Dorsal striatal circuit mechanisms contributing to astrocyte modulation of alcohol-related behaviors

Abstract

Background:

The dorsolateral striatum (DLS) is a key site for coordinating the alcohol-induced stimulant response, a behavioral marker predictive of future alcohol use disorder. Although ethanol (EtOH) affects all brain cells, little is known about the contribution of non-neuronal DLS cell types to EtOH-induced stimulation.

Methods:

We used ex vivo two photon calcium imaging, in vivo fiber photometry of astrocyte and neuronal GCaMP, and astrocyte-specific manipulations in mice to determine DLS astrocyte contributions to EtOH-induced stimulation and voluntary EtOH drinking behavior. Using fiber photometry of GRAB-ACh sensors and cell-type specific chemogenetics, we also assessed the role of cholinergic signaling in observed astrocyte EtOH effects.

Results:

As expected, intraperitoneal EtOH injections (0.5–2g/kg) evoked a stimulant response, evidenced by increased locomotion compared to saline. In parallel, EtOH dose-dependently decreased astrocyte calcium activity but had minimal effects on direct and indirect pathway neuronal activity. Mimicking this reduction with astrocyte-specific expression of CalEX, a calcium extruding pump, facilitated EtOH stimulation compared to mice expressing a control fluorophore. Hence, EtOH-induced suppression of DLS astrocyte activity contributes to stimulation. Astrocyte calcium signaling is a well-known target of neuromodulation. Fiber photometry recordings of extracellular acetylcholine (ACh) levels via GRAB-ACh imaging showed inhibition of ACh release by acute EtOH. We virally expressed the excitatory chemogenetic actuator hM3Dq in striatal cholinergic interneurons to assess whether artificially increasing ACh release blocks EtOH-induced inhibition of astrocytic calcium activity. Despite facilitating ACh release, this manipulation did not impact astrocyte calcium activity under control (saline) or EtOH conditions. Together, this work identifies DLS astrocytes as key contributors to EtOH-induced stimulation and highlights the importance of considering astrocyte-neuron interactions in evaluating alcohol effects.

Introduction

AUD is a multi-faceted disorder with few FDA-approved treatment options available. Current FDA-approved pharmacotherapies for AUD target neuronal mechanisms to reduce alcohol intake [1]. However, alcohol impacts neurons and non-neuronal glia cells, including astrocytes [2]. Astrocytes have non-overlapping spatial territories and express a variety of neurotransmitter receptors which make them both anatomically and functionally capable of responding to synaptic transmission and neuromodulation with intracellular calcium elevations [3]. In turn, astrocyte calcium activity regulates diverse functions like neurotransmitter uptake and ionic homeostasis, neurovascular coupling, and direct neuronal modulation through release of bioactive substances [4, 5]. The fine processes of a single astrocyte can contact thousands of neuronal synapses, making astrocytes uniquely positioned to modulate synaptic transmission and integration [6, 7]. Growing evidence suggests that astrocyte calcium signaling plays a role in regulating behavior by modulating communication between ensembles of neurons [8–10]. Importantly, astrocytes also modulate alcohol-related behaviors. For example, cortical astrocytes regulate alcohol intake in mice [11] and activation of astrocytes in the basolateral amygdala decreases binge-like drinking [12]. Further, astrocytes in the nucleus accumbens have been implicated in behavioral flexibility after chronic alcohol exposure [13]. Despite these advancements, how alcohol affects in vivo astrocyte calcium activity and the underlying mechanisms remain largely unclear.

Alcohol produces divergent and heterogenous behavioral responses across individuals [14, 15]. Humans with heightened stimulant and decreased sedative responses to alcohol are at higher risk for developing alcohol use disorder (AUD) [16–21], and prevailing theory suggests a relationship between the initial stimulant response to alcohol and future alcohol intake [22, 23]. This subjective response to alcohol has recently been shown to also predict alcohol self-administration in heavy drinkers [24]. In rodents, alcohol-induced stimulation and sedation are commonly measured through the assessment of locomotor activity following injected alcohol [25–28]. In mice, the stimulant response to alcohol correlates with AUD-relevant behaviors like alcohol consumption and aversion-resistant drinking [25]. Relatedly, a lower sedative response correlates with alcohol consumption [29]. Like in humans, these preclinical findings show significant interindividual variability in alcohol behavioral sensitivity that correlates with AUD-relevant drinking behaviors. Hence, alcohol stimulation represents an avenue for reverse translation into pre-clinical models amenable to mechanistic dissection of neurobiological processes that jointly contribute to alcohol stimulation and associated AUD vulnerability.

Human neuroimaging studies and preclinical work both implicate the dorsolateral striatum (DLS) as an important regulator of alcohol’s stimulant effects [25, 30]. The DLS is a major basal ganglia input nucleus that integrates cortical and thalamic glutamatergic innervation under strong neuromodulatory control from midbrain dopaminergic neurons and local cholinergic interneurons [31–33]. Subgroups of striatal projection neurons (SPNs) differentially express dopamine D1 and D2 receptors, defining the direct and indirect pathways of the striatum, respectively. Although dopamine signaling in DLS SPNs modulates alcohol stimulation [25], the role of cholinergic signaling in this process is not known. Similarly, the contribution of astrocyte calcium activity to alcohol stimulation and sedation is not clear. Alcohol inhibits cerebellar and cortical astrocyte calcium activity via inhibition of norepinephrine (NE) release [34]. A similar mechanism in the DLS is unlikely given sparse NE innervation of the striatum. Astrocytes express multiple subtypes of cholinergic receptors, including GPCR coupled muscarinic and ionotropic nicotinic receptors, and respond to cholinergic neuromodulation with intracellular calcium increases [35–38]. In the current study, we use a range of ex vivo and in vivo approaches to determine how ethanol (EtOH) impacts astrocyte calcium signaling in the DLS and the relevance of this modulation for EtOH stimulation and consumption behaviors. We find that EtOH induced an expected locomotor stimulant response, along with an inhibitory effect on astrocyte calicum signaling and extracellular ACh levels in-vivo. Molecular genetic inhibition of astrocyte calcium signaling potentiated the EtOH-induced stimulant response, suggesting a role for DLS astrocytes in modulating this behavior. We also evaluated whether and how cholinergic activity modulates astrocyte alcohol effects. Together, our experiments highlight DLS astrocyte and cholinergic signaling as key coordinators of alcohol stimulation.

Materials and methods

Mice

All experimental procedures performed on mice were approved by the Rutgers University Institutional Animal Care and Use Committee (Protocol #202000004) and conformed with NIH guidelines. Progenitor mice were purchased from vendors (see Supplementary Table 1 for details), bred in-house, and used for experiments. A few experiments were performed with C57BL/6J mice purchased from Jax, which are noted in Supplementary Table 2. When appropriate, offspring were genotyped using Transnetyx services. All mice were at least 8 weeks of age prior to performing stereotaxic surgeries. Following surgery, mice were housed in a reversed 12-h light/dark cycle in standard-sized cages and provided ad libitum food (LabDiet 5R53) and water.

Stereotaxic surgery

Viral infusions, head bar implants, and fiber optic implant surgeries were conducted as described by our lab previously [39, 40]. Specific viral infusion volumes and placements for each experiment can be found in Supplementary Table 2. For fiber optic surgeries targeting the dorsolateral striatum (DLS) and dorsomedial striatum (DMS), ceramic black fibers from RWD (907–03019-00), 400 μm core, 0.5 NA, 4 mm length were used. For fiber optic surgeries targeting the bed nucleus of the stria terminalis (BNST), the same fibers were used at a 5 mm length.

Slice imaging using two photon microscopy

Brain slice preparation

C57BL/6 mice (n=6; 3M, 3F) received bilateral stereotaxic DLS (A/P: +0.3, ML: ± 2.5, DV: −3.3, [41]) injections of gfaABC1D-cyto-GCaMP6F. Following recovery and viral expression, mice were deeply anesthetized with isoflurane prior to decapitation. The brain was immediately removed and placed within an ice chilled cutting solution composed of 194 mM Sucrose, 30 mM NaCl, 4.5 mM KCl, 1 mM MgCl₂, 26 mM NaHCO₃, 1.2 mM NaH₂PO₄, 10 mM Glucose. 250 μm thick coronal slices containing the dorsal striatum were cut with a Leica VT1200 vibrating microtome. The vibratome chamber was ice-chilled prior to cutting and was filled with oxygenated cutting solution. Slices were incubated at 32 °C in oxygenated artificial cerebrospinal fluid (aCSF). The aCSF contained 124 mM NaCl, 4.5 mM KCl, 1 mM MgCl₂, 26 mM NaHCO₃, 1.2 mM NaH₂PO₄, 10 mM Glucose, and 2 mM CaCl₂. The cutting solution and aCSF were prepared with deionized water (18.2 MΩ-cm), filtered (0.22 μm), and bubbled with 95% O₂ and 5% CO₂ for at least 15 min prior to use and throughout slice preparation and recording. After incubation, slices were maintained at room temperature until 2 photon calcium imaging, which was performed in the same aCSF formulation used for incubation.

Brain slice two photon calcium imaging & analysis

GCaMP6f fluorescence from astrocyte somas was imaged through a 16x/0.8 NA objective (Nikon) using resonance-galvo scanning with a Ultima Investigator two-photon microscopy system (Bruker). Image frames were 512 × 512 pixel resolution acquired at 64 Hz with 64-frame averaging resulting in an effective frame rate of 1 Hz. Emitted light was filtered using a dichroic mirror and collected with GaAsP photomultiplier tubes (Hamamatsu). GCaMP6f-expressing astrocytes were imaged with 2x optical zoom. Astrocyte soma fluorescence over time was extracted using ImageJ, and a $\mathrm{\Delta }\mathrm{F}/{\mathrm{F}}_0$ (DFF) was calculated as

$$\frac{\mathrm{\Delta }\mathrm{F}}{F_0}=\frac{F\left(t\right)-F_0}{F_0}$$

Where ${\mathrm{F}}_0$ represents the 8^th percentile of the fluorescence values recorded in the surrounding two minutes. The “scipy.signal.find_peaks” function was utilized to identify calcium events in the DFF trace with minimum peak prominence of 0.4 DFF. From the identified peaks, both inter event interval (IEI) and average peak amplitude were determined. All recordings were visually examined for drift in the imaging plane. Where such drift occurred transiently (<30 seconds) the affected period was excluded from analysis, in other cases the entire recording was excluded from analysis.

Fiber photometry recordings of bulk astrocyte and neuronal calcium activity

We used fiber photometry to assess how acute EtOH impacts astrocytic calcium signaling during ethanol (EtOH)-induced stimulant response open field testing. Striatal astrocytes of C57BL/6J mice were unilaterally transfected with the genetically encoded Ca²⁺ indicator AAV-gfaABC1D-cyto-GCaMP6F (n = 7; 4F, 3M, DLS) or pZac2.1-gfaABC1D-lck-GCaMP6f (n = 11; 5F, 6M, [AP: +1.00, ML: ±1.5, DV: −2.6; DMS]). See Supplementary Table 1 for detailed information for each construct. Calcium activity from cell-type specific striatal neurons was recorded using pGP-AAV-Syn-FLEX-jGCaMP8m-WPRE stereotaxically injected into the DLS of transgenic mice (Drd1-Cre: n = 5; 3F, 2M; A2a-Cre: n = 6; 4F, 2M). For all fiber photometry experiments, at least 3 weeks was given following surgery to allow for viral expression. Fiber recordings were conducted using the Tucker-Davis Technologies system as described by our lab previously [40].

Fiber photometry analysis

Fiber photometry data were analyzed using custom MATLAB scripts, as previously described by our lab [40]. Briefly, raw fiber photometry data was imported using the opensource file “TDTbin2mat.m,” which is available for download on the TDT website. Data from emission at 465 nm (signal) and 405 nm (isobestic control) were both down sampled to 20 Hz. Then a linear fit of the isobestic 405 nm channel to the 465 signal was calculated using the ‘polyfit’ MATLAB function. Lastly, a commonly used ΔF/F calculation was performed to control for photobleaching and movement artifacts, where ΔF/F = (signal − fitted control) / fitted control. Event rate and event prominence were calculated from these ΔF/F traces using the ‘findpeaks’ MATLAB function.

EtOH-induced open field stimulant response testing

Open field testing was conducted using a 40 × 40 × 40 cm arena (60101, Stoelting). For all open-field tests, mice received 3 days of habituation where they received an intraperitoneal injection of sterile 0.9% saline and were allowed to acclimate to the open-field arena for 30 minutes. Mice were then randomly assigned to receive one dose of EtOH prior to testing (saline, 0.5, 1, 1.5, or 2 g/kg) on a given day. EtOH was injected at a volume of 10 mL/kg (i.e., for 0.5 g/kg EtOH the concentration was 6.25% v/v; for 1 g/kg it was 12.5% to ensure injection volumes were consistent with body weight across testing days regardless of dose). Saline injections were volume-matched with EtOH injections. Note that doses were given in a randomly assigned order and mice only received one EtOH injection per day to allow them to metabolize the dose completely before the next testing session, which typically occurred as subsequent testing days. EtOH solutions were prepared from 200 proof EtOH (Koptec) diluted in sterile 0.9% saline. An overhead camera (Microsoft) captured open field video to assess movement and fiber recordings were conducted as described above if applicable. For astrocytic and neuronal calcium recordings (Figures 1 and 3) and CalEX open field experiments (Figure 2), mice were injected with EtOH and immediately placed into the center of the open field for 30 minutes of testing. For neuromodulator recordings, a 10-minute baseline fiber recording conducted in the open field preceded the EtOH injection. For DREADD stimulation of cholinergic neurons, a drug pre-treatment of Clozapine-N-oxide (CNO; HelloBio) or Deschloroclozapine (DCZ; Sigma Aldrich) and 10-minute baseline fiber recording preceded EtOH injection (described in “Chemogenetic manipulation of cholinergic interneurons (CINs)” in more detail below).

Figure 1. Alcohol acutely inhibits DLS astrocytic calcium signaling.

A) Mice were given a stereotaxic viral injection of GCaMP6f with an astrocyte-specific promoter along with an optical fiber to enable in-vivo fiber photometry recordings. B) A representative image showing the DLS fiber placement and GCaMP6f viral expression (left) and zoomed-in image of the GCaMP6f expression (right). The dotted box on the left image represents the zoomed-in portion shown on the right image. C) Mice received saline or 0.5 – 2 g/kg EtOH injected intraperitoneally and then were placed in an open field box for 30 minutes while overhead video was recorded to assess locomotor activity simultaneously with astrocyte calcium activity gathered via fiber photometry. This panel was created using Biorender. D) Representative traces of astrocyte calcium activity from the first 5 minutes of an open-field recording across all doses tested. Red dots represent detected events. Note that all traces are from the same mouse and the scale bars from saline apply to all doses. E) In the first 5 minutes, alcohol inhibited astrocyte calcium events; F(2.380, 17.26) = 5.518, p = 0.011 (Greenhouse-Geisser corrected one-way repeated measures mixed-effects analysis). Tukey’s multiple comparison post-hoc tests indicated no significant differences between any two given doses; but a trend was identified in the post-hoc comparison of saline to 2 g/kg EtOH (p = 0.0819). F) No significant impact was observed on event prominence in the first 5 minutes; F(1.172, 6.740) = 0.518; p > 0.05 (Greenhouse-Geisser corrected one-way repeated measures mixed-effects analysis). G) For saline, raster plots (top) are shown where each row represents an individual mouse, and each vertical line represents an identified astrocyte calcium event across the 30-minute open field session. Average astrocyte calcium events binned in one-minute bins (middle), and average distance traveled with the same binning (bottom) allow for visualization of concurrent astrocyte calcium and locomotor behavior. H) The same information shown in G, but when mice were given 2 g/kg EtOH. I) Alcohol inhibited astrocyte calcium events across the full 30-minute session, F(2.774, 15.95) = 4.938, p = 0.0144 (Greenhouse-Geisser corrected one-way repeated measures mixed-effects analysis). Tukey’s multiple comparison post-hoc tests indicated a significant difference between saline and 2 g/kg EtOH (p < 0.05). J) Event prominence was decreased when events from the full 30-minute session were considered, F(2.593, 14.91) = 5.230, p = 0.0139 (Greenhouse-Geisser corrected one-way repeated measures mixed-effects analysis). Tukey’s multiple comparison post-hoc tests indicated no significant differences between any two given doses. K) Overview of experimental design for open field tests created with cartoon mouse modified from SciDraw.io. L) Representative track plots in the first 5 minutes after mouse was given saline or M) 1 g/kg EtOH. N) Alcohol stimulation, as measured by distance travelled, peaked in the early part of the open field session during fiber photometry recordings; significant main effect of minute, F(1.738, 29.54) = 13.65, p = 0.0001, dose, F(1.000, 17.00) = 6.099, p = 0.0244, and dose x minute interaction, F(3.197, 47.32) = 9.170, p < 0.0001; Šídák's multiple comparisons post-hoc tests indicated more movement at the 5 minute bin when mice were given 1 g/kg EtOH as compared to saline (p = 0.0076) (repeated measures mixed effects analysis). Note that this graph also includes behavioral data from mice recorded in the DMS (n = 18 total; DMS recording data shown in Figure S2). All error bars represent standard error of the mean (SEM).

Figure 3. Acute alcohol exhibited minimal effects on striatal SPN activity.

A) To record direct pathway neuron calcium activity, D1-Cre mice were given a stereotaxic injection of a Cre-dependent GCaMP virus with accompanying fiber optic implant for in-vivo fiber photometry calcium recordings (top); representative viral expression (bottom). B) Example direct pathway activity when mice were given saline or C) 1 g/kg EtOH. Red dots represent detected neuronal events. D) Acute alcohol did not have a significant impact on D1 neuron event rate; F(2.387, 9.549) = 1.700, p > 0.05, η²_p = 0.298 (Greenhouse-Geisser corrected repeated measures one-way ANOVA); E) but did decrease event prominence at higher doses; F(1.795, 7.180) = 6.180, p = 0.029, η²_p = 0.61 (Greenhouse-Geisser corrected repeated measures one-way ANOVA). F) To record indirect pathway neuron calcium activity, A2a-Cre mice were given a stereotaxic injection of a Cre-dependent GCaMP virus with accompanying fiber optic implant for in-vivo fiber photometry (top); representative viral expression (bottom). G) Example indirect pathway activity when mice were given saline or H) 1 g/kg EtOH. Red dots represent detected neuronal events. I) Acute alcohol did not have a significant impact on indirect pathway neuronal event rate; F(2.544, 12.72) = 3.01, p = 0.075, η²_p = 0.376 (Greenhouse-Geisser corrected repeated measures one-way ANOVA); J) or prominence; F(2.161, 10.81) = 0.813, p = 0.478, η²_p = 0.139 (Greenhouse-Geisser corrected repeated measures one-way ANOVA). All error bars represent standard error of the mean (SEM).

Figure 2. DLS astrocyte calcium activity inhibition increases the EtOH-induced stimulant response.

A) Representative images of GCaMP6f-lck (control) and B) CalEX DLS expression. C) Representative track plots in the first 5 minutes when control or D) CalEX mice are given 1 g/kg EtOH. E) When given saline, viral groups moved the same in the first 5 minutes, t(18) =0.4144, p = 0.683, d = −0.188 (unpaired t-test). F) When given 1 g/kg EtOH, CalEX mice mounted a greater EtOH-induced stimulant response in the first 5 minutes, t(18) =2.228, p = 0.038, d = −0.985 (unpaired t-test). G) Across doses, this effect was most prominent at 1 g/kg EtOH, trend for a main effect of dose, F(3, 54) = 2.581, p = 0.0629, η²_p = 0.125 (two-way repeated measures ANOVA). H) Mice displayed considerable variability in which EtOH dose produced the maximum EtOH-induced stimulant response (as determined using distance traveled in first 5 minute normalized to saline data); Fisher’s exact test comparing CalEX and control distributions: p = 0.2972. I) Maximum response regardless of EtOH dose did not differ between CalEX and control mice; t(18) = 0.9623, p = 0.3487 (unpaired t-test). J) Movement patterns under saline and 1 g/kg EtOH conditions reveal that control mice do display an EtOH-induced stimulant response; red stars indicate significant FDR-corrected t-tests for the 10-second time bins. K). CalEX mice mounted a greater stimulant response that also lasted longer across the 30-minute open field session. All error bars represent standard error of the mean (SEM).

Open field video analysis

Open field video data were analyzed using DeepLabCut [42]. A model was trained to track six key points on the mouse: nose, right ear, left ear, body center, end of the tail, and base of the tail. Distance traveled was computed using the body center key point. Open field box corners were also tracked. Like analysis conducted by our lab previously [40], these points were used to calculate the length of each side of the box in pixels. X-Y pixel coordinates of the body center key point were converted to cm using the known cm length of each box side (40 cm). Points tracked with a < 99% probability were eliminated, and the remaining data were interpolated using the “interp1” MATLAB function. X-Y coordinates were LOESS filtered using the function “smooth” with a window size of 10 frames (0.5s). Coordinates were then used to calculate the frame-by-frame distance, which was summed to calculate the total distance traveled across various time bins.

Effect of CalEX astrocyte expression on EtOH stimulation and other EtOH-related behaviors

CalEX: EtOH-induced stimulant response testing

To assess if DLS astrocyte calcium signaling plays a role in modulating the EtOH-induced stimulant response, we used CalEX to inhibit DLS astrocyte calcium signaling. CalEX is a calcium extrusion hPMCA pump, which works to constitutively extrude cytosolic calcium from the astrocyte soma [43]. This tool has been extensively validated [44–47], including in the striatum [43, 47]. In our first CalEX experiment, C57BL/6J mice were given bilateral stereotaxic DLS injections of gfaABC1D-GCaMP6f-lck as a control virus (n = 11; 6F, 5M) or CalEX (n = 9; 6F, 3M). Following recovery and time for viral expression (~4 weeks), mice underwent EtOH-induced open field stimulant response testing as described above.

Using CalEX, we also assessed the role of DLS astrocytes in modulating the following additional EtOH-related behaviors: two-bottle choice home cage EtOH drinking, EtOH metabolism, drinking-in-the-dark (DID), aversion-resistant EtOH drinking, and loss of righting reflex (LORR). Please see Supplementary Note 1 for detailed methods and results of these additional behavioral assays.

Neuromodulator release recordings during tail suspension test & open field locomotion

To assess how acute EtOH impacts striatal acetylcholine (ACh) and norepinephrine (NE) release, fiber photometry recordings were conducted as described above. C57BL/6J mice (n = 6, 5F, 1M) received a stereotaxic injection of AAV-hSyn-GRAB_ACh3.0 and fiber implant in the DLS. As a control for the GRAB-ACh, AAV-hSyn-ACh4.3_mut (n = 2M mice targeting DLS) was used as a null sensor. To measure NE release, C57BL6J (n = 4F) received a stereotaxic injection of rAAV-hSyn-GRAB-NE2m(3.1)-WPRE-pA and fiber in the DLS. Another group of C57BL6J mice (n = 3; 1F, 2M) targeting the BNST [AP = +0.14, ML = ± 0.88, DV = −4.24] [48]) was included as a control brain region which receives substantial noradrenergic input [49].

In addition to open field EtOH-induced stimulant response testing, these mice also underwent a tail suspension test with concurrent fiber recordings the week after stimulant response testing. This test was performed as 5 trials that lasted 60 seconds each with a minimum of 60 seconds in between each trial [50]. The tail suspension test occurred in the open field because this apparatus was already outfitted for fiber photometry recordings.

Chemogenetic manipulation of cholinergic interneurons (CINs)

Impact on ACh-release:

ChAT-Cre mice (n = 5; 2F, 3M) received an injection of AAV-hSyn-GRAB_ACh3.0 in the cortex (AP: +2.1, ML: +0.4, DV: − 2.3). One week later, mice received a stereotaxic injection of a Cre-dependent G_q-DREADD virus (AAV-hSyn-DIO-hM3D(Gq)-mCherry) along with an optical fiber in the DLS. By expressing the GRAB-ACh sensor in cortical axons projecting to the the striatum instead of injecting it locally in striatum, we would overcame known issues co-expressing two viruses. This facilitated the ability to activate CINs with simultaneous measurement of ACh release. These mice received a pre-treatment with CNO (1 mg/kg) or saline 50-minutes before a 10-minute baseline fiber recording, following which mice were given an additional i.p. injection of saline. CNO was prepared as a 0.1 mg/mL solution in 0.9% sterile saline.

Impact on astrocyte calcium activity:

A separate group of ChAT-Cre mice (n = 10, 1F, 9M) received a stereotaxic co-injection of a Cre-dependent G_q-DREADD virus (AAV-hSyn-DIO-hM3D(Gq)-mCherry) mixed 1:1 with an astrocyte-specific GCaMP virus (AAV-gfaABC1D-cyto-GCaMP6F) along with an optical fiber in the DLS. This facilitated the ability to activate CINs with simultaneous measurement of DLS astrocyte calcium. Another group of ChAT-Cre mice (n = 4; 2F, 2M) served as the viral control, receiving a Cre-dependent mCherry (AAV-hSyn-DIO-mCherry) instead of the DREADD along with the astrocytic GCaMP and DLS fiber. All mice underwent EtOH-induced stimulant response testing with a 30-minute DCZ (0.2 mg/kg, [51]) or saline pre-treatment and 10-minute baseline fiber recording preceding the EtOH or saline injection (1 g/kg EtOH). Then, half of the DREADD mice underwent the same protocol, but now testing for 2 g/kg EtOH. These mice then underwent subsequent LORR testing as described above. A subset of the mice underwent head-fixed motorized wheel testing, as described by our lab previously [40]. Mice were first acclimated to a head-fixed voluntary running wheel for 3 days. Then, on a given testing day, they underwent 25 motorized locomotion trials (10s duration) with at least 50 seconds of rest in between. On these testing days, the mice received an i.p. injection of DCZ (0.2mg/kg) or saline 30-minutes prior to the start of the session and a subsequent i.p. injection of saline immediately before testing began.

Perfusion, immunofluorescent staining, and histological imaging

At the conclusion of all experiments, mice were transcardially perfused with heparinized 1X phosphate buffered saline (PBS) followed by 4% paraformaldehyde (PFA). Brains were extracted and stored in 4% PFA in a 4°C refrigerator overnight and then transferred to 1X PBS the following day where they remained in the refrigerator until they were sliced using a vibrating microtome (Compresstome, Precisionary). Brain slices containing the targeted brain regions were assessed for fiber optic and viral placement, or viral placement only, and were sliced at 100 μM. Brains that were used for immunofluorescent staining (described below) were sliced at 60 μM. All brain slices were imaged using a Widefield Leica DMI8 epifluorescence microscope with automated tile stitching. Mice that were identified to have misplaced viral and/or optical fiber placements were excluded from all analyses and are not reflected the reported sample sizes throughout the manuscript.

For mice that received GRAB sensor (or control) stereotaxic injections, brains were counterstained with GFP to boost the signal. All steps were completed with slices shaking at room temperature unless otherwise noted. Brains were sliced at 60 μM and then washed 3 times in 1X phosphate-buffered saline (PBS) for 15 minutes each wash. Then, slices were incubated in 50% methanol in 1X PBS for 30 minutes. Slices were then incubated in 1% hydrogen peroxide in 1X PBS for 15 minutes then washed 3 times in 1X PBS for 15 minutes each wash. Slices were then washed with 1X PBS + 0.4% Triton-X for 30 minutes and then blocked in 10% normal goat serum (NGS) in 1X PBS + 0.2% Triton-X for 90 minutes. Following blocking, slices were incubated in the primary antibody (1:500 of GFP Polyclonal Antibody from ThermoFisher, #A11122; in NGS +0.02% Triton-X) overnight at 4°C while shaking. The next day, brains were washed 4 times in 1X PBS for 10 minutes each wash. Then, they were incubated in secondary antibody for 2 hours (Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 647, ThermoFisher, # A-21245; 1:1000 dilution in 1X PBS). Slices were then washed 3 times in 1X PBS for 10 minutes each wash and then mounted on glass slides and dried in the dark overnight. Once dry, slides were cover slipped and imaged using a Widefield Leica DMI8 epifluorescence microscope.

Analyses and consideration of sex effects

For slice 2P imaging and fiber photometry, events were detected as described above. Event rate and amplitude or prominence was analyzed across EtOH conditions using one-way repeated measures ANOVAs (or mixed effects analysis in the case of missing data; e.g., a documented missed EtOH injection) using GraphPad Prism 11.0.0. For consideration of the EtOH-induced stimulant response, locomotor data were processed using DeepLabCut as described above and analyzed across EtOH and viral groups using one-way or two-way repeated measures ANOVAs as applicable. T-tests were utilized for some comparisons (i.e., LORR dependent variables comparing CalEX vs. viral control groups). Throughout the manuscript, post-hoc analyses were conducted when warranted and Greenhouse-Geisser corrections were applied to analyses where appropriate.

Where possible, sex effects were considered and reported in the results section if found. Some fiber photometry experiments have a low sample size between sexes, which precluded analysis by sex (see Supplementary Table 2 for sample sizes for each experiment). If there were no interaction effects with sex, results are shown collapsed on sex and null sex-based analyses are reported in Supplementary Table 3.

Results

Acute EtOH inhibits in vivo dorsolateral striatal astrocyte calcium signaling

We used two-photon calcium imaging to begin assessing the impact of acute EtOH on striatal astrocyte calcium activity. We virally expressed GCaMP6f in the DLS under the control of an astrocyte-specific promoter and prepared acute ex vivo brain slices for imaging (Fig. S1A). Consistent with previous work, astrocytes showed robust intracellular calcium dynamics in the absence of any experimental stimuli under baseline conditions (Fig. S1B, C). We bath applied 20 and 50 mM EtOH followed by a washout to determine how EtOH impacts this spontaneous calcium activity. EtOH had no significant effect on astrocyte activity metrics quantified by detecting individual calcium transients (red dots in Fig. S1C). Event frequency and amplitude were similar across baseline, EtOH, and washout conditions (Figure S1D-G). These findings show that EtOH has minimal effects on astrocyte calcium activity recorded in a reduced preparation lacking the full complement of DLS synaptic inputs, neuronal activity and neuromodulation dynamics.

In vivo fiber photometry was then used to capture population-level bulk activity dynamics in the intact striatal circuit. This strategy also allowed us to assess EtOH -induced stimulation in parallel with bulk activity changes. We first report how acute EtOH changes in vivo astrocyte activity followed by analysis of how these changes relate to EtOH-induced stimulation. We virally expressed in DLS astrocytes the calcium sensor gfaABC1D-GCaMP6f and implanted an optic fiber above the viral injection site (Figure 1A, B). Mice were given intraperitoneal (i.p.) injections of saline (control) or EtOH (0.5, 1, 1.5, and 2 g/kg) in separate daily sessions and placed immediately in an open field box with an overhead video camera (Figure 1C). We restricted our initial activity analysis to the first 5 mins after the i.p. injection since previous work shows EtOH stimulation effects in this same time window [25, 52]. Astrocytes showed robust calcium dynamics under baseline saline conditions in freely moving animals (Figure 1D), similar to our recent observations in head-fixed mice [40]. Examining astrocyte calcium activity traces recorded across doses suggested suppression of astrocyte activity with EtOH (Figure 1D). We identified individual calcium events and quantified event frequency and prominence (amplitude relative to surrounding events) as a measure of astrocyte activity. In contrast to our observations in brain slices, EtOH acutely inhibited in vivo striatal astrocyte calcium activity by reducing event frequency (Figure 1E), but event prominence was unaffected (Figure 1F).

Given that our observed astrocyte activity inhibition and the previously reported stimulant response both occur during the first 5 mins after EtOH injection, we examined the temporal relationship between these EtOH-induced effects in our data. We used DeepLabCut to track mouse location in fiber photometry open field videos. We plotted identified astrocyte calcium events as raster plots and quantified average events per minute across the full 30-minute open-field session. When these data were plotted alongside distance traveled a temporal relationship between these two variables emerged, such that when astrocyte calcium activity was inhibited by EtOH, the mice tended to show acute EtOH stimulation (i.e., they moved more than when given saline) (Figure 1G, H). Across the full 30-minute session, astrocyte calcium events and prominence were inhibited by EtOH (Figure 1I, J). Notably, performing this same protocol in the DMS produced similar results (Figure S2), suggesting that the EtOH-induced astrocyte calcium signaling inhibition observed in vivo is robust across the striatum and not DLS-specific. We focused our further studies in the DLS given the well-established role of this striatal subregion in EtOH stimulation. Taken together, these results show that EtOH reduces astrocyte activity in the intact striatal circuit.

EtOH induces behavioral stimulation in parallel with astrocyte activity inhibition

EtOH exposure recruits a stimulant response that manifests as increased locomotion in rodents [25, 53]. We further examined the stimulant response under our fiber photometry recording conditions. Because the DLS and DMS recording cohorts are the same mouse strain (C57BL/6J) and the behavioral testing was conducted identically, we pooled the behavioral data from both groups to increase the sample size. Indeed, distance traveled in the first 5 minutes normalized to saline locomotion across all tested doses (Figure S3A) indicated no significant main [F(1,16) = 0.423, p = 0.52] or interaction effect with brain region [F(1.848, 26.49) = 0.269, p = 0.75]; Greenhouse-Geisser corrected two-way (brain region x dose) mixed effects analysis. Locomotion was quantified as the open field distance traveled as a function of time from injection (see Methods). Consistent with previous findings in untethered animals, there was higher locomotion in response to an i.p. injection of 1g/kg EtOH relative to saline injections (Figure 1K - N). This stimulation effect peaked in the first 5 mins after EtOH injection, during the same time when astrocyte activity inhibition was observed.

EtOH-induced stimulation, quantified as the distance traveled in first 5 minutes normalized to saline locomotion, was highly variable across mice (Figure S3A), which is consistent with previous reports in both humans [14, 19] and rodents [25]. Due to this variability, there was no significant difference in EtOH-induced stimulant response across doses (Figure S3A). Relatedly, we found that there was no single dose where mice as a group had the maximal or minimal EtOH-induced stimulant response (Figure S3B), further highlighting variability in individual EtOH responses. Taken together, our behavioral analyses show that tethered mice undergoing fiber photometry recordings show a robust EtOH stimulation response with considerable interindividual differences. Given the important role of astrocyte calcium signaling in multiple homeostatic and neuromodulatory processes [54], these results raise the possibility that EtOH’s behavioral and neurophysiological effects are modulated by its suppression of astrocyte activity. We addressed this possibility in our next series of experiments.

DLS astrocyte calcium inhibition facilitates the EtOH-induced stimulant response

To establish how astrocyte calcium activity contributes to EtOH stimulation, we virally overexpressed the molecular genetic tool CalEX selectively in this cell type. This virus encodes a human plasma membrane Ca²⁺-ATPase that decreases astrocyte calcium activity [43] and thus mimics the suppressive effect of EtOH on astrocytes in a constitutive manner. C57BL/6J mice were given bilateral stereotaxic DLS injections of gfaABC1D-CalEX or gfaABC1D-GCaMP6f-lck as a fluorophore control (Figure 2A, B). These experiments were performed similarly as above, except no mouse tethering for calcium recording was required. Inspecting 2D locomotion traces of individual mice suggested that EtOH induces a greater stimulant response in CalEX expressing animals as compared to viral controls (Figure 2C, D). Viral groups did not differ in baseline locomotion as there was no difference in total distance traveled in the first 5 mins after saline injections (Figure 2E). However, when given 1 g/kg EtOH, CalEX expressing mice showed greater EtOH-induced stimulation (Figure 2F). When considering the stimulant response across the range of tested doses, CalEX expression seemed to shift peak stimulation to a lower dose compared to controls (Figure 2G). To address if this shift is due to changes in EtOH sensitivity, we quantified for each viral group the number of mice showing their maximal stimulant response at each dose. Mice in both viral groups showed considerable variability in the dose which produced a maximally stimulating response, and there was no difference in the distribution of maximally stimulating dose between groups (Figure 2H). This suggests that the observed enhancement in stimulation is unlikely due to changes in EtOH sensitivity. In agreement, comparing the maximal stimulant response observed for each animal regardless of the dose at which it occurred showed no difference between control and CalEX mice (Figure 2I). Taken together, these findings suggest that attenuating astrocyte calcium activity with CalEX potentiates the stimulant response at submaximal doses without affecting the dose at which the maximal stimulant response is produced.

Finally, we tested the impact of astrocyte CalEX expression on the duration of the stimulant response. We quantified locomotion across the full recorded 30 mins for saline and 1g/kg EtOH sessions. Control mice displayed the expected EtOH-induced stimulant response (Figure 2J), however, in CalEX-expressing mice, EtOH stimulation was more pronounced and lasted for longer (Figure 2K). Together, these results indicate that DLS astrocyte calcium activity modulates the intensity and duration of the EtOH-induced stimulant response.

Given the above results, in 2 new cohorts of mice we tested whether striatal astrocyte CalEX expression impacts other EtOH-related behaviors (see Supplementary Note 1 for full description). In the first cohort, we tested mice in home-cage intermittent 2-bottle choice (2BC) drinking using automated Promethion (Sable) behavioral monitoring cages (Figure S4). In the next cohort, we assayed whether CalEX and control viral groups differ in EtOH metabolism, drinking-in-the-dark (DID) home cage drinking, aversion-resistant drinking, and loss of righting reflex (Figure S5). Importantly, there were no differences between viral groups in EtOH metabolism (Figure S5D), suggesting that our findings related to EtOH-induced stimulation are not driven by differences in metabolic rate. Overall, these experiments showed minimal effects of CalEX astrocyte expression on EtOH sedation and EtOH consumption behaviors, suggesting that DLS astrocyte calcium signaling specifically modulates EtOH stimulation.

EtOH exerts minor effects on bulk striatal projection neuron activity

Our ex vivo two photon calcium imaging experiments showed minimal effects of EtOH on spontaneous astrocyte calcium activity, yet we observed robust inhibition in our fiber photometry experiments. In vivo astrocyte calcium activity is strongly dependent on local neuronal activity and neuromodulatory inputs [54, 55], which are compromised in the ex vivo brain slice system. We reasoned that the discrepancy between our ex vivo and in vivo findings suggest the involvement of EtOH-sensitive circuit-level processes in the observed inhibition of astrocyte activity. Striatal astrocytes express a variety of neurotransmitter receptors [54, 56, 57], including GPCR-coupled GABA_B receptors which enable these cells to respond to GABA release from local SPNs with intracellular calcium activity [58]. Therefore, the EtOH-induced astrocyte activity inhibition we observed could reflect changes in SPN activity. To consider this possibility, transgenic Drd1-Cre and A2a-Cre mice were given a stereotaxic injection of a Cre-dependent GCaMP8m calcium sensor in the DLS to label direct and indirect pathway SPNs, respectively (Figure 3). We performed fiber photometry open field recordings to assess the effect of EtOH on SPN activity. There was no effect of EtOH on the event rate of direct pathway neuron activity (Figure 3A-D) but event prominence was mildly reduced (Figure 3E). In contrast, we could not detect an effect of EtOH on indirect pathway neuron activity (Figure 3F-J). Hence, EtOH has minimal effects on DLS SPN activity, which is unlikely to explain the observed astrocyte suppression.

Acute EtOH inhibits DLS acetylcholine release

Given the minimal effect of EtOH on SPN activity, we next tested the involvement of neuromodulatory inputs. EtOH inhibits cerebellar and cortical astrocyte calcium activity via inhibition of norepinephrine (NE) release [34]. Though the DLS receives only sparse NE innervation [59], we considered if the mechanism of EtOH-induced astrocyte calcium signaling could be the same as previously described. We virally expressed the genetically encoded GRAB-NE2m sensor [60] and used fiber photometry to measure NE release. Given potential concerns with crosstalk of GPCR based NE sensors [59], we first validated our use of the GRAB-NE sensor in a control brain region known to receive strong NE innervation (bed nucleus of stria terminalis, BNST) [49]. We virally expressed GRAB-NE and implanted a fiber in the BNST (Figure S6A, B) or in the DLS (Figure S6C, D). Tail suspension leads to high NE release [50]. Therefore, we determined NE release dynamics in the BNST and the DLS during tail suspension (Figure S6E). As expected, recordings from the BNST displayed robust NE release during tail suspension (Figure S6F). In contrast, we could not detect substantial NE release in the DLS in response to the same manipulation. (Figure S6G). Accordingly, direct comparison of recordings in the two brain regions showed significantly higher NE release in the BNST as compared to the DLS. (Figure S6H). Overall, this series of experiments establish a minimal NE tone in the DLS. Hence, the observed inhibition of astrocyte calcium activity likely involves a mechanism specific to the neuromodulatory landscape of the striatal circuit.

Striatal activity is under strong neuromodulatory control from local cholinergic interneurons (CINs) that tonically release acetylcholine (ACh) across the striatum [61]. Astrocytes in the dorsal striatum are found in close proximity to CINs [62], which may represent a structural basis for functional communication. Recent work shows acute EtOH inhibits ACh release in the dorsomedial striatum [63], where EtOH also suppressed astrocyte calcium activity (Figure 4). We tested the possibility that astrocyte inhibition by EtOH involves cholinergic signaling. We began by assessing how acute EtOH impacts in vivo ACh release in the DLS, which previous work did not address. Mice were given a stereotaxic DLS injection of GRAB-ACh3.0 (Figure 4A, B) or the null (ACh4.3_mut) sensor as control (Figure 4C) along with an optical fiber implant for fiber photometry recordings. Mice were placed in an open field for 10 mins to record baseline GRAB-ACh activity, after which they received i.p. injections of saline or EtOH. Acute injected EtOH led to a pronounced decrease in DLS ACh release across the range of tested doses (Figures 4D-F). This phenomenon was not observed in mice expressing the control null sensor (Figure 4G), indicating that EtOH exposure is not associated with non-specific changes in the fiber photometry fluorescence signal. As an additional fluorophore control, we tested mice expressing GRAB-NE in the DLS with saline and EtOH injections. Acute EtOH injections did not produce changes in the GRAB-NE signal (Figure S6 I-K), further validating the described astrocyte and ACh release EtOH effects.

Figure 4. Acute alcohol inhibited DLS acetylcholine release.

A) C57BL/6J mice received a stereotaxic viral injection of GRAB-ACh3.0 or null control sensor (AAV-hSyn-ACh4.3_mut) and an optical fiber for recording. B) Representative image of DLS GRAB-ACh and fiber placement. C) Representative image showing null control sensor expression. D) In the first 5 minutes, EtOH inhibited ACh release as measured by average DFF, F(2.175, 10.88) = 7.488, p = 0.0081 (Greenhouse-Geisser corrected repeated-measures one-way ANOVA). E) Example ACh DFF trace when a mouse received saline, or F) 1 g/kg EtOH. G) Mice expressing a control null ACh sensor in the DLS did not show as pronounced of a decrease in avg. DFF in the first 5 minutes from saline to the highest dose of EtOH tested (2 g/kg); GRAB-ACh: t(5) = 3.216, p = 0.0235; Null = t(1) = 1.402, p = 0.3944 (paired t-tests within virus). All error bars represent standard error of the mean (SEM).

Chemogenetic stimulation of DLS cholinergic interneurons has no effect on astrocyte calcium activity

Our results so far show that EtOH inhibits both DLS astrocyte calcium activity and ACh release, raising the possibility that astrocyte activity suppression is driven by the concomitant decrease in extracellular ACh. Such a mechanism implies a facilitatory effect of ACh release on astrocyte calcium activity under basal conditions. We tested this prediction by first examining how G_q-DREADD stimulation of CINs affects ACh release. ChAT-Cre mice received a stereotaxic injection of Cre-dependent hM3Dq-mCherry in the DLS. We virally delivered GRAB-ACh to the cortex so it is expressed in corticostriatal axons (Figure S7A, B). This strategy allowed us to measure ACh release by monitoring fluorescence changes in axonally expressed GRAB-ACh while avoiding potential issues with viral co-expression within the striatum itself (Figure S7B). Mice were given an i.p. injection of the DREADD agonist clozapine-n-oxide (CNO, 1mg/kg) and 50 mins later put in an open field box for recordings. G_q-DREADD stimulation of CINs with CNO did not change the frequency or prominence of ACh release events as compared to the control saline injection (frequency: t(4) = 0.5928, p = 0.5852; prominence: t(4) = 0.3572, p = 0.7390). Surprisingly, we observed that a subsequent i.p. injection, which involved scruffing the animal, increased the ACh release event rate in mice previously injected with CNO compared to saline (Figure S7G). These findings suggest that G_q-GPCR based stimulation of CINs does not appreciably increase ACh release under basal conditions but sensitizes ACh release to subsequent stimuli like scruffing.

Given that we could increase ACh release with DREADD stimulation of CINs, we sought to determine how this manipulation affects DLS astrocyte calcium activity. In a new group of ChAT-Cre mice, we delivered into the DLS a viral mixture to co-express astrocytic GCaMP and Cre-dependent G_q-DREADD-mCherry (or AAV-hSyn-DIO-mCherry as control) in CINs, and implanted an optical fiber for photometry recordings (Figure 5A, B). Mice were given an i.p. injection of the DREADD agonist DCZ (0.2 mg/kg) or saline as control and 50 mins later placed into an open field box for fiber photometry recordings. After recording baseline astrocyte calcium activity for 10 mins, mice were given an i.p. injection of saline (Figure 5) or EtOH (Figure 6) and placed back into the box to continue recording. We analyzed activity after the second i.p. injection since this is when we observed increased ACh release in the previous experiment (Figure S7). There was no difference in astrocyte event frequency (Figure 5C – E) or prominence when comparing the same mice pretreated with DCZ or saline (prominence data not shown; sal-sal, mean prominence = 1.312, std. dev = 0.3493; DCZ-sal, mean prom. = 1.356, std. dev. = 0.3694; t(9) = 0.4946, p = 0.6327, paired t-test). We also did not detect any changes in astrocyte activity in control mice expressing mCherry in CINs (Figure 5E). These results suggest that G_q-DREADD stimulation of CINs has minimal effects on astrocyte calcium activity during open field exploration.

Figure 5. Stimulation of cholinergic interneurons using DREADDs did not impact astrocyte calcium activity during spontaneous or evoked locomotion.

A) ChAT-Cre mice received a stereotaxic co-injection of gfaABC1D-cyto-GCaMP6f with either a Cre-dependent Gq-DREADD (AAV-hSyn-DIO-hM3D(Gq)-mCherry) or control virus (AAV-hSyn-DIO-mCherry) in the DLS. An optical fiber was placed above the injection site to measure astrocytic calcium activity. On a given day, mice received DCZ to activate cholinergic interneurons (CINs) (or received saline as a drug control) 30 minutes prior to 1 g/kg EtOH (or saline). Ten minutes prior to 1 g/kg EtOH (or saline), mice received a 10-minute baseline recording in the open field. Mice then received their EtOH or saline injection, and astrocytic calcium and overhead video were subsequently recorded for 30-minutes in the open field. A subset of these mice were also tested the following week using this same protocol, but now instead with 2 g/kg EtOH. Note that this figure is only displaying data from the saline treatment, which is pulled from this larger experiment. The EtOH data is featured in the next figure. B) Representative image showing fiber placement, expression of DREADD (red) and GCaMP6f-cyto (green), DAPI (blue), and these three merged. In the merged image, example CINs and astrocytes are labeled. C) Representative astrocyte calcium traces when mice received saline-saline or D) DCZ-saline. E) Stimulation of CINs via DCZ did not impact astrocytic calcium activity event rate in the first 10 minutes of the open field session in DREADD or viral control mice; a significant virus x treatment interaction was found, F(1, 12) = 6.073, p = 0.0298, η²_p = 0.336 (two-way ANOVA), but post-hoc comparisons revealed no significant differences between and across conditions. F) To assess if CIN stimulation could modulate astrocyte calcium activity during evoked locomotion, a subset of mice (mice that did not undergo 2 g/kg EtOH testing) were head-fixed and tested using a motorized wheel, which prompted 25 10-second-long movement bouts. G) Percent DFF of astrocyte calcium activity during saline-saline vs. DCZ-saline testing indicated that CIN stimulation had no prominent impact on astrocytic activity. H) Quantification of area under the curve (AUC) from wheel on to wheel off (time 0 to 10 seconds on graph G) indicated no significant difference during CIN stimulation (DCZ-saline); t(4) = 0.322, p = 0.7635 (paired t-test). All error bars represent standard error of the mean (SEM). Panels A and F were created using BioRender.

Figure 6. CIN stimulation does not reduce alcohol-induced astrocyte calcium activity inhibition.

A) Astrocyte calcium traces from a representative mouse when receiving saline-EtOH (1 g/kg), saline-EtOH (2 g/kg), DCZ-EtOH (1 g/kg), and DCZ-EtOH (2 g/kg). B) In the first 10 minutes, no significant difference in event rate was noted across treatment groups, F(2.698, 24.28) = 1.186, p = 0.3329, η²_p = 0.116 (Greenhouse-Geisser corrected repeated measures one-way ANOVA). C) Event rate in the first 10 minutes was normalized to within-treatment saline (i.e., saline-EtOH was normalized to saline-saline and DCZ-EtOH was normalized to DCZ-saline) to facilitate comparison across pre-treatment conditions. This analysis indicated that there was no significant difference in normalized event rate between the pre-treatment groups for mice expressing the active DREADD, t(9) = 0.8836, p = 0.3999 (paired t-test). D) Null results were also observed in control mice expressing only mCherry, t(3) = 0.8376, p = 0.4638 (paired t-test). E) In the subset of mice tested for 2 g/kg EtOH, we also observed no significant difference in event rate across treatment combinations in the first 10 minutes, F(1.956, 7.823) = 3.380, p = 0.0881, η²_p = 0.2279 (Greenhouse-Geisser corrected repeated measures one-way ANOVA). F) No significant difference was observed between pre-treatment conditions when normalized to within-treatment saline, t(4) = 0.1008, p = 0.9245 (paired t-test). G) J) The subset of mice that underwent 2 g/kg EtOH testing subsequently underwent a LORR assessment to measure the impact of CIN stimulation on EtOH-induced sedation. H) No differences were observed in latency to LORR between saline and DCZ pre-treatments, t(4) = 0.2860, p = 0.7891 (paired t-test). I) No differences were observed in LORR duration, t(4) = 0.1810, p = 0.8651 (paired t-test). J) No differences were observed in total length of LORR test, t(4) = 0.1915, p = 0.8575 (paired t-test). All error bars represent standard error of the mean (SEM). Note: these are the same mice from which data are shown in Figure 5, now focusing on EtOH treatment.

Previous work from our lab [40] and others [64–67] shows robust recruitment of astrocyte calcium responses during locomotion and movement more generally. We did not find any changes in astrocyte calcium activity as a result of CIN stimulation during spontaneous locomotion in the open field. Behavior in the open field is dominated by diverse head and other body movements, which makes it challenging to specifically isolate the effect of CIN stimulation on locomotion-related astrocyte responses. We trained a subset of the same mice on a head-fixed motorized wheel paradigm (see Methods) and recorded astrocyte calcium signals via fiber photometry (Figure 5F). Mice were pretreated with DCZ or saline 30 mins before a subsequent i.p. injection of saline. Motorized locomotion elicited robust calcium responses that were similar with either DCZ or saline pre-treatment (Figure 5G – H). Together, these experiments show that under basal conditions, CIN stimulation has no measurable effects on astrocyte activity that are detectable by bulk fiber photometry.

CINs are tonically active, leading to a high ACh release rate in the striatum. One explanation for not observing astrocyte effects of increasing ACh release is that under basal conditions astrocyte cholinergic signaling is saturated. In this scenario, EtOH-induced suppression of ACh release could lead to astrocyte activity suppression. We reasoned that if this were the case, then augmenting ACh release with chemogenetic CIN stimulation should mitigate EtOH-induced changes in astrocyte activity. The same mice as described in Figure 5 were pretreated with either DCZ or saline and then subsequently injected with 1g/kg and, in a subset of mice, 2g/kg EtOH. We performed fiber photometry and compared astrocyte calcium activity across the tested conditions (Figure 6A). At 1 g/kg EtOH, we did not observe any significant differences across conditions (Figure 6B, C). This was also true in our viral control mice (Figure 6D). At 2 g/kg EtOH, there was more robust astrocyte calcium activity inhibition in the saline-EtOH group, however, CIN stimulation via DCZ pretreatment did not modulate this effect (Figure 6E, F).

Next, we considered the behavioral effects of CIN stimulation. We note that this experiment was designed primarily to address physiological effects of CIN manipulations on astrocytes and hence employed unilateral (and not bilateral) manipulations. With this caveat in mind, we determined how CIN stimulation affects EtOH stimulation. Examining locomotion behavior after i.p. injections of saline and 1g/kg EtOH suggested differences between the DCZ and saline pretreatment conditions (Figure S8A-C). We quantified EtOH stimulation as the locomotion observed with EtOH normalized by saline for each pretreatment condition. 1g/kg produced a variable stimulant response as before (Figure S8D). There was a trend for lower EtOH stimulation in the DCZ pretreatment condition as compared to saline (Figure S8D). Given these results, we also evaluated the impact of CIN stimulation on EtOH-induced sedation using the LORR test. We did not observe any impact of CIN stimulation on the tested LORR dependent variables (Figure 6G – J). The observed trend for reduced EtOH stimulation with CIN neuron activation is broadly consistent with previous pharmacological data showing potentiation of EtOH stimulation with muscarinic antagonism [68].

Discussion

We found that acutely injected EtOH produced an expected locomotor stimulant response, which corresponded with an EtOH-induced inhibition of DLS astrocyte calcium signaling (Figure 1). Over-expression of CalEX in DLS astrocytes (which inhibits astrocytic calcium signaling [43] and thus mimic’s alcohol’s suppressive effect on these cells), resulted in a greater EtOH-induced stimulant response to 1 g/kg EtOH as compared to mice expressing a control virus (Figure 2). This finding suggests that DLS astrocytes play an important role in regulating EtOH-induced stimulation. Astrocyte CalEX expression did not change other alcohol-related behaviors including 2-bottle choice (2BC) home-cage drinking, drinking-in-the-dark (DID; a measure of binge alcohol drinking), aversion-resistant drinking, or loss of righting reflex, overall suggesting a somewhat specific role for DLS astrocytes in modulating EtOH-induced stimulation (Figures S4, S5).

These findings on DLS astrocytes are broadly consistent with known roles of the DLS itself in the development of alcohol drinking and alcohol use disorder. In addition to an established role in facilitating alcohol-induced stimulation [30], the DLS plays a key role in habitual alcohol drinking and facilitation of a shift from habitual to compulsive alcohol use [69, 70]. The drinking-in-the-dark and two-bottle choice drinking paradigms leveraged in the current study may not have provided a long enough alcohol exposure or induced high enough blood ethanol concentrations (BECs) to substantially recruit DLS involvement over these behaviors. DID decreases GABAergic transmission in the DLS, as measured by a decrease in mIPSC frequency, but does not alter DLS SPN spine density or spontaneous glutamatergic transmission [71]. In contrast, dendritic hypertrophy has been observed in DLS neurons along with CB1 receptor-mediated long-term depression following chronic intermittent alcohol vapor exposure, a paradigm which induces nearly twice the BEC level of DID, highlighting that some DLS alcohol drinking-induced neuroadaptations are only observed following much higher levels of alcohol exposure [72]. Therefore, it is possible that DLS astrocytes may modulate alcohol-drinking under higher alcohol-exposure conditions. Additionally, recent work has shown that the use of CalEX to inhibit astrocyte calcium signaling in the cortex decreases 2BC drinking [11], and mice overexpressing CalEX in nucleus accumbens show increased aversion-resistant drinking following a two-bottle choice drinking history [47]. Together, these results highlight the circuit-specific roles of astrocytes in alcohol stimulation and voluntary alcohol consumption, which is congruent with prevailing theories of the development of alcohol use disorder [73].

Astrocyte calcium activity is thought to in part signal behavioral state information and accordingly provides contextual modulation of circuit-specific neuronal activity and behavioral output [10, 74]. Locomotion in the open field is partly driven by a drive to explore the environment. Interestingly, we found astrocyte CalEX expression to have minimal effects on baseline locomotion (Figure 2E), yet alcohol delivery revealed a hyperlocomotion phenotype (Figure 2F, G). This is unlike that observed in the medial entorhinal cortex where astrocyte CalEX expression reduced baseline exploratory locomotion [75]. These findings suggest that astrocyte contribution to exploratory behavior is both circuit and context specific. Interestingly, our work also suggests a strong dependence of astrocyte behavioral contributions on the neuromodulatory landscape even within the same circuit. In previous work, we found that augmenting DLS astrocyte calcium signaling with chemogenetic stimulation in a mouse model of Parkinson’s disease increased open field locomotion, but only during the early phase when exploratory drive is high [40]. Hence, coordinating exploratory behavior under different contexts may be a key function of DLS astrocytes. At the same time, it is perplexing that opposing manipulations of the same cell type in the same circuit produces the same results – increased locomotion. There are pronounced differences in the striatal neuromodulatory landscape under alcohol and parkinsonian conditions. While alcohol is expected to increase dopamine release [76], in parkinsonian animals dopamine release is severely curtailed. In contrast, alcohol decreases ACh release (Figure 4), whereas dopamine loss is associated with increased DLS cholinergic signaling [77]. Collectively, these results suggest that the impact of astrocyte calcium signaling on behavioral output is not static but rather highly context-dependent and shifts with neuromodulatory conditions in different contexts. A noted limitation of the current research is that we only assessed astrocyte calcium signaling, but it is known that astrocytes use diverse additional second messengers for communication with neurons [78]. Future work will assess how additional astrocytic signals, such as cyclic adenosine monophosphate (cAMP) and adenosine, are impacted by alcohol.

Astrocytes express a variety of neurotransmitter receptors, which allows them to “listen in” to the neuronal activity around them and respond with elevations in intracellular astrocyte calcium [54]. Because astrocyte activity reflects an integration of local neuronal and neuromodulatory signals, it was important to record local DLS neuronal activity to get a more complete view of alcohol effects in the DLS circuit. We found that indirect pathway SPN activity was not significantly impacted by acute alcohol at any dose tested (Figure 3F-J), but we observed a reduction in the prominence of calcium events recorded from direct pathway neurons (Figure 3E). This reduction could reflect the astrocyte calcium inhibition that we observed. Alternatively, it is possible that both cell types show inhibition because of alcohol’s effect on an input common to both cell types. Following this rationale, we recorded neuromodulator release activity from the DLS during alcohol exposure. Previous work has shown that alcohol stimulation is under dorsal striatal control and heavily influenced by dopaminergic signaling [25]. One goal of the current work was to assess if other striatal neuromodulators contribute to the alcohol-induced stimulant response. Cerebellar and cortical astrocytes are inhibited by alcohol via inhibition of norepinephrine release [34]. To assess a contribution of NE to astrocyte calcium inhibition in the DLS, we recorded NE release from the DLS and BNST (a control brain region chosen because it has strong norepinephrine inputs [49]). During tail suspension recordings, the BNST, but not DLS, showed robust NE release (Figure S6F - H). These findings confirmed that the DLS has a low NE-tone and suggested a striatum-specific mechanism for inhibition of astrocyte calcium activity by alcohol.

We found that acute alcohol inhibited DLS ACh release (Figure 4D - G). This finding is similar to other recent results demonstrating that alcohol decreases the spontaneous firing rate of DMS CINs in ex vivo brain slices and reduces ACh release in vivo as measured using fiber photometry with the iAChSnFR sensor [63]. Astrocytes express both nicotinic [79] and muscarinic ACh receptors (AChRs) [80, 81], which support astrocyte calcium activity [82, 83]. Given this relationship, we tested whether a reduction in ACh release following EtOH drives alcohol-induced astrocyte activity inhibition. We expressed the G_q-coupled excitatory chemogenetic actuator hM3D in CINs and tested how artificially increasing ACh release impacts alcohol effects on astrocyte activity. Although this manipulation augmented ACh release (Figure S7), it did not impact astrocyte calcium activity under basal or alcohol conditions (Figures 5, 6, respectively). Given that CINs are tonically active in the striatum [84], continually and locally releasing ACh, it is possible that further increasing ACh release is ineffective due to a competing saturation effect wherein we are unable to further modulate downstream impact on astrocyte activity. In future studies, it will be important to employ loss-of-function tools that decrease ACh release to test this possibility. Notwithstanding these concerns, our current experiments rule out one direction of the potential interaction between ACh release and astrocyte calcium. We have not addressed the possibility that alcohol’s inhibition of astrocyte calcium signaling drives the inhibition of ACh release. Recent work shows that optogenetic activation of striatal astrocytes results in the modulation of CIN activity [62], suggesting that this option is not only possible but may even be likely. It is also possible that the observed alcohol-induced suppression of ACh release and astrocyte calcium inhibition are unrelated and that astrocytes and CINs independently regulate alcohol-related behaviors. More work is needed to establish the exact relationship between alcohol stimulation, striatal astrocyte calcium activity and ACh release.

We found an interesting trend where mice showed lower alcohol stimulation with chemogenetic CIN stimulation (Figure S8), suggesting that striatal ACh release curtails alcohol stimulation. These findings also suggest that the EtOH-induced suppression of ACh release that we observed contributes to alcohol stimulation. Our collective findings are broadly consistent with the idea that CIN activity during alcohol exposure controls the level of alcohol-induced stimulation and astrocytes “fine tune” the response. CalEX expression did not lead to overall higher alcohol-induced stimulation but potentiated the stimulant response at the moderate dose of 1g/kg (Figure 2), suggesting a modulatory role. Moreover, CalEX expressing astrocytes are presumably already inhibited, which would preclude further inhibition by alcohol. Since enhanced alcohol stimulation observed under these conditions is unlikely to be driven by further inhibition of astrocyte calcium, these findings suggest that the residual astrocyte activity observed after alcohol delivery in our fiber photometry experiments plays a protective role against excessive stimulation. In this scenario, when astrocyte calcium signaling is inhibited by CalEX, these cells are unavailable to adaptively fine tune the neuronal processes which govern alcohol stimulation and as a result mice become more susceptible to producing a maladaptive behavioral response (i.e., a greater alcohol-induced stimulation). It is well established that acute alcohol stimulates striatal dopamine release and that alcohol stimulation is heavily regulated by dopaminergic inputs [25, 27, 85]. Striatal dopamine-ACh dynamics are carefully co-regulated to produce complex behaviors [86, 87]. One additional hypothesis raised by the current findings is that ACh inputs serve as a gating signal, activating a permissive state which by itself doesn’t modulate calcium activity but allows astrocytes to respond to other neurotransmitters, such as dopamine, to fine tune alcohol stimulation. Recent work has shown GPCR gating of astrocyte neurotransmitter reponses, supporting this possibility [88].

In summary, we have identified that alcohol inhibits both DLS astrocyte calcium activity and ACh release. Further, we have shown that inhibition of DLS astrocytes using a molecular genetic manipulation resulted in a more pronounced alcohol-induced stimulant response, suggesting that DLS astrocytes play an important role in modulating this behavior. We have also established that SPN activity, NE release, and ACh release are unlikely to contribute to alcohol-induced astrocyte activity suppression. Our experiments uncover the multifaceted effects of acute alcohol on the neuronal, non-neuronal, and neuromodulatory elements of the DLS circuit. Given the clinical relevance of the alcohol-induced stimulant response in predicting the future development of AUD, it is critically important to understand the biological mechanisms which drive this and other risk factors for this disease.