Sexually dimorphic effects of monoacylglycerol lipase inhibitor MJN110 on stress‐related behaviour and drinking in Marchigian Sardinian alcohol‐preferring rats

The endocannabinoid (eCB) system plays an important homeostatic role in the regulation of stress circuits and has emerged as a therapeutic target to treat stress disorders and alcohol use disorder (AUD). Extensive research has elucidated a role for the eCB anandamide (AEA), but less is known about 2‐arachidonoylglycerol (2‐AG) mediated signalling.

Sex-specific eCB-based approaches should be considered in the clinical development of therapeutics.

K E Y W O R D S
2-AG, alcohol use disorder, endocannabinoid system, monoacylglycerol lipase, sex differences, stress

| INTRODUCTION
Stress-related disorders, including anxiety disorders and post-traumatic stress disorder (PTSD), affect over 30% of adults in the United States and represent one of the most prevalent mental health morbidities in the world (Kessler et al., 2012).Symptoms of these disorders include hyperarousal, sleep disorders, excessive fear response even in the absence of an immediate threat, and irritability (Ressler et al., 2022).
Additionally, stress-related disorders are often intertwined with, and possibly a risk factor for, substance use disorders such as alcohol use disorder (AUD).AUD often is co-diagnosed with anxiety and stress disorders, whereby alcohol is used for self-medication purposes to alleviate anxiety, distress and negative affect state.In AUD patients, alcohol intake progressively increases, initiating a vicious feed-forward cycle with subsequent worsening of anxiety symptoms and dysphoria that results in comorbidity (Kushner et al., 2000).
The endogenous cannabinoid, or endocannabinoid (eCB), system plays an important homeostatic role in the regulation of stress circuits (Worley et al., 2018) and has emerged as a putative therapeutic target to treat stress, anxiety, fear-related behaviours and AUD (Bedse et al., 2019;Natividad et al., 2017;Shonesy et al., 2014).
The eCB system is a retrograde signalling system that regulates multiple physiological functions in the central and peripheral nervous systems.N-arachidonoylethanolamine (anandamide [AEA]) and 2-arachidonoylglycerol (2-AG) are the two most studied lipid-derived endogenous ligands that exert biological effects via the activation of cannabinoid receptors type 1 and 2 (CB 1 and CB 2 ).AEA and 2-AG are mainly degraded by two serine hydrolases, fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL), respectively.Preclinical evidence supports the hypothesis that serine hydrolase inhibitors augment brain eCB levels and thereby may therapeutically reduce symptoms of stress and anxiety (Pavon et al., 2021;Zhang et al., 2015;Zhong et al., 2014).
Recent development of pharmacological tools to selectively inhibit MAGL activity has made it possible to study the contribution of 2-AG signalling in mediating these responses (Blankman & Cravatt, 2013).2-AG is produced on demand in postsynaptic neurons by the hydrolytic activity of diacylglycerol-lipase alpha (DAGLa) and canonically acts as a retrograde transmitter by activating G i/o -coupled CB receptors located at the presynaptic level to reduce neurotransmitter release (Kano et al., 2009;Piomelli, 2003;Stella et al., 1997).
Preclinical studies have demonstrated that augmenting 2-AG signalling can reduce anxiety-like behaviours and treat stress-related symptoms (Bedse et al., 2017;Bluett et al., 2017;Bosch-Bouju et al., 2016;Ivy et al., 2020;Lisboa et al., 2017;Lutz et al., 2015;McLaughlin et al., 2014;Patel & Hillard, 2009), whereas genetic or pharmacological inhibition of 2-AG synthesis increases anxiety-like behaviours, hinders extinction of conditioned fear, and increases susceptibility to stress-induced anxiety (Bluett et al., 2017;Cavener et al., 2018;Shonesy et al., 2014).In addition, studies have shown that stress induces bidirectional changes in brain eCB levels.Stress exposure reduces amygdala AEA levels, and its reduction after stress correlates with anxious-like behaviours (Bluett et al., 2014).In contrast, stress can increase amygdala 2-AG levels, which determines termination and adaptation to stress, as well as changes in emotional memory (Hill et al., 2010;Patel et al., 2009Patel et al., , 2005)), suggesting that 2-AG regulates stress signalling once the stress response is activated.Similarly, effects of alcohol exposure on eCB content have been described and the interactions between alcohol and eCB system at molecular, synaptic and behavioural levels have been reviewed recently (Kunos, 2020;Wolfe et al., 2022).For instance, acute alcohol exposure has been associated with decreased 2-AG in prefrontal cortex (PFC) (Rubio What is already known • The endocannabinoid system is a target for treatment of stress-related and alcohol use disorders (AUD).
• 2-AG signalling is a key modulator of the stress response.

What does this study add
• MJN110 significantly inhibits the 2-AG degrading enzyme MAGL and elevates 2-AG in relevant brain regions.
• MJN110 alleviates stress-related symptoms in a sex-specific manner.

What is the clinical significance
• This study highlights a distinct role for 2-AG in the modulation of stress-related disorders.
• Sex-specific endocannabinoid-based approaches should be considered in the clinical development of therapeutics. et al., 2007), while chronic intermittent alcohol has been associated with reduction of baseline 2-AG levels in the central nucleus of the amygdala (central amygdala [CeA]) (Serrano et al., 2018).
Genetically selected Marchigian Sardinian alcohol-preferring (msP) rats display comorbid symptoms of excessive alcohol drinking and enhanced anxiety compared with their genetic counterpart Wistar rats, and they resemble a post-dependence phenotype (Ciccocioppo et al., 2006(Ciccocioppo et al., , 1999)).It has been hypothesized that elevated alcohol intake in msP rats is driven, at least in part, by the attempt to attenuate anxiety-like behaviours and 'self-medicate'.The msP rats display pronounced aberrations in brain stress signalling, including genetically determined overexpression of corticotropin-releasing factor receptor 1 (CRF 1 ) in various brain regions (Ayanwuyi et al., 2013;Hansson et al., 2007Hansson et al., , 2006)).More recently, the discovery of dysregulated eCB signalling in msP rats raises the possibility that the eCB system critically underlies the comorbid expression of behavioural anxiety and excessive alcohol drinking (Natividad et al., 2017(Natividad et al., , 2021)).Additionally, previous work has shown that FAAH inhibitors can be successfully employed to increase AEA levels and regulate drinking behaviours (Stopponi et al., 2018), but much less is known about the role of 2-AG in stress-related disorders.Lastly, sex-dependent effects of eCB modulation of stress-related disorders have started to be elucidated (Henricks et al., 2017;Morena et al., 2021), but more studies are needed to examine potential sex-specific effects of MAGL inhibition.
Here, we used systemic administration of the brain-permeant MAGL inhibitor, MJN110, to pharmacologically block MAGL enzymatic activity and test the hypothesis that modulation of 2-AG can decrease voluntary alcohol drinking and stress-related behaviours in male and female msP rats.First, we assessed MJN110 efficacy to significantly increase brain 2-AG levels using liquid chromatographymass spectrometry (LC-MS/MS).Then, we assessed the effects of augmenting the 2-AG signalling on voluntary alcohol drinking and stress-induced behavioural adaptations such as novelty-induced anxiety behaviours, irritability-like behaviours and cued fear expression in a rodent model resembling post-dependent states and affective impairments (Hansson et al., 2007(Hansson et al., , 2006)).

| Animals
We used a total of 292 rats.Adult male (n = 64, $450 g) and female (n = 81, $250 g) msP rats, obtained from the School of Pharmacy, University of Camerino (Camerino, Italy), were bred at The Scripps Research Institute (La Jolla, CA, USA).Adult male (n = 66, $450 g) and female (n = 81, $250 g) Wistar rats from which msP rats originate were purchased from Charles River Laboratories (Wilmington, MA, USA).Rats were housed in a temperature-and humiditycontrolled vivarium on a 12-h reverse light/dark cycle (lights off at 8:00 AM), with food and water available ad libitum.Rats were pairhoused, separated by a perforated clear plexiglass divider to habituate them to the behavioural test conditions, while also reducing isolation stress (Steinman et al., 2021).The rats were randomly assigned to the different treatment groups.The in vivo experimental system in rat replicates aspects of the human AUD and stress mechanisms (Borruto et al., 2021).We conducted all procedures in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and with The Scripps Research Institute Institutional Animal Care and Use Committee (IACUC) policies (Protocol Number 09-0006).Animal studies are reported in compliance with the ARRIVE guidelines (Percie du Sert et al., 2020) and with the recommendations made by the British Journal of Pharmacology (Lilley et al., 2020).

| Materials and drug administration
MJN110 was synthesized as described in Niphakis et al. (2013) and provided by the Cravatt laboratory at The Scripps Research Institute.
The doses and pretreatment time were chosen based on previous work demonstrating that the acute systemic administration of MJN110 at 5 mgÁkg À1 dose yields maximal MAGL inhibition in the rat brain, while having no effect on other measured eCBs, including AEA (Burston et al., 2016;Niphakis et al., 2013).

| Two-bottle choice alcohol access procedure
The two-bottle choice (2-BC) procedure (free choice between water and 10% v/v alcohol) was used to measure voluntary alcohol drinking and preference (Leeman et al., 2010).Animals were given free access to water and alcohol (10% v/v) 24 hÁday À1 , for 15 days to establish a stable drinking baseline and preference for alcohol (80%-90% preference alcohol vs. water in msP rats).Once the baseline was reached, the access to alcohol was then decreased to 2 h, starting from 2 h into the animals' dark active phase, for 4 days.After reaching a stable 2-h-drinking baseline, on the 5th day, rats were injected with either vehicle or MJN110 (10 or 5 mgÁkg À1 ), 3 h prior to the beginning of the 2-BC session (2 h).Fluids were offered through graduated drinking tubes equipped with metallic spouts and intake was measured by weighing the tubes at the end of each drinking session.The drinking tubes were switched daily to avoid the development of side preference.Animals had free access to food.Alcohol intake was calculated as absolute value of consumption at each time interval and expressed as grams per kilogram of body weight to control for body weight variability (Devgun & Dunbar, 1990).
Exposure to a novel environment elicits a stressful reaction in rodents that can interfere with normal behaviour, including food consumption, because the animals face a choice between engaging in feeding and the fear of novelty (Bechtholt et al., 2007).Here, 24 h before undergoing testing, the animals received home cage exposure to a novel palatable food (50% sucrose, chocolate-flavoured pellets, 45 mg, 3.4 kcalÁg À1 , 5TUL, Test Diets, St. Louis, MO, USA) in their housing room during their dark phase, to familiarize the rats with the food pellets.Upon exposure to the palatable food ($1 g), rats were left undisturbed and allowed to consume the whole amount, while monitored by the experimenter to confirm that each rat tasted the novel food.
The next day, the rats were treated either with MJN110 (5 mgÁkg À1 ) or vehicle and left undisturbed for 3 h prior to evaluation under novel testing conditions that were perceived as stressful (i.e., white lights, unfamiliar double-size cage, 60-dB background noise and different room).Rats were placed in the novel environment and allowed to explore.Treatment-blind experimenters measured the latency to first approach the food, latency to consume the chocolate pellets, and the total intake of the novel palatable food in the unfamiliar environment over a 10-min trial.Rats that did not approach the food within the 10-min trial were assigned 600-s latencies.Rats that did not eat any pellet were assigned intake of 0 gÁkg À1 .We also performed a separate control experiment on a separate set of animals where rats were exposed to the relevant test diet (i.e., highly palatable chocolate pellets) and tested in the familiar home cage 24 h after the first home cage exposure to the novel food.Intake data are expressed as grams of pellet intake per kilogram of body weight.

| Irritability test
Irritability-like behaviour was measured using the bottle-brush procedure (Kimbrough et al., 2017).The testing was performed in a behavioural room under dim red lights, 3 h after treatment with MJN110 (5 mgÁkg À1 ) or vehicle.Each individual rat was placed in a clear plastic cage (27 Â 48 Â 20 cm) containing clean bedding, facing the back of the cage.For each trial, a treatment-blind experimenter inserted a bottle brush into the cage and rotated it in five phases, each one lasting 3 s: (1) rotating towards the rat, (2) forcing interaction by touching the whiskers, (3) withdrawing from the rat, (4) rotating upright and (5) holding upright without rotation.The testing consists of 10 trials per rat with 10-s intervals between each trial.The following were scored as aggressive-like responses: biting, boxing, siding, following and mounting.The following were scored as defensive-like responses: avoiding, digging/burying, freezing, jumping, startling and vocalizing.
Grooming, rearing and exploring were additionally recorded during the trial.Data are expressed as the sum of all the aggressive-like behaviours and the sum of all the defensive-like behaviours.

| Fear conditioning
Rats were placed in 30.5 Â 25 Â 28-cm chambers (Med Associates, St. Albans, VT, USA), which were cleaned using 70% isopropanol before and in between rats.Chambers were equipped with an infrared camera for behavioural observation and recordings.On conditioned fear acquisition days (Days 1 and 2), rats were acclimatized first to the chamber for 5 min without lights or any stimuli.This 5-min period was used to examine initial freezing behaviour before the presentation of the cued stimulus.Rats were then presented with six conditioned stimulus (CS; light)-unconditioned stimulus (US; foot shock) pairings that were separated by variable (60 ± 20 s) intertrial intervals (ITIs).
Each light cue lasted 20 s and co-terminated with a 2-s 0.6-mA electric foot shock.At the end of the conditioning session, rats were left in the dark chamber for an additional 2 min.On the conditioned fear expression day (Day 3), half of the rats of each sex and genotype were treated with vehicle and the other half with MJN110 (5 mgÁkg À1 ) 3 h before being placed in the same chamber as per conditioned fear acquisition Days 1 and 2. After 5 min in the chamber, rats were presented with 18 CS presentations (20 s) (no shock) separated by a 60 ± 20-s ITI, followed by final 2 min without any stimulus presentation.Animals were videorecorded during testing, and freezing behaviour was manually scored, blind to drug treatment, by measuring the absence of any movement except for respiratory-related movements during the CS + ITI duration.

| Tissue collection
A separate batch of rats (n = 40) was used for brain tissue collection and LC-MS/MS analysis.Rats were injected intraperitoneally either with vehicle or MJN110 (5 mgÁkg À1 ) and 3 h later deeply anaesthetized with isoflurane and rapidly decapitated while unconscious using a guillotine.Brains were removed and flash frozen in dry ice-cold isopentane.Prelimbic (PrL) cortex (1 mm), infralimbic (IL) cortex (1 mm) and CeA (0.8 mm) punches were dissected from cryostat-sectioned slices (400 μm).

| Lipid extraction and analyses
Rat brain punches were thawed on ice and homogenized using a glass Dounce homogenizer or a mechanical mortar and pestle gun in 500 μl of ice-cold Dulbecco's phosphate-buffered saline (DPBS) (Gibco, Thermo Fisher Scientific, Waltham, MA, USA, Catalogue #14190144) in a cold room.Dounce homogenized samples were then probe sonicated briefly (8 pulses, 10% power).An aliquot (10 μl) of each homogenized sample was reserved for protein concentration quantification using DC (detergent compatible) Protein Assay (Bio-Rad, Hercules, CA, USA, Catalogue #5000112).The remaining homogenized lysate was then immediately transferred into 3 ml of ice-cold chloroform:methanol (2:1) spiked with the internal standards 2-AG-d5 (500 pmol, Cayman, Ann Arbor, MI, USA, Catalogue #362162) and AEA-d4 (100 pmol, Cayman, Catalogue #10011178), and an additional 500 μl of DPBS was added to the extraction mixture.Samples were vortexed for 30 s and centrifuged at 1400 g for 3 min at 4 C, and the lower organic phase was collected.Additional 1 ml of chloroform and 100 μl of 3 N HCl were added to the remaining aqueous fraction; samples were then vortexed for 30 s and centrifuged at 1400 g for 3 min at 4 C, and the lower organic mixture was collected and combined with the first organic fraction.Samples were dried down using a nitrogen stream, resuspended in 100 μl of chloroform:methanol (2:1) and transferred into LC-MS/MS vials.

| Experimental design overview
This report consists of four behavioural assessments (2-BC alcohol drinking, irritability, NIH and conditioned fear expression) and one lipidomic study, that were conducted in five separate cohorts of rats.
Cohort 1 (lipid measurement): 10 male and 10 female Wistar rats and 10 male and 10 female msP rats were subjects.Rats were euthanized as described above (Subsection 2.7) 3 h after MJN110 or vehicle administration, and brains were collected for LC-MS/MS analysis (Figures 1 and S1).
Cohort 3 (irritability and NIH): 16 male and 16 female Wistar rats and 14 male and 16 female msP rats were subjects (Figures 3 and 4).
Cohort 5 (conditioned fear expression): 20 male and 20 female Wistar rats and 20 male and 20 female msP rats were subjects (Figure 5).

| Statistical analyses
2-BC alcohol drinking, irritability, NIH, and lipid measures were analysed using two-way analysis of variance (ANOVA) with treatment (vehicle vs. MJN110) and genotype (Wistar vs. msP) as betweensubject factors.Significant interaction effects were followed by Tukey's multiple comparisons test.Fear conditioning data containing repeated CS were analysed using a repeated measure (RM) two-way or three-way ANOVA with genotype (two-way) or genotype and drug treatment (three-way) as between-subject factors, and trial (e.g., CS + US pairing or CS presentation) as repeated, within-subject factors.
Significant interaction effects were followed by Bonferroni's multiple comparisons test.Freezing during the pre-CS acclimation period (context-only, no cue exposure) was also monitored and analysed separately using unpaired t-test or two-way ANOVA, as appropriate.
To better evaluate any difference in drug effects in the early or late The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology (Curtis et al., 2018(Curtis et al., , 2022)).All statistical analyses were performed, and all graphs were generated using Graph-Pad Prism V9 (GraphPad, San Diego, CA, USA).

| Single systemic administration of MJN110 increases 2-AG in PFC subregions and CeA
First, we wanted to determine whether the MJN110 dosage (5 mgÁkg À1 ) and timepoint (3 h) we selected would elevate 2-AG levels in cortical regions associated with stress, negative affect and alcohol drinking behaviour, such as the medial prefrontal cortex (mPFC), which includes the PrL and IL subregions (Klenowski, 2018), and the CeA (Gilpin et al., 2015).We focused on the 5 mgÁkg À1 MJN110 dose, which had been previously shown to decrease motivated behaviours in rats (Feja et al., 2020).The results revealed a main effect of the MJN110 treatment in significantly increasing 2-AG in the PrL cortex (Figure 1a,b) and IL cortex (Figure 1c,d) of both female and male rats, regardless of the genotype.Similarly, following a single MJN110 administration, 2-AG levels were significantly elevated in the CeA of both female (Figure 1e) and male (Figure 1f) rats.
No changes were observed in AEA levels across the brain regions analysed, sex or genotype, suggesting a selective inhibition of MAGL, and not FAAH, by MJN110 administration (Figure S1).The 5 mgÁkg À1 dose was selected and used for all the other behavioural assessments.

| Single systemic administration of MJN110 does not affect voluntary alcohol intake
To elucidate the acute role of 2-AG signalling in the regulation of voluntary alcohol drinking behaviour, we tested the MAGL inhibitor MJN110 in the 2-BC free access procedure in non-selected Wistar and genetically selected msP rats.Consistent with previous work (Ciccocioppo et al., 2006), we found that msP rats displayed significantly higher alcohol intake and preference than their Wistar controls in female rats (Figure 2a,b) and male rats (Figure 2c,d).Systemic administration of MJN110 (5 or 10 mgÁkg À1 , i.p.) did not alter the total alcohol intake (Figure 2a,c) or preference (Figure 2b,d) in either female or male, and Wistar or msP rats.

| Single systemic administration of MJN110 reduces anxiety-like behaviours in female rats
Consistent with our previous work (Vozella et al., 2021), female and male msP rats displayed a greater latency to approach the palatable food (Figure 3a,d) and greater latency to consume the palatable food (Figure 3b,e) under novelty stress conditions as compared with their Wistar counterparts.Similar genotype-dependent effects were observed with total food pellet intake when compared with Wistar rats (Figure 3c,f).Higher latency and lower pellet intake indicate higher sensitivity to novelty stress (Bechtholt et al., 2007;Carr et al., 2011).Notably, a single administration of MJN110 (5 mgÁkg À1 , i.p.) did not alter the latency to approach or latency to eat the palatable food in either sex or genotype but induced an overall increase of pellet intake in females, regardless of genotype (Figure 3c).To rule out generalized appetite-related action of MJN110 (Feja et al., 2020), in a separate cohort of females, we performed a control experiment.Rats were tested in their home cage (familiar environment) 24 h after the first exposure to the novel palatable food.Acute administration of MJN110 did not change total home cage palatable food pellet intake in either Wistar or msP female rats (Figure S2).

| Single systemic administration of MJN110 reduces irritability-like behaviours in msP rats
Acute treatment with MJN110 produced a robust reduction of total aggressive-like signs (i.e., biting, boxing, siding, following and mounting) in female msP rats (P < 0.05) but not Wistar rats (Figure 4a).MJN110 also exerted a main effect in elevating total defensive-like behaviours (i.e., avoiding, digging/burying, freezing, jumping, startling and vocalizing) in females (Figure 4b).When administered in male animals, MJN110 did not have any effect on aggressive signs (Figure 4c); however, it significantly increased the number of total defensive events in male msP rats (P < 0.05) but not in Wistar rats (Figure 4d).These results indicate that pharmacological inhibition of MAGL alleviates irritability-like behaviours in rats showing innate hypersensitivity to stress and anxiety but with different coping responses between female and male rats.For a summary of individual behaviours, refer to Table S1.

| Single systemic administration of MJN110
does not suppress cue-induced fear expression but reduces fear in the absence of cues in male msP rats On conditioned fear acquisition Day 1, we did not observe any genotype (Wistar vs. msP) difference in the freezing response during the chamber pre-conditioning (pre-CS) 5-min period in both females (Figure 5a) and males (Figure 5c).However, during the conditioned stimulus-unconditioned stimulus (CS-US) trial presentations, we found a significant Genotype Â CS-US trial interaction in both females and males.Post hoc comparisons indicated significantly higher freezing behaviour in female msP rats as compared with female Wistars in response to CS-US 2-6 (Figure 5b).Male msP rats showed significantly higher freezing behaviour as compared with Wistar rats in response to CS-US 1-6 (Figure 5d).
On conditioned fear acquisition Day 2, we did observe a significant genotype difference in the freezing response during the initial, pre-CS 5-min period in males (Figure 5g), but not in females (Figure 5e).During the CS-US trial presentations, we found a significant Genotype Â CS-US trial interaction in females (Figure 5f) and a significant main effect of genotype in males (Figure 5h).Post hoc comparisons indicated that female msP rats showed significantly higher freezing behaviour as compared with Wistar rats during presentations of CS-US 1-6 (Figure 5f).
On the next day, we tested whether a single injection of MJN110 (5 mgÁkg À1 ) suppressed conditioned fear expression.Rats were treated with MJN110 3 h prior to testing for conditioned fear expression where, after the pre-CS 5-min period, 18 CS were presented but in absence of the US (shock).MJN110 did not alter the freezing response during the 5-min pre-CS period in female Wistar rats, but it exerted a trending, nonsignificant effect in female msP rats (Figure 5i).
Notably, MJN110 significantly decreased the freezing response during the pre-CS period in msP males (P = 0.0128), but not in Wistar males (P = 0.9267) (Figure 5k).MJN110 did not significantly reduce the mean duration of freezing in the ITI following CS presentations in either sex, suggesting that, in the present model, an acute MAGL inhibition does not interfere with a cue-conditioned fear memory recall but might ameliorate non-cued fear responses (i.e., contextual conditioned fear and non-associative freezing).
Interestingly, we found a significant Genotype Â CS-US trial interaction in both females and males during the CS presentations.
Post hoc comparisons indicated that female msP rats showed a greater fear extinction than female Wistar rats as indicated by a lower freezing response by the last trial block, CS 13-18 (P < 0.05) (Figure 5j).Similar behaviour was observed in male msP rats, which showed progressively reduced freezing response across the three trial blocks (Figure 5l), even though their freezing response remained significantly higher than Wistar rats across all CS trials.

| DISCUSSION
In the present study, we sought to determine whether acute MAGL inhibition using MJN110 could suppress the innate vulnerabilities to stress and alcohol drinking in genetically selected msP rats compared with their counterpart Wistar rats.We found that MJN110 ameliorates the stress-related responses in msP rats, but not voluntary alcohol drinking, in a sex-dependent manner.Recent evidence has shown that eCB signalling regulates synaptic plasticity within the amygdala-PFC circuit under stressful experiences (Marcus et al., 2020) and fundamental sex differences have been identified within the amygdala-PFC fear circuit (Gruene et al., 2015).Here, we used a single systemic administration of the selective MAGL inhibitor MJN110 to first assess whether the dose (5 mgÁkg À1 ) and timepoint  et al., 2015;Zhang et al., 2021), and a growing body of work has demonstrated that eCB signalling in the amygdala is involved in the regulation of emotional states (Lim et al., 2016;Marsicano et al., 2002;Morena et al., 2019).After we treated male and female Wistar and msP rats with a single dose of MJN110 (5 mgÁkg À1 , 3 h), we measured 2-AG and AEA levels in PrL cortex, IL cortex and CeA.We were able to quantify 2-AG and AEA in each dissected brain region and found a general and consistent elevation of 2-AG levels in each brain region analysed, but no changes in AEA levels, for Wistar and msP rats.
MJN110 exhibits good potency for rat MAGL compared with the JZL184 inhibitor (Niphakis et al., 2013) and, consistent with previous reports, it produced a 5-to 10-fold increase of 2-AG at the 5 mgÁkg À1 dose.Additionally, because MJN110 achieves partial inhibition of MAGL, it does not lead to drug tolerance through CB 1 receptor desensitization.We then selected this 5 mgÁkg À1 dose to test MJN110 efficacy in ameliorating symptoms that are often associated with alcohol drinking or stress-related behaviours.
Previous studies have characterized the role of the eCB AEA/FAAH signalling system on alcohol-related behaviours, showing, for instance, that local injection of FAAH inhibitor URB597 into the CeA or the basolateral amygdala reduced alcohol self-administration in msP rats (Stopponi et al., 2018).Additionally, a single systemic administration of MAGL inhibitor MJN110 reduced selfadministration in alcohol-dependent rats and the MAGL inhibitor JZL184 reduced voluntary alcohol drinking in dependent mice (Serrano et al., 2018).Although we predicted that MJN110 would attenuate voluntary alcohol drinking and preference in the 2-BC procedure in the high alcohol-preferring msP rat model, we observed no effects of MJN110 on reducing 2-BC drinking in either msP or Wistar rats, even when we tested a higher dose (10 mgÁkg À1 ).This discrepant finding may be due to differences in drinking procedures (voluntary 2-BC drinking vs. operant self-administration), to the higher MJN110 dose (20 mgÁkg À1 ) and, most importantly, to previous alcohol history (dependent vs. non-dependent).Consistent with Serrano and coauthors, none of the doses tested reduced alcohol intake in control, non-dependent rats.
Stress disrupts 2-AG signalling, with acute and repeated stress exposure causing a delayed increase of 2-AG levels in the PFC and amygdala, which contributes to stress response termination (Morena et al., 2016;Sumislawski et al., 2011).Moreover, chronic stress is generally associated with a down-regulation of CB 1 receptors in most brain regions except for the mPFC (McLaughlin et al., 2014).We assessed anxiety-like behaviours using the NIH procedure under stressful-perceived novel environmental conditions in male and female Wistar and msP rats treated with a single dose of MJN110 (Burston et al., 2016).We found no significant effect of the treatment on the latency to approach or to start eating the familiar palatable food.However, the increased pellet intake elicited by MJN110 in female rats exposed to a novel, stressful environment is consistent with the hypothesis that inhibition of MAGL might alleviate anxiogenic-like behaviour selectively in females, possibly by acting on the maintenance of the feeding behaviour under anxiogenic-like conditions rather than on its initiation.We also confirmed that appetite or feeding per se was not playing a role (Feja et al., 2020) by testing a separate group of female rats in their own familiar home cage (Figure S2).
The mPFC is involved in executive functions, such as the ability to attend to or ignore stimuli, and cognitive and behavioural flexibility.
Structural and functional abnormalities of the mPFC have been linked to aggressive disorders (Miczek et al., 2022); thus, we examined the effects of MJN110 on irritable-like behaviours.We treated male and female Wistar and msP rats with a single dose of MJN110 and assessed the efficacy of pharmacological MAGL inhibition on modulating the number of aggressive and defensive behaviours in the bottle-brush test.We have previously used the bottle-brush test (Cruz et al., 2022;Kimbrough et al., 2017;Kirson et al., 2021;Somkuwar et al., 2017;Spierling et al., 2020;Steinman et al., 2021) to assess withdrawal or PTSD irritability-like behaviours.As expected, the compound did not have any effect in Wistar controls, but it did exert a significant effect in msP rats.Interestingly, we identified a clear sex difference in the eCB regulation of stress-coping strategies in msP rats.In female msP rats, MJN110 reduced aggressive-like behaviours, which can be classified as a more active-coping behaviour.
In male msP rats, however, MJN110 increased defensive behaviours, which is indicative of increased passive-coping behaviours.
Importantly, we observed that 5 mgÁkg À1 MJN110 did not reduce general locomotor activity, which serves as an indication that general active behaviours were not negatively impacted (Table S1).Our findings in male rats are in line with recent work (Pavon et al., 2021) showing that MAGL inhibition induces different changes in stresscoping behaviours compared with FAAH inhibition and increases passive-coping behaviours in male mice (Pavon et al., 2021).Therefore, our irritability assessments with the bottle-brush test build up upon previous findings and expand our understanding of sexdependent mechanisms associated with eCB-mediated stress responses.
Lastly, we examined the effect of MAGL inhibition on cueinduced fear expression.A separate cohort of male and female Wistar and msP rats was fear conditioned in response to six light cue stimuli (CS) paired with 0.6-mA electric shocks (US) across 2 consecutive days.On the third day, rats were tested for conditioned fear expression after a single systemic administration of MJN110.Our experimental design allowed us to observe a significant fear conditioning in msP versus Wistar rats of both sexes.Importantly, the freezing behaviour and the absence of any movements should not be attributed to locomotor activity differences between msP and Wistar rats (Cannella et al., 2016).We choose relatively mild conditions for our protocol (0.6-mA shocks) because msP rats are particularly vulnerable to stress.
On the conditioned fear acquisition Day 2, msP females showed a significantly higher freezing percentage versus Wistars across the 6 CS + US, reaching 70% freezing by CS-US 6.Similar behaviour was observed in males, with the conditioned freezing reaching 60%.Notably, on conditioned fear expression test day (Day 3), we found a significant effect of MJN110 in reducing freezing response during the pre-conditioning period, selectively in male msP rats.We did observe a similar trend in females, although not statistically significant.Interestingly, acute 2-AG elevation seemed to affect only the pre-CS expression of fear but not the contingent, cue-induced expression of fear (i.e., light CS presentation), indicating a selective role in the regulation of contextual conditioning or possibly non-associative fear versus cue-induced expression of fear memory.Notably, while the MAGL inhibitor JZL184 induces hypomotility and decreases mobility time in male mice, MJN110 does not affect the overall time spent mobile (Ignatowska-Jankowska et al., 2015).Our results partially align with recent studies (Morena et al., 2021) showing that MJN110 administered prior to fear extinction tests did not reduce cued fear memory expression and extinction in male rats but acutely reduced freezing at the last CS presentations.Interestingly, in the same paradigm, MJN110 affected darting behaviour in females, a sign of increased active-coping behaviour.Overall, this evidence points to the hypothesis that increased 2-AG signalling promoted active over passive fear and stress-coping responses in female rats.This shift from passive to active forms of acute coping might be explained by the established role of CB 1 receptors on modulating glutamatergic neurons (Metna-Laurent et al., 2012), suggesting that, in females, elevated 2-AG signalling may preferentially engage this signalling to promote this behavioural transition.Our results, however, are opposed to another study in a male mice line of high anxiety where inhibition of 2-AG hydrolysis increased active, but not passive, responses (Heinz et al., 2017).
We acknowledge some limitations of our studies that will be addressed in future work.First, female rats were left to cycle freely, and the oestrous cycle was not monitored at the time of testing.
Given that stress-coping responses and freezing behaviour during fear expression vary with oestrous phases (Gruene et al., 2015;Wiersielis et al., 2016) and oestrous cycle modulates CB 1 receptor density and affinity (De Fonseca et al., 1994), as well as AEA and 2-AG levels across different brain regions (Bradshaw et al., 2006;Gonzalez et al., 2000), future studies will address this point.Additionally, only acute effects of MAGL inhibition are reported herein, but the overall effects of the elevation of eCBs are determined by cannabinoid receptor expression and post-receptor signalling, which were not assessed in the present study.Future studies that involve repeated or chronic administration will be translationally informative because the longterm effects of sustained eCB elevation also will depend on potential downstream changes in CB 1 expression and signalling cascades.Second, we used only one dose of MJN110 in most of our behavioural studies (anxiety, irritability and conditioned fear expression).Future studies will evaluate MJN110 across a wider range of doses, which might uncover dose-dependent effects and more genotypeand sex-dependent differences.Also, we will perform brain region site-specific administrations of MJN110 to examine potential interactions between local 2-AG elevations and other co-factors (e.g., corticosterone/glucocorticoid receptor signalling) that would help tease apart and identify target regions or circuits involved in our behavioural assessments.Lastly, our conditioned fear experiment design did not allow us to discriminate whether the nature of increased pre-CS freezing we observed in the msP rats (and its reversal by MJN110) is contextual or non-associative learning.

| CONCLUSIONS
Our results showed that pharmacological enhancement of 2-AG levels did not reduce voluntary alcohol drinking in the 2-BC procedure.Importantly, MAGL inhibition by MJN110 reduced irritabilitylike behaviours, with potential opposing coping strategies between males and females, and ameliorated contextual or non-associative fear expression but without affecting cue-conditioned fear expression.These effects were selectively observed in msP rats, a genetic rat model of innate and inheritable vulnerability to stress.Altogether, our findings suggest that MAGL inhibition is efficacious in attenuating some stress-related maladaptive behaviours, pointing to a selective role of 2-AG in specific aspect of stress symptoms, but not 0.4 mlÁmin À1 ; 5.01-20 min, 60%-100% Buffer B, 0.4 mlÁmin À1 ; 20-20.01 min, 100%-0% Buffer B, 0.5 mlÁmin À1 ; 20.01-33 min, 100% Buffer B, 0.5 mlÁmin À1 ; 33-34 min, 100%-0% Buffer B, 0.5 mlÁmin À1 ; and 34-35 min, 0% Buffer B, 0.5 mlÁmin À1 .Eluted lipids were detected using the Agilent 6470 Triple Quadrupole Liquid Chromatography Mass Spectrometer System (Agilent, RRID:SCR_019421) via multiple reaction monitoring (MRM) using an electrospray ionization (ESI) source in positive mode.MS analysis was performed using ESI with the following parameters: gas temperature, 350 C; gas flow, 9 LÁmin À1 ; nebulizer, 35 psi; sheath gas temperature, 375 C; sheath gas flow, 11 LÁmin À1 ; capillary, 4000 V; and nozzle voltage/charging, 500 V. MRM transitions were specific for each lipid.2-AG precursor ion: 379 m/z; product ion: 287 m/z; dwell: 100; fragmentation (F): 100 V; collision (C): 8 V; collision acceleration (CA): 7 V. 2-AG-d5 precursor ion: 384 m/z; product ion: 287 m/z; dwell: 100; F: 100 V; C: 8 V; CA: 7 V. AEA precursor ion: 348.3 m/z; product ion: 62.1 m/z; dwell: 100; F: 100 V; C: 40 V; CA: 7 V. AEA-d4 precursor ion: 352.3 m/z; product ion: 66.1 m/z; dwell: 100; F: 100 V; C: 40 V; CA: 7 V. Lipids were quantified by Agilent MassHunter Quantitative Analysis software, Version 10 (Agilent, RRID: SCR_015040), by integrating their peak area and normalizing relative to the peak area of the internal standard, as follows: pmol metabolite ¼ area of metabolite peak area of internal standard peak pmol internal standard ð Þ phases of the conditioned fear expression sessions, we divided the 18 CS into 3 blocks of 6 consecutive CS each (Block 1: CS 1-6; Block 2: CS 7-12; Block 3: CS 13-18).We scored the time spent freezing per each block.Data are expressed as a percentage of time spent freezing over the total duration of the block.Because the main goal of our study was to investigate MJN110 effects in a genetic rat model of high alcohol preference and high anxiety, and because of the underlying sexual dimorphism of the eCB system (for review, see L opez, 2010, andCraft et al., 2013), male and female rats were herein analysed separately.Studies were designed to generate groups of comparable size, using randomization and blinded analysis.Group size, indicated for each experiment in the figure legends, is the number of independent values (i.e., individual rats) and the statistical analysis was done using these independent values.Sample sizes are based on experience from our laboratory and the Cravatt laboratory(Niphakis et al., 2013) for similar studies.All datasets were derived from ≥7 rats (behavioural experiments) or 3-5 rats (biochemical experiment;Niphakis et al., 2013) from each sex and genotype.While processing the samples for lipid quantification, two MJN110-treated female Wistar CeA samples and one MJN110-treated female msP CeA sample were damaged, reducing the number of final quantifiable samples to three and four, respectively.All data are presented as mean ± standard error of the mean (SEM).The significance level was determined at P < 0.05.Outliers in the lipid quantification were determined using Grubbs' outlier test and excluded.Two female msP rats per group were excluded in the 2-BC procedure due to bottle leakage on the test day.
(3 h) we selected were sufficient to achieve significant elevation of 2-AG in brain regions involved in the regulation of drinking behaviours and stress response: mPFC (PrL and IL subregions) and CeA.The i.p. route of administration was chosen here to reduce stress confounds of oral gavage administration, relevant with the stresssensitive msP rats, and to facilitate comparison with previous behavioural work, most of which used i.p. administration.Because MJN110 is reportedly orally active, future translationally oriented work will compare effects of oral administration.mPFC regulates emotions and plays an important role in the processing of and behavioural responses to stressful or threatening stimuli.The amygdala regulates physiological and behavioural responses to stress (Gilpin voluntary drinking behaviours.These data open new avenues of investigation of MAGL inhibitors as potential therapies for a subset of patients suffering from AUD and stress-related disorders.Results from this study may inform future research aiming at investigating eCB regulation of stress-dependent disorders within specific brain regions and neuronal circuits and with a focus on sex-specific therapeutic approaches.AUTHOR CONTRIBUTIONS Valentina Vozella: Conceptualization (lead); data curation (lead); formal analysis (lead); investigation (lead); methodology (lead); validation (lead); visualization (lead); writing-original draft (lead); writingreview and editing (lead).Bryan Cruz: Data curation (supporting); formal analysis (supporting); investigation (supporting); methodology (supporting); writing-review and editing (supporting).Hannah C. Feldman: Data curation (supporting); formal analysis (supporting); investigation (supporting); methodology (supporting); writing-review and editing (supporting).Ryan Bullard: Data curation (supporting); writing-review and editing (supporting).Paula C. Bianchi: Investigation (supporting); writing-review and editing (supporting).Luis A. Natividad: Investigation (supporting); writing-review and editing (supporting).Benjamin F. Cravatt: Resources (supporting); supervision (supporting); writing-review and editing (supporting).Eric P. Zorrilla: Conceptualization (supporting); funding acquisition (lead); investigation (supporting); resources (supporting); supervision (supporting); writing-review and editing (supporting).Roberto Ciccocioppo: Conceptualization (supporting); funding acquisition (lead); investigation (supporting); resources (supporting); supervision (supporting); writing-review and editing (supporting).Marisa Roberto: Conceptualization (equal); data curation (supporting); funding acquisition (lead); investigation (supporting); project administration (lead); resources (lead); supervision (lead); writing-original draft (supporting); writingreview and editing (equal).ACKNOWLEDGEMENTS This is Manuscript Number 30227 from The Scripps Research Institute.This study was supported by the National Institutes of Health (National Institute on Alcohol Abuse and Alcoholism) grants: AA030609 (VV), AA017447 (MR and RC), AA027700 (MR and EZ), AA021491 (MR), AA006420 (MR and EZ), AA029841 (MR), AA013498 (MR) and AA025393 (LAN); the Schimmel Family Chair (MR); The Pearson Center for Alcoholism and Addiction Research; and the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) postdoctoral fellow 2021/12978-1 (PCB).We thank Dr. Daisuke Ogasawara for synthesizing MJN110 and for technical assistance with the compound.The conducted research was not preregistered with an analysis plan in an independent, institutional registry.agencies, publishers and other organizations engaged with supporting research.ORCID Roberto Ciccocioppo https://orcid.org/0000-0003-3126-9240