Tyramine synthesis, vesicular packaging, and the SNARE complex function coordinately in astrocytes to regulate Drosophila alcohol sedation

Abstract

Identifying mechanisms underlying alcohol-related behaviors could provide important insights regarding the etiology of alcohol use disorder. To date, most genetic studies on alcohol-related behavior in model organisms have focused on neurons, leaving the causal roles of glial mechanisms less comprehensively investigated. Here, we report our studies on the role of Tyrosine decarboxylase 2 (Tdc2), which converts tyrosine to the catecholamine tyramine, in glial cells in Drosophila alcohol sedation. Using genetic approaches that drove transgene expression constitutively in all glia, constitutively in astrocytes and conditionally in glia during adulthood, we found that knockdown and overexpression of Tdc2, respectively, increased and decreased the sensitivity to alcohol sedation in flies. Manipulation of the genes tyramine β-hydroxylase and tyrosine hydroxylase, which respectively synthesize octopamine and dopamine from tyramine and tyrosine, had no discernable effect on alcohol sedation, suggesting that Tdc2 affects alcohol sedation by regulating tyramine production. We also found that knockdown of the vesicular monoamine transporter (VMAT) and disruption of the SNARE complex in all glia or selectively in astrocytes increased sensitivity to alcohol sedation and that both VMAT and the SNARE complex functioned downstream of Tdc2. Our studies support a model in which the synthesis of tyramine and vesicle-mediated release of tyramine from adult astrocytes regulates alcohol sedation in Drosophila. Considering that tyramine is functionally orthologous to norepinephrine in mammals, our results raise the possibility that gliotransmitter synthesis release could be a conserved mechanism influencing behavioral responses to alcohol as well as alcohol use disorder.

1 |. INTRODUCTION

Alcohol abuse affects and is affected by central nervous system (CNS) function.¹ There is consequently a continuing effort to use model organisms to identify mechanisms underlying alcohol-related behaviors to better understand the role of the CNS in alcohol abuse. Although the central nervous system contains glia and neurons as principal cell types, the preponderance of research to date has focused on the role of neuronal mechanisms in alcohol-related behaviors. Interestingly, though, the largest number of differentially expressed genes in post-mortem alcohol-dependent human brains were in glia,² suggesting that much is likely to be learned by investigating the role of glia in behavioral responses to alcohol.

Recent and pioneering studies have begun to offer insights into the effect of alcohol on glia and how glia contribute to alcohol-related behavior. For example, calcium signaling genes are upregulated in astrocytes after chronic alcohol administration in mice,^3,4 expression of the glial cytoskeletal protein GFAP is altered by alcohol exposure in rodents and humans,^5,6 and alcohol exposure correlates with increased hemichannel opening in mouse hippocampal astrocytes.⁷ Consistent with the possibility that glia are directly involved in behavioral responses to alcohol, increasing intracellular calcium in astrocytes via designer receptors exclusively activated by designer drugs (DREADDs) decreased motivation for alcohol after a 3-week abstinence period in rats⁸ and serotonin 2C receptor expression in astrocytes regulates alcohol intake in mice.⁹ Additionally, mutation of the gene moody in flies, which is expressed in CNS surface glia, alters alcohol-induced loss of postural control,¹⁰ Drosophila perineural glia have been associated with alcohol tolerance,¹¹ and expression of Cysteine proteinase 1 in adult brain cortex glia is required for normal alcohol sedation.¹² Thus, there is good evidence that CNS glia respond to, and are involved in, behavioral responses to alcohol.

Here, we report results of studies using the fruit fly, Drosophila melanogaster, to further explore the role of glia in alcohol-related behavior. Flies are well-suited for such studies for many reasons, including (i) they have conserved behavioral responses to alcohol,¹³ (ii) there is a large suite of genetic tools available to manipulate gene expression in flies,^14,15 and (iii) many genes that affect fly alcohol behavior have been implicated in various aspects of alcohol abuse in humans.^16,17 Importantly, flies also have several glial cell subtypes in the central nervous system (astrocytes, cortex glia, ensheathing cells, perineural glia, and subperineural glia) that collectively share many morphological and functional attributes of mammalian glia. For example, fly astrocytes maintain ion homeostasis, remove neurotransmitters from the synapse, have hemichannels and gap junctions, produce Ca²⁺ oscillations, and release gliotransmitters like their mammalian counterparts.^18–20 Additionally, glia regulate multiple behaviors in both flies and mammals including circadian rhythm,^21–23 locomotion,²⁴ chemotaxis, and startle responses²⁵ as well as sleep regulation.²⁶ Flies are therefore well-suited for investigating the molecular contribution of glia generally, and astrocytes more specifically, to behavioral responses to alcohol.

We performed a targeted ethanol sedation screen of genes shown to be expressed in Drosophila CNS glial cells through genomic and genetic studies.^21,22 We identified the gene tyramine decarboxylase 2 (Tdc2) as a key regulator of alcohol sedation in flies via its function in astrocytes. Tdc2 is a brain-specific enzyme that converts the amino acid tyrosine to the invertebrate catecholamine tyramine.²⁷ Manipulation of Tdc2 expression in all glia, selectively in astrocytes, or conditionally in glia during adulthood altered alcohol sedation. Additionally, manipulation of the vesicular monoamine transporter (VMAT) and the SNARE complex in all glia, astrocytes, and in adult glia also influenced alcohol sedation and did so by functioning downstream of Tdc2. Our studies support a model in which Tdc2 activity in astrocytes produces tyramine, which is packaged into a synaptic vesicle and released via the SNARE complex, thereby modulating alcohol sedation in Drosophila. Tyramine is functionally orthologous to mammalian norepinephrine, and adult human astrocytes express dopamine β-hydroxylase,²⁸ the enzyme responsible for norepinephrine synthesis, raising the possibility that astrocyte-mediated synthesis and release of norepinephrine might influence behavioral responses to ethanol in other species as well as alcohol use disorder.

2 |. MATERIALS AND METHODS

All flies were reared under standard conditions as described previously.²⁹ Flies were grown in 6-ounce polypropylene Drosophila bottles (Fisher Scientific, Hampton, NH) with food medium containing 10% sucrose, 3.3% cornmeal, 2% yeast, 1% agar, 0.2% Tegosept, and antibiotics (0.1 g/L ampicillin, 0.02 g/L tetracycline, 0.125 g/L chloramphenicol) with active dry yeast on top. Flies were housed in an environmental chamber at 25°C and 60% relative humidity with a 12-h light/dark cycle. All comparisons between groups were based on studies with flies grown, handled, and tested together.

2.1 |. Fly stocks

Detailed genotypes are in Table S5. Citations for Gal4 drivers are provided in the main text. All RNAi and UAS transgenic flies were obtained from the Bloomington Drosophila Stock Center (BDSC; Bloomington, IN) or the Vienna Drosophila Resource Center (VDRC, Vienna, Austria), unless noted otherwise. w¹¹¹⁸ reference stocks were obtained from the VDRC to control for the genetic background of all flies obtained from this center (GD transgenic control: stock number 60000; shRNA transgenic control: 60200). Within Table S5, all mini-w stocks marked with an asterisk were backcrossed to our standard reference strain, w[A] (w¹¹¹⁸ in an isogenic background; BDSC, stock number 5905) for seven generations to normalize the genetic background. All y⁺ stocks marked with a ^$ were backcrossed to a y¹w¹ strain (stock number 1495, BDSC) for seven generations to normalize the genetic background. All double transgenic flies were created via standard crosses, and we performed PCR to confirm that all transgenes were present.

2.2 |. Ethanol sedation

The day prior to behavioral testing, adult flies (4 days post-eclosion) were collected under CO₂ anesthesia, and 11 adult female flies were placed into fresh non-yeasted food vials (standard food medium without active dry yeast on top). Flies recovered in food vials stored upside down (food side up) overnight in the environmental chamber (described above).

Ethanol sedation experiments were performed 3–7 h after lights on at 23–25°C and 55–65% relative humidity essentially as described.^29,30 After acclimating in the behavioral testing room for 1–2 h, flies were transferred to empty polystyrene food vials (VWR, Radnor, PA). A cellulose acetate Flug (FlyStuff, San Diego, CA), was inserted approximately 2 cm into the bore of the vial. The number of inactive flies was recorded for each vial (typically 0–1 flies/vial) prior to ethanol administration (time point zero). One milliliter of freshly prepared 85% ethanol was applied to the top of each Flug, and the vials were sealed with a rubber stopper. Every 6 min, each vial was gently tapped on the table 3 times and the number of sedated flies was recorded 30 s later. Flies were recorded as sedated if they were motionless at the bottom of the vial. The ethanol sedation experiment ended when all flies in each vial were sedated, approximately 60–90 min. The percentage of active flies over time was calculated for each vial, and the time required for 50% of the flies in each vial to become sedated (sedation time 50, ST50) was interpolated from sigmoidal curve fits using Excel (Microsoft, Redwood, WA).^29,30 Each vial of 11 flies corresponded to n = 1, and up to 24 vials were tested in each ethanol sedation behavioral experiment.

2.3 |. GeneSwitch induction

One hundred microliters of 1-mM Mifepristone (RU486; Sigma Aldrich, St. Louis, MO) or vehicle (100% ethanol) were added to the surface of solidified food in vials and allowed to dry overnight in an environmental chamber. Flies were collected and placed on food topped with RU486 (induced) or vehicle (control) for 6 days total. Flies were transferred to fresh drug- or vehicle-treated food vials after 3 days. Pre-treatment of food vials with ethanol vehicle did not affect ST50 values (no ethanol, 31.8 ± 1.15 min; ethanol pre-treated, 30.5 min ± 1.97 min; unpaired two-tailed t test, p = 0.6109, n = 8).

2.4 |. Internal ethanol

Flies were exposed to 85% ethanol as described for measuring ethanol sedation sensitivity as described above. After exposure to ethanol for approximately 15 or 30 min, flies were transferred to 1.5 ml snap-cap tubes and frozen at −80°C. Frozen flies were homogenized in ice-cold ddH₂O and centrifuged at 14,000 rpm at 4°C for 20 min. The internal ethanol concentration of the supernatant was determined using Alcohol Reagent Set (Pointe Scientific Inc., Canton, MI) according to the manufacturer’s instructions or via gas chromatography as described previously.^31,32

2.5 |. Whole brain imaging and immunodetection

Whole brains from adult (4 days post-eclosion) female flies were dissected in 0.3% Phosphate buffer Triton X-100 (PSBT) under a dissecting microscope. Dissected brains were mounted to a polyleucine cover slip, fixed in 4% paraformaldehyde for 40 min at room temperature and blocked with 5% normalized goat and donkey serum overnight. Brain were incubated with primary antibodies diluted in block at 4°C for approximately 48 h. Brains were washed with 0.3% PBST three times for 20 minutes and incubated for approximately 48 h with secondary antibodies diluted in block at 4°C. Brains were then washed three times with 0.3% PBT for 20 min, dehydrated with increasing concentrations of ethanol (30, 50, 70, 90, 100%) and xylenes for 12 to 18 min, and mounted onto a glass slide with DPX mounting medium.

The following primary antibodies at the indicated concentrations from the listed sources were used: rabbit anti-Tdc2 (1:200; Covalab, Villeurbanne, France), chicken anti-GFP (1:1000, Aves Labs Davis, CA, USA), and mouse anti-NC82 (1:50; Developmental Studies Hybridoma Bank, Iowa City, Iowa). The following secondary antibody was used: donkey anti-chicken Alexa 488, donkey anti-mouse RRX, and donkey anti-rabbit Alexa 647 (1:400; ThermoFisher, Waltham, MA).

All images were collected using a Zeiss LSM 800 confocal microscope (Carl Zeiss Microscopy, LLC, Thornwood, NY) housed in the Doe laboratory at the University of Oregon. Confocal images using a pin hole of 1 Airy disc unit and Nyquist sampling were collected from each adult brain. Images were taken with a 40× oil immersion objective, resulting in approximately 400× magnification of the tissue. The laser power, gain, and offset values were kept constant for all images compared within an experiment. The Z-slice thickness was kept constant for all samples (z = 1.04 μm), as was the overall thickness of each brain region analyzed between samples and genotypes (50 Z-slices). Individual images of each brain region (antennal lobes, central complex, and subesophageal zone, identified as described previously³³) were captured, and quantification of colocalization between Tdc2 and GFP was determined using the colocalization feature in Imaris image analysis software. In Imaris, all samples underwent automatic thresholding, and all colocalization analyses were confined to the brain via boundaries defined by staining neuropil with anti-NC82, indicated by a red dashed line in Figure S2G–I. A colocalization channel was created to visualize colocalization throughout the sample (as seen in Figures 2 and S2), and Manders colocalization coefficient is reported in Figure S3 and in the text. The representative images in Figures 2 and S2 are Z-stack three-dimensional reconstructions. All quantification in Figure S3 and in the text was performed on individual Z-slices within the Z-stack, averaged together to for an N = 1 for each brain region.

FIGURE 2.

Endogenous Tdc2 colocalizes with GFP expressed in astrocytes in three brain regions. Endogenous Tdc2 visualization (white, first column; magenta, fourth column), expression of mCD8::GFP (membrane bound) in astrocytes via alrm-Gal4 (white, second column; green, fourth column), and colocalization between endogenous Tdc2 and astrocyte GFP (white, third and fourth columns). Images are compiled Z-slices (3D reconstructions) within the (A) antennal lobes, (B) central complex and (C) subesophageal zone (representative images, 40× objective oil immersion). Additional images provided in Figure S2

2.6 |. Statistical analyses and data presentation

All statistical analyses (one-way and two-way ANOVA with Bonferroni’s multiple comparison tests) were performed with Prism 8.4.3 (GraphPad Software, San Diego, CA, USA). Chi-square analyses were performed with the Chi-Square Calculator from Social Science Statistics (soscistatistics.com). Numerical data in graphs are shown as individual measurements, means (represented by light grey bars), and error bars for SEM. Numerical data in tables are show as mean and SEM. The number of replicates (N) is indicated for each data set in the corresponding legend.

3 |. RESULTS

3.1 |. Tdc2 functions in CNS glia, specifically astrocytes, to regulate alcohol sedation

To identify candidate genes that could influence alcohol-related behaviors by functioning in glia, we compiled a list of genes known to be expressed in Drosophila CNS glial cells through previous computational, microarray, and TRAP RNA-seq gene expression analyses as well as forward genetic studies^21,22 and selected 33 candidate genes we believed had plausible connections to alcohol behavior. We then determined whether manipulation of these 33 genes individually in all CNS glia influenced alcohol sedation in Drosophila. These results suggested that Tdc2 was important for alcohol sedation, the focus of this manuscript. Other results from this screen have been reported elsewhere.¹²

For the studies described here, flies with pan-glial Gal4 (via repo-Gal4³⁴) driven expression of two unique Tdc2 RNAi transgenes (JF01910 and v330541, hereafter Tdc2 RNAi #1 and #2, respectively) became sedated significantly faster in the presence of alcohol compared with controls containing the repo-Gal4 or either of the RNAi transgenes alone (Figure 1A,B). This is evidenced by a significantly decreased sedation time 50 (ST50, the time required for 50% of the flies to become sedated).

FIGURE 1.

Manipulating Tdc2 expression constitutively in all glia or specifically in astrocytes alters alcohol sedation. (A, B) ST50 values were reduced in flies with the pan-glial driver repo-Gal4 and a Tdc2 RNAi transgene (blue symbols) compared with control flies with either repo-Gal4 alone (grey symbols) or the respective RNAi transgene alone (grey symbols) (Panel A, Tdc2 RNAi#1: one-way ANOVA, p < 0.0001; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 8; Panel B, Tdc2 RNAi #2: one-way ANOVA, p < 0.0001; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 16). (C, D) ST50 values were decreased in flies expressing the astrocyte-specific driver alrm-Gal4 and a Tdc2 RNAi transgene (blue symbols) compared with control flies containing either the astrocyte Gal4 driver (grey symbols) or the respective RNAi transgene (grey symbols) alone (Panel C, Tdc2 RNAi #1: one-way ANOVA, p < 0.0001; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 8; Panel D, Tdc2 RNAi #2: one-way ANOVA, p < 0.0001; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 16)

In principle, Tdc2 could influence alcohol sedation through the collective effect of its role in all CNS glia or by a role in a single glial cell subtype. Therefore, we used a panel of extensively characterized Gal4 drivers that each principally express in astrocytes (alrm-Gal4³⁵), ensheathing cells (TIFR-Gal4³⁶ and Mz0709-Gal4³⁵), cortex glia (Np2222-Gal4³⁷), perineural glia (Indy-Gal4¹¹), or subperineural glia (Gli-Gal4³⁸) to address whether manipulating Tdc2 expression in specific subsets of glia altered ST50 values. In these studies, a meaningful effect of RNAi expression would manifest as (i) a significant overall effect of group detected by a one-way ANOVA and (ii) significant differences between flies expressing RNAi (Gal4/RNAi) and the two relevant controls (Gal4/+ and RNAi/+). Additionally, all data within each individual figure panel are derived from animals reared, collected, and tested side-by-side, all statistical comparisons are restricted to data within a panel, and interpretations from our experiments are based on these statistical comparisons. Any differences in ethanol sedation behavior of genotypes between panels are presumably due to the widely appreciated day-to-day variability in behavior.

Expression of Tdc2 RNAi in astrocytes (using alrm-Gal4) significantly decreased ST50 values compared with controls (Figure 1C,D; Table S1). Expression of Tdc2 RNAi in all other glial cell subtypes did not alter ST50 values (Table S1). Additionally, Tdc2 RNAi expression in neurons did not alter ST50 values compared with controls (Figure S1A). Taken together, these data indicate that Tdc2 plays a role in alcohol sedation by functioning within astrocytes, but likely not within other glial cell subtypes or neurons.

Tdc2 is known to be expressed in Drosophila neurons.^27,39,40 Importantly, though, Jackson and co-workers found through Translating Ribosome Affinity Purification followed by genome-wide expression profiling via RNA sequencing (TRAP-Seq) that Tdc2 is also expressed in Drosophila astrocytes.²² Additionally, the mammalian ortholog of Tdc2 (dopamine β-hydroxylase) is expressed in adult human astrocytes.²⁸ To confirm the previously reported astrocyte expression of Tdc2 in flies,²² we used immunofluorescence to detect endogenous Tdc2 and adult Drosophila astrocytes labeled with membrane bound GFP (UAS-mCD8::GFP driven by alrm-Gal4) in whole adult brains. Consistent with previous reports,^22,28 we found that endogenous Tdc2 was expressed broadly in the brain (not shown). Visual inspection suggested that colocalization between endogenous Tdc2 and GFP-labeled astrocytes was greatest in the antennal lobes (Figure 2A.i), central complex (Figure 2B.i), and subesophageal zone (SEZ, Figure 2C.i), focusing our studies in these three areas. Additional representative images are provided in Figure S2A,D,G. Five to 15% of the Tdc2 signal in these three brain regions colocalized with GFP expressed in astrocytes (Figure S3). Other major anatomical areas including the optic lobes, mushroom bodies, and superior neuropils had no obvious colocalization (not shown). The most parsimonious interpretation of the data in Figures 1, 2, S2, and S3, in alignment with a previously report,²² is that Tdc2 influences alcohol sedation in Drosophila by functioning in astrocytes.

We also assessed whether expression of Tdc2 RNAi transgenes disrupted Tdc2 localization in astrocytes. Expression of two independent Tdc2 RNAi transgenes in astrocytes significantly decreased colocalization of endogenous Tdc2 and GFP expressed in astrocytes via alrm-Gal4 in the antennal lobes (Figure S2A–C), central complex (Figure S2D–F), and subesophageal zone (Figure S2H–I) (quantification of colocalization within compiled Z-stacks from each brain region: Figure S3). The disruption of Tdc2 localization in astrocytes by two independent RNAi transgenes within three different brain regions strongly suggests that these Tdc2 RNAi transgenes knockdown endogenous Tdc2 in astrocytes.

3.2 |. Tdc2 regulates alcohol sedation sensitivity in CNS glia during adulthood

To determine if Tdc2 expression in glia during adulthood is important for alcohol sedation, we utilized a steroid (mifepristone, RU486) inducible pan-glial Gal4 driver, GliaGS,^12,41 thereby allowing conditional expression of UAS-transgenes in glia during adulthood. Compared with vehicle, induction of the Tdc2 RNAi transgenes in all glia during adulthood (i.e., in flies with GliaGS and a Tdc2 RNAi transgene fed RU486) significantly decreased ST50 values (Figure 3A,B). Conversely, overexpression of Tdc2 via two previously validated³⁹ transgenes in all glia during adulthood (i.e., in flies with GliaGS and a UAS-Tdc2 transgene fed RU486) significantly increased ST50 values (Figure 3C). Compared with vehicle, treatment with RU486 did not significantly alter ST50 values in control genotypes, which had the GliaGS driver, the Tdc2 RNAi transgenes or the UAS-Tdc2 transgenes alone (Figure 3). (Note that the key comparisons in studies of this design are between vehicle and RU486-treated flies within each genotype because those comparisons assess experimentally manipulated expression of Tdc2. Also note that comparisons between genotypes are not as informative because they are unrelated to explicit manipulation of Tdc2 expression.) Taken together, these data indicate that knocking down Tdc2 in all glia during adulthood decreases ST50, and overexpressing Tdc2 in all glia during adulthood increases ST50. Unfortunately, Tdc2 overexpression constitutively in all glia (via repo-Gal4) and in astrocytes (via alrm-Gal4) caused developmental lethality, precluding assessment of ethanol sedation in adults in these animals.

FIGURE 3.

Manipulating Tdc2 expression in all glia during adulthood influences alcohol sedation. (A, B) Compared with vehicle (grey circles), treatment with RU486 (blue squares) decreased ST50 values in flies with the GliaGS driver and a Tdc2 RNAi transgene, but not in control flies with either GliaGS or the respective RNAi transgene alone (Panel A, Tdc2 RNAi #1: two-way ANOVA; RU486, p = 0.0569; genotype, p < 0.0001; interaction, p = 0.1353; *Bonferroni’s multiple comparisons between vehicle and RU486, p < 0.05; n = 8; Panel B, Tdc2 RNAi #2: two-way ANOVA; RU486, p = 0.4055; genotype, p < 0.0001; interaction, p = 0.0456; *Bonferroni’s multiple comparisons between vehicle and RU486, p < 0.05; n = 8). (C) Compared with vehicle (grey circles), treatment with RU486 (blue squares) increased ST50 values in flies with the GliaGS driver and a UAS-Tdc2 transgene, but not in control flies with either GliaGS or either of the UAS-Tdc2 transgenes alone (two-way ANOVA; RU486, p = 0.0012; genotype, p = 0.0004; interaction, p = 0.0003; *Bonferroni’s multiple comparisons between vehicle and RU486, p < 0.05; n = 6–7)

3.3 |. Manipulation of Tdc2, but not tyrosine hydroxylase or tyramine β-hydroxylase, affects alcohol sedation

The amino acid tyrosine is converted to the catecholamines tyramine by Tdc2 and dopamine by tyrosine hydroxylase (Th). Tyramine can be converted to octopamine by tyramine β-hydroxylase (Tbh).²⁷ We postulated that manipulation of Tdc2 might either influence alcohol sedation by altering tyramine levels directly, or through secondary effects on dopamine or octopamine synthesis. To test this, we targeted Tbh and Th expression in glia during adulthood using GliaGS and previously validated RNAi and UAS transgenes.^42–46 Induction of Tbh RNAi or UAS transgenes (Figure S4) and induction of Th RNAi or UAS transgenes (Figure S5) individually in all glia during adulthood did not alter ST50 values. Although additional studies would be needed to fully rule out a role for glial Tbh and Th, and therefore octopamine and dopamine, in alcohol sedation, our results support the hypothesis that manipulation of Tdc2 expression in glia influences alcohol sedation by altering tyramine levels.

3.4 |. Tdc2 expression within glia regulates alcohol sedation through a pharmacodynamic mechanism

To determine whether glial Tdc2 was altering alcohol sedation via a pharmacodynamic or pharmacokinetic mechanism, we measured alcohol levels in flies expressing Tdc2 RNAi via repo-Gal4 and control genotypes after prescribed periods of alcohol exposure. As expected,^31,32 internal alcohol levels significantly increased the longer flies were exposed to alcohol (Table S2, Time factor). Importantly, though, genotype did not affect internal alcohol levels (Genotype factor), indicating that constitutive pan-glial Tdc2 knockdown did not discernably alter alcohol metabolism (Table S2). Additionally, relative to vehicle-treated controls, adult-specific Tdc2 knockdown or overexpression in glia (via GliaGS) did not alter internal alcohol levels (Table S3). Thus, manipulating Tdc2 in glia constitutively or specifically during adulthood altered ST50 values without influencing the net uptake or metabolism of alcohol. The level of Tdc2 expression in adult glia is therefore a key regulator of alcohol sedation in flies via a pharmacodynamic mechanism.

3.5 |. Alcohol sedation is influenced by vesicular packaging and release machinery in glia

In Drosophila neurons, tyramine is released through vesicular exocytosis into the synaptic space, where it functions as a neurotransmitter.^25,39,47 If tyramine production and release in glia mediates alcohol sedation, we predicted that inhibiting the vesicular exocytosis machinery in glia might have the same effect on alcohol sedation as Tdc2 knockdown. The vesicular monoamine transporter (VMAT) packages monoamines, including tyramine, into vesicles in neurons⁴⁸ and is expressed in mammalian and fly astrocytes,^22,49–51 leading us to postulate that this transporter might influence alcohol sedation. Constitutive expression of two unique VMAT RNAi transgenes (v4856 and v104072, hereafter VMAT RNAi #1 and VMAT RNAi # 2, respectively) in all glia via repo-Gal4 significantly decreased ST50 (Figure 4A,B). Similarly, flies expressing VMAT RNAi in astrocytes (via alrm-Gal4) had significantly decreased ST50s compared with controls (Figure 4C,D). RU486-induced expression of VMAT RNAi in all glia during adulthood significantly decreased ST50s compared with vehicle control within its genotype, whereas treatment with RU486 had no effect in control genotypes compared with their respective vehicle control (Figure 4E, F). All VMAT RNAi transgenes have been previously validated to knockdown VMAT expression,^52,53 and VMAT expression in fly astrocytes has been previously demonstrated by immunofluorescence,⁵⁰ single-cell RNA sequencing,⁵¹ and TRAP-sequencing.²² Thus, knocking down VMAT either constitutively in all glia, constitutively in astrocytes, and in all glia during adulthood decreased ST50 in flies.

FIGURE 4.

VMAT expression in glia is required for normal alcohol sedation. (A, B) ST50 values were reduced in flies with the pan-glial driver repo-Gal4 and a VMAT RNAi transgene (blue symbols) compared to control flies with either repo-Gal4 alone (grey symbols) or the respective RNAi transgene alone (grey symbols) (Panel A, VMAT RNAi #1: one-way ANOVA, p = 0.0035; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 8; Panel B, VMAT RNAi #2: one-way ANOVA, p = 0.0020; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 8). (C, D) ST50 values were decreased in flies expressing the astrocyte-specific driver alrm-Gal4 and a VMAT RNAi transgene (blue symbols) compared with control flies containing either the astrocyte Gal4 driver (grey symbols) or the respective RNAi transgene (grey bars) alone (Panel C VMAT RNAi #1: one-way ANOVA, p < 0.0001; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 8; Panel D, VMAT RNAi #2: one-way ANOVA, p = 0.0014; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 8). (E, F) Compared with vehicle (grey circles), treatment with RU486 (blue squares) decreased ST50 values in flies with the GliaGS driver and a VMAT RNAi transgene, but not in control flies with either GliaGS or the respective RNAi transgene alone (Panel E, VMAT RNAi #1: two-way ANOVA; RU486, p = 0.0420; genotype, p = 0.2389; interaction, p = 0.0039; *Bonferroni’s multiple comparisons between vehicle and RU486, p < 0.05; n = 8; Panel F, VMAT RNAi #2: two-way ANOVA; RU486, p = 0.0507; genotype, p = 0.2847; interaction, p = 0.0189; *Bonferroni’ multiple comparisons between vehicle and RU486, p < 0.05; n = 8)

We measured internal alcohol to determine whether expression of VMAT RNAi might reduce ST50 values by increasing internal alcohol concentrations. Internal ethanol concentrations increased with exposure time in repo-Gal4/VMAT RNAi and control flies as expected (Table S2, Time factor). There was also an overall effect of genotype on internal alcohol in these studies (Table S2, Genotype factor), but internal ethanol levels in flies expressing VMAT RNAi were not consistently different than in control flies. Specifically, internal alcohol in repo-Gal4/UAS-VMAT RNAi#1 and repo-Gal4/+ controls was indistinguishable after 15 and 30 min of exposure, whereas it was higher in repo-Gal4/UAS-VMAT RNAi#1 compared with the UAS-VMAT RNAi#1 control at both time points (Table S2, Multiple comparisons). In contrast, while internal ethanol was indistinguishable in repo-Gal4/VMAT RNAi#2 and the UAS-VMAT#2 control, it was lower in repo-Gal4/VMAT RNAi#2 flies than in the repo-Gal4/+ control (Table S2, Multiple comparisons). Additionally, expression of VMAT RNAi #1 in all glia during adulthood did not change internal ethanol, whereas expression of VMAT RNAi #2 significantly decreased internal ethanol levels compared with vehicle-treated controls (Table S3). Given that internal ethanol in flies expressing VMAT RNAi was not changed in five of 10 pairwise comparisons to control groups, was increased in two of 10 comparisons, and was decreased in three of 10 comparisons, the somewhat variably altered internal ethanol concentrations cannot explain the consistently reduced ST50 values seen in flies expressing VMAT RNAi (Figure 4A,B,E,F). Our interpretation of these data is that manipulation of VMAT in glia alters the pharmacodynamic response of flies to alcohol.

Given that manipulation of Tdc2 and VMAT in glia produced similar changes in alcohol sedation, we postulated that these two genes might function in the same pathway. To address this possibility, we assessed alcohol sedation in flies with Tdc2 over-expressed and VMAT knocked down using three approaches: pan-glial expression via repo-Gal4, astrocyte-specific expression using alrm-Gal4, and adult-induced expression in all glia using GliaGS. Over-expression of Tdc2 in all glia and in astrocytes was lethal (missing bars in Figure 5A,B). Expression of VMAT RNAi in all glia (Figure 5A) or in astrocytes (Figure 5B) significantly decreased ST50 values as described above (Figure 4). Interestingly, expression of VMAT RNAi suppressed the lethality associated with Tdc2 over-expression (quantified for repo-Gal4 in Table S4). More importantly, the ST50 values in flies over-expressing Tdc2 in conjunction with the VMAT RNAi were statistically indistinguishable from flies expressing only VMAT RNAi (Figure 5A, all glia; Figure 5B, astrocytes), suggesting that VMAT functions downstream of Tdc2. Lethality due to constitutive overexpression of Tdc2 was not suppressed by co-expression of GFP in glia via repo-Gal4 (Table S4), consistent with previous reports showing that transgene expression in the Gal4-UAS system is not affected by the presence of two UAS-transgenes.^12,54 The simplest interpretation of our data is that VMAT has a bona fide genetic interaction with overexpression of Tdc2. Taken together, our results strongly support a model in which knockdown of VMAT is epistatic to over-expression of Tdc2, thereby placing VMAT biochemically downstream of Tdc2. Additionally, our results suggest that alcohol sedation might be influenced by tyramine packaging into vesicles within glia, specifically within astrocytes.

FIGURE 5.

VMAT functions downstream of Tdc2. (A) ST50 values were decreased in flies with the pan-glial driver repo-Gal4 and a VMAT RNAi transgene (blue symbols) and in flies with repo-Gal4, the VMAT RNAi and UAS-Tdc2 transgenes (red symbols) compared with control flies (grey symbols). ST50 values were not different in flies with the repo-Gal4 and VMAT RNAi transgenes (blue symbols) and in flies with the repo-Gal4, VMAT RNAi and UAS-Tdc2 transgenes (red symbols) (one-way ANOVA, p = 0.0003; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 8). Flies with the pan-glial driver repo-Gal4 and the UASTdc2 transgene did not emerge as adults (no symbol). (B) ST50 values were decreased in flies with the astrocyte-specific driver alrm-Gal4 and the VMAT RNAi transgene (blue symbols) and in flies with alrm-Gal4, the VMAT RNAi and UAS-Tdc2 transgenes (red symbols) compared with control flies (grey symbols). ST50 values were not different between flies with the alrm-Gal4 and VMAT RNAi transgene (blue symbols) and flies with the alrm-Gal4, VMAT RNAi and UAS-Tdc2 transgenes (red symbols) (one-way ANOVA, p < 0.0001; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 8). Flies expressing both alrm-Gal4 and UASTdc2 transgenes did not emerge as adults (no symbol)

Synaptic vesicles loaded with transmitters dock to the plasma membrane prior to releasing their contents into the synapse.⁵⁵ This process is, in part, mediated by the SNARE complex in both glia and neurons.^55–57 We therefore reasoned that the SNARE complex in glia might be required for normal alcohol sedation, and that this complex might also function downstream of Tdc2. We consequently used tetanus toxin, which cleaves synaptobrevin, to inhibit the SNARE complex and block synaptic transmission.⁵⁸ We used two expression strategies to test this theory: constitutive expression of the tetanus toxin in astrocytes via alrm-Gal4 and RU486-induced expression in glia during adulthood via GliaGS. Expression of the tetanus toxin in all glia caused immediate post-eclosion lethality, so this manipulation was not pursued further. Expression of tetanus toxin (UAS-TeTx) in astrocytes significantly decreased ST50 values (Figure 6A). Additionally, induction of tetanus toxin expression in adult glia significantly decreased ST50 values compared with vehicle-treated controls of the same genotype (Figure 6B). Internal ethanol levels were not altered by expressing UAS-TeTx constitutively in astrocytes (Table S2) or conditionally in glia during adulthood (Table S3), suggesting that the SNARE complex in glia (i) does not significantly affect internal alcohol levels and therefore (ii) influences alcohol sedation through a pharmacodynamic mechanism.

FIGURE 6.

Expression of tetanus toxin in glia alters alcohol sedation and is epistatic to Tdc2. (A) ST50 values were decreased in flies with alrm-Gal4 and the UAS-TeTx transgene (blue symbols) compared to control flies with either arlm-Gal4 (grey symbols) or the UAS-TeTx transgene (grey symbols) alone (one-way ANOVA, p = 0.0004; *Bonferroni’s multiple comparison versus controls, p < 0.05; n = 15–16). (B) Compared with vehicle (grey circles), treatment with RU486 (blue squares) decreased ST50 values in flies with the GliaGS driver and the UAS-TeTx transgene, but not in control flies with either GliaGS or the UAS-TeTx transgene alone (two-way ANOVA; RU486, p = 0.0002; genotype, p < 0.0001; interaction, p = 0.0009; *Bonferroni’s multiple comparisons between vehicle and RU486, p < 0.05; n = 8). (C) ST50 values were decreased in flies with alrm-Gal4 and the UAS-TeTx transgene (blue symbol) and in flies with alrm-Gal4, the UAS-TeTx and UAS-Tdc2 transgenes (red symbols) compared with control flies. ST50 values were not different between flies alrm-Gal4 and the UAS-TeTx transgene (blue symbols) and flies with alrm-Gal4, the UAS-TeTx and UASTdc2 transgenes (red symbols) (one-way ANOVA, p < 0.0001; *Bonferroni’ multiple comparison versus controls, p < 0.05; n = 8). Flies expressing both alrm-Gal4 and UAS-Tdc2 transgenes did not emerge as adults (no symbols)

The effect of expressing tetanus toxin, and therefore inhibition of the SNARE complex, in astrocytes and adult glia was similar to that of knocking down Tdc2 (Figures 1 and 2), suggesting that the SNARE complex might function in the same pathway as Tdc2. To formally address this possibility, we assessed alcohol sedation in flies with Tdc2 over-expression and tetanus toxin expression. Expression of tetanus toxin in astrocytes suppressed the lethality due to Tdc2 over-expression, and these flies had significantly decreased ST50s. Additionally, the ST50 value of these flies was indistinguishable from flies that expressed tetanus toxin alone in astrocytes (Figure 6C). Taken together, these data strongly support a role for the SNARE complex in regulating alcohol sedation by functioning within astrocytes and adult glia. Additionally, these data argue that SNARE-dependent vesicle-mediated release is functionally downstream of Tdc2 in glia, specifically astrocytes, within the context of alcohol sedation.

4 |. DISCUSSION

A more detailed understanding of the genes and mechanisms that influence behavioral responses to alcohol could ultimately facilitate the development of novel diagnostic and treatment options for individuals that abuse the drug. Much of the genetic analysis of alcohol behavior in model organisms (mainly mice, flies, and worms) has focused on genes that function in neurons, leaving mechanisms driven by other cell types less comprehensively studied. Our studies on Tdc2, glia/astrocytes, and fly alcohol sedation help fill this gap. Here, we show that (i) knockdown and overexpression of Tdc2 in glia makes flies sensitive and resistant, respectively, to alcohol sedation, (ii) VMAT and the SNARE complex influence alcohol sedation by functioning in glia, (iii) VMAT and the SNARE complex alter alcohol sedation by functioning downstream of Tdc2 in glia, and (iv) these findings map to astrocytes and adulthood specifically. Our data support a model in which astrocytes, during adulthood, influence alcohol sedation by synthesizing and releasing tyramine into the synapse through SNARE-dependent vesicular exocytosis.

Tyramine in flies is functionally homologous to norepinephrine in mammals,²⁷ and adult human astrocytes express dopamine β-hydroxylase,²⁸ the enzyme responsible for norepinephrine synthesis, suggesting that our findings in flies may be translatable to humans. Given that resistance to alcohol responses in humans is linked to the propensity to abuse it,⁵⁹ our findings raise the possibility that astrocytes and tyramine/norepinephrine may be key contributors to AUD and problematic alcohol consumption through their role in mediating alcohol sensitivity as a gliotransmitter.

Interestingly, mutations in Tdc2 alter cocaine sensitivity in flies,³⁹ but the cell type responsible was not identified. Our data raise the possibility that Tdc2 function in glia may also be mediating cocaine-related behaviors. This also suggests that Tdc2 function, and therefore tyramine synthesis, may broadly regulate behavioral responses to drugs of abuse.

Although synaptic vesicle exocytosis is a slower process in astrocytes than in neurons, the SNARE complex is used by both cell types to release synaptic vesicle contents.⁶⁰ Whether vesicular exocytosis is a physiologically relevant mechanism in astrocytes, however, is somewhat controversial. Our studies on VMAT and the SNARE complex strongly suggest that synaptic vesicle loading and release within astrocytes are required for normal alcohol sedation in flies, thereby supporting the hypothesis that synaptic vesicle exocytosis in astrocytes could have important physiological roles.

Our data suggest that a fraction of astrocytes expresses Tdc2 and synthesizes/releases tyramine, potentially as a gliotransmitter. The presence of, or level of, expression of Tdc2 could therefore represent astrocyte heterogeneity, and these functional differences between astrocytes may contribute to alcohol sedation behavior. Given that the fly brain contains approximately 4,600 astrocytes total,⁶¹ it is intriguing to speculate how a fraction of such a small number of cells could affect alcohol sedation in an organism whose brain contains roughly 100,000 neurons.⁶² One possibility is that the astrocytes engaged in tyramine synthesis could be physically associated with numerous neurons involved in regulating alcohol sedation, and the tyramine released from these astrocytes could bind to G protein-coupled tyramine receptors on neurons,⁶³ thereby influencing the response of those neurons to alcohol. Another possibility is that tyramine released from a fraction of astrocytes could permeate the brain as a whole, thereby influencing the physiological properties of nearby, as well as distant, neurons and neural circuits. Another possibility is that tyramine released from a fraction of astrocytes could function as an autocrine/paracrine factor by activating octopamine/tyramine receptors on nearby astrocytes, regulating intracellular calcium signaling within these cells leading to release of ATP, and thereby influence the activity of dopaminergic neurons involved in alcohol sedation.²⁵ Yet another possibility is that astrocytes could express direct pharmacological targets of alcohol, and the binding of ethanol to these targets could alter the release of tyramine, which would influence alcohol sedation. It is possible in all of these models that tyramine levels in and therefore release from astrocytes could be regulated by astrocyte-specific degradation enzymes such as monoamine oxidases and dehydrogenases/reductases like Nazgul.⁶⁴ Although these models are somewhat speculative (and not mutually exclusive), they highlight several possible avenues for future experiments and emphasize the need for additional studies to better understand the role of astrocytes in behavioral responses to ethanol.

Glia in flies and rodents, as well as in human alcoholic post-mortem tissue, are molecularly and morphologically altered by the presence of alcohol.^{6,10,11,65–67} In flies, surface glia and cortex glia can regulate initial alcohol sedation and rapid tolerance development.^10–12 In rodents, blocking astrocyte hemichannels, increasing astrocyte intracellular calcium, and increasing astrocyte cytokine release has been associated with changes in alcohol-related behaviors.^8,68 Our study is the first to identify a mechanism within astrocytes that directly influences alcohol-related behavior in any species. Our data stress the importance of considering mechanisms within glia when investigating molecular-genetic contributions to alcohol-related behaviors and, potentially, AUD.