Abstract
Ethanol tolerance is the first type of behavioral plasticity and neural plasticity that is induced by ethanol intake, and yet its molecular and circuit bases remain largely unexplored. Here, we characterize the following three distinct forms of ethanol tolerance in male Drosophila: rapid, chronic, and repeated. Rapid tolerance is composed of two short-lived memory-like states, one that is labile and one that is consolidated. Chronic tolerance, induced by continuous exposure, lasts for 2 d, induces ethanol preference, and hinders the development of rapid tolerance through the activity of histone deacetylases (HDACs). Unlike rapid tolerance, chronic tolerance is independent of the immediate early gene Hr38/Nr4a. Chronic tolerance is suppressed by the sirtuin HDAC Sirt1, whereas rapid tolerance is enhanced by Sirt1. Moreover, rapid and chronic tolerance map to anatomically distinct regions of the mushroom body learning and memory centers. Chronic tolerance, like long-term memory, is dependent on new protein synthesis and it induces the kayak/c-fos immediate early gene, but it depends on CREB signaling outside the mushroom bodies, and it does not require the Radish GTPase. Thus, chronic ethanol exposure creates an ethanol-specific memory-like state that is molecularly and anatomically different from other forms of ethanol tolerance.
SIGNIFICANCE STATEMENT The pattern and concentration of initial ethanol exposure causes operationally distinct types of ethanol tolerance to form. We identify separate molecular and neural circuit mechanisms for two forms of ethanol tolerance, rapid and chronic. We also discover that chronic tolerance forms an ethanol-specific long-term memory-like state that localizes to learning and memory circuits, but it is different from appetitive and aversive long-term memories. By contrast, rapid tolerance is composed of labile and consolidated short-term memory-like states. The multiple forms of ethanol memory-like states are genetically tractable for understanding how initial forms of ethanol-induced neural plasticity form a substrate for the longer-term brain changes associated with alcohol use disorder.
Introduction
Alcohol use disorder (AUD) is a chronic, recurring medical condition that causes extraordinarily long-term changes to brain function. Understanding how ethanol affects the brain is complicated by the relative nonspecificity of the molecular targets of ethanol, the long time course from the first experience to developing AUD, the variation in definitions of AUD, and the myriad ways that genetics and experience interact with ethanol intake patterns. Multiple forms of behavioral adaptations to ethanol are defined operationally, and neither their relative importance for AUD nor their interconnectedness is clear. A useful approach to sorting out this complexity is to start from simpler, early forms of adaptation, where effects are more reproducible, a stimulus–response relationship is clearer, and a complete description is more feasible. Some early forms of adaptation to ethanol are likely a basis on which longer-term forms build.
Tolerance is an early form of behavioral plasticity induced by ethanol intake. Tolerance is defined as the acquired resistance to ethanol's negative effects and sensitization to its positive or rewarding effects, facilitating increased intake (Fadda and Rossetti, 1998). Ethanol tolerance is classically divided into the following three forms: acute (acquired within a drinking session); rapid (expressed after the first drink is metabolized); and chronic. Some molecular mechanisms are known and are distinct for each form of tolerance, suggesting that the dose, time, and pattern of ethanol exposure can engage different plasticity mechanisms. It is not known if different forms of ethanol tolerance colocalize in brain circuits, if they share some common plasticity mechanisms, or their relative contribution to the progression towards AUDs.
Drosophila is useful for defining the mechanisms of ethanol tolerance. Acute, rapid, and chronic tolerance all exist in flies, and they are separable genetically (Berger et al., 2004). Molecular parallels to early forms of ethanol plasticity in mammals indicate potential deep conservation of the basic mechanisms (Cowmeadow et al., 2005; Morozova et al., 2006; Kong et al., 2010; Ghezzi et al., 2013; Engel et al., 2016; Park et al., 2017; Ranson et al., 2019). Tolerance in Drosophila is functional, because of adaptive changes in behavior (pharmacodynamic), and not dispositional, because of changes in ethanol metabolism (pharmacokinetic; Scholz et al., 2000; Berger et al., 2004).
We show that the pattern of initial ethanol intake engages distinct molecular pathways for plasticity that are encoded into different neural circuits. Rapid tolerance requires the induction of the immediate early gene (IEG) Hr38, an ortholog of mammalian Nr4a1-3, and the histone deacetylase (HDAC) Sirt1 in the α/β lobes of the adult mushroom bodies, the major learning and memory center of the Drosophila brain (Engel et al., 2016; Adhikari et al., 2019). Chronic tolerance, induced by a prolonged exposure to a low concentration of ethanol, is Hr38 independent and is inhibited by Sirt1 in the γ lobes of the adult mushroom bodies. Chronic tolerance enables the induction of the c-fos-like IEG kayak by subsequent ethanol exposures, whereas rapid tolerance does not. Chronic exposure specifically forms a long-term memory (LTM)-like state that is distinct from both appetitive and aversive forms of LTM, whereas rapid tolerance exhibits characteristics of intermediate-term memory that is composed of labile and consolidated components. We also discovered an additional form of chronic tolerance, induced by repeated exposure to moderately inebriating doses of ethanol, that results in a third pattern of gene expression and lasting behavioral outcomes.
Materials and Methods
Drosophila culturing and strains
Culturing.
All strains used in this study were outcrossed for at least five generations to the Berlin genetic background carrying the w1118 genetic marker mutation. This genetic background strain was used as an experimental control for loss-of-function mutants. Mutants of intermediate-term memory were FM7-balanced during outcross. Flies were cultured on standard cornmeal/molasses/yeast medium at 25°C and 60% relative humidity under a 12 h light/dark schedule. Male flies were used for all experiments. Flies were collected 2–4 d after eclosion, and they were rested from CO2 collection for at least 1 d before any behavioral paradigms.
Strains.
The following fly strains were used: Sirt12A-7–11 [catalog #8838, Bloomington Drosophila Stock Center (BDSC)]; UAS-Sirt1.IR (catalog #32481, BDSC); elav(c155)-Gal4 (catalog #458, BDSC); tub-Gal80ts (catalog #7019, BDSC); 17d-Gal4 (catalog #51631, BDSC); repo-Gal4 (catalog #7415, BDSC); UAS-TeTx (a gift from Sean Sweeney, University of York, York, UK); UAS-Kir2::eGFP (catalog #6595, BDSC); R13F02-Gal4 (catalog #48571, BDSC); R11D09-Gal4 (catalog #48456, BDSC); R35B12-Gal4 (catalog #49822, BDSC); R28H05-Gal4 (catalog #49472, BDSC); Hr38y214 (a gift from Carl Thummel; University of Utah); rad1 (catalog #79209, BDSC); amn1 (catalog #5954, BDSC); nSyb-Gal4 (a gift from Julie Simpson; University of California, Santa Barbara); and UAS-Creb2b (catalog #7219, BDSC).
Ethanol behavior
Rapid tolerance.
Ethanol sensitivity and rapid tolerance were measured as previously described (Engel et al., 2016). Briefly, groups of 20 genetically identical flies (n = 1) were exposed to 55% ethanol vapor or 100% humidified air. The 55% ethanol is an intermediate ethanol dose that results in submaximal rapid tolerance and in 50% sedation in 12–20 min (Kong et al., 2010). The number of flies that lost the righting reflex were counted at 6 min intervals. The time to 50% sedation (ST50) was calculated for each group (E1). Flies were allowed to rest for 3.5 h and were re-exposed to an identical concentration of ethanol vapor (E2). Rapid tolerance was calculated as the difference in ST50, E2 – E1.
Chronic tolerance.
Flies in perforated 50 ml conical tubes with 5 ml of food were placed in a temperature-controlled and light-controlled chamber that was perfused with 16% ethanol vapor or 100% humidified air continuously for 48 h. The flies were then allowed to rest in the rearing incubator for 3–72 h. Chronic tolerance was measured between groups pre-exposed to ethanol minus pre-exposed to air, with random assignment of ethanol and air pre-exposure group pairing within the daily cohort. The experiment was repeated across days to achieve a final n of 15–30 groups per treatment and genotype.
Repeated tolerance.
Flies were exposed to 42% ethanol vapor or 100% humidified air for 20 min every 24 h for 4 d. Repeated tolerance was measured between groups pre-exposed to ethanol minus pre-exposed to air, with random assignment of ethanol and air pre-exposure group pairing within a day. The experiment was repeated across days to achieve a final n of 15–30 groups per treatment and genotype.
Ethanol preference.
The capillary feeding assay (CAFÉ) was used as previously described (Ja et al., 2007; Devineni and Heberlein, 2009; Peru y Colón de Portugal et al., 2014). Groups of eight genetically identical flies were exposed to ethanol vapor at different concentrations and lengths of time or to 100% humidified air alone. After 16 h of recovery, flies were placed into the CAFÉ chamber, which consists of empty Drosophila wide culture vials with 5 µl capillary tubes (catalog #53432–706, VWR) containing liquid food, with or without 15% ethanol, embedded in the vial plug through adaptors made from 200 µl pipette tips. The 15% ethanol resulted in intermediate preference values in prior reports (Devineni and Heberlein, 2009; Engel et al., 2016). The preference index was measured as the volume of food consumed over 1 night from the ethanol capillaries minus that consumed from the no-ethanol capillaries, over the total volume consumed, corrected for evaporation by measuring the volume lost in tubes with no flies.
Cold shock anesthesia
Flies underwent cold shock disruption 30 min after pre-exposure to acute ethanol, chronic ethanol, or humidified air. Standard fly vials or perforated 50 ml conical tubes housing 20 flies were sheathed in aluminum foil and placed in an ice bath for 3 min. Locomotion was quickly lost, then regained within a minute on return to 25°C.
Drug treatments
Nicotinamide (70 mm; Sigma-Aldrich) or Trichostatin A (TSA; 8 µm; Sigma-Aldrich) was fed to flies dissolved in 5% sucrose/2% yeast extract on Whatman filter paper for 24 h.
RNA measurement
RNA was extracted from heads using TRIzol (Thermo Fisher Scientific), treated with DNase (Promega), and reverse transcribed using MultiScribe (Applied Biosystems). Quantitative PCRs were performed on the qTOWER³84 machine (Genesee Scientific) using the SYBR Green method (BIO-RAD) and custom-designed primers (Integrated DNA Technologies). RpL32 was used to normalize Ct values, the expression of genes of interest was calculated using the ΔΔCt method, and the mean expression was calculated from multiple independent biological replicates.
Oligonucleotide primer sequences were as follows: Hr38: Forward (F), GAGTGGCTCAACGACATCAT; Reverse (R), CGTTCTGTGATCAGGGTTAGG; Stripe (Sr): F, CCGAGTATGCCGCTCAATTA; R, GGCGTATGGTGGTGATAAGG; Jra: F, GTTCCCACCCACTGATTGA; R, GCTTGTTCTTGGCACTCTTG; kayak (kay): F, CCGATACTTCAAGTGCCCATAC; R, CCAGGACATTGGAGAAGTTGTT; Sirt1: F, CGGTGGCCGTTACTGAGGAGGA; R, TACTCATCCGGCAGTCCCTCGC; RpL32: F, GTTCGATCCGTAACCGATGT; R, CCAGTCGGATCGATATGCTAA.
Ethanol absorption and metabolism
Groups of ∼20 flies underwent typical chronic pre-exposure: continuous humidified air or 16% ethanol for 2 d, then 1 d of rest. Flies were exposed to 30 min of 20% ethanol to avoid sedation, then were frozen in liquid nitrogen either immediately for “absorption” samples, or 30 min later for “metabolism” samples. After homogenization in 50 mm Tris-HCl, pH 7.5, ethanol concentrations were measured using the nicotinamide adenine dinucleotide (NAD)-alcohol dehydrogenase (ADH) Reagent Kit (catalog #N7160, Sigma-Aldrich) following the manufacturer protocol. Specifically, the acquired 340 nm absorbance values were converted to millimolar concentration and adjusted by the estimated 1 µl volume of an average fly to calculate ethanol concentrations.
Water content
Exposure conditions were identical to those used for chronic tolerance. Groups of 25 flies were exposed for 48 h to either 16% ethanol vapor or an equal flow rate of humidified air in vials containing 5 ml of standard fly food. The flies were weighed (wet weight) immediately after exposure cessation, and then desiccated at 55°C for 24 h and weighed again (dry weight). Total water content was wet weight minus dry weight divided by the number of flies. The experiment was repeated for six biological replicates.
Experimental design and statistical analysis
For all experiments, experimental and genetically matched or treatment-matched controls were tested in the same session in a balanced experimental design. Experiments were repeated across days with progeny from repeat parental crosses, and data from all days and crosses were collated together without between-day adjustments. Untransformed (raw) data were used for statistical analysis. Where the experimental group was compared with two or more control groups, significance was only interpreted when all controls were different from the experimental. GraphPad Prism 8.4.3 was used for unpaired t test, one-sample t test, one-way ANOVA with Tukey's post hoc test for normally distributed data, Kruskal–Wallis test with Dunn's post hoc test for nonparametric data, and Brown–Forsythe test with Dunnett's post hoc test for data that fails the Shapiro–Wilk normality test. The full results of the statistical tests are presented in the figure legends. Significance indicators on the figures indicate the results of t tests or post hoc tests for significant effects by ANOVA (****p ≤ 0.0001; ***p ≤ 0.001; **p ≤ 0.01; *p ≤ 0.05; and ns, p > 0.05). Error bars represent the SEM.
Results
Different ethanol exposure paradigms induce different forms of behavioral plasticity
Flies were given ethanol exposures of different length, concentration, or pattern (or matched humidified air controls), and they were then tested for ethanol behavioral responses with a uniform challenge exposure (Fig. 1A). The uniform challenge exposure was 55% ethanol vapor, to determine the ST50 for a population of ∼20 flies; ST50 typically occurs at 12–20 min for ethanol-naive flies. Rapid tolerance is induced by a just sedating ethanol exposure, which typically causes 70–80% of flies to lose the righting response after 30 min of 55% ethanol vapor. Rapid tolerance expression is measured 4 h later, after the initial dose is fully metabolized, with an identical exposure (Scholz et al., 2000; Kong et al., 2010). This level of ethanol exposure is operationally similar to the tolerance-inducing effects of binge drinking in humans (Schuckit, 1994). Chronic tolerance is induced by prolonged exposure to a low dose of ethanol; it is intended to mimic aspects of maintenance drinking by individuals with alcohol use disorder (Berger et al., 2004). Finally, we also gave flies repeated exposures to an inebriating, but not sedating, ethanol dose (42% ethanol vapor for 20 min), once per day for 4 consecutive days, similar to limited-access paradigms with mice (Rhodes et al., 2005).
Types of ethanol tolerance in Drosophila. A, Ethanol exposure schemes to induce and measure ethanol sensitivity and tolerance. Challenge doses (C) are 55% ethanol unless noted otherwise. Gene expression measurements were sampled 1 h after the ethanol challenge dose. B, Dose response for chronic ethanol pre-exposure. Chronic tolerance is ST50, chronic minus acute. C, Ethanol absorption and metabolism with air or 16% chronic ethanol pre-exposure. Flies were exposed to 30 min of 20% ethanol to avoid sedation. Absorption was measured immediately afterward, and metabolism was measured 30 min later. D, Left, Chronic ethanol pre-exposure caused resistance to sedation. Right, Chronic ethanol exposure induces tolerance that lasts for 2 d. E, Chronic ethanol exposure interfered with the subsequent development of rapid tolerance. Flies were either given 48 h of chronic ethanol exposure or humidified air, and 24 h later were subjected to a rapid tolerance test, two 30 min 60% EtOH exposures, E1 and E2, 4 h apart. Rapid tolerance is ST50, E2–E1. F, Chronic ethanol exposure induced ethanol preference, measured in the CAFÉ two-choice assay. G, Left, Repeated inebriating doses of ethanol caused resistance to ethanol sedation. Right, Repeated ethanol exposure induced tolerance that lasts for at least 1 d. Tolerance is from data shown in the left panel, repeated minus acute. H, Repeated ethanol exposure inhibited the subsequent development of rapid tolerance. I, Repeated ethanol exposure induced ethanol aversion in the CAFÉ two-choice assay. B–I, Quantitative data are the mean ± SEM. C, Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; unexposed air versus unexposed EtOH, p = 0.3892; absorption air versus absorption EtOH, p = 0.9948; metabolism air versus metabolism EtOH, p = 0.9966. D, Left, Mann–Whitney test; air versus EtOH, ****p < 0.0001. Right, One-sample t test (theoretical mean = 0); 3 h, ****p < 0.0001; 24 h, ***p = 0.0005; 48 h, **p = 0.0090; 72 h, p = 0.2719. E, Unpaired t test; air versus EtOH, ***p = 0.0007. F, Wilcoxon signed-rank test (theoretical mean = 0); air, p = 0.6511; acute: ****p < 0.0001; chronic, ***p = 0.0007. G, Left, Welch's t test; air versus EtOH, ****p < 0.0001. Right, One-sample t test (theoretical mean = 0); repeated, ****p < 0.0001. H, Welch's t test; air versus EtOH, *p = 0.0155. I, Wilcoxon signed-rank test (theoretical mean = 0); air, p = 0.2087; acute, **p < 0.0019; repeated, *p = 0.0347.
Chronic ethanol exposure induced resistance to ethanol sedation (chronic tolerance). Increasing concentrations of ethanol during the chronic exposure increased the extent of tolerance (Fig. 1B). Pre-exposure with chronic ethanol did not alter ethanol absorption or metabolism, measured with a challenge dose, indicating that tolerance is not because of altered ethanol handling by the fly (Fig. 1C). Chronic exposure also did not alter the water content of the flies (1.157 ± 0.1108 µl for ethanol exposed; 1.174 ± 0.0593 µl for humidified air exposed; n = 6 biological replicates). Chronic tolerance dissipated over time, lasting for at least 48 h (Fig. 1D). Chronic ethanol exposure also did not cause tissue damage. Prior research showed that repeated brief exposures to high-concentration ethanol specifically causes necrosis of the third antennal segment olfactory organ, with little to no damage to other tissues (French and Heberlein, 2009). Chronic exposure to 21% ethanol vapor, the highest concentration we tested, resulted in necrosis (blackening) of 0 of 56 antennae, or 0% necrosis. To determine whether rapid and chronic tolerance could co-occur in flies, we subjected flies serially to a chronic exposure and 24 h later to a rapid tolerance induction paradigm (Fig. 1E). Rapid tolerance was decreased in flies previously given a chronic ethanol exposure, suggesting a mechanistic link between rapid and chronic tolerance. The link between chronic and rapid tolerance may reflect a ceiling effect on tolerance development in Drosophila, or it may be because of interactions between molecular or neural circuit encoding mechanisms. Ethanol preference, measured in the CAFÉ two-choice assay, is an acquired behavior that can be induced by an acute inebriating ethanol dose (Ja et al., 2007; Peru y Colón de Portugal et al., 2014). Binge-like acute pre-exposure and chronic pre-exposure to ethanol both potentiated ethanol preference (Fig. 1F). These results indicate that there exist behavioral and functional links between rapid and chronic ethanol tolerance, and that chronic exposure can induce a positive valence toward ethanol.
Repeated inebriation and withdrawal (chronic intermittent exposure) is a common exposure paradigm used in rodents to induce AUD-like states, which has not yet been explored for its effects on ethanol tolerance (Läck et al., 2007; Goltseker et al., 2019). We tested a repeated ethanol exposure paradigm where flies were given an inebriating but not sedating ethanol exposure (42% ethanol vapor for 20 min) once per day for 4 d, and then given a challenge dose 24 h later (Fig. 1A). Repeated ethanol exposure reduced sensitivity to ethanol sedation (repeated tolerance), measured after a 24 h recovery period (Fig. 1G). Like chronic exposure, repeated exposure decreased subsequent rapid tolerance development (Fig. 1H), suggesting a link between rapid and repeated tolerance mechanisms. However, repeated exposure induced mild ethanol aversion instead of ethanol preference (Fig. 1I). Repeated tolerance perdured for 24 h, whereas rapid tolerance is extinguished by 24 h (Scholz et al., 2000). Thus, repeated ethanol exposure likely creates a third distinct form of ethanol tolerance.
Rapid and chronic tolerance are transcriptionally distinct
We asked whether tolerance-inducing ethanol exposure paradigms cause distinct transcriptional responses by measuring the expression levels of several IEG transcriptional regulators that are implicated in neural plasticity. We specifically focused on IEGs that are transcription factors, because their induction may drive programs of gene expression to alter neuron function in response to ethanol. We found that there are different IEG response profiles for rapid and chronic tolerance (Fig. 2A). The Nr4a1-3 ortholog Hr38 and the Egr1-4 ortholog Sr were selectively induced by the rapid tolerance-inducing acute exposure, whereas the c-Fos ortholog kay was selectively induced by chronic tolerance. By contrast, Hr38 remained inducible by ethanol following repeated inebriating ethanol exposures. Interestingly, two sedating ethanol doses delivered in the rapid tolerance protocol blocked Hr38 inducibility, possibly indicating molecular distinctions between inebriation and sedation, or a ceiling to rapid tolerance. Thus, rapid, chronic, and repeated tolerance are molecularly distinct forms of alcohol plasticity.
Immediate early gene induction by different tolerance paradigms. A, Quantitative PCR of transcription factor IEGs induced by acute dose, or a challenge dose following acute (rapid), chronic, or repeated exposure, expressed as the fold change versus humidified air control exposure. Left, Different ethanol pre-exposures induce different IEG response profiles. Note: kay is significantly induced when acute and chronic exposures are directly compared (p = 0.0221, Mann–Whitney test). Right, Hr38 is inducible following acute and repeated ethanol exposure conditions. B, Left, Chronic tolerance is unaffected in Hr38 mutants. Right, Chronic ethanol pre-exposure causes sedation resistance in control and Hr38 mutants. Ethanol sensitivity is unaffected in Hr38 mutants. A, B, Quantitative data are the mean ± SEM. A, Left: Hr38: Kruskal–Wallis ANOVA with Dunn's multiple-comparisons test; acute versus rapid, ***p = 0.0008; acute versus chronic, **p = 0.0012; rapid versus chronic, p > 0.9999. Sr: Kruskal–Wallis ANOVA with Dunn's multiple-comparisons test; acute versus rapid, p = 0.3239; acute versus chronic, **p = 0.0087; rapid versus chronic, p = 0.7261. Jra: One-way ANOVA with Tukey's multiple-comparisons test; acute versus rapid, p = 0.0961; acute versus chronic, p = 0.5183; rapid versus chronic, p = 0.5289. kay: Kruskal–Wallis ANOVA with Dunn's multiple-comparisons test; acute versus rapid, p > 0.9999; acute versus chronic, p = 0.1467; rapid versus chronic, p = 0.0552. Right: Wilcoxon signed-rank test (theoretical mean = 1); acute, **p = 0.0039; repeated, **p = 0.0039. B, Left: Unpaired t test; control versus Hr38−, p = 0.0906. Right: Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; control air versus control EtOH, ****p < 0.0001; control air versus Hr38− air, p = 0.8031; Hr38− air versus Hr38− EtOH: **p = 0.0048; control EtOH versus Hr38− EtOH, p = 0.2679.
Hr38 induction by acute ethanol exposure is necessary and sufficient for rapid tolerance development (Adhikari et al., 2019). Because Hr38 is no longer inducible by ethanol following chronic exposure, we asked whether it was required for chronic tolerance. Chronic tolerance was normal in Hr38 loss-of-function mutants, indicating that Hr38 functions in rapid but not in chronic tolerance (Fig. 2B).
Histone deacetylation maintains chronic tolerance and sets IEG inducibility
The distinct patterns of IEG induction in rapid and chronic tolerance paradigms suggests that chronic exposure may create a specific chromatin state to set a transcriptional response pattern, one that is likely a plasticity-encoding mechanism for ethanol and addiction (Walker et al., 2015). Rapid tolerance in flies is encoded into chromatin: histone acetylation is quickly increased by a sedating ethanol exposure, and rapid tolerance is blocked by chemical inhibition of the NAD-dependent sirtuin (Sirt) class of HDACs, and by genetic deletion of the HDAC Sirt1 (Engel et al., 2016). To test whether there may be a chromatin state component to chronic tolerance, we fed flies two broadly acting HDAC inhibitors that act similarly in flies and mammals (Bitterman et al., 2002; Foglietti et al., 2006). TSA inhibits Class I/II HDACs, and nicotinamide inhibits the NAD-dependent sirtuins, including Sirt1, which promotes rapid tolerance (Engel et al., 2016). Flies were given a chronic ethanol exposure, transferred to HDAC inhibitor-containing food during the withdrawal period, and then given an ethanol challenge dose (Fig. 3). TSA erased and nicotinamide reduced chronic tolerance without affecting ethanol sensitivity, suggesting that multiple HDACs may encode the experience of chronic ethanol exposure (Fig. 3A,B). Interestingly, TSA treatment after chronic ethanol exposure restored Hr38 inducibility by ethanol, whereas nicotinamide did not (Fig. 3C). Thus, Class I/II HDACs may be responsible for blocking Hr38 induction by ethanol following chronic exposure. Because chronic ethanol pre-exposure decreases the expression of rapid tolerance (Fig. 1E), we asked whether TSA treatment erased this interaction between the forms of tolerance. TSA treatment during withdrawal from chronic ethanol exposure restored the expression of normal rapid tolerance (Fig. 3D). We conclude that chronic ethanol exposure likely alters chromatin through the actions of Class I/II HDACs to impede rapid tolerance development. Chronic ethanol exposure may promote chromatin compaction through histone deacetylation to encode the state of chronic ethanol tolerance.
Histone deacetylase inhibitors reveal deacetylation maintains encoding of chronic tolerance. A, Left, TSA, which inhibits Class I/II histone deacetylases, trends toward decreased chronic tolerance. Right, TSA does not affect ethanol sensitivity. B, Left, Nicotinamide, which inhibits NAD-dependent sirtuin class histone deacetylases, decreases chronic tolerance. Right, Nicotinamide does not affect ethanol sensitivity. C, TSA but not nicotinamide restores Hr38 ethanol inducibility by an ethanol challenge following chronic ethanol exposure, measured by quantitative PCR. D, TSA restores rapid tolerance following a chronic exposure. A–D, Quantitative data are the mean ± SEM. A, Left, Mann–Whitney test; no Rx versus TSA, p = 0.0752. Right, One-way ANOVA with Sidak's multiple-comparisons test; no Rx air versus no Rx EtOH, ***p = 0.0007; no Rx air versus TSA air, p > 0.9999; no Rx EtOH versus TSA EtOH, **p = 0.0062; TSA air versus TSA EtOH, p = 0.9249. B, Left, Unpaired t test; no Rx versus nicotinamide, *p = 0.0446. Right, Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; no Rx air versus no Rx EtOH, ***p = 0.0002; no Rx air versus nicotinamide air, p = 0.3244; no Rx EtOH versus nicotinamide EtOH, **p = 0.0013; nicotinamide air versus nicotinamide EtOH, *p = 0.0180. C, No Rx: unpaired t test; air versus EtOH, **p = 0.0050; TSA: unpaired t test; air versus EtOH, p = 0.7319. nicotinamide: Mann–Whitney test; air versus EtOH, *p = 0.0499. D, No Rx: unpaired t test; air versus EtOH, *p = 0.0342; TSA: unpaired t test; air versus EtOH, p = 0.8264.
Sirt1 limits chronic tolerance development in the mushroom bodies
The likely NAD-dependent sirtuin encoding of chronic tolerance prompted us to ask whether a sirtuin may regulate chronic tolerance development. We chose to test the role of Sirt1 because it is strongly regulated by a rapid tolerance-inducing acute ethanol exposure and because it promotes rapid tolerance development (Kong et al., 2010; Engel et al., 2016). In contrast to rapid tolerance, Sirt1-null mutants exhibited increased chronic tolerance (Fig. 4A,A′). Thus, Sirt1 inhibits chronic tolerance, whereas it promotes rapid tolerance, providing further evidence that chronic and rapid tolerance are molecularly distinct.
Sirt1 acts in the mushroom body γ lobes to limit the expression of chronic tolerance. A, Left, Sirt1-null mutant flies develop more chronic tolerance. Right, Sirt1-null mutant flies develop less rapid tolerance. A′, Sirt1-null mutants have decreased sensitivity to ethanol sedation. B, Reduction of Sirt1 by RNAi specifically in all postmitotic neurons increases chronic tolerance. B′, Reduction of Sirt1 expression in all neurons causes decreased ethanol sensitivity. C, Reduced Sirt1 expression in all glia does not affect chronic tolerance. C′, Reduced Sirt1 expression in all glia does not affect ethanol sensitivity. D, Sirt1 RNAi in all neurons of the mushroom bodies (R13F02>Sirt1.IR), or specifically in the mushroom body γ lobes (R11D09>Sirt1.IR), increases chronic tolerance. Reduction of Sirt1 in either the mushroom body α′/β′ lobes (R35B12>Sirt1.IR) or the α/β lobes (R28H05>Sirt1.IR) does not affect chronic tolerance. D′, Reduced Sirt1 in the mushroom bodies or in each lobe causes decreased sensitivity. E, Chronic tolerance was unaffected when Sirt1 was reduced in 17d-Gal4 neurons, the site of action for Sirt1 in rapid tolerance. E′, Reduced Sirt1 in 17d-Gal4 neurons causes decreased sensitivity. A–E′, Quantitative data are the mean ± SEM. A, Left, Welch's t test; control versus Sirt1−, *p = 0.0162. Right, Welch's t test; control versus Sirt1−, **p = 0.0025. A′, Mann–Whitney test; control versus Sirt1−, ****p < 0.0001. B, Kruskal–Wallis ANOVA with Dunn's multiple-comparisons test; Gal4/UAS versus Gal4, **p = 0.0097; Gal4/UAS versus UAS, ****p < 0.0001. B′, Kruskal–Wallis ANOVA with Dunn's multiple-comparisons test, Gal4/UAS versus Gal4, ****p < 0.0001; Gal4/UAS versus UAS, ***p = 0.0004. C, One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, p = 0.9243; Gal4/UAS versus UAS, p = 0.0640. C′, Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; Gal4/UAS versus Gal4, ****p < 0.0001; Gal4/UAS versus UAS, p = 0.1240. D, R13F02 panel: One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, ***p = 0.0001; Gal4/UAS versus UAS, ***p = 0.0006. R11D09 panel: Kruskal–Wallis ANOVA with Dunn's multiple-comparisons test; Gal4/UAS versus Gal4, **p = 0.0056; Gal4/UAS versus UAS, *p = 0.0108. R35B12 panel: Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; Gal4/UAS versus Gal4, p = 0.7905; Gal4/UAS versus UAS, p = 0.4612. R28H05 panel: One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, p = 0.0913; Gal4/UAS versus UAS, **p = 0.0022. D′, R13F02 panel: One-way ANOVA with Dunnett's multiple comparisons; Gal4/UAS versus Gal4, ****p < 0.0001; Gal4/UAS versus UAS, ****p < 0.0001. R11D09 panel: One-way ANOVA with Dunnett's multiple comparisons, Gal4/UAS versus Gal4, ***p = 0.0001; Gal4/UAS versus UAS, ****p < 0.0001. R35B12 panel: One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, ****p < 0.0001; Gal4/UAS versus UAS, ****p < 0.0001. R28H05 panel: One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, ****p < 0.0001; Gal4/UAS versus UAS, ****p < 0.0001. E, One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, p = 0.7108; Gal4/UAS versus UAS, p = 0.8961. E′, Kruskal–Wallis ANOVA with Dunn's multiple-comparisons test; Gal4/UAS versus Gal4, ****p < 0.0001; Gal4/UAS versus UAS, **p = 0.0014.
We used RNAi to decrease Sirt1 in specific tissues and cell types to map its site of action. The Sirt1 RNAi transgene used in this study was previously proven effective at decreasing Sirt1 expression, and when it was expressed in the nervous system, it phenocopied the Sirt1-null mutant for rapid tolerance (Engel et al., 2016). Sirt1 RNAi in all neurons, but not in all glia, resulted in increased chronic tolerance, the same behavioral phenotype as in the Sirt1-null mutant (Fig. 4B–C′). The mushroom bodies are the major learning and memory centers in the fly brain, and they are critical for multiple forms of ethanol-induced behavioral plasticity, including rapid tolerance, ethanol preference, and ethanol reward. We used a panel of GAL4 driver transgenes that express in the entire mushroom bodies or specifically in one of the three mushroom body lobes (Noyes et al., 2020). Sirt1 RNAi in all mushroom body lobes or specifically in the mushroom body γ lobes increased chronic tolerance (Fig. 4D). Sirt1-dependent sensitivity to acute ethanol inebriation mapped to neurons in all three mushroom body lobes (Fig. 4D′). In rapid tolerance, Sirt1 functions in the mushroom body α/β lobes (Engel et al., 2016). Thus, ethanol sensitivity is likely encoded through mechanisms that are fundamentally distinct from either form of tolerance. To verify that rapid and chronic tolerance map to distinct neurons of the mushroom bodies, we also decreased Sirt1 in the α/β lobes with the 17d-Gal4 driver that was used previously to localize rapid tolerance (Engel et al., 2016). Sirt1 RNAi in the 17d-Gal4 α/β neurons did not affect chronic tolerance development (Fig. 4E,E′). Thus, both rapid and chronic ethanol tolerance require Sirt1 in the mushroom bodies, with Sirt1 limiting chronic tolerance in the γ lobes and promoting rapid tolerance in the α/β lobes. We conclude that there exist distinct mushroom body circuits for different forms of ethanol tolerance.
The adult mushroom bodies require Sirt1 and neuronal activity for chronic tolerance
To determine when Sirt1 acts to limit chronic tolerance, we used a well characterized temperature-sensitive form of the GAL4 inhibitor GAL80 to limit the expression of Sirt1 RNAi to adults. Adult-specific Sirt1 RNAi in all mushroom body neurons increased chronic tolerance, indicating a regulatory role of Sirt1 on ethanol plasticity in adult flies (Fig. 5A). There was no effect of adult-specific Sirt1 RNAi on sensitivity to an acute inebriating dose of ethanol (Fig. 5A′). Thus, Sirt1-dependent inhibition of chronic tolerance is an adult function of the HDAC, whereas Sirt1-dependent promotion of ethanol sensitivity is likely to arise during development.
Chronic tolerance development is adult-specific and activity-dependent in the mushroom bodies. A, Sirt1 is required in the adult mushroom bodies for chronic tolerance development. GAL4 was suppressed by temperature-sensitive GAL80 throughout development (Dev) by rearing the flies at 18°C (GAL80 on, GAL4 blocked), and shifting them to 29°C (GAL80 off, GAL4 active) after eclosion (Adult). A 50% ethanol challenge dose was used to account for increased ethanol sedation at 29°C. A′, Adult-specific decrease of Sirt1 in the mushroom bodies did not affect sensitivity to a 50% ethanol challenge, independent of temperature during adulthood. B, Hyperpolarization of the mushroom bodies in adults with the inwardly rectifying potassium channel Kir2.1 blocks chronic tolerance development. A 50% ethanol challenge dose was used. B′, Adult-specific hyperpolarization of the mushroom bodies had no effect on sensitivity to a 50% ethanol challenge. C, Blocking synaptic vesicle release [tetanus toxin light chain (TeTx)] in the mushroom bodies in Sirt1-null mutants results in decreased chronic tolerance. C′, Blocking synaptic vesicle release in the mushroom bodies in Sirt1-null mutant flies increased ethanol sensitivity. A–C′, Quantitative data are the mean ± SEM A, Left, One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, *p = 0.0416; Gal4/UAS versus UAS, **p = 0.0019. Right, One-way ANOVA with Holm–Sidak's multiple-comparisons test; Gal4/UAS versus Gal4, p = 0.9685; Gal4/UAS versus UAS, p = 0.6880. A′, Left, One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, p = 0.1940; Gal4/UAS versus UAS, p = 0.0508. Right, Kruskal–Wallis ANOVA with Dunn's multiple-comparisons test; Gal4/UAS versus Gal4, *p = 0.0492; Gal4/UAS versus UAS, p = 0.6952. B, One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, *p = 0.0406; Gal4/UAS versus UAS, *p = 0.0246. B′, One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, **p = 0.0066; Gal4/UAS versus UAS, p = 0.2596. C, One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, *p = 0.0384; Gal4/UAS versus UAS, *p = 0.0192. C′, One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, ****p < 0.0001; Gal4/UAS versus UAS, **p = 0.0055.
We next asked whether neuronal activity of the mushroom bodies is necessary to produce chronic tolerance in adults. Flies expressing the inward rectifying potassium channel Kir2.1 to hyperpolarize the adult mushroom bodies showed decreased chronic tolerance development (Fig. 5B,B′). Thus, mushroom body neurons promote chronic tolerance via neuronal activity. Considering that chronic tolerance is negatively regulated by Sirt1 and positively regulated by mushroom body activity, we asked whether these two factors interact. Silencing mushroom body neurons in Sirt1-null mutant flies decreased chronic tolerance (Fig. 5C,C′). Thus, neuronal activity is critical for Sirt1-dependent inhibition of chronic tolerance development.
Chronic tolerance is a persistent, noncanonical memory-like state of ethanol
We asked whether chronic tolerance shares molecular features with forms of Drosophila memory. It is possible to distinguish different types of labile and consolidated memories using specific mutants and tests for anesthesia sensitivity. Drosophila intermediate-term memories are composed of anesthesia-sensitive memory (ASM) and anesthesia-resistant memory (ARM) traces (Dubnau and Tully, 2001). Drosophila also form longer-lasting forms of memory as LTM. The gene radish encodes a GTPase that sustains ARM, a quickly consolidated form of memory, as well as appetitive, but not aversive, LTM (Folkers et al., 1993; Krashes and Waddell, 2008). Mutant flies expressing a truncated Radish protein displayed decreased ethanol sensitivity, yet chronic tolerance was unaffected (Fig. 6A,A′). The amnesiac (amn) gene encodes a pituitary adenylate cyclase-activating polypeptide (PACAP)-like neuropeptide that is required to form labile ASM. Flies mutant for amn showed decreased sensitivity to ethanol sedation, yet chronic tolerance was unaffected (Fig. 6A,A′). The Sirt1, amn, and Hr38 mutants were previously shown to have no effect on ethanol absorption (Moore et al., 1998; Engel et al., 2016; Adhikari et al., 2019). The rad1 mutant also showed normal ethanol absorption and metabolism [control = 4.82 mm (n = 3), 11.40 mm (n = 3), 9.47 mm (n = 3); rad1 = 3.17 mm (n = 4), 8.15 mm (n = 4), 4.74 mm (n = 4) for pre-exposure, absorption, and metabolism; Kruskal–Wallis test, p = 0.0077; Dunn's multiple-comparisons test for control vs rad1 pre-exposure, p = 0.7153; absorption, p > 0.9999; metabolism, p = 0.1653]. Thus, chronic tolerance does not use the same genetic pathways as ASM or ARM.
Chronic and rapid tolerance are distinct memory-like states in the brain. A, Mutants that abolish anesthesia-resistant memory (rad1) and anesthesia-sensitive memory (amn1) develop normal chronic tolerance at 24 h. A′, Decreased ethanol sensitivity in mutants for early forms of memory. B, Cold-shock anesthesia, blocking anesthesia-sensitive memories, does not affect chronic tolerance. B′, Increased ethanol sensitivity following cold-shock anesthesia. C, Chronic tolerance does not include 3 h anesthesia-sensitive memory-like or 3 h anesthesia-resistant memory-like states. C′, Decreased ethanol sensitivity in radish mutants, no effect of cold shock. D, Rapid tolerance is composed of 3 h anesthesia-sensitive memory-like and 3 h anesthesia-resistant memory-like states. D′, radish mutants have decreased ethanol sensitivity, as measured from the rapid tolerance inducing exposure. A–D′, Quantitative data are the mean ± SEM. A, Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; control versus rad1, p = 0.8360; control versus amn1, p = 0.5176. A′, Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; control versus rad1, ****p < 0.0001; control versus amn1, ****p < 0.0001. B, Unpaired t test; untreated versus cold shock, p = 0.4243. B′, Unpaired t test; untreated versus cold shock, *p = 0.0439. C, Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; control versus cold shock, p = 0.0844; control versus rad1, p = 0.5030; control versus rad1 cold shock, p = 0.2429. C′, Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; control versus cold shock, p = 0.9743; control versus rad1, ****p < 0.0001; control versus rad1 cold shock, ****p < 0.0001. D, One-way ANOVA with Sidak's multiple-comparisons test; control versus cold shock, ***p = 0.0007; control versus rad1, ****p < 0.0001; rad1 versus rad1 cold shock, ****p < 0.0001. D′, Unpaired t test; control versus rad1, ****p < 0.0001.
An orthogonal test of ASM performs memory disruption via cold-induced anesthesia. Control flies cold shocked after chronic ethanol showed a mild increase in sedation sensitivity, indicating sufficient cold-shock conditions, but chronic tolerance measured at 24 h was unaffected (Fig. 6B,B′). We also directly tested for labile and consolidated intermediate-term memory pathways, ASM and ARM, by measuring tolerance after a 3 h recovery, when ASM and ARM are typically tested. Again, there was no effect of either cold-shock anesthesia or mutation of radish, providing further evidence that chronic exposure does not use the genetic and molecular pathways of ASM and ARM (Fig. 6C,C′). In contrast, rapid tolerance was partially affected by both manipulations and was nearly abolished with the combination (Fig. 6D,D′). Thus, a binge-like ethanol exposure that forms rapid tolerance consists of pathways shared by labile ASM and consolidated ARM.
Finally, we tested whether chronic tolerance depended on CREBB, a transcriptional regulator of LTM that is required in the mushroom bodies for LTM (Yin et al., 1994). We expressed an inhibitor form, CREB2b, in all adult neurons, or specifically in all mushroom body neurons. Brain-wide, but not mushroom body-specific, suppression of CREBB interfered with chronic tolerance (Fig. 7A,A′). Thus, chronic ethanol exposure engages CREBB-dependent transcription outside the mushroom bodies, potentially creating an interplay of mushroom body epigenetic and other brain circuit memory-like states. A model for how rapid and chronic tolerance are separably encoded into the Drosophila brain is shown in Figure 7B.
Chronic ethanol creates a non-canonical LTM-like state. A, Inhibition of CREBB blocks chronic tolerance when expressed in all adult neurons (left), but not when expressed in all adult mushroom body neurons (right). A′, Adult-specific inhibition of CREBB signaling in all neurons (left), but not in mushroom body neurons (right), decreased sensitivity to a 50% ethanol challenge. B, Summary diagram of the encoding of chronic and rapid tolerance in the adult Drosophila mushroom bodies. A, A′, Quantitative data are the mean ± SEM, A, Left, Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; Gal4/UAS versus Gal4, *p = 0.0114; Gal4/UAS versus UAS, *p = 0.0137. Right, One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, p = 0.8458; Gal4/UAS versus UAS, p = 0.7429. A′, Left, Brown–Forsythe ANOVA with Dunnett's T3 multiple-comparisons test; Gal4/UAS versus Gal4, ****p < 0.0001; Gal4/UAS versus UAS, ****p < 0.0001. Right, One-way ANOVA with Dunnett's multiple-comparisons test; Gal4/UAS versus Gal4, **p = 0.0064; Gal4/UAS versus UAS, *p = 0.0385.
Discussion
Chronic and rapid ethanol tolerance are encoded by distinct molecular programs in distinct neural circuits. Moreover, repeated tolerance is another distinct form of tolerance in Drosophila. Thus, the pattern of initial ethanol intake selects different molecular encoding mechanisms of neural plasticity, with different durations and circuitry. Here, we discuss the properties of chronic tolerance, how chronic and rapid tolerance may contribute to the progression toward AUD, and evidence that chronic tolerance is an ethanol-specific form of long-term memory.
Chronic ethanol exposure creates a unique, ethanol-specific memory-like state, differentiating ethanol experience from other experience-dependent memories. Here we compare chronic tolerance to experience dependent memories in flies. We note that we do not explicitly pair chronic ethanol experience with a neutral cue: the chronic ethanol-exposed flies may form a lasting nonassociative memory, or they may use environmental cues, such as ethanol olfactory cues, to form a lasting associative memory. The perdurance of chronic tolerance, 2–3 d, overlaps with two types of memories created by associative conditioning paradigms: the shorter ARM, and the longer LTM (Margulies et al., 2005). Chronic tolerance shares some features of ARM and LTM, but differs from both in critical ways. First, our chronic exposure paradigm more closely resembles massed training, where animals are given multiple training trials without rest resulting in ARM. However, both chronic ethanol tolerance and spaced training LTM depend on de novo protein synthesis, whereas ARM does not (Berger et al., 2004). Like LTM, chronic ethanol tolerance is dependent on CREB signaling in the brain. However, unlike in LTM, CREB functions outside the mushroom bodies for chronic tolerance, indicating that additional as yet undiscovered neural circuitry likely harbors a site of learning for chronic ethanol (Miyashita et al., 2018). Moreover, kayak (c-fos) is selectively induced following the chronic challenge dose. In classical aversive conditioning, kayak is induced by spaced training but not massed training, and kayak is required for LTM expression (Miyashita et al., 2018). We do not yet know whether kay is required for the expression of chronic tolerance. We suggest that chronic ethanol exposure in flies primes neural circuits for new learning via licensing of kayak inducibility and supports the progress of the ethanol experience toward the longer-term debilitating effects of ethanol. In support of this, chronic intermittent ethanol exposure in rats causes changes in subsequent nonethanol learning and memory (Shields and Gremel, 2021).
Ethanol likely creates complex memory-like states. For example, ethanol in flies and mammals is both appetitive and aversive, relative to the time, pattern, and dose of the ethanol experience (Koob and Le Moal, 2001; Lynch and Carroll, 2001; Kaun et al., 2011; Nunez et al., 2018). Moreover, both appetitive and aversive states can be created by the same ethanol experience. Unlike most stimuli that elicit learning and memory, ethanol has direct access to the entire nervous system, where it can exert its pharmacological and metabolic effects. Chronic ethanol exposure does create an appetitive memory-like state since it induces ethanol preference. Appetitive associative conditioning in flies, with food when hungry or water when thirsty, forms protein synthesis-dependent LTM after a single associative training (Krashes and Waddell, 2008; Shyu et al., 2017). However, appetitive LTM is distinct from chronic ethanol tolerance as well as aversive LTM in that it requires radish. Thus, the appetitive memories created by ethanol and other natural rewards are distinct. Together, chronic ethanol encoding shares features with, but is distinct from, appetitive and aversive LTM.
Additional complexity of ethanol memory-like states is hinted at by the distinct forms of memory-like states in rapid tolerance compared with chronic tolerance: rapid tolerance is composed of two shorter-term states that share features of labile ASM and consolidated ARM. Drinking patterns in humans can be complex, mixing binge drinking and chronic intake over relatively short time periods (Kaprio et al., 1987; Sudhinaraset et al., 2016). Thus, ethanol may create multiple interacting memory-like states that can shape future intake.
Epigenetic mechanisms, important for both rapid and chronic ethanol tolerance expression, likely ensure that distinct patterns of ethanol intake are separably encoded and expressed as tolerance. Histone acetylation state and thus chromatin structure may maintain the experience of chronic ethanol exposure, with differential involvement of Class I/II HDACs and the sirtuins. Chromatin state encoding of chronic tolerance is manifest in the altered inducibility of immediate early genes, the selective suppression of Hr38 inducibility by Class I/II HDACs (as suggested by treatment with the drug TSA), and by the active occlusion of rapid tolerance development. Epigenetic suppression of Hr38 inducibility may be a molecular mechanism by which chronic ethanol exposure suppresses rapid tolerance development, since rapid tolerance development requires Hr38 induction and chronic tolerance development does not (Adhikari et al., 2019). An as yet unidentified Class I/II HDAC, and not Sirt1, is likely the effector of Hr38 suppression by chronic ethanol exposure, based on the ability of TSA treatment to restore Hr38 inducibility. Adult Sirt1 promotion of rapid tolerance and inhibition of chronic tolerance underscores the fundamentally different states created by different patterns of initial ethanol exposure. Moreover, our findings with Sirt1 broadly fit with prevailing mammalian models of increased chromatin accessibility with initial ethanol inebriation and chromatin compaction with prolonged ethanol use and with ethanol withdrawal (Berkel and Pandey, 2017). Additional studies targeted to chronic ethanol tolerance-encoding circuitry will be needed to reveal the specific tolerance encoding roles of chromatin states.
Chronic ethanol suppression of rapid tolerance may be because of circuit-level communication from the γ lobes (chronic tolerance) to the α/β lobes (rapid tolerance) of the mushroom body learning and memory centers. Chronic tolerance is limited by Sirt1 specifically in the mushroom body γ lobes, and neural activity in the mushroom bodies is required for chronic tolerance development. Thus, we predict that Sirt1 dampens mushroom body γ lobe neural activity in response to chronic low-dose ethanol. Although the mechanism is currently not understood, a possible mechanism is that Sirt1 regulates GABA receptor function in the γ lobe in response to chronic ethanol exposure to effect a homeostatic maintenance of mushroom body activity through the broadly mushroom body-innervating anterior paired lateral GABAergic neurons; chronic ethanol exposure in rodents causes marked changes in GABAergic neurotransmission (Petrie et al., 2001; Roberto et al., 2003; Liu and Davis, 2009). By contrast, rapid tolerance is promoted by Sirt1 in the mushroom body α/β lobes, and neural activity in the mushroom bodies is required for rapid tolerance development. Thus, we predict that Sirt1 promotes mushroom body α/β lobe neural activity in response to acute inebriating ethanol. Moreover, the mushroom bodies are the site for ethanol reward memories produced from a spaced associative training protocol (Kaun et al., 2011). The association is established in the γ lobes, consolidation occurs in the α′/β′ lobes, and expression requires the α/β lobes. Thus, circuit-level communication between mushroom body lobes is a feature of ethanol reward learning and memory. Interestingly, short-term rapid tolerance and ethanol reward memory expression colocalize in the α/β lobes, and long-term chronic tolerance and ethanol reward learning colocalize in the γ lobes. Furthermore, the initiating ethanol exposures for rapid and chronic tolerance both potentiate ethanol preference, indicating they include appetitive memory-like states that may indicate reward-like actions. Thus, ethanol tolerance uses some of the same brain circuitry as ethanol reward, but tolerance likely uses the circuitry differently. Recent advances in memory mapping indicate that there exists some functional specialization of the mushroom body intrinsic neurons; distinct, as yet uncharacterized lobe subregions may encode tolerance versus other forms of memory (Lee et al., 2020). Alternatively, there may be a physiological ceiling effect that limits the maximal amount of any kind of tolerance expression. If a ceiling to tolerance exists, however, it can be circumvented in animals that are mutant for specific genes (Berger et al., 2004, 2008).
Repeated tolerance in Drosophila is induced by a paradigm that is similar to the regular intermittent drinking in humans that the National Institute on Alcohol Abuse and Alcoholism defines as heavy alcohol use. Repeated or intermittent ethanol exposure is widely used in vertebrate models of AUD to study ethanol-induced plasticity, and it has features that support opponent process models of addiction (Koob and Le Moal, 1997; Carnicella et al., 2014; Nimitvilai et al., 2016; Morales et al., 2018). Repeated exposure to inebriating, but not sedating, ethanol doses in Drosophila resulted in molecular and behavioral responses that are distinct from both rapid and chronic tolerance. For example, rapid and chronic pre-exposures both prime flies for ethanol preference, whereas repeated pre-exposure primes flies for mild ethanol aversion. Thus, repeated exposure appears to create a third form of ethanol plasticity in flies. In rodents, the length of the interstimulus interval determines whether ethanol experience is appetitive or aversive (Cunningham et al., 1997). Optimization of Drosophila repeated exposures may reveal conditions that favor ethanol preference and facilitate the development of new models of AUD in Drosophila.
Footnotes
This work was supported by a National Science Foundation Graduate Research Fellowship to C.L., and by National Institutes of Health (NIH)/National Institute on Alcohol Abuse and Alcoholism Grants 5R21-AA-028352 and 5R21-AA-026066 to F.W.W. We thank Jesus Rascon for help with the experiments.
The authors declare no competing financial interests.
- Correspondence should be addressed to Fred W. Wolf at fwolf{at}ucmerced.edu