Synchrony between midbrain gene transcription and dopamine terminal regulation is modulated by chronic alcohol drinking

Abstract

Alcohol use disorder is marked by disrupted behavioral and emotional states which persist into abstinence. The enduring synaptic alterations that remain despite the absence of alcohol are of interest for interventions to prevent relapse. Here, 28 male rhesus macaques underwent over 20 months of alcohol drinking interspersed with three 30-day forced abstinence periods. After the last abstinence period, we paired direct sub-second dopamine monitoring via ex vivo voltammetry in nucleus accumbens core with RNA-sequencing of the ventral tegmental area. We found persistent augmentation of dopamine transporter function, kappa opioid receptor sensitivity, and putative dynorphin release - all inhibitory regulators which act to decrease extracellular dopamine. Surprisingly, though transcript expression was not altered, the relationship between gene expression and functional readouts of these encoded proteins was highly dynamic and altered by drinking history. These results outline the long-lasting synaptic impact of alcohol use and suggest that assessment of transcript-function relationships is critical for the rational design of precision therapeutics.

Fig. 1. Diagram of experimental design and rationale.

A Rhesus macaques (17 alcohol and 11 calorically yoked or housing controls) underwent a drinking (or housing control) protocol designed to uncover individual differences in drinking phenotypes between subjects. Briefly, a schedule induced polydipsia procedure was used to induce voluntary alcohol consumption, followed by 12 months of 22 hour/day open access drinking, a month of forced abstinence, 3 months of open access alcohol reintroduction, a month of forced abstinence, 3 months of open access alcohol reintroduction, and one month of forced abstinence at the end of which subjects were necropsied. B–E Drinking data from each epoch of the self-administration paradigm is shown, with color indicating the same subject throughout. B Cumulative alcohol intake was calculated over the first period of open access. C Average daily alcohol intake during each of the two alcohol reintroduction periods. D Lifetime intake in g/kg was calculated for each of the 17 alcohol-exposed subjects. E Blood alcohol concentration (in milligram percent [i.e. mg/dL]) was collected weekly, 7 hours after session start, for each subject and was strongly positively associated with the alcohol intake on the same day (two-tailed Pearson’s correlation). F After necropsy, the brain was blocked in coronal sections including the nucleus accumbens (NAc) core and the ventral tegmental area (VTA); dopamine dynamics were recorded from the NAc with fast-scan cyclic voltammetry, and gene expression from the upstream VTA region was assessed via bulk RNA-seq. Stimulation parameters and pharmacological manipulations were used to assess different features of dopamine terminal release in the NAc. These effects were then correlated with gene expression measures from the VTA to assess the relationship between terminal function and upstream transcription. Unless otherwise indicated, values indicate mean ± SEM. (drinkers: n = 17) Created in BioRender. Siciliano, C. (2025) https://BioRender.com/l92m987.

Fig. 2. Minimal changes to the expression of individual genes or coexpression networks in midbrain following chronic alcohol intake and protracted abstinence.

Deep sequencing was performed on the ventral tegmental area (VTA) from 28 macaques (paired-end 150 bp, roughly 150 million read pairs per sample). Reads were aligned to the Macaca mulatta genome (Mmul_10) and read counts were calculated and normalized for each subject. A A volcano plot of gene expression showing log₂(fold change) between drinkers and controls and the raw p-value. After Benjamini Hochberg FDR <0.05 correction of p-values, zero genes were significantly different between drinkers and controls. While no genes passed false-discovery rate correction, the transcripts with the five lowest p values are labeled and highlighted in green. B Dimensionality reduction via principal component analysis did not produce appreciable segregation between drinkers and controls across pairwise visualizations of the top 2 principal components. C Adjacency plot of weighted co-expression between the top 500 genes defining modules after WGNA analysis. D Tree dendrogram of weighted gene co-expression. Module eigengenes that reached a threshold of 0.99 or above were included for network analysis. E Heatmap showing module-trait relationships by group from WGCNA analysis. Color axis indicates Pearson’s r value. F Genes and connectivity, and significant GO terms of interest of the dark grey module most correlated with controls. G Genes and connectivity, and significant GO terms of interest of the sienna 3 module most correlated with alcohol drinkers. All statistical tests were two-tailed. (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001) (controls: n = 11; drinkers: n = 17).

Fig. 3. Chronic voluntary alcohol consumption retricts dynamic range of accumbal dopamine release.

A An input-output curve showing the peak dopamine release (nA) evoked by single pulse stimulations across ascending intensities (50–900 µA). Curves were fit with a 4-parameter sigmoidal regression, and best-fit values are shown with 95% confidence band and half-maximal excitatory amperage (EA₅₀) indicated (controls: EA₅₀ = 316.4 µA; drinkers: EA₅₀ = 315.9 µA). B–D Comparison of best-fit values between drinkers and controls. B Subjects with ethanol history have an attenuated upper asymptote compared to controls, indicative of decreased maximal dopamine release magnitude (unpaired t-test; t₂₆ = 2.710, p = 0.0118). C There is no difference between groups at the bottom of the curve plateau (unpaired t-test; t₂₆ = 1.033, p = 0.3110). D Drinkers show a decreased span of the input-output curve [upper plateau minus lower plateau] compared to control subjects suggesting a lower dynamic range of dopamine release (unpaired t-test; t₂₆ = 2.365, p = 0.0258). E There was no difference in the EA₅₀ between drinkers and controls (unpaired t-test; t₂₆ = 0.01, p = 0.99). (controls: n = 3; drinkers: n = 7) All statistical tests were two-tailed and values indicate mean ± SEM. (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001).

Fig. 4. Chronic drinking induced long-lasting changes in dopamine reuptake concomitant with synchrony between upstream transcription and downstream dopamine release dynamics.

A Representative normalized concentration versus time traces, pseudo-color plots, and current by voltage traces for a drinker and a control subject. B There was no difference in dopamine release between drinkers and controls (unpaired t-test; t₄₄ = 1.13, p = 0.2647). C Alcohol history increased the rate of dopamine reuptake (V_max) (unpaired t-test; t₄₄ = 2.239, p = 0.0302). D–K The best-fit linear regression is plotted and Pearson’s correlation coefficient r and p values are reported as an inset. Covariance between expression and function emerged following drinking: r values of 0.6 or greater were observed for (D) the dopamine receptor 2 (DRD2), (E) the kappa opioid receptor (OPRK1), (F) the dopamine transporter (DAT), (G) and the vesicular monoamine transporter 2 (VMAT2). H For DRD2, there was a trend towards a positive correlation between gene expression and V_max in drinkers, not controls. I Expression of OPRK1 was not correlated with V_max in either group. V_max and transporter expression did not correlate in controls, but in drinkers, V_max showed a positive correlation with the expression of DAT (K) and VMAT2 (J). L The mean slope of gene expression over terminal function was greater in drinkers (unpaired t-test; t₁₄ = 4.14, p = 0.001). The mean slope differed from zero in drinkers (one sample t-test; t₇ = 4.18. p = 0.004), but not in controls (one sample t-test; t₇ = 1.01, p = 0.35). M, N The best-fit linear regression is shown with a 95% confidence band. Inset: Spearman’s r- and p-values. M There was no correlation between expression of homosynaptic regulators and terminal release in control subjects. N Alcohol history induced a positive correlation between gene expression and function. All statistical tests were two-tailed and unless otherwise indicated, values indicate mean ± SEM. (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001, ^††p ≤ 0.01 vs 0) (controls: n = 8 [23 slices]; drinkers: n = 8 [23 slices]).

Fig. 5. Chronic drinking drove persistent upregulation of kappa opioid receptor control of dopamine release and altered function-transcription relationships.

A Representative dopamine release (1 µM U50,488). B U50,488 decreased dopamine release to a greater extent in drinkers (repeated measures two-way ANOVA; dose: F_{1, 14} = 79.28, p < 0.0001; group: F_{1, 14} = 4.81, p < 0.05; dose x group: F_{1, 14} = 0.01, p = 0.91). C BaCl₂ increased dopamine release to a greater extent in drinkers (unpaired t-test; t₅ = 3.41, p = 0.02; controls: n = 3, drinkers: n = 4). D Nalfurafine decreased dopamine release in both groups (two-way ANOVA; concentration: F_{1, 6} = 30.76, p = 0.0015; group: F_{1, 6} = 0.10, p = 0.76; concentration x group: F_{1, 6} = 0.75, p = 0.42; Šídák’s test: baseline vs. nalfurafine; controls: t₆ = 4.06, p = 0.01, n = 3; drinkers: t₆ = 3.82, p = 0.02, n = 5). E–O Best-fit linear regression. Inset: Pearson’s r and p-values. E Change in dopamine release by U50,488 (300 nM) was correlated with OPRK1 expression in drinkers, not controls. F U50,488 potency was not correlated with PDYN, (G) ARRB1, or (H) ARRB2 expression in either group. I There was no association between OPRK1 expression and efficacy of 1 µM U50,488 in controls. J Efficacy was not associated with PDYN or (K) ARRB1 in either group. L ARRB2 and efficacy were not correlated in controls, but negatively correlated in drinkers. M Mean slope of the regressions was greater in controls (unpaired t-test; t₁₄ = 3.08, p = 0.008). The mean slope for controls (one sample t-test; t₇ = 3.45, p = 0.01), but not drinkers (one sample t-test; t₇ = 1.52, p = 0.17), was greater than zero. N In controls, ranked expression was positively correlated with rank sensitivity. O This association was reversed in drinkers. All statistical tests were two-tailed. Unless otherwise indicated, values indicate mean ± SEM. (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001, ^†p ≤ 0.05 vs 0) (unless otherwise noted, controls: n = 8; drinkers: n = 8).

Fig. 6. Chronic alcohol consumption increased dynorphin release probability in long-term abstinence.

A Representative traces indicating dopamine release in controls at 10 Hz and 60 Hz normalized to 1 pulse dopamine release before and after NorBNI. B Dopamine release in controls with and without NorBNI across stimulation frequencies. C There was no effect of a history of ethanol intake, but there was an effect of NorBNI, on the area under the curve (AUC) of dopamine release across frequencies (mixed model two-way ANOVA: ethanol history: F_{1, 14} = 0.03, p = 0.88; drug: F_{1, 14} = 6.152, p = 0.03). NorBNI had no effect on AUC of dopamine release in controls (Šídák’s test: baseline vs. NorBNI in controls: t₁₄ = 0.66, p = 0.77). D Representative traces indicating dopamine release in controls at 10 Hz and 60 Hz normalized to 1 pulse dopamine release with and without NorBNI. E Dopamine release in drinkers with and without application of NorBNI across stimulation frequencies. AUC of dopamine release is visually represented by figure shading. F NorBNI administration increased the AUC of normalized dopamine release in drinkers suggesting significant dynorphin release at these stimulation intensities after chronic alcohol self-administration (mixed model two-way ANOVA: ethanol history: F_{1, 14} = 0.03, p = 0.88; drug: F_{1, 14} = 6.152, p = 0.03; Šídák’s test: baseline vs. NorBNI in drinkers: t₁₄ = 2.85, p = 0.03). G, H Upstream expression of OPRK1 and PDYN were correlated within-subject with the proportion of change in AUC of dopamine release. The best-fit linear regression is shown for each group. Inset: Pearson’s r and p-values. G Expression of OPRK1 in the ventral tegmental area was positively correlated with the change in dopamine release with NorBNI administration in controls, not drinkers. H PDYN expression was also only correlated with the dynorphin release probability in controls, not drinkers. All statistical tests were two-tailed. Unless otherwise indicated, values indicate mean ± SEM. (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001) (controls: n = 8; drinkers: n = 8).