycles revealed significant main effects of Group [F(1,34) = 16.79, p = 0.0002, partial η2 = 0.3306], Test [F(3,102) = 11.83, p < 0.0001, partial η2 = 0.2581], and the Group × Test interaction [F(3,102) = 4.11, p = 0.0085, partial η2 = 0.1077]. Post-hoc analysis indicated that wheel-running was lower in the WH-CIE group compared to WH-CTL mice during the last two test cycles (ps < 0.01) (Supplemental Fig. 4). Wheel-running did not appear to influence alcohol metabolism, as suggested by the similar blood alcohol levels produced during each of the CIE exposure cycles (Supplemental Table 1). Mice with wheel access registered lower body weights that those without at the end of baseline and this persisted throughout all test weeks. NWH-CIE mice showed a decrease in weight compared to the NWH-CTL group during test weeks 1, 3, and 4 (Group × Test interaction: [F(12,304) = 8.54, p < 0.0001, partial η2 = 0.2520; ps < 0.05] (Supplemental Table 2).

CIE exposure resulted in escalation of alcohol consumption compared to CTL groups, but providing scheduled opportunity to exercise attenuated this increase in drinking (Fig. 4A). Analysis of alcohol intake data revealed a significant main effect of Group [F(3,76) = 25.06, p < 0.0001, partial η2 = 0.4972] due to CIE-exposed groups consuming overall more alcohol compared to CTL groups (ps < 0.05). ANOVA also revealed a significant Group × Test interaction [F(12,304) = 9.14, p < 0.0001, partial η2 = 0.2651]. Post-hoc comparisons indicated that alcohol intake was similar for all groups at the end of the baseline period. Alcohol intake was significantly higher for the NWH-CIE group compared to their baseline level of intake and consumption in the NWH-CTL group during each Test cycle (NWH-CIE > NWH-CTL; ps < 0.05). Exercise significantly reduced this effect during Tests 2, 3, and 4 (NWH-CIE > WH-CIE; ps < 0.05). Alcohol intake remained relatively stable across Baseline and Test cycles for the NWH-CTL group, and wheel-running alone (WH-CTL) did not alter alcohol drinking throughout the study.

A similar profile of results was obtained when data were expressed as a change from each animal’s baseline level of intake (Fig. 4B). Analysis of these data indicated a significant main effect of Group [F(3,76) = 28.35, p < 0.0001, partial η2 = 0.5281] as well as the Group × Test interaction [F(9,228) = 2.53, p = 0.0087, partial η2 = 0.0909]. NWH-CIE mice evidenced an escalation of drinking, with alcohol intake increasing from 66% during Test 1 to more than doubling (105%) during Test 4 (ps < 0.05). This escalated drinking was significantly attenuated in CIE-exposed mice that had scheduled daily access to a running wheel in their home-cage during baseline and each test cycle (NWH-CIE > WH-CIE, ps < 0.05). Alcohol intake remained relatively unchanged throughout the study in NWH-CTL and WH-CTL groups.

Wheel-Running Reverses Chronic Alcohol-Induced Deficits in mPFC Bdnf mRNA Expression

A subset of mice from each group were randomly selected to receive a 5th cycle of CIE (or air) exposure and were then sacrificed to determine Bdnf mRNA expression in mPFC (N = 7–11/group) (Fig. 5). ANOVA revealed significant group differences for Bdnf exon I [F(3,31) = 21.50, p < 0.0001, partial η2 = 0.6754], exon IV [F(3,31) = 33.48, p < 0.0001, partial η2 = 0.7642], and exon IX: [F(3,31) = 27.64, p < 0.0001, partial η2 = 0.7279]). Post-hoc comparisons indicated that overall, CIE exposure alone (NWH-CIE group) produced a deficit in Bdnf mRNA levels relative to the NWH-CTL condition (exon I: p = 0.0361; exon IV: p = 0.0036]; exon IX: p = 0.0001). Exercise alone (WH-CTL group) significantly elevated Bdnf mRNA expression in mPFC (WH-CTL > NWH-CTL) (exon I: p = 0.0001; exon IV: p = 0.0001; exon IX: p = 0.0020). In turn, reduced Bdnf mRNA expression in the mPFC produced by CIE exposure was ‘normalized’ when mice were given the opportunity to exercise and this served to completely block the deficit in Bdnf mRNA levels produced by CIE exposure (WH-CIE > NWH-CIE) (exon I; p = 0.0166; exon IV: p = 0.0031; exon IX: p = 0.0004). This suggests that BDNF activity plays an important role in mediating escalation of drinking associated with chronic alcohol exposure.

TrkB Receptor Antagonism Blocks the Ability of Exercise to Attenuate CIE-Induced Escalated Drinking

To more directly investigate whether increased BDNF activity at TrkB receptors contributes to the ability of exercise to reduce chronic alcohol-induced escalated drinking, we examined the effects of a TrkB receptor antagonist, ANA-12, in the CIE Drinking model. ANA-12 is a non-competitive TrkB receptor-selective antagonist ANA-12 [64] and it has been shown to reverse various BDNF-mediated effects relevant to addiction-related behaviors [65, 66]. Additionally, ANA-12 was shown to block the BDNF-related ameliorative effects of exercise [67–69]. Mice were separated into NWH-CTL, NWH-CIE, WH-CTL, and WH-CIE groups and treated as described in the previous study (N = 8–11/group). After the 2nd CIE/Air exposure cycle, the groups were further separated into drug treatment conditions, receiving injections of Vehicle, 0.5, or 1.0 mg/kg ANA-12 immediately following the 2-hr daily wheel-running sessions.

Analysis of alcohol consumption during Test 2 indicated a significant main effect of Group [F(3,102) = 34.33, p < 0.0001, partial η2 = 0.5024] and a significant Group × ANA-12 Dose interaction [F(6,102) = 3.09, p < 0.0081, partial η2 = 0.1537] (Fig. 6). Post-hoc comparisons revealed higher alcohol intake in CIE-exposed mice compared to CTL mice with no access to running wheels (NWH-CIE > NWH-CTL) (p < 0.05), and ANA-12 treatment did not alter this effect. Replicating our previous finding, exercise (wheel-running) significantly reduced alcohol intake in CIE-exposed mice that received Vehicle (NWH-CIE > WH-CIE; p < 0.05). However, the dampening effects of exercise on CIE-induced elevated drinking was eliminated in mice that were treated with the TrkB antagonist ANA-12. That is, alcohol intake in WH-CIE mice treated with either 0.5 mg/kg or 1.0 mg/kg ANA-12 did not differ from the corresponding NWH-CIE groups and these mice consumed significantly more alcohol in comparison to the NWH-CTL group (ps < 0.05). These results support the notion that exercise attenuates chronic alcohol-induced escalation of drinking by increasing BDNF activity at TrkB receptors.

DISCUSSION

Results from this study demonstrate that providing mice with the opportunity to exercise 2-hr/day in their home-cage over several weeks is sufficient to elevate BDNF mRNA and protein expression in the hippocampus (dentate gyrus) and medial PFC. Chronic alcohol (CIE) exposure produced the expected escalation of alcohol drinking and, as we previously showed, this was accompanied by a significant reduction in Bdnf mRNA levels in mPFC. However, daily scheduled exercise attenuated escalated drinking in CIE-exposed mice without significantly altering alcohol intake in nondependent (CTL) mice. Further, the daily scheduled exercise regimen mitigated chronic alcohol-induced reduction in Bdnf mRNA levels in mPFC. Finally, to more directly implicate a role for BDNF-TrkB receptor activity in this effect, the TrkB receptor antagonist ANA-12 was shown to block the ability of exercise to reduce CIE-induced escalated alcohol drinking.

It has been previously reported that continuous (unlimited) home-cage access to running wheels significantly increased BDNF mRNA and protein expression in the DG, an exercise responsive region [59, 60]. Complementing these findings, results from this study show that scheduled limited daily access (2-hr/day) to wheel-running over 6 weeks significantly increased BDNF mRNA and protein expression in the DG as well as the mPFC. The activity wheels were provided in the home-cage 1-hr following a 2-hr period of free-choice alcohol drinking. The temporal order and separation of these behaviors was implemented to avoid any immediate influence exercise may have on alcohol drinking due to the need for more calories or a general increase in thirst. Wheel-running did not alter alcohol consumption during a 6-week baseline period. Further, access to alcohol prior to wheel running did not affect running behavior (which remained relatively stable over 6 weeks), and alcohol intake did not influence the ability of exercise to elevate BDNF transcriptional activity in target brain regions.

The limited access exercise regimen selectively reduced excessive levels of alcohol drinking produced by CIE exposure while not altering more moderate stable levels of alcohol intake in non-dependent (CTL) mice. While unlimited access to voluntary wheel-running has generally been shown to reduce alcohol consumption [30–35], these studies have typically involved concurrent unlimited access to wheel running and alcohol. Other studies have examined effects of exercise on alcohol consumption under limited access conditions or when wheel-running was restricted to periods of abstinence, with mixed results [33, 35, 37]. Differences in study outcomes are likely related to various biological variables (e.g., species, strain, sex) as well as a host of experimental procedural factors (e.g., duration and temporal relationship of both exercise and alcohol access) [28, 38–40]. All these studies have examined the effects of physical activity on alcohol consumption using drinking models in nondependent animals. Here we demonstrate that scheduled limited access to wheel-running attenuates escalated alcohol drinking in dependent (CIE-exposed) mice. To our knowledge, there is only one other study that has examined the effects of wheel-running in dependent animals. Somkuwar et al. [70]reported that wheel-running access following CIE exposure in rats reduced alcohol relapse-like behavior using an operant reinstatement model. As such, these results complement findings of the present study and provide support for the potential therapeutic value of exercise in reducing excessive alcohol drinking and relapse vulnerability.

Studies have demonstrated a negative relationship between alcohol drinking behavior and BDNF expression in hippocampus [19], striatum [20, 21], and amygdala [22]. In this study we show that increased alcohol consumption following chronic alcohol (CIE) exposure was associated with decreased Bdnf mRNA expression in mPFC. This aligns with our previous findings and those of others showing reduced BDNF expression in mPFC following chronic alcohol exposure [9, 23–25]. Here we demonstrate that daily limited access to wheel-running not only elevated BDNF expression in mPFC but also reversed the chronic alcohol-induced reduction in Bdnf mRNA levels and attenuated escalated voluntary alcohol drinking. These results support the notion that reversing the deficit in BDNF expression following chronic alcohol exposure can reduce alcohol consumption, and exercise may be an effective means to produce this effect.

Our previous work showed that replenishing BDNF levels by direct infusion of the neurotrophin into the mPFC reduced CIE-related escalation of drinking, Further, viral-mediated overexpression of BDNF in the mPFC was effective in preventing the emergence of escalated drinking in CIE-exposed mice [26]. In this study, physical activity (wheel-running) produced a similar effect in reducing CIE-induced elevated alcohol consumption. To further interrogate the hypothesized link between exercised-induced elevated BDNF expression and reduced alcohol drinking, we examined whether this effect of exercise could be reversed by administrating the TrkB receptor antagonist ANA-12. ANA-12 was shown to be effective in reversing the ability of exercise to reduce CIE-induced excessive drinking. ANA-12 treatment did not alter drinking in non-dependent mice. Similarly, elevating BDNF levels in mPFC did not alter alcohol intake in non-dependent mice [26]. Together, these results suggest that BDNF-TrkB signaling in the mPFC may be most relevant in modulating alcohol consumption under conditions that engender excessive levels of intake (i.e., models of chronic alcohol exposure that produce deficits in BDNF expression). However, since ANA-12 was systemically administered, it cannot be assumed that this outcome was due to blocking BDNF-TrkB receptor activity in the mPFC. Future studies will need to target pharmacological antagonism or genetic deletion of the receptor in specific brain sites to address this issue.

A limitation of this work is that only male subjects were evaluated in these experiments. Sex-related differences in the ability of physical activity to modulate alcohol/drug self-administration have been reported [38, 40]. Sex has also been shown to influence the ability of exercise to induce central BDNF transcriptional activity [34, 71]. Thus, it will be important to determine whether the effects of exercise on BDNF expression and CIE-related escalated alcohol drinking observed in males in this series of studies are similar or differ in females.

Finally, the present set of studies did not control for the presence of activity wheels in the home-cage. The presence of activity wheels may be viewed as providing environmental enrichment, a manipulation that is known to influence alcohol consumption [72] as well as increasing BDNF levels in brain [73]. Some studies have used a locked wheel to address this issue. These studies have produced mixed results, with some reports showing that presenting a locked wheel to animals did not alter alcohol intake [33, 36] while other studies showed an increase in alcohol intake [30, 74]. Future studies will need to investigate whether blocking access to physical activity (locked wheel) alters BDNF levels and excessive alcohol consumption associated with dependence.

In summary, daily limited access to physical activity (wheel-running) was effective in significantly reducing escalation of alcohol drinking in chronic alcohol (CIE) exposed mice. This exercise regimen was shown to elevate BDNF levels in brain as well as reverse chronic alcohol-induced deficits in Bdnf mRNA expression in mPFC. Additionally, systemic administration of the TrkB receptor antagonist ANA-12 blocked the ability of exercise to attenuate CIE-related excessive alcohol drinking, thereby implicating increased BDF-TrkB receptor activity as a significant contributor to this effect. Together, these results support the emerging evidence that exercise may be an effective adjunct to treatment for individuals with alcohol use disorder.

Supplementary Files

This is a list of supplementary files associated with this preprint. Click to download.

26WHCIEBDNFSupplFinal.docx

26TPSuppFig1.tif

26TPSuppFig2.tif

26TPSuppFig3.tif