The First Alcohol Drink Triggers mTORC1-Dependent Synaptic Plasticity in Nucleus Accumbens Dopamine D1 Receptor Neurons

Abstract

Early binge-like alcohol drinking may promote the development of hazardous intake. However, the enduring cellular alterations following the first experience with alcohol consumption are not fully understood. We found that the first binge-drinking alcohol session produced enduring enhancement of excitatory synaptic transmission onto dopamine D1 receptor-expressing neurons (D1+ neurons) in the nucleus accumbens (NAc) shell but not the core in mice, which required D1 receptors (D1Rs) and mechanistic target of rapamycin complex 1 (mTORC1). Furthermore, inhibition of mTORC1 activity during the first alcohol drinking session reduced alcohol consumption and preference of a subsequent drinking session. mTORC1 is critically involved in RNA-to-protein translation, and we found that the first alcohol session rapidly activated mTORC1 in NAc shell D1+ neurons and increased synaptic expression of the AMPAR subunit GluA1 and the scaffolding protein Homer. Finally, D1R stimulation alone was sufficient to activate mTORC1 in the NAc to promote mTORC1-dependent translation of the synaptic proteins GluA1 and Homer. Together, our results indicate that the first alcohol drinking session induces synaptic plasticity in NAc D1+ neurons via enhanced mTORC1-dependent translation of proteins involved in excitatory synaptic transmission that in turn drives the reinforcement learning associated with the first alcohol experience. Thus, the alcohol-dependent D1R/mTORC1-mediated increase in synaptic function in the NAc may reflect a neural imprint of alcohol's reinforcing properties, which could promote subsequent alcohol intake. Significance statement: Consuming alcohol for the first time is a learning event that drives further drinking. Here, we identified a mechanism that may underlie the reinforcing learning associated with the initial alcohol experience. We show that the first alcohol experience induces a persistent enhancement of excitatory synaptic transmission on NAc shell D1+ neurons, which is dependent on D1R and mTORC1. We also find that mTORC1 is necessary for the sustained alcohol consumption and preference across the initial drinking sessions. The first alcohol binge activates mTORC1 in NAc D1+ neurons and increases levels of synaptic proteins involved in glutamatergic signaling. Thus, the D1R/mTORC1-dependent plasticity following the first alcohol exposure may be a critical cellular component of reinforcement learning.

Keywords: addiction; alcohol; dopamine; mTOR; plasticity.

Figure 1.

Alcohol increases the AMPAR/NMDAR ratio in D1+ neurons but not D2+ neurons selectively in the NAc shell. Patch-clamp recordings in D1+ and D2+ neurons in the NAc shell 24 h after an alcohol challenge. Drd1-Cre/Ai14 and Drd2-Cre/Ai14 mice underwent a single 24 h two-bottle choice (2BC) 20% alcohol session (Alcohol) or water only (Water; B), or were administered a single intraperitoneal administration of alcohol (2 g/kg) or saline (C, D). A, An example of two patched neurons, D1+ (left) and D1− (right). Top left image is bright field, top right image is TdTomato (Ai14+) and the bottom is a merged image. Arrow on bottom image points to pipette that has formed a seal on a D1− neuron. B, Top left, Time course showing cumulative amount of alcohol consumed by Drd1-Cre/Ai14 and Drd2-Cre/Ai14 mice (n = 5) within the 24 h voluntary drinking session. Top right, Bar graphs show average AMPAR/NMDAR ratio in D1+ and D2+ neurons following voluntary alcohol session. Two-way ANOVA, Bonferroni multiple comparison, *p < 0.05; D1+: water versus alcohol; n (neurons) = D1+: 7 water, 9 alcohol; D2+: 6 water, 9 alcohol. Bottom, Representative AMPAR and NMDAR currents in D1+ (black) and D2+ neurons (gray) after water or alcohol. C, D, Left, Bar graphs show average AMPAR/NMDAR ratio in D1+ and D2+ neurons following alcohol injection. Two way-ANOVA, Bonferroni multiple comparison, *p < 0.05; D1+: saline versus alcohol; n (neurons): NAc Shell (C) = D1+: 7 saline, 10 alcohol; D2+: 8 saline, 8 alcohol. NAc Core (D) = D1+: 9 saline, 9 alcohol; D2+: 8 saline, 8 alcohol. Right, Representative current traces in D1+ (black) and D2+ neurons (gray) after saline or alcohol. Calibration: 20 ms, 50 pA.

Figure 2.

Alcohol increases mEPSC current amplitudes and alters the AMPAR rectification index in D1+ neurons but not D2+ neurons. Patch-clamp recordings in NAc shell D1+ and D2+ neurons 24 h after subjects received a single intraperitoneal alcohol injection (2 g/kg). A, Top left, Bar graph showing effect of alcohol on mEPSC amplitude. Two-way ANOVA, Bonferroni multiple comparison, **p < 0.01; D1+: saline versus alcohol; n (neurons) = D1+: 7 saline, 7 alcohol, D2+: 6 saline, 7 alcohol. Bottom left, Effect of alcohol on mEPSC frequency. Right, Representative traces showing mEPSC events from D1+ neurons (black) and D2+ neurons (gray) after saline or alcohol. Calibration 500 ms, 25 pA. B, Top, Bar graph shows average AMPAR PPR following alcohol. Bottom, Representative traces of evoked AMPAR currents evoked 50 ms apart. Stimuli artifacts are removed for clarity. Calibration: 25 ms, 50 pA. n (D1+ neurons) = 5 saline, 7 alcohol. C, Top, Bar graphs show average AMPAR rectification index following alcohol. Two-way ANOVA, Bonferroni multiple comparison, **p < 0.01; D1+: saline versus alcohol. n (neurons) = D1+: 8 saline, 9 alcohol, D2+: 8 saline, 9 alcohol. Bottom, Representative AMPAR traces at holding potentials of +40 (outward current), 0, and −70 mV (inward current) from D1+ neurons (black) and D2+ neurons (gray) after saline or alcohol. Calibration: 25 ms, 50 pA.

Figure 3.

mTORC1 is required for the alcohol-induced enhancement of the AMPAR/NMDAR ratio in D1+ neurons and for reinforcement learning triggered by the first alcohol drinking session. A, Patch-clamp recordings from D1+ neurons in the NAc shell 24 h following an alcohol (2 g/kg) or saline injection. Subjects were pretreated with either vehicle (3% DMSO) or rapamycin (10 mg/kg) 3 h before, or D1R antagonist SCH-23390 (0.1 mg/kg) 30 min before alcohol injection. Bar graphs depict the average AMPAR/NMDAR ratio after alcohol challenge. Two-way ANOVA, Bonferroni multiple comparison, **p < 0.01; Vehicle: saline versus alcohol; n (D1+ neurons) = 6 vehicle/saline, 11 vehicle/alcohol, 6 rapamycin/saline, 8 rapamycin/alcohol, 5 SCH23390/saline, 7 SCH23390/alcohol. B, Alcohol consumption and preference during the first two 4 h sessions of two-bottle choice for alcohol and water. Top, Schematic representation of the experimental procedure. Mice received intraperitoneal administration of vehicle or rapamycin (10 mg/kg) 3 h before a 4 h two-bottle choice session (black), and then 24 h later, mice were given another 4 h two-bottle choice session. Middle bottom, Bar graphs show alcohol consumption (middle) and alcohol preference (bottom) on days 1 and 2 of vehicle (white) and rapamycin (gray) groups. **p < 0.01; n = 12 vehicle, 15 rapamycin.

Figure 4.

A single alcohol drinking session activates mTORC1 and increases protein expression in the synaptic fraction. Western blot analyses of the total homogenate (B, D) and synaptic fraction (C) immediately after a single 4 h two-bottle choice session for alcohol (black) or water only (white). B, right, and C, D, bottom, Representative images of the bar graph analyses. A, Scatter plot showing the relationship between alcohol consumption (g/kg) and BAC (g/dL) from the 4 h alcohol two-bottle choice session. BAC assay was run in duplicate or triplicate (black points). Mean consumption and BAC (±SEM) is signified by the gray point. Centerline is the linear regression and outer curved lines are the 95% confidence intervals. B, Bar graphs depict the changes in phosphorylation levels of Akt, GSK3β, Erk2 (lower band), 4E-BP, S6K, and S6 in the total homogenate following alcohol access. Data are expressed as the optical density ratio of phosphoprotein to total protein, normalized to water control. Data are expressed as mean ± SEM. C, Bar graphs show the average changes in the phosphorylation of S6 (left) or total levels of GluA1, Homer, and PSD-95 (right) in the synaptic fraction in response to alcohol. For total protein levels, data are expressed as the optical density ratio of total protein to GAPDH, normalized to control. D, Bar graph shows average change in phospho-GluA1 (Ser485; left), and the change in total GluA1 (right), following alcohol drinking. A, R² = 0.637; slope is significantly different from zero: F_(1,17) = 29.9; p < 0.0001. B, ***p < 0.001, **p < 0.01, *p < 0.05. C, **p = 0.0052, *p < 0.05. D, *p = 0.0484. A, n = 8 mice. B, C, 4E-BP and S6K: n = 4 water, 5 alcohol; all others: n = 6 water, 8 alcohol. D, n = 3 water, 3 alcohol.

Figure 5.

A single voluntary alcohol session increases pS6 immunoreactivity only in D1+ neurons in the NAc shell. Immunohistochemical analyses of phospho-S6 levels in D1+ and D2+ neurons in the medial NAc shell and NAc core following a 4 h two-bottle choice alcohol session (Alcohol) or water only (Water). A, Representative coronal section that contains the medial portion of the NAc shell (shaded black) and NAc core (red). Image courtesy of Allen Mouse Brain Atlas (Lein et al., 2007). B, C, Bar graphs depict the average changes in phospho-S6 immunoreactivity as a ratio of total neurons (left) or of D1+ and D2+ neurons (right) following alcohol access. Data are expressed as mean ± SEM. B, Left, *p < 0.05; Right, Mixed ANOVA, Bonferroni multiple comparison, **p < 0.01 D1+: water versus alcohol. D, E, Representative images from the NAc shell (D) and core (E) with labeling of pS6 (green) and D1+ neurons (red) in a Drd1-Cre/Ai14 subject that received only water (top) or alcohol (bottom). All images also contained NeuN immunoreactivity, but that channel was removed for clarity. Yellow arrow indicates colocalization of pS6 and TdTomato (D1+ neuron). White arrow indicates pS6 alone. Calibration: 25 μm. B, C, n: 7 water, 5 alcohol.

Figure 6.

The D1 agonist SKF-81927, but not the D2 agonist quinpirole, activates Akt and mTORC1 in the NAc. Western blot analyses 30 min after intraperitoneal injection of 3% DMSO (vehicle) or 5 mg/kg SKF-81927 (A), or 0.9% saline (vehicle) or 5 mg/kg quinpirole (B). Right, Representative blots of bar graph analyses. The vertical line that separates groups in quinpirole experiment indicates that blots were from the same gel but were not run in adjacent lanes. A, B, Bar graphs depict the average changes of Akt, GSK3β, Erk2 (lower band), 4E-BP, S6K, and S6 phosphorylation in response to SKF (A) or quinpirole (B) treatment. Data are expressed as the optical density ratio of phosphoprotein to total protein and normalized to vehicle control, and are expressed as mean ± SEM. A, **p < 0.01, ***p < 0.001. A, 4E-BP: n = 4 vehicle, 5 SKF; S6K: n = 4 vehicle, 5 SKF; All others: n = 6 Vehicle, 6 SKF. B, n = 5 Vehicle, 4 Quinpirole.

Figure 7.

D1R activation promotes the translation of GLUA1, HOMER2, and PSD-95. RT-PCR of polysomal RNA (B) and total RNA (C) in the NAc 30 min after intraperitoneal SKF (5 mg/kg) or vehicle (3% DMSO). A, Isolation of polysomes by sucrose gradient centrifugation. Left, Seventy fractions from the 15–45% sucrose gradient were analyzed by absorbance at 254 nm to measure RNA content. The addition of EDTA disrupts polysome formation and shifts the elution profile. Right, Map of approximate location of each fraction along the sucrose gradient. B, C, Bar graphs show average changes of GLUA1, HOMER2, and PSD-95 following SKF treatment. Data are expressed as a ratio to total GAPDH, normalized to vehicle group, and are expressed as mean ± SEM. B, **p < 0.01. *p < 0.05. B, C, n = 3 vehicle, 3 SKF.

Figure 8.

D1R-dependent increases in GluA1 and Homer protein levels require mTORC1 activity. Western blot analyses in the NAc 30 min after intraperitoneal administration of SKF or vehicle (3% DMSO) from subjects that were pretreated 3 h prior with rapamycin (i.p., 10 mg/kg) or vehicle (3% DMSO). A, Top, Representative blots from groups following vehicle or rapamycin injection and then vehicle or SKF injection. Bottom, Bar graphs display the average changes in GluA1, Homer, or PSD-95 following pretreatment of rapamycin or vehicle, followed by treatment of SKF or vehicle. Data are expressed as an optical density ratio of total protein to GAPDH, normalized to the vehicle/vehicle group, and are expressed as mean ± SEM. GluA1: two-way ANOVA, Bonferroni multiple comparison, **p < 0.01 veh/veh versus veh/SKF; *p < 0.05 veh/SKF versus rapamycin/SKF. Homer: two-way ANOVA, Bonferroni multiple comparison, **p < 0.01 veh/veh versus veh/SKF and veh/SKF versus rapamycin/SKF. PSD-95: two-way ANOVA Bonferroni multiple comparison, *p < 0.05 veh/veh versus veh/SKF and rapamycin/veh versus rapamycin/SKF. B, Left, Bar graphs show the average changes in GluA1, Homer, and PSD-95 levels in response to quinpirole. Data are expressed as the optical density ratio of total protein to GAPDH, normalized to saline control, an expressed as mean ± SEM. Right, Representative blots of bar graph analyses. The vertical line between groups indicates that blots were from the same gel but were not run in adjacent lanes. A, n = 4 vehicle/vehicle, 4 vehicle/SKF, 4 rapamycin/vehicle, 5 rapamycin/SKF. B, n = 5 saline, 4 quinpirole.