Abstract
Alcohol use disorder is complex and multifaceted, involving the coordination of multiple signaling systems across numerous brain regions. Previous work has indicated that both the insular cortex and dynorphin (DYN)/kappa opioid receptor (KOR) systems contribute to excessive alcohol use. More recently, we identified a microcircuit in the medial aspect of the insular cortex that signals through DYN/KOR. Here, we explored the role of insula DYN/KOR circuit components on alcohol intake in a long-term intermittent access (IA) procedure. Using a combination of conditional knock-out strategies and site-directed pharmacology, we discovered distinct and sex-specific roles for insula DYN and KOR in alcohol drinking and related behavior. Our findings show that insula DYN deletion blocked escalated consumption and decreased the overall intake of and preference for alcohol in male and female mice. This effect was specific to alcohol in male mice, as DYN deletion did not impact sucrose intake. Further, insula KOR antagonism reduced alcohol intake and preference during the early phase of IA in male mice only. Alcohol consumption was not affected by insula KOR knockout in either sex. In addition, we found that long-term IA decreased the intrinsic excitability of DYN and deep layer pyramidal neurons (DLPNs) in the insula of male mice. Excitatory synaptic transmission was also impacted by IA, as it drove an increase in excitatory synaptic drive in both DYN neurons and DLPNs. Combined, our findings suggest there is a dynamic interplay between excessive alcohol consumption and insula DYN/KOR microcircuitry.
SIGNIFICANCE STATEMENT The insular cortex is a complex region that serves as an integratory hub for sensory inputs. In our previous work, we identified a microcircuit in the insula that signals through the kappa opioid receptor (KOR) and its endogenous ligand dynorphin (DYN). Both the insula and DYN/KOR systems have been implicated in excessive alcohol use and alcohol use disorder (AUD). Here, we use converging approaches to determine how insula DYN/KOR microcircuit components contribute to escalated alcohol consumption. Our findings show that insula DYN/KOR systems regulate distinct phases of alcohol consumption in a sex-specific manner, which may contribute to the progression to AUD.
Introduction
Excessive alcohol consumption has a broad impact on human health, generating high personal and economic costs (Sacks et al., 2015). As a leading preventable cause of death in the United States, a large body of work has been devoted to identifying the brain structures and signaling mechanisms that drive excessive drinking and contribute to the development of alcohol use disorder (AUD). Preclinical research has shown the kappa opioid receptor (KOR) and its endogenous ligand dynorphin (DYN) are one of several key neurochemical systems involved in AUD (Crowley and Kash, 2015). For example, repeat variants in the genes that encode for KOR and DYN have been associated with an increased risk of AUD (Xuei et al., 2006; Edenberg et al., 2008; Karpyak et al., 2013). The involvement of DYN/KOR signaling in alcohol consumption and dependence has been further indicated by studies in rats showing that KOR antagonism increases alcohol self-administration (Mitchell et al., 2005) while agonism decreases volitional intake (Lindholm et al., 2001). Interestingly, the effects of KOR appear to reverse following alcohol dependence, as KOR antagonism can reduce excessive alcohol self-administration in postdependent rats only (Walker and Koob, 2008; Walker et al., 2011). This finding suggests that chronic alcohol exposure shifts KOR signaling dynamics in a manner that drastically alters its function in alcohol intake.
As a multisensory hub, the insula plays a critical role in integrating internal and external stimuli to generate complex internal states and subsequent actions (Gehrlach et al., 2020). Connectivity patterns of the insula are extensive, with inputs and outputs to this region spanning across cortical and subcortical domains. Among the most predominate inputs to the insula, specifically to the medial and posterior portions, are the amygdala, thalamus, and sensory cortex (Gehrlach et al., 2020). Notably, the reciprocal connectivity between the amygdala and insula has been well documented (Allen et al., 1991; Augustine, 1996; McDonald et al., 1999; Santiago and Shammah-Lagnado, 2005), and this connection is believed to underlie the role of the insula in tastant reinforcement and gustatory valence encoding (Lavi et al., 2018; Schiff et al., 2018; Wang et al., 2018).
Human imaging studies have also implicated the insula as a region that is altered in AUD (Claus et al., 2011; Ihssen et al., 2011; Grodin et al., 2017), with findings supported by multiple rodent studies (Seif et al., 2013; Jaramillo et al., 2018c; Centanni et al., 2019; Chen and Lasek, 2020; Marino et al., 2021). In humans, PET imaging has shown that there is reduced KOR availability in the insula of patients with AUD compared with control subjects (Vijay et al., 2018). Most notably, recent work has indicated that reductions in alcohol drinking and craving induced by the nonspecific opioid antagonist naltrexone are, in part, associated with KOR availability in the insula (de Laat et al., 2019). Combined, these findings strongly suggest that the insula may serve as a critical locus for DYN/KOR action in AUD.
Beyond this work, there have been no comprehensive explorations of how DYN/KOR signaling in the insula can directly impact alcohol consumption. Here, we explored the role of insula DYN/KOR systems in both male and female mice using converging genetic and pharmacological approaches. In addition, we assessed how long-term alcohol consumption can impact insula neuronal function using slice physiology.
Materials and Methods
Mice.
Male and female mice, 9–16 weeks of age at the start of the procedures, were singly housed and maintained on a reverse 12 h light/dark cycle in a temperature-controlled colony room. Throughout the experimental procedures, mice were provided continuous access to food and water. In addition to PdynIRES-Cre and PdynIRES-Cre × Gt(ROSA26)SorloxSTOPlox-L10-GFP(PdynGFP) mice (Krashes et al., 2014), Pdynlox/lox (Bloodgood et al., 2021) and Oprk1lox/lox (Crowley et al., 2016) conditional knock-out mice were generated as previously described and bred in-house. Wild-type C57BL/6J mice were purchased from The Jackson Laboratory. All experiments were performed in accordance with the National Institutes of Health guidelines for animal research and with the approval of the Institutional Animal Care and Use Committee at the University of North Carolina at Chapel Hill.
Alcohol and tastant drinking procedures.
For drinking experiments, mice were singly housed and maintained on an Isopro RMH 3000 (LabDiet) diet, which has been shown to produce high levels of alcohol consumption (Marshall et al., 2015). To evaluate the effect of genetic knockout and KOR antagonism on alcohol consumption and preference, an intermittent access (IA) to alcohol procedure was run as previously described (Bloodgood et al., 2021). Briefly, mice were given 24 h home-cage access to a bottle containing a 20% w/v alcohol solution alongside a water bottle. Alcohol bottles were introduced 3 h into the dark cycle, and bottles were weighed at the start and end of each 24 h session. Drinking sessions occurred three times per week with no less than 24 h and no more than 48 h between sessions (e.g., Monday, Wednesday, and Friday). The amount of fluid lost because of passive drip was determined by placing alcohol and water bottles in an empty cage. Drip values were then calculated for each solution, and total consumption was normalized to these values. To prevent the development of a side bias, the placement of alcohol bottles (left or right) was counterbalanced across sessions. To determine the selectivity of treatment effects on alcohol consumption, mice were given a sucrose challenge where they received 2 h of access to a bottle of 3% sucrose in addition to water. Sucrose drinking experiments occurred 48 h after the final IA to alcohol session.
Elevated plus maze.
To assay anxiety-related behavior, an elevated plus maze (EPM) task was used. The EPM apparatus consisted of two open (75 × 7 cm) and two closed (75 × 7 × 25 cm) arms arranged in a plus configuration wherein the arms of each type (open or closed) are positioned opposite to one another and connected via a central open zone (7 × 7 × 25 cm). During EPM testing, ambient illumination was maintained at ∼15 lux. Mice were placed in the center zone of the apparatus at task onset, then allowed to explore freely for 5 min. Video recordings were obtained and analyzed by a blinded observer and with Ethovision 9.0 (Noldus) to determine the time spent in the open/closed arms and the total number of open arm entries.
Surgical procedure.
Mice were anesthetized by injection (1.5 ml/kg, i.p.) of ketamine (Ketaset [National Drug Code (NDC) EA2489-564])/xylazine [AnaSed (NDC 59399-111-50)], then secured in a stereotaxic frame (Kopf Instruments) for intracranial viral and drug infusions. The insula (measured from bregma in mm: anteroposterior, +0.86; mediolateral, ±3.59; dorsoventral, −3.9) was targeted using standard mouse brain atlas coordinates (Franklin and Paxinos, 2008), and substances were microinjected using a 1 µl Neuros Syringe (33 gauge needle; Hamilton) controlled by an infusion pump. For conditional deletion, 300 nl/side AAV5-CAMKIIα-Cre-eGFP [2.3 × 1013 viral genomes (vg)/ml; University of North Carolina (UNC) Vector Core] or 200 nl/side AAV5-CMV-HI-eGFP-Cre-WPRE-SV4 (≥1 × 1013 vg/ml; Addgene) was injected at a rate of 100 nl/min into the insula of Pdynlox/lox and Oprk1lox/lox mice, respectively. After infusion, injectors were left in place for 5 min to allow for viral diffusion. To pharmacologically block KOR, norbinaltorphimine (nor-BNI; 5 µg/µl, 0.5 µl/side in PBS) was infused into the insula of C57BL/6J mice over a period of 5 min. In a subset of mice, a green fluorescent protein (GFP)-tagged virus (AAV5-CMV-eGFP; 50 nl/side) was added to the nor-BNI mix to assess infusion spread. To minimize postoperative discomfort, meloxicam [5 mg/kg, i.p.; Metacam (NDC 0010-6013)] was administered at the time of surgery.
Histology.
Mice were deeply anesthetized with Avertin (250 mg/kg, i.p.) and transcardially perfused with chilled 0.01 m PBS, pH 7.4, followed by 4% paraformaldehyde (PFA). Brains were removed and immersed in 4% PFA overnight, then stored in 30% sucrose/PBS before 45 µm coronal sections were taken on a vibratome (model VT1000 S, Leica Biosystems). Free-floating sections were processed for immunofluorescence to amplify GFP signal. Briefly, insular cortex (IC)-containing sections (from bregma in mm: anteroposterior, +1.54 to +0.38) were washed in PBS, then were blocked and permeabilized in 5% normal donkey serum/0.3% Triton X-100/PBS for 45 min. Tissue was incubated overnight with gentle agitation at 4°C with a chicken polyclonal anti-GFP antibody (1:2000, Aves Labs) in blocking solution. Sections were rinsed, then blocked for 45 min before 2 h incubation at room temperature in Alexa Fluor 488-conjugated donkey anti-chicken IgG (1:400 in blocking solution; Jackson ImmunoResearch). Sections were rinsed in PBS after the final incubation and mounted with Vectashield Hardset Mounting Medium with DAPI (Vector Laboratories). Slides were imaged on an Olympus BX43 microscope with attached optiMOS sCMOS camera (QImaging) or a KEYENCE BZ-X800 microscope.
Fluorescent in situ hybridization.
To validate Pdyn (prodynorphin) and Oprk1 (opioid receptor kappa 1) deletion and compare basal levels of these transcripts between sexes, mice were anesthetized with isoflurane and rapidly decapitated, and brains were rapidly extracted and snap frozen. Brains were then stored in a −80°C freezer before slicing. Sections (12 µm) were obtained using a cryostat (model CM3050 S, Leica Microsystems) mounted onto Superfrost Plus slides (Thermo Fisher Scientific) and stored at −80°C before in situ hybridization. RNAscope was performed according to the manufacturer instructions (Advanced Cell Diagnostics). Briefly, sections were fixed using 4% PFA, dehydrated, and washed with intravenous protease solution (Advanced Cell Diagnostics) before incubation with probes for Pdyn and Oprk1 (Advanced Cell Diagnostics). Slides were then coverslipped using ProLong Gold Antifade Mountant with DAPI and stored at −80°C. Images were captured using a Zeiss 800 upright confocal microscope (Hooker Imaging Core, UNC at Chapel Hill). For examination of basal sex differences in levels of Pdyn, Oprk1, and Vgat mRNA, tiled z-stacks of bilateral insula were captured at 20× magnification (three slices/animal), and maximum projection intensity images were generated using Zen (blue edition) software (Zeiss). Analysis was performed using subcellular particle detection in QuPath (Bankhead et al., 2017). Cells were considered Pdyn, Oprk1, and/or Vgat-positive if they contained five or more puncta in the respective channel of each probe. For examination of Pdyn and Oprk1 knockdown, validation of adeno-associated virus (AAV)-Cre injection into the insula was performed by confirming the presence of GFP at the target site, and only sections with GFP were used for imaging and subsequent analysis. Tiled z-stacks of bilateral insula (two slices/mouse) were captured and converted to maximum projection intensity images as described above. Mean fluorescence intensity per square millimeter was calculated using QuPath (Bankhead et al., 2017).
Electrophysiology recordings.
Whole-cell patch-clamp recordings were obtained from the insula of PdynGFP mice. After rapid decapitation under isoflurane anesthesia, brains were quickly extracted and immersed in a chilled and carbogen (95% O2/5% CO2)-saturated sucrose artificial CSF (aCSF) cutting solution as follows (in mm): 194 sucrose, 20 NaCl, 4.4 KCl, 2 CaCl2, 1 MgCl2, 1.2 NaH2PO4, 10 d-glucose, and 26 NaHCO3. Coronal slices (300 μm) containing the insula were prepared on a vibratome then transferred to a holding chamber containing the following heated oxygenated aCSF (in mm): 124 NaCl, 4.4 KCl, 1 NaH2PO4, 1.2 MgSO4, 10 d-glucose, 2 CaCl2, and 26 NaHCO3. After equilibration (≥30 min), slices were placed in a submerged recording chamber superfused (2 ml/min) with oxygenated aCSF warmed to ∼30–35°C. Neurons were visualized under a 40× water-immersion objective with video-enhanced differential interference contrast, and a mercury arc lamp-based system was used to visualize fluorescently labeled pDynGFP neurons. Recording pipettes (2–4 MΩ) were pulled from thin-walled borosilicate glass capillaries. Signals were acquired using an Axon Multiclamp 700B amplifier (Molecular Devices), digitized at 10 kHz, filtered at 3 kHz, and analyzed in pClamp 10.7 or Easy Electrophysiology. Series resistance (Rs) was monitored without compensation, and data were discarded from recordings where changes in Rs exceeded 20%.
Intrinsic properties and action potentials (APs) were recorded using the following potassium gluconate-based intracellular solution (in mm): 135 K-gluconate, 5 NaCl, 2 MgCl2, 10 HEPES, 0.6 EGTA, 4 Na2ATP, and 0.4 Na2GTP, at pH 7.3 and 289–292 mOsm. Input resistance and sag ratio were obtained in voltage-clamp mode, and sag ratio was obtained from a negative −100 mV current step and defined as the sag divided by the minimum voltage deflection from baseline. Current-clamp recordings were obtained from insula Pdyn and deep layer pyramidal neurons (DLPNs) at resting membrane potential (RMP). Current injection-evoked action potentials were evaluated by measuring the following: (1) action potential threshold (voltage at which cell first fired) and rheobase (minimum current required to evoke an action potential), both determined using a linearly increasing 120 pA/s ramp protocol; and (2) the number of spikes fired at increasing current steps (50 pA increments, 0–650 pA). Action potential kinetics including (1) peak amplitude, (2) half-width, (3) rise time, and (4) decay time were obtained from the current step protocol, and action potential thresholds were estimated using method 2 (Sekerli et al., 2004) and analyzed using Easy Electrophysiology.
Spontaneous synaptic events were recorded in voltage-clamp mode with the following cesium methanesulfonate-based intracellular solution (in mm): 135 cesium methanesulfonate, 10 KCl, 1 MgCl2, 10 HEPES, 0.2 EGTA, 4 MgATP, 0.3 Na2GTP, and 20 phosphocreatine, at pH 7.3 and 285–290 mOsm with 1 mg/ml QX-314. The excitatory (E)/inhibitory (I) ratio transmission was measured by isolating glutamate [holding potential (Vhold) = −55 mV] and GABA currents (Vhold = +10 mV) within individual neurons.
Experimental design and statistical analysis.
Data were analyzed by mixed-model three-way (sex × group × session/time; sex × treatment × time/current) or two-way (sex × group/treatment; group/treatment × session/time/current) ANOVA, where group is GFP/Cre or vehicle/nor-BNI and treatment is H2O/IA. Additional analyses were performed using one-way ANOVA or t tests to examine the effects of group and sex on Pdyn and Oprk1 transcript levels, alcohol intake/preference, total fluid intake, sucrose intake, time spent in open/closed arms of EPM or the effects of treatment on RMP, rheobase, sEPSC/sIPSC frequency/amplitude, and E/I ratio. Where data violated assumptions of normality or homoscedasticity, a Mann–Whitney U test or Welch's correction was used, respectively, for analysis. Data are presented as the mean ± SEM and were analyzed using JASP 0.14 (JASP Team) or Prism 9 (GraphPad Software). Significance threshold was set at α = 0.05, and post hoc pairwise comparisons were multiplicity adjusted.
Results
Knockout of Pdyn in the insula decreases alcohol drinking in male and female mice
To define the role of insula Dyn neurons in alcohol intake, we conditionally deleted Pdyn in male and female mice before initiation of an 8 week IA drinking procedure (Fig. 1A). Anxiety-like behavior was also assessed in the EPM test 24 h following the final IA session. To assess the specificity of knockout on alcohol intake, mice were given a 2 h sucrose drinking challenge 24 h after EPM testing. Site-specific knockout of Pdyn was achieved by infusing AAV5 with CaMKIIa promoter-driven expression of a GFP-fused Cre recombinase (n = 8 male/n = 9 female) or a GFP-only control (n = 6 male/n = 7 female) into the insula of Pdynlox/lox mice. Figure 1B shows an overlay of viral expression mapped at + 0.98 mm (from bregma) for male and female GFP (blue) and Cre (magenta) group mice. Representative photomicrographs of GFP expression and Pdyn mRNA expression in the insula of GFP (top) and Cre (bottom) mice are shown in Figure 1, C and D, respectively. In situ hybridization showed that, compared with GFP controls, injection of a Cre encoding vector led to a significant reduction in insula Pdyn mRNA expression as indexed by mean fluorescent intensity (n = 2 mice/sex, t(6) = 3.33, p = 0.016; Fig. 1E). Additionally, we found no difference in basal Pdyn mRNA expression in naive male or female mice (n = 3/sex, t(4) = 1.60, p = 0.186; Fig. 1F). Knockout of Pdyn decreased alcohol intake and preference in males (Fig. 1G,H; group × session interaction: F(23,276) = 1.97, p = 0.006, intake/F(23,276) = 2.62, p < 0.001, preference; group main effect: F(1,12) = 4.62, p = 0.053, intake/F(1,12) = 6.59, p = 0.025, preference; session main effect: F(23,276) = 3.89 p < 0.001, intake/F(23,276) = 3.96, p < 0.001, preference) and females (Fig. 1I,J; group × session interaction: F(23,322) = 1.61, p = 0.039, intake/F(23,322) = 1.33, p = 0.143, preference; group main effect: F(1,14) = 11.5, p = 0.004, intake/F(1,14) = 9.39, p = 0.008, preference; session main effect: F(23,322) = 9.33, p < 0.001, intake/F(23,322) = 7.92, p < 0.001, preference). Knockout of Pdyn in the insula also blocked escalation of alcohol intake (Fig. 1K,N), as average intake levels significantly increased from weeks 1–4 to weeks 5–8 in controls (GFP; males: t(5) = 4.42, p = 0.007; females: t(6) = 3.56, p = 0.012) but not in Pdyn knock-out (Cre) mice (males, p = 0.378; females, p = 0.083).
Genetic deletion of insula Pdyn blocks intermittency-induced escalation of alcohol intake in male and female mice. A, Experimental timeline: following AAV injections and a 3 week delay for expression, male and female Pdyn floxed mice were run through an 8 week 2-bottle choice IA to alcohol procedure. Twenty-four hours after the final alcohol exposure, mice were tested for anxiety-like behavior in the EPM, then given a 2 h sucrose challenge 24 h thereafter. B, Map of viral spread at +0.98 mm (from bregma) in male and female mice in GFP (blue) and Cre (magenta) groups. C, Representative images showing expression following injection of AAV-CAMKIIa-GFP control vector (top) and AAV-CAMKIIa-Cre-GFP (bottom) in the mouse insula. Cl, Claustrum; AI, agranular insula. Scale bars, 100 μm. D–F, In situ hybridization was used for validation of Pdyn knockout and comparisons of basal Pdyn mRNA expression in male and female mice. D, Representative images of Pdyn mRNA expression in a GFP (control) mouse and a Cre (knock-out) mouse. E, Compared with GFP controls, mice injected with an AAV encoding for Cre recombinase showed a significant reduction in Pdyn mRNA expression in the insula, as indexed by mean fluorescence intensity. F, In naive male and female mice, there was no difference in basal Pdyn mRNA expression. G–J, Insula Pdyn knockout significantly reduced alcohol consumption and preference in male (G, H) and female (I, J) mice. K, F, This effect was most prominent during the final 4 weeks of drinking (highlighted blue), during which male (K) and female (F) GFP control mice consumed significantly more alcohol compared with the first 4 weeks of IA. K, N, Cre mice did not exhibit this escalated pattern of intake and showed lower average intake levels compared with GFP controls, with males (K) consuming significantly less alcohol at weeks 5–8 and females (N) at weeks 1–4 and 5–8. L, M, In male mice, the effect of Pdyn deletion was specific to alcohol, as it did not affect mean fluid intake (L) or intake of a 3% sucrose solution (M). O, P, In female mice, Pdyn deletion did not impact mean fluid intake (O) while it decreased sucrose consumption (P). Q–T, There was no effect of insula Pdyn knockout on anxiety-like behavior as measured in the EPM test. Q, S, Representative heat maps showing time spent in EPM arms in male (Q) and female (S) mice in GFP (left panel) and Cre (right panel) groups. R, S, In both male (R) and female (S) mice, Cre and GFP groups spent a similar amount of time in the open and closed arms of the EPM. Group × sex interaction, #p < 0.05; group main effect, $p < 0.01; *p < 0.05; **p < 0.01; ***p < 0.001.
Total fluid intake was not affected by Pdyn knockout in male (p = 0.249) or female (p = 0.737; Fig. 1L,O) mice. Interestingly, Pdyn deletion decreased sucrose consumption in females (t(14) = 3.17, p = 0.007; Fig. 1P), but this effect was absent in male mice (p = 0.737; Fig. 1M). Combined, these findings show that Pdyn deletion in the insula selectively reduces alcohol intake in male mice and suggest that insula Pdyn may play a key role in driving escalated alcohol consumption. We next tested the impact of Pdyn deletion on anxiety-like behavior by exposing mice to an EPM test 24 h following their final drinking session (Fig. 1Q–T). We found no effect of Pdyn knockout in either sex on time spent in the open (males, p = 0.671; females, p = 0.252) or closed (males, p = 0.188; females, p = 0.843) arms of the EPM 24 h following their final drinking session (Fig. 1R,T).
Local infusion of nor-BNI decreases early-phase alcohol drinking in male, but not female, mice
Previous studies have found that the KOR antagonist nor-BNI can reduce alcohol consumption when given systemically (Walker et al., 2011; Anderson et al., 2016). To assess insula KOR involvement in alcohol drinking, we locally infused the long-acting antagonist nor-BNI (Munro et al., 2012) or PBS vehicle into male (n = 12 nor-BNI/n = 10 PBS) and female (n = 10 nor-BNI/n = 10 PBS) C57BL/6J mice before the start of an 8 week IA procedure (Fig. 2A,B). To determine the spread of nor-BNI infusion within the insula, a subset of female mice were coinjected with AAV5-CMV-eGFP (50 nl/side). Representative images showing GFP expression and cannula tracts along with reconstructed maps of injector tip placements are shown in Figure 2, O and P. Previous findings show that a single injection of nor-BNI blocks KOR for up to 3 weeks (Horan et al., 1992; Bruchas et al., 2007). Thus, the prolonged action of nor-BNI allowed us to pharmacologically assess the role of insula KOR in alcohol consumption over multiple weeks. As above, mice were also assessed for anxiety-like behavior (EPM) and sucrose intake at 24 and 48 h, respectively, following the final drinking session. As shown in Figure 2B, female mice were given one additional assay, a saccharine challenge, 24 h after the sucrose challenge.
Pharmacological blockade of insula KOR by nor-BNI decreases alcohol intake in early phases of drinking in male mice and increases the intake of sweet solutions in female mice. A, B, Experimental timeline: The KOR antagonist nor-BNI (2.5 μg/side) or PBS (vehicle) was microinjected into the insula of C57BL/6J male (A) and female (B) mice ∼3 d before exposure to an 8 week two-bottle choice IA to alcohol procedure. Twenty-four hours after their final alcohol exposure, mice were tested for anxiety-like behavior in the EPM, then given a 2 h sucrose challenge 24 h later, and female mice received a saccharine challenge 24 h thereafter. C, E, G, I, In male (C, G) but not female (E, I) mice, nor-BNI (NBI) produced a transient decrease in alcohol consumption (C, E) and preference (G, I) during the first 3 weeks or 9 drinking sessions (highlighted blue). D, H, During the first 9 sessions of IA in male mice, average alcohol intake (D) and preference (H) levels were significantly lower in NBI mice compared with vehicle (Veh) controls. F, J, In female mice, there was no impact of NBI on alcohol consumption (F) and preference (J) during the first 9 weeks of IA. K, There was no effect of NBI on sucrose intake in males, suggesting that KOR antagonism selectively reduced the consumption of alcohol. M, In female mice, NBI increased both sucrose and saccharine intake compared with Veh. L, N, NBI did not impact EPM performance in either sex. Group × sex interaction, #p < 0.05; group main effect, $p < 0.01; *p < 0.05; **p < 0.01; ***p < 0.001. O, Representative micrographs showing the location of the injector tract (rectangular dashed outline) and its most ventral point of termination, which was used to map the injector tip placement illustrated for Veh-treated (left) and NBI-treated (right) mice. Numbers indicate the location of each slice in the anteroposterior (AP) plane (in mm from bregma). Dashed lines segment each division of the insula shown. Insular cortex divisions: DI, dysgranular; GI, granular; AID, dorsal agranular; AIV, ventral agranular. P, Reconstruction of injector tip placements for nor-BNI-treated and Veh-treated female mice that were coinjected with 50 nl of an AAV encoding for GFP. Numbers indicate distance from bregma (in mm) for each coronal section.
Our findings show that in male mice, insula KOR antagonism by nor-BNI reduced alcohol consumption (Fig. 2C; group × session: F(23,460) = 2.08, p = 0.003; main effects of group: F(1,20) = 8.27, p = 0.009; session: F(23,460) = 8.07, p < 0.001) and alcohol preference (Fig. 2G; group × session: F(23,460) = 1.45, p = 0.082; main effects of group: F(1,20) = 6.98, p = 0.016; session: F(23,460) = 6.93, p < 0.001). Conversely, in female mice, we found no effect of KOR antagonism on intake (Fig. 2E; group × session: p = 0.281; group main effect: p = 0.647; session main effect: F(23,414) = 8.78, p < 0.001) or preference (Fig. 2I; group × session: p = 0.378; group main effect: p = 0.692; session main effect: F(23,414) = 16.3, p < 0.001). The reduction in intake by nor-BNI in males appeared to occur predominately during the first nine sessions or 3 weeks of the IA procedure. Thus, we evaluated this by collapsing alcohol intake over the first nine sessions of IA and found that nor-BNI decreased average alcohol intake (Fig. 2D; t(22) = 4.56, p < 0.001) and alcohol preference (Fig. 2H; t(20) = 3.58, p = 0.002) in male mice but not female mice (Fig. 2F,J; intake, p = 0.578; preference, p = 0.245). Nor-BNI did not affect intake of a 3% sucrose solution in male mice (Fig. 2K; p = 0.691), but, led to a significant increase in sucrose intake in female mice (Fig. 2M; t(18) = 2.52, p = 0.022). As a follow-up in females, we tested whether nor-BNI would affect the consumption of saccharin, a noncaloric sweet solution. Female mice were given access to a 0.15% saccharin solution, and intake was measured at the end of a 2 h session. As with sucrose, insula KOR antagonism increased saccharin intake, as female nor-BNI mice consumed significantly more of the sweetened solution than did vehicle controls and the noncaloric sweetener saccharine at a 0.15% concentration (Fig. 2M; t(18) = 2.73, p = 0.014). Thus, while insula KOR antagonism can selectively decrease alcohol intake in males, it may drive increased consumption of palatable solutions in females. Finally, we tested anxiety-like behavior in the EPM and found that nor-BNI did not affect time spent in the open arms or closed arms of the apparatus in male mice (Fig. 2L; open arm, p = 2.80; closed arm, p = 0.213) or in female mice (Fig. 2N; open arm, p = 0.213; closed arm, p = 0.179).
In a follow-up experiment, we administered nor-BNI 7 weeks into a drinking procedure to (1) test whether the transient reduction in alcohol intake it produced in male mice was because of a waning effect of the drug and (2) whether nor-BNI would impact females at this later phase of intake. Here, male (n = 10 nor-BNI/n = 11 PBS) and female (n = 8 nor-BNI/n = 9 PBS) C57BL/6J mice underwent 7 weeks of IA to alcohol before nor-BNI was delivered locally into the insula (Fig. 3A). After recovery, mice were given 3 additional weeks (nine sessions) of IA followed by a 3% sucrose challenge. When the final nine sessions of IA were assessed, we found there was no effect of the nor-BNI on alcohol intake or preference in males [intake (Fig. 3B): group × session interaction, p = 0.594; group main effect, p = 0.803; session main effect: F(8,152) = 4.86, p < 0.001; preference (Fig. 3D): group × session interaction, p = 0.356; group main effect, p = 0.564; session main effect: F(8,152) = 2.41, p = 0.018] and females [intake (Fig. 3F): group × session interaction, p = 0.717; group main effect, p = 0.458; session main effect: F(8,120) = 4.42, p < 0.001; preference (Fig. 3H): group × session interaction, p = 0.219; group main effect, p = 0.464; session main effect: F(8,120) = 3.69, p < 0.001]. Similarly, when data were collapsed across the final nine sessions of IA, we found no effect of insula nor-BNI infusion on average levels of alcohol intake and preference in males or females (t < 1, for all analyses).
Pharmacological blockade of insula KOR by nor-BNI does not affect alcohol drinking in male or female mice during later stages of intake. A, Experimental timeline: to test the effect of KOR antagonism on the later phase of alcohol drinking, mice were first exposed to 7 weeks of IA before receiving intrainsula injections of nor-BNI (NBI; 2.5 μg/side) or PBS [vehicle (Veh)]. Following recovery, IA was resumed for an additional 3 weeks. In both male (B–E) and female (F–I) mice, NBI did not affect intake (B, F) or preference (D, H) for alcohol when injected after 7 weeks of IA. C, E, G, I, When the final 3 weeks (8 sessions, postinjection) were analyzed for male (C, E) and female (G, I) mice, there was no difference in average alcohol intake (C, G) or preference (E, I) between NBI and Veh groups. J, Location of intracranial drug infusions. Reconstruction of injector tip placements for nor-BNI-treated and Veh-treated female mice that were coinjected with 50 nl of an AAV encoding for GFP. Numbers indicate distance from bregma (in mm) for each coronal section.
Knockout of KOR in the insula does not alter alcohol drinking in male or female mice
Our nor-BNI data suggest that in male mice, KOR localized within the insula can regulate aspects of alcohol consumption. However, KOR can be expressed at multiple sites within a given brain region, including on presynaptic nerve terminals. Given that nor-BNI can also act presynaptically at KOR-expressing afferent terminals, we wanted to more selectively target locally expressed KOR in the insula. Using an Oprk1lox/lox mouse line, we selectively deleted KOR in the insula by injecting AAV5 encoding for a GFP-tagged Cre recombinase (n = 15 male/n = 16 female) or a GFP only control (n = 13 male/n = 15 female; Fig. 4A). Following recovery, mice underwent 8 weeks of IA before being tested for anxiety-like behavior (EPM) and sucrose intake. Overlay of viral expression mapped at +0.98 mm (from bregma) for male and female GFP (blue) and Cre (magenta) group mice is illustrated in Figure 4B. Representative photomicrographs of viral expression and Oprk1 mRNA expression in the insula of GFP (top) and Cre (bottom) mice are shown in Figure 4, C and D, respectively. Using in situ hybridization, we found that viral-mediated delivery of Cre recombinase significantly reduced Oprk1 mRNA expression compared with a GFP control, as indexed by mean fluorescent intensity (n = 2–3 mice/sex, t(9) = 3.30, p = 0.009; Fig. 4E). We found no difference in basal Oprk1 mRNA in naive male or female mice (n = 3/sex, t(4) = 1.70, p = 0.164; Fig. 4F), indicating that males and females express similar KOR transcript levels. In female mice, knockout of insula KOR had no effect on alcohol intake (Fig. 4I; group × session interaction, p = 0.123; group main effect, p = 0.734; session main effect: F(23,667) = 8.99, p < 0.001) or alcohol preference (Fig. 4J; group × sex interaction, p = 0.331; group main effect, p = 0.519; session main effect: F(23,667) = 9.27, p < 0.001) across the 8 week IA procedure. In male mice, analyses yielded a significant group × session interaction on intake (Fig. 4G; F(23,575) = 1.56, p = 0.047) in addition to a main effect of session (F(23,575) = 2.79, p < 0.001) but not group (p = 0.097). This finding may be due, in part, to the lack of escalation in GFP control mice and the increased intake in KOR knock-out mice during the final week of IA (sessions 22–24). However, we found no effect of KOR deletion on alcohol preference in males (Fig. 4H; group × session interaction, p = 0.484; group main effect, p = 0.162; session main effect: F(23,575) = 2.33, p < 0.001). To further assess intake and preference, we examined the first and latter halves (weeks 1–4 and 5–8) of the IA procedure. We found that from weeks 1–4 to weeks 5–8, female mice escalated their average alcohol intake (Fig. 4N; paired t tests; GFP: t(14) = 3.98, p < 0.001; Cre: t(15) = 4.42, p < 0.001), whereas male mice did not (Fig. 4K; GFP, p = 0.298; Cre, p = 0.700). In both males and females, KOR knockout did not affect total fluid intake [males (Fig. 4L), p = 0.587; females (Fig. 4O), p = 0.654] or 3% sucrose intake [males (Fig. 4G), p = 0.810; females (Fig. 4P), p = 0.594]. Additionally, there was no difference in anxiety-like behavior between GFP and Cre groups, as indexed by time spent in the open and closed arms of the EPM apparatus in both males (Fig. 4R; open arm, p = 0.318; closed arm, p = 0.117) and females (Fig. 4T; open arm, p = 0.711; closed arm, p = 0.654). Representative heat map images illustrating time spent in the open and closed arms of the EPM for GFP (left) and Cre (right) male and female mice are shown in Figure 4, Q and S, respectively.
Genetic deletion of insula KOR does not impact alcohol intake in male or female mice. A, Experimental timeline: 3 weeks following AAV injections, male and female floxed Oprk1 mice were run through an 8 week 2-bottle choice IA to alcohol procedure. Twenty-four hours after the final alcohol exposure, mice were tested for anxiety-like behavior in the EPM, then given a 2 h sucrose challenge 24 h thereafter. B, Map of viral spread at +0.98 mm (from bregma) in male and female mice in GFP (blue) and Cre (magenta) groups. C, Representative images showing expression following injection of AAV-CMV-GFP control vector (top) and AAV-CMV-Cre-GFP (bottom) in the insula. Cl, Claustrum; AI, agranular insula. Scale bars, 100 μm. D–F, In situ hybridization was used for validation of Oprk1 knockout and comparisons of basal Oprk1 mRNA expression in male and female mice. D, Representative images of Oprk1 mRNA expression in a GFP (control; top) mouse and a Cre (knock-out; bottom) mouse. E, Compared with GFP controls, mice injected with an AAV encoding for Cre recombinase showed a significant reduction in the number of Oprk1+ cells in the insula. F, In naive male and female mice, there was no difference in basal Oprk1 mRNA expression. G–J, Compared with GFP mice, Cre-mediated deletion of insula KOR did not affect alcohol consumption (G, I) or preference (H, J) in male (G, H) or female (I, J) mice. K–P, In males (K–M) and females (N–P), there was no difference in alcohol intake (K, N), mean fluid intake (L, O), or sucrose intake (M, P) between GFP and Cre groups (G). Female (N) but not male (K) mice showed escalated alcohol intake between weeks 1–4 and 5–8. Q–T, There was no effect of insula KOR knockout on anxiety-like behavior as measured in the EPM test. Q, S, Representative heat maps showing time spent in EPM arms in male (Q) and female (S) mice in GFP (left panel) and Cre (right panel) groups. R, S, In both male (R) and female (S) mice, Cre and GFP groups spent a similar amount of time in the open and closed arms of the EPM.
Long-term alcohol drinking exerts a sex-dependent effect on insula neuronal function
Previously, we identified a microcircuit in the medial agranular insular cortex that is modulated by KOR (Pina et al., 2020). Within this microcircuit, Pdyn neurons are densely clustered in layer 2/3 and KOR expression is localized to deep layer 5/6. Functionally, activation of insula KOR by Dyn dampens local inhibitory tone, which leads to a disinhibition of and an increase in excitatory synaptic drive in DLPNs. Thus, based on our above findings, we wanted to determine the effect of long-term alcohol drinking on neuronal function in insula Pdyn and DLPNs. To assess the effect of long-term alcohol drinking on Pdyn neuronal function, adult male and female PdynGFP mice were exposed to 8 weeks of IA to a 20% alcohol solution or a water control before tissue was collected for whole-cell patch-clamp electrophysiology 24 h after the final IA session (Fig. 5). We first assessed the intrinsic properties of Pdyn neurons in H2O controls and IA male and female mice, including RMP, input resistance membrane input resistance (Rm), and sag ratio (Fig. 5A–C). We found a sex-dependent effect of IA on RMP (Fig. 5A), as indicated by a significant sex × treatment interaction (F(1,36) = 7.13, p = 0.011) but no effect of sex (p = 0.908) or treatment (p = 0.591). Follow-up analyses revealed that IA significantly increased RMP in female mice (U = 23.5, p = 0.025) but not in male mice (p = 0.151). In both male (EtOH: n = 9 cells/n = 3 mice; H2O: n = 10 cells/n = 4 mice) and female (EtOH: n = 11 cells/n = 5 mice; H2O: n = 10 cells/n = 4 mice) mice, IA did not alter input resistance (Fig. 5B; treatment × sex interaction, p = 0.343; treatment main effect, p = 0.321) or sag ratio (Fig. 5C; treatment × sex interaction, p = 0.147; treatment main effect, p = 0.599) of insula Pdyn neurons. There was no effect of sex on sag ratio (p = 0.192); however, in female mice Pdyn neurons exhibited a lower input resistance as indicated by a significant main effect of sex (F(1,36) = 6.72, p = 0.014). We next assessed the impact of IA on Pdyn neuron excitability and found that 8 weeks of IA decreased the excitability of these neurons in male but not female mice (Fig. 5D–H). This was supported by a significant decrease in firing to increasing current steps (Fig. 5D) in males (treatment × current interaction: F(10,170) = 7.75, p < 0.001; treatment main effect: F(1,17) = 9.69, p = 0.006; current main effect: F(10,170) = 40.90, p < 0.001) but not in females (treatment × current interaction, p = 0.999; treatment main effect, p = 0.993; current main effect: F(10,190) = 18.5, p < 0.001). Additionally, the rheobase (minimum current to elicit firing) of Pdyn neurons was significantly higher in male IA mice compared with H2O controls (Fig. 5G; t(15) = 2.19, p = 0.042), and there was no effect of IA in female mice (p = 0.859). There was no effect of IA on AP threshold in either sex (Fig. 5F; treatment × group interaction, p = 0.403; treatment main effect, p = 0.186; sex main effect, p = 0.549). Representative traces show Pdyn neuron firing to a 600 pA current step (Fig. 5E) and a 120 pA/s ramp protocol (Fig. 5H) in male (left, blue) and female (right, purple) H2O and IA mice. Finally, we analyzed action potential kinetics and found no effects of IA on AP peak amplitude (Fig. 5I; F values < 1 for treatment × sex interaction, and treatment and sex main effects), half-width (Fig. 5J; F values < 1 for treatment × sex interaction and treatment main effect; sex main effect, p = 0.058), rise (Fig. 5K; F values < 1 for treatment × sex interaction and treatment main effect; sex main effect, p = 0.073), and decay (Fig. 5L; F values < 1 for treatment × sex interaction and treatment main effect; sex main effect, p = 0.037). Thus, the only significant effect found on Pdyn neuron AP kinetics was a main effect of sex on AP decay time, as female mice exhibited faster decay times compared with males.
Insula Pdyn neurons are differentially altered by intermittent alcohol access in male and female mice. Whole-cell patch-clamp recordings were obtained from Pdyn neurons in the insula of male and female PdynGFP mice following 8 weeks of intermittent alcohol (EtOH) access or water (H2O) control to assess the effects of alcohol on the intrinsic properties (A–C), excitability (at RMP; D–L), and synaptic transmission (N–R) of insula Pdyn neurons. Alcohol increased the RMP of Pdyn neurons in female mice, but not male mice (A), and had no effect on input resistance (B) or sag ratio (C) in either sex. D, H, In male but not female mice, alcohol decreased the excitability of Pdyn neurons, as indicated by decreased firing to increasing current steps (D) and increased rheobase (minimum current needed to elicit firing; H). G, There was no significant effect of alcohol on Pdyn AP threshold. E, H, Representative traces from male (left, blue) and female (purple, right) mice show total firing elicited by a 600 pA current step (E) and the amount of current required to fire an action potential (rheobase) determined using a 120 pA/1 s ramp protocol (H). I–L, There was no effect of alcohol on the AP threshold or on any measures of AP kinetics assessed, including amplitude (I), half-life (J), rise time (K), or decay time (L). L, There was a main effect of sex on AP decay time, as Pdyn neurons of female mice exhibited shorter decays times. M–R, Alcohol did not affect synaptic transmission in Pdyn neurons in either sex aside from producing a decrease in sIPSC amplitude in female, but not in male, mice. There was no difference between EtOH and H2O groups in sEPSC frequency (M), sEPSC amplitude (N), or sIPSC frequency (O) in male and female mice. Sex differences in spontaneous synaptic transmission in Pdyn neurons were observed, as female mice exhibited higher sEPSC amplitude (N), lower sIPSC frequency and amplitude (O), and higher E/I ratio compared with male mice, and EtOH exposure increased the ratio of excitatory to inhibitory input onto Pdyn neurons (Q). R, Representative traces of sIPSC (top trace) and sEPSC (bottom trace) obtained from individual Pdyn neurons in male and female water-exposed (light blue and purple, respectively) and alcohol-exposed (dark blue and purple, respectively) mice. *p < 0.05, **p < 0.01, #p < 0.05 main effect of sex.
Next, we evaluated the impact of IA on synaptic transmission in Pdyn neurons (Fig. 5M–R). We found no effect of IA or sex on sEPSC frequency (males: EtOH and H2O, n = 9 cells/n = 4 mice; females: EtOH, n = 11 cells/n = 5 mice; H2O, n = 9 cells/n =4 mice; sex × treatment interaction, p = 0.401; sex main effect, p = 0.280; treatment main effect, p = 0.260), suggesting that the frequency of glutamatergic synaptic events is similar in male and female mice and is not altered by long-term IA. When we assessed GABAergic transmission, we found that females received less inhibitory input onto insula Pdyn neurons than male mice, as supported by a main effect of sex on sIPSC frequency (F(1,34) = 4.75, p = 0.036). As with sEPSCs, we found no effect of IA on sIPSC frequency (sex × treatment interaction, p = 0.819; treatment main effect: F(1,36) = 4.71, p = 0.159). Interestingly, when we assessed the E/I ratio onto insula Pdyn neurons, we found significant main effects of sex (F(1,34) = 5.08, p = 0.031) and treatment (F(1,34) = 6.10, p = 0.018) but no group × sex interaction (p = 0.754). This finding indicates that there is greater excitatory synaptic drive in Pdyn neurons of male mice, and that IA increases E/I ratio in this neuronal population in both males and females. When event amplitude was assessed, we found that SE/IPSC amplitudes were lower in male Pdyn neurons as compared with females, as indicated by a significant main effects of sex (sEPSC: F(1,34) = 18.5, p < 0.001; sIPSC: F(1,34) = 12.1, p < 0.001). Whereas there was no effect of IA on sEPSC amplitude in male and female mice (sex × treatment, p = 0.802; treatment main effect, p = 0.222), sIPSC amplitude was decreased in IA-exposed animals, as revealed by a significant main effect of treatment (F(1,34) = 4.23, p = 0.047) but no interaction of sex × treatment (p = 0.330).
We subsequently examined the effect of long-term IA on intrinsic properties, excitability, and synaptic transmission in insula DLPNs of male mice (EtOH, n = 11 cells/n = 4 mice; H2O, n = 10 cells/n = 4 mice) and female mice (EtOH, n = 12 cells/n = 5 mice; H2O, n = 8 cells/n = 4 mice), as shown in in Figure 6. In both male and female mice, we found that some intrinsic properties of insula DLPNs were altered by IA (Fig. 6A–C). Specifically, 8 weeks of IA decreased RMP (Fig. 6A) in male but not female mice (IA vs H2O: males: t(37) = 2.75, p = 0.009; females: p = 0.135; treatment × sex interaction, p = 0.431; treatment main effect: F(1,37) = 9.03, p = 0.005; sex main effect, p = 0.737), and input resistance (Fig. 6B) in male and female mice (IA vs H2O: males: t(37) = 2.88, p = 0.013; females: t(37) = 3.09, p = 0.008; treatment × sex interaction, p = 0.812; treatment main effect: F(1,37) = 17.8, p < 0.001; sex main effect: F(1,37) = 4.67, p = 0.037). There was no effect of IA on sag ratio (Fig. 6C) in either sex (treatment × sex interaction, p = 0.228; treatment main effect, p = 0.460; sex main effect, p = 0.069). Long-term IA also reduced the excitability of insula DLPNs (Fig. 6D–H) in male mice as we found a significant decrease in the number of action potentials fired at increasing current steps (Fig. 6D) in males (treatment × current interaction: F(10,190) = 2.92, p = 0.002; treatment main effect: F(1,19) = 10.4, p = 0.005; current main effect: F(10,190) = 92.0, p < 0.001), but not in females (treatment × current interaction, p = 0.052; treatment main effect, p = 0.557; current main effect: F(10,170) = 62.3, p < 0.001). This was further supported by an increase in rheobase (Fig. 6G) in IA-exposed male, but not in female, mice (IA vs H2O: males: t(19) = 2.43, p = 0.020; females, p = 0.059; sex × treatment interaction, p = 0.788; treatment main effect: F(1,37) = 9.53, p = 0.004; sex main effect, p = 0.103). There was no effect of IA on AP threshold (Fig. 6F; F values < 1 for treatment × sex interaction, and main effects of treatment and sex main effect). Representative traces of insula DLPN firing to a 600 pA current step (Fig. 6E) and a 120 pA/s ramp protocol (Fig. 6H) in male (left, blue) and female (right, purple) H2O and IA mice. Additionally, we found several effects of IA on action potential kinetics in male mice, including an effect of IA on AP peak amplitude (Fig. 6I; treatment × sex interaction: F(1,34) = 9.81, p = 0.004; main effects of treatment: F(1,34) = 5.65, p = 0.023; and sex: F(1,34) = 4.83, p = 0.035) and on AP rise time (Fig. 6K; treatment × sex interaction: F(1,34) = 6.98, p = 0.012; treatment main effect: F(1,34) = 20.7, p < 0.001; and sex main effect, p = 0.094). Follow-up analyses showed that in male mice only IA produced a significant increase in AP peak amplitude (males: t(34) = 3.92, p < 0.001; females: p > 0.999) and a significant decrease in AP rise time (males: t(34) = 5.11, p < 0.001; females, p = 0.378) compared with H2O controls. Analyses also revealed that on both measures there were significant sex difference in H2O control animals, as AP peak amplitude was lower (t(34) = 3.58, p = 0.002) and rise time higher (t(34) = 2.94, p = 0.012) in H2O males compared with females. There were no effects of IA or sex on AP half-width (Fig. 6J; F values < 1 for treatment × sex interaction and sex main effect; treatment main effect, p = 0.095) or decay (Fig. 6L; treatment × sex interaction: F < 1; treatment main effect, p = 0.094; sex main effect, p = 0.175).
Insula DLPNs are differentially altered by intermittent alcohol access in male and female mice. Whole-cell patch-clamp recordings were obtained from insula DLPNs in male and female mice following 8 weeks of intermittent alcohol (EtOH) access or water control (H2O). A–R, The effects of alcohol on the intrinsic properties (A–C), excitability (at RMP; D–L), and synaptic transmission (M–R) of insula DLPNs were assessed. A, B, Intrinsic properties of insula DLPNs were altered by alcohol, as it produced a decrease in DLPN RMP in male mice (A), and a decrease in input resistance in males and females (B). C, There was no effect of alcohol on DLPN sag ratio. D, G, Alcohol decreased DLPN excitability in male but not female mice, as indicated by decreased firing to increasing current steps (D) and increased rheobase (minimum current needed to elicit firing; G). F, There was no significant effect of alcohol on DLPN AP threshold. E, H, Representative traces from males (left, blue) and females (purple, right) show total firing elicited by a 600 pA current step (E) and the amount of current required to fire an action potential (rheobase) determined using a 120 pA/1 s ramp protocol (H). I–L, In male mice, alcohol altered AP kinetics of insula DLPNs, as it (I) increased AP peak amplitude and (K) decreased AP rise time. There was no effect of alcohol in female mice on these measures. In H2O controls, APs elicited by insula DLPNs in female mice had a higher peak amplitude (I) and shorter rise time (K) compared with male mice. J, L, There was no effect of alcohol and no sex differences found for AP half-width (J) or decay time (L). M–R, Alcohol produced sex-specific effects on synaptic transmission in insula DLPNs. M, N, Alcohol increased sEPSC frequency (M) but not amplitude (N) in DLPNs of male mice and had no effect in female mice. Excitatory events were lower in frequency in EtOH females and higher in amplitude in H2O females compared with males within each group. O, P, Alcohol had no effect on sIPSC frequency (O) or amplitude (P) in either sex. In female mice, inhibitory events were lower in frequency and higher in amplitude compared with males. Q, Alcohol increased the E/I ratio events in male mice, and there was no difference in E/I ratio by sex. R, Representative traces of sIPSC (top trace) and sEPSC (bottom trace) obtained from individual DLPNs in the insula of male and female water (light blue and purple, respectively) and alcohol (dark blue and purple, respectively) exposed mice. *p < 0.05, **p < 0.01, ***p < 0.001.
Synaptic transmission was next assessed in DLPNs of IA-exposed and water control male and female mice (Fig. 6M–R). In these experiments, we found that long-term IA produced a sex-dependent increase in excitatory synaptic transmission. This was revealed by a significant sex × treatment interaction on sEPSC frequency (males: EtOH, n = 10 cells/n = 5 mice; H2O, n = 13 cells/n = 4 mice; females: EtOH, n = 12 cells/n = 5 mice; H2O, n = 8 cells/n = 4 mice; F(1,39) = 6.60, p = 0.014) as well as a significant main effect of sex (F(1,39) = 11.7, p = 0.001) but not group (p = 0.117). Post hoc analyses revealed that IA increased DLPN sEPSC frequency in male mice (p = 0.023) but not in female mice (p > 0.999), which subsequently resulted in a significant increase in sEPSC frequency in IA-exposed males compared with females (p < 0.001). Conversely, there was no effect of long-term alcohol intake on inhibitory synaptic transmission in insula DLPNs, as sIPSC frequency did not differ between IA mice and water controls. This finding was demonstrated by the absence of both a sex × treatment interaction (p = 0.442) and a main effect of treatment (p = 0.776). Analyses yielded a main effect of sex on sIPSC frequency (F(1,39) = 25.3, p < 0.001), and post hoc comparisons revealed that H2O-exposed and IA-exposed males exhibited higher sIPSC frequency compared with females of the same treatment group (H2O, p = 0.006; IA, p < 0.001). We next assessed E/I ratio in insula DLPNs and found that IA significantly increased synaptic drive in this neuronal population. This was demonstrated by a significant main effect of treatment (F(1,39) = 5.13, p = 0.029) but not of sex (p = 0.881), and no sex × treatment interaction (p = 0.141). Follow-up analyses showed that there was a significant difference in E/I ratio between IA-exposed and water-exposed males (p = 0.008) but not females (p = 0.608), indicating that IA increased excitatory synaptic drive in male mice only. Moreover, we found sex-related differences in event amplitude, as sEPSC and sIPSs were higher in female mice. This was illustrated by a significant main effect of sex on sEPSC/sIPSC amplitude (sEPSC: F(1,39) = 23.6, p < 0.001; sIPSC: F(1,39) = 36.5, p < 0.001) and post hoc analyses showing significant differences in sEPSC between male and female H2O mice (p < 0.001) and sIPSC between male and female H2O (p < 0.001) and IA mice (p = 0.004).
Discussion
In this study, we used converging pharmacological and genetic approaches to gain insight into how DYN/KOR signaling in the insula contributes to excessive alcohol consumption in male and female mice. In addition, we used slice electrophysiology to probe physiological changes in insula neuronal populations, focusing on DYN-expressing neurons and layer 5 pyramidal neurons. Across these studies, we identified numerous sex differences at the behavioral and physiological levels. This adds to a growing literature demonstrating sex differences in how KOR systems can modulate behavior and neuronal function.
Divergence between pharmacology and genetic approaches
In this study, there were several key differences between the site-directed pharmacological (blocking KOR) and genetic (deleting KOR/Pdyn) approaches we used to probe the impact on drinking. In male mice, we saw an early impact on alcohol consumption that diminished over time. We hypothesized that could be because of the waning effects of nor-BNI or a shift to a different state that is insensitive to local KOR antagonism. To test this, we then infused nor-BNI in another set of male mice that had already escalated their alcohol intake and found no effect on consumption. This suggested that in males only, local KOR antagonism in the insula could reduce early alcohol consumption, but following escalation, it was no longer sensitive. This result was surprising, as in many other studies, KOR antagonist effects emerge as animals become dependent or escalate their consumption (Walker et al., 2011; Erikson et al., 2018). However, it is notable that KOR antagonists can impact drinking in more acute models of escalated consumption, such as Drinking in the Dark (Anderson et al., 2019; Haun et al., 2020). The antagonist approach does have limitations, including difficulty in accurately quantifying the spread and potential off-target actions of the compound.
To address these limitations, we used a genetic approach, similar to our previous work in the central nucleus of the amygdala (CeA; Bloodgood et al., 2021). In contrast to local inhibition, we found that deletion of KOR had no effect on any measures of alcohol intake. This divergence is likely because of the dissociation between the receptors that the local antagonist can block compared with the target population of KOR ablated by a conditional deletion approach. It is likely that the insula receives inputs that presynaptically express KOR on their terminals, and this population of KOR can be blocked by local pharmacology but not by deletion of KOR from insula neurons. Another possibility is that the deletion of KOR from insula neurons also leads to removal of presynaptically expressed KOR in regions downstream from the insula. Thus, in these output regions, KOR signaling may play a distinct role in alcohol consumption that differs from that of the insula. Finally, we cannot exclude the possibility that some of the differences between approaches may be because of issues related to knock-out efficiency and viral expression/spread.
We also examined how deletion of insula Pdyn could impact alcohol drinking and found that it led to a robust reduction of alcohol consumption in both males and females, with some sex-related differences. Specifically, in female mice there was a reduction in alcohol intake in both early and late phases of the IA procedure. In males, the reduction in alcohol intake was only apparent in the later phases of IA, where mice typically escalate their consumption. Deletion of Pdyn in females also reduced sucrose intake, suggesting that insula Pdyn may more broadly regulate consumption of rewarding solutions in female mice. This is especially interesting as recent studies have implicated the insula in feeding-related behaviors (Stern et al., 2021). However, the experiments in the study by Stern et al. (2021) were performed in male mice and implicated a distinct population of central amygdala-projecting insula neurons in feeding and satiety signaling. Thus, we are cautious to speculate whether our observations may extend to other signaling systems or neuronal populations within the insula. Interestingly, we found that Pdyn and KOR manipulations differentially impacted alcohol consumption, specifically in terms of the phase of intake that was impacted by our manipulations. One possible reason for this may be that when we delete Pdyn, we are altering both local and distal neuropeptide release. There are a number of insula targets that could contribute to this effect, such as the CeA, which has been implicated in KOR regulation of alcohol consumption. Further, the deletion of pDyn may have additionally impacted other peptides, such as leu-enkephalin, which is derived from precursor molecules pDyn and proenkephalin (Akil et al., 1984; Evans et al., 1988). Thus, in using this genetic approach, we cannot exclude the possibility that other peptides may have contributed to our findings. Another potential caveat to this work is that the mouse lines we have used to perform these genetic studies are not fully backcrossed on to a C57BL/6J background. We observed similar differences in baseline consumption in an earlier study, investigating the role of Dyn/KOR in the CeA (Bloodgood et al., 2021). In the future, CRISPR-based tools may provide a means to circumvent this concern.
Notably, none of the manipulations we performed that altered alcohol consumption altered avoidance behavior, assessed via the elevated plus maze. This suggests that the role of the KOR/DYN system may be distinct from the regulation of avoidance behavior. Previous work has shown that the insula and its output to the extended amygdala are engaged by and contribute to negative affective states in mice (Centanni et al., 2019; Luchsinger et al., 2021). However, these experiments examined the role of insula during the distinct state of forced abstinence from alcohol or during an explicit behavior, specifically struggling during restraint stress. Although we did not explore insula DYN/KOR during forced abstinence or restraint stress, this would be an interesting direction for future work, and may further illuminate the impact of alcohol-induced DYN/KOR signaling shifts in affective outcomes. Additional work from Seif et al. (2013), Chen et al. (2015), Chen and Lasek (2020), Darevsky and Hopf (2020), and De Oliveira Sergio et al. (2021) has shown that the insula is associated with aversion-resistant and compulsive-like alcohol intake. We did not explore that phenotype in our model, and it represents another path for exploration.
Alcohol consumption has complex effects in distinct insula cell types
Given that Pdyn deletion decreased alcohol intake, we next wanted to evaluate the impact of long-term alcohol consumption on the properties of insula Pdyn neurons. We found that with the excitability of these neurons, there were divergent effects depending on sex. In males, we found a reduction in current-evoked firing and an increase in rheobase, consistent with reduced excitability. In female mice, we found that IA increased the resting membrane potential in Pdyn neurons only. There was a main effect of alcohol drinking on the E/I ratio across both sexes; however, there was reduced basal GABAergic tone in insula Pdyn neurons from females, suggesting possible differences in the function of insula interneurons or external GABA inputs on insula Pdyn neurons. Together, the results suggest that IA induces long-term synaptic plasticity in insula Pdyn neurons in males and are possibly indicative of a synaptic scaling process. This is especially interesting considering recent results suggesting that synaptic scaling plays a role in learning in the closely related gustatory cortex (Wu et al., 2021). The increase in RMP and synaptic balance in insula Pdyn neurons from female mice could be linked to the greater effect of DYN deletion in females; however, our data do not explicitly address this possibility. It is noteworthy that these findings, in particular the sex differences, are similar to our recent study exploring the impact of binge-like alcohol drinking on DYN neurons in the CeA (Bloodgood et al., 2021), and align with recent data from the Sparta laboratory demonstrating sex-specific effects of alcohol on insula neuronal plasticity (Marino et al., 2021). This raises the important issue that sex is a critical factor to consider when evaluating the insula and KOR ligands in preclinical models, as has been noted by Chartoff and Mavrikaki (2015).
We also evaluated the impact of long-term alcohol consumption on layer 5 pyramidal neurons in the IC, as outputs from the insula have been implicated in a range of complex cognitive and physiological processes (Gogolla, 2017) and found them to be altered following alcohol exposure (McGinnis et al., 2020). A similar pattern emerged with these layer 5 neurons showing reductions in excitability, with an increase in E/I ratio from males, which is suggestive of a scaling-like phenomenon, but with no differences in the properties of layer 5 insula from female mice. This further demonstrates that alcohol drinking induces a sex-dependent plasticity and highlights the need for additional work to resolve how this may impact the behaviors we observed. Previous studies by Jaramillo et al. (2018a,b,c) and Seif et al. (2013) have shown that an insula–nucleus accumbens pathway is important for alcohol self-administration and aversion-resistant alcohol drinking. This may indicate one possible mechanism for how the cellular changes we observed may promote increased alcohol consumption. De Oliveira Sergio et al. (2021) also found a role for insula outputs to brainstem in punished alcohol drinking but not alcohol-only drinking, suggesting that while this path is important, it may not be related to the drinking phenotypes seen. However, there is also intriguing work that insula outputs to the BLA can play a critical role in learning (Yiannakas et al., 2021). This is especially exciting when taken with recent data supporting the role of the BLA in alcohol reinforcement (Faccidomo et al., 2021). Future studies should focus on investigating plasticity in these distinct outputs, as it can provide important insight into how alcohol can impact insular cortex circuits.
As summarized in our working model in Figure 7, our combined data suggest that KOR/DYN signaling in the insula can regulate alcohol consumption and that there are key sex differences in the mechanism. Pharmacology suggests that in males there is an early KOR signal in the insula that plays a role in the induction of, but not the maintenance of, escalated alcohol drinking. In contrast, Pdyn deletion can reduce alcohol consumption during the later phases of drinking, suggesting a shift to KOR sensitivity in downstream structures, potentially the CeA. In females, Pdyn deletion appears to robustly impact drinking in both the early and late IA. Given that sucrose intake was concomitantly decreased, it is possible that Pdyn deletion may more generally affect consummatory or appetitive responses in female mice. There are similar sex-dependent changes on neuronal function; however, our data do not specifically identify how they are related to escalated consumption. Notably, our findings underscore the importance of accounting for sex as a biological variable in alcohol-related behavior, and specifically the mechanisms by which KOR regulates alcohol consumption. As a potential target for the treatment of AUD, it is important to identify these mechanisms and determine how they relate to sex differences in clinical populations.
Working model of insular cortex DYN/KOR involvement in drinking behavior in male and female mice. Summary of experimental outcomes highlighting how different manipulations of KOR signaling can impact behavior.
Footnotes
This work was supported by National Institutes of Health Grants P60-AA-011605 (to T.L.K. and M.N.), F32-AA-026485 (to M.M.P.), K99/R00-AA-028298 (to M.M.P.), U01-AA-020911 (to T.L.K.), R01-AA-025582 (to T.L.K.), and R01-AA-025582-S1 (to T.L.K. and M.M.P.).
The authors declare no competing financial interests.
- Correspondence should be addressed to Melanie M. Pina at mpina{at}som.umaryland.edu