Abstract
There are five cloned muscarinic acetylcholine receptors (M1–M5). Of these, the muscarinic type 5 receptor (M5) is the only one localized to dopamine neurons in the ventral tegmental area and substantia nigra. Unlike M1–M4, the M5 receptor has relatively restricted expression in the brain, making it an attractive therapeutic target. Here, we performed an in-depth characterization of M5-dependent potentiation of dopamine transmission in the nucleus accumbens and accompanying exploratory behaviors in male and female mice. We show that M5 receptors potentiate dopamine transmission by acting directly on the terminals within the nucleus accumbens. Using the muscarinic agonist oxotremorine, we revealed a unique concentration–response curve and a sensitivity to repeated forced swim stress or restraint stress exposure. We found that constitutive deletion of M5 receptors reduced exploration of the center of an open field while at the same time impairing normal habituation only in male mice. In addition, M5 deletion reduced exploration of salient stimuli, especially under conditions of high novelty, yet had no effect on hedonia assayed using the sucrose preference test or on stress-coping strategy assayed using the forced swim test. We conclude that M5 receptors are critical for both engaging with the environment and updating behavioral output in response to environment cues, specifically in male mice. A cardinal feature of mood and anxiety disorders is withdrawal from the environment. These data indicate that boosting M5 receptor activity may be a useful therapeutic target for ameliorating these symptoms of depression and anxiety.
SIGNIFICANCE STATEMENT The basic physiological and behavioral functions of the muscarinic M5 receptor remain understudied. Furthermore, its presence on dopamine neurons, relatively restricted expression in the brain, and recent crystallization make it an attractive target for therapeutic development. Yet, most preclinical studies of M5 receptor function have primarily focused on substance use disorders in male rodents. Here, we characterized the role of M5 receptors in potentiating dopamine transmission in the nucleus accumbens, finding impaired functioning after stress exposure. Furthermore, we show that M5 receptors can modulate exploratory behavior in a sex-specific manner, without affecting hedonic behavior. These findings further illustrate the therapeutic potential of the M5 receptor, warranting further research in the context of treating mood disorders.
Introduction
Disruption of both dopamine and acetylcholine transmission has been implicated in preclinical models of depression and anxiety as well as in individuals with mood and anxiety disorders (Nutt et al., 1998; Higley and Picciotto, 2014; Small et al., 2016; Belujon and Grace, 2017). In rodents, genetic, viral, or chemogenetic manipulations that reduce the firing of cholinergic interneurons in the nucleus accumbens (NAc) produce depression-like behavior (Warner-Schmidt et al., 2012; Hanada et al., 2018; Cheng et al., 2019). Likewise, manipulations that have inhibited or disrupted normal dopamine transmission in the NAc produce similar depression-like behavior (Tye et al., 2013; LeBlanc et al., 2020). In contrast, a significant number of studies has also linked elevations in accumbal acetylcholine with depression-like behavior. Likewise, it has been shown that muscarinic antagonists reversed or attenuated the effects of chronic stress (Chau et al., 2001; Chen et al., 2012; Navarria et al., 2015). The latter has been specifically linked to activation of M1 receptors. Indeed, one explanation for these disparate findings is the diversity of both nicotinic and muscarinic receptors, highlighting the need for receptor-specific manipulations using gene deletion and receptor-specific pharmacology.
Acetylcholine and dopamine have a reciprocal modulatory relationship that has been studied widely in the literature. It has been shown that dopamine can induce transient inhibition or pauses in striatal cholinergic interneuron firing via activation of dopamine D2 receptors localized to cholinergic interneurons (Ding et al., 2010; Chuhma et al., 2014; Helseth et al., 2021). Acetylcholine regulation of dopamine signaling is more complex. Focusing on the NAc, it has been demonstrated that synchronous release of acetylcholine via electrical or optogenetic stimulation of cholinergic interneurons can trigger the release of dopamine by activating β2-containing nicotinic acetylcholine receptors (nACh-Rs) on en passante synaptic boutons from dopamine axons within the striatum (Zhou et al., 2001; Rice and Cragg, 2004; Cachope et al., 2012; Threlfell et al., 2012). Acetylcholine also modulates dopamine release and uptake through activation of both Gi/o-coupled and Gq-coupled muscarinic receptors. Activation of M2/M4 autoreceptors can suppress dopamine transmission, likely by suppressing both acetylcholine release and the amount of β2-nACh-R activation on dopamine terminals (Threlfell et al., 2010, 2012).
Strong evidence points to the role of muscarinic Gq-coupled M5 receptors in potentiating dopamine transmission in the ventral striatum (Yamada et al., 2003; Shin et al., 2015). The mechanisms underlying this potentiation are not fully understood but likely involve both an increase in the release probability of dopamine as well as a suppression of dopamine reuptake as M5 activation results in a slowing of the decay kinetics of dopamine transients (Shin et al., 2015; Underhill and Amara, 2021). Endogenous release of acetylcholine has been shown to potentiate dopamine transmission via these mechanisms because ambinomium, an inhibitor of acetylcholinesterase, mimics the potentiation of dopamine transmission observed by muscarinic receptor agonists (Shin et al., 2015). A previous study further demonstrated that the stress-associated neuropeptide corticotropin-releasing factor (CRF) increases cholinergic interneuron firing, which in turn potentiates dopamine transmission via activation of M5 receptors, again indicating that endogenous acetylcholine can also activate M5 receptors to potentiate dopamine transmission (Lemos et al., 2019).
Reduction or ablation of M5 receptor function leads to reduced stimulant-induced psychomotor activation, drug seeking, and drug taking (Basile et al., 2002; Fink-Jensen et al., 2003; Thomsen et al., 2005; Gunter et al., 2018; Gould et al., 2019). This appears to be the case for psychostimulants, alcohol, and opioids, revealing a broad potential for these muscarinic receptors as therapeutic targets for substance use disorders. Furthermore, suppression of M5 function in the ventral tegmental area (VTA) disrupts food reward reinforcement behavior (Yeomans et al., 2000; Wang et al., 2008). These findings point to a potential role for disruption of M5 receptor signaling in depression-like or anxiety-like states, which far fewer studies have investigated. In this study, we examined the impact of M5 receptor deletion on exploratory and hedonic behaviors, two behaviors classically used to assess the expression of anxiety-like or depression-like behaviors in male and female mice.
Here, we examined the M5-dependent potentiation of dopamine transmission within the NAc that takes place independent of the actions of M5 at the somata of dopamine neurons. We examined M5 function in both male and female mice and across a range of concentrations of the nonselective muscarinic agonist oxotremorine (OXO-M) applied in the presence of the nicotinic antagonist dihydro-β-erythroidine (DhβE) to produce conditions that favor M5 activation. Maintaining slices in DhβE effectively removes the cholinergic component of the composite dopamine transient, thereby isolating the component of the dopamine transient that is driven by direct dopamine fiber stimulation. By removing the cholinergic component of the composite dopamine transient, we are negating the impact of M2/4 autoreceptors on dopamine transmission. This is necessary as no selective M5 receptor agonists are currently available, forcing us to use a nonselective muscarinic receptor agonist. Our findings show that M5-dependent potentiation of dopamine transmission in the NAc is similar between males and females. Interestingly, the concentration dependence of the M5 response was atypical in that increasing concentrations of OXO-M increased the rate at which a maximal effect was achieved but not the size of the maximal effect. Moreover, increasing concentrations of OXO-M produced a secondary inhibition that may be because of recruitment of other receptors or desensitization of the M5 receptor. Behaviorally, M5 deletion disrupts normal exploratory behavior in a context-specific fashion without having an impact on hedonia. This study points to a novel and critical function of M5 receptor activation, which we believe will contribute to development of M5 receptor-targeted therapeutics.
Materials and Methods
Animals
All procedures were performed in accordance with guidelines from the Institutional Animal Care and Use Committee at the University of Minnesota. Male and female mice (postnatal day (P)60–180) were group housed and kept under a 12:12 h light/dark cycle (6:00 A.M. on/18:00 P.M. off) with food and water available ad libitum. Breeding was done under a summer photoperiod (14:10 hr light/dark cycle (6:00 A.M. on/20:00 P.M off)), and following weaning, animals were transferred to our Investigator Managed Housing Area and kept under a 12:12 h light/dark cycle for at least a week before use. The M5−/− (knock-out) mouse line was a gift from Jürgen Wess. M5 wild-type (WT, M5+/+), heterozygous (HET, M5+/−), and knock-out (KO, M5−/−) littermates were generated using a M5 HET crossed with M5 HET breeding scheme that produced ∼25% WT, 50% HET, and 25% KO. The M5 HET crossed with M5 HET (C57BL/6J background) breeding strategy was used to ensure uniform maternal care during development. However, a caveat to this strategy is that there was a non-uniform number of WT, HET, and KO mice in each cage. We found that this disrupted normal social interaction behavior, and therefore we excluded social interaction experiments from our study. The low yield of WT and KO mice also made it prohibitive to parse estrous-cycle-dependent effects, although this is of interest for future studies.
For neuroanatomy and behavior experiments, we used M5 WT littermates only. For voltammetry experiments, it was not possible to exclusively use M5 WT littermates given the small percentage produced. Thus, we combined M5 WT littermates with C57BL/6J mice bred in house under identical conditions. We found no difference between these groups and therefore pooled the data.
Genotyping was outsourced to Transnetyx, an automated genotyping service. Original development of the M5 KO mice was completed through insertion of a neomycin-resistance cassette (NRC) replacing the start of the Chrm5 gene, first 750 base pairs (bp) that comprise the first 250 amino acids including the translation start site (Yamada et al., 2001). Mice were classified as WT if they were positive for Chrm5 and negative for NRC, HET if mice were positive for both Chrm5 and NRC, and KO if mice were negative for Chrm5 and positive for NRC (Chrm5, forward CCATCACAAGACCACTGACATACC, reverse GCCATGCCAAGCCGATCA; NRC, forward GGGCGCCCGGTTCTT, reverse CCTCGTCCTGCAGTTCATTCA). Genotyping was performed on tail-tip samples collected on the day of weaning (P21).
Fluorescent in situ hybridization using RNAscope
Brains were rapidly dissected and flash frozen in isopentane on dry ice. Brains were kept in a −80°C freezer until they were sectioned. Coronal sections (16 µm) containing the NAc or VTA were thaw mounted onto Superfrost plus slides (Electron Microscopy Sciences) using a Leica CM 1860 cryostat maintained at −20°C. Before sectioning, brains were equilibrated in the cryostat for at least 2 h (overnight is optimal). Slides were cleaned with RNAseZap RNase Decontamination Solution to prevent mRNA degradation. Slides were stored at −80°C. RNAscope in situ hybridization (ISH) was conducted according to the Advanced Cell Diagnostics (ACD) user manual. Briefly, slides were fixed in 10% neutral buffered formalin for 20 min at 4°C. Slides were washed 2 × 1 min with 1× PBS, then dehydrated with 50% ethanol (1 × 5 min), 70% ethanol (1 × 5 min), and 100% ethanol (2 × 5 min). Slides were incubated in 100% ethanol at −20°C overnight. The following day, slides were dried at room temperature (RT) for 10 min. A hydrophobic barrier was drawn around the sections using a hydrophobic pen and allowed to dry for 15 min at RT. Sections were then incubated with Protease IV pretreat solution for 20 min at RT. Slides were washed with ddH2O (2 × 1 min) before being incubated with the appropriate probes for 2 h at 40°C in the HybEZ oven (ACD). Our preliminary data demonstrated that the commercially available ACD probe for Chrm5 (catalog #495301) had an unacceptable amount of signal in the M5 KO mouse. Thus, we collaborated with ACD to design a custom Chrm5 probe directed at the first 750 bp of the Chrm5 gene, which was replaced with an NRC in the M5 KO (Mm-Chrm5-O2, catalog #1052411, ACD). This probe was multiplexed with the probe directed at tyrosine hydroxylase (Th; catalog #317621, ACD). Following incubation with the appropriate probes, slides were subjected to a series of amplification steps at 40°C in the HybEZ oven with 2 × 2 min washes (with agitation) in between each amplification step at RT. Amplification steps were conducted as follows: Amp 1 at 40°C for 30 min, Amp 2 at 40°C for 15 min, Amp 3 at 40°C for 30 min, Amp 4-Alt B at 40°C for 15 min. A DAPI-containing solution was applied to sections (one slide at a time) at RT for 20 s. Finally, slides were coverslipped using ProLong Gold Antifade mounting media (Invitrogen) and stored at 4°C until imaging.
Image analysis and quantification for RNAscope
Sections were imaged using a Keyence BZ-X710 epifluorescent microscope and corresponding software. Unique 20× images of the VTA were acquired from M5 WT male and female littermates using the same software and hardware settings. The settings were titrated for each specific experimental probe. Quantification was done using Fiji/ImageJ software and a pipeline using MATLAB and CellProfiler software. In FIJI/ImageJ, numbers of DAPI and TH+ cells were automatically generated using the particle counter function in ImageJ. A TH+ cytosolic mask was generated by subtracting a DAPI+ mask from the original TH+ mask. This TH+ cytosolic mask was then used to assess M5/TH coexpression within the VTA and puncta number. Thresholding was kept consistent across images. In the pipeline, an in-house MATLAB code was used to convert all images to greyscale to show raw intensity values and allowed for the selection and application of a freehand region of interest (ROI) to overlay images of each image set and their corresponding DAPI+, TH+, and M5+ channels. All pixel values outside each unique ROI were transformed to zero values using a binary mask function to exclude fluorescence outside the VTA. The image sets were passed through an in-house CellProfiler pipeline to detect the amount of TH+ and M5+ puncta found in generated, approximated cell outlines created by correlating the objects in the DAPI+ image against its corresponding overlay image. Preliminary numbers of DAPI+, M5+, and TH+ cells and puncta from CellProfiler were processed further in an in-house MATLAB code. The code calculated the following for each channel in an image set: nonobject/background area, area/puncta, total area of puncta/cell, nonspecific binding rate (sum of puncta area/nonobject area), and probability of off-target binding for each puncta based on a binomial distribution. A cell was counted as M5+ and/or TH+ when the amount of puncta in each cell was found to be greater than the cell size times nonspecific binding rate/channel. A Bonferroni correction was applied to each p value of all counted cells in an image (p < 0.01/total number of cells). Total counts of M5+, TH+, and M5+/TH+ cells and percentage of each compared total cells were generated for each image set. All settings of object detection and thresholds were kept the same across all images.
Fast scan cyclic voltammetry
Coronal slices (240 µm) containing the NAc core were prepared from M5 WT, M5 HET, or M5 KO mice. To minimize potential subregion heterogeneity, we restricted our electrode placements to a few spots around the anterior commissure in coronal sections from bregma +1.2 to −0.86 mm. Slices were cut in ice-cold cutting solution containing the following (in mm): 225 sucrose, 13.9 NaCl, 26.2 NaHCO3, 1 NaH2PO4, 1.25 glucose, 2.5 KCl, 0.1 CaCl2, 4.9 MgCl2, and 3 kynurenic acid. Slices were maintained in oxygenated artificial CSF (ACSF) containing the following (in mm): 124 NaCl, 2.5 KCl, 2.5 CaCl2, 1.3 MgCl2, 26.2 NaHCO3, 1 NaH2PO4, and 20 glucose (∼310–315 mOsm) at RT following a 1 h recovery period at 33°C. Carbon fiber (7 µm diameter, Goodfellow) electrodes were fabricated with a glass capillary (catalog #602000, A-M Systems) using a Sutter P-97 puller, and fiber tips were hand cut to 100–150 µm past the capillary tip. Immediately before the experiments, they were filled with 3 m KCl internal solution. Recordings were conducted at 31–33°C, maintained with a Warner inline heating system. The carbon-fiber electrode was held at −0.4 V, and a voltage ramp to and from 1.3 V versus Ag/AgCl (400V/s) was delivered every 100 ms (10 Hz). Before recording, electrodes were conditioned by running the ramp at 60 Hz for 10 min and at 10 Hz for another 10 min. Dopamine transients were evoked by electrical stimulation delivered through a glass microelectrode filled with ACSF. For most experiments, a single monophasic pulse (4 ms, 300 µA) was delivered to the slice in the absence or presence of the nACh-R antagonist DhβE (1 µm) every 2 min. To investigate M5 effects on stimulation trains, we used the stimulation parameters used by Rice and Cragg (2004), separated by 3 min intervals, 1 pulse (p), 1 p, 5 p/5 Hz, 5 p/10 Hz, 5 p/25 Hz, 5 p/100 Hz. Data were acquired using a Dagan headstage and amplifier and National Instruments PCI boards. Data were acquired and analyzed using the Demon Voltammetry and Analysis software package (Yorgason et al., 2011). Experiments were rejected when the evoked current did not have the characteristic electrochemical signature of dopamine assessed by a current-voltage plot.
Behavior
Before the start of behavioral experiments, mice were handled and acclimated to the testing room for 5 d. Behavioral testing chambers for open field (OF), novel object (NO), or food exploration, light/dark box, and elevated zero maze experiments were contained in individual custom-built sound-attenuated chambers that were equipped with an overhead light (with dimmer switch) and a white noise fan. Mice were video monitored during behavioral testing, and data were acquired and analyzed using Noldus EthoVision (version 14 or 15) software., For OF, NO exploration, and novel palatable food (NF) exploration, animals were placed in a circular arena (50 cm diameter, 40 cm height) for 60 min on day 1. On day 2, mice were placed in the OF chamber and allowed to habituate for 60 min. Mice were then placed back in their home cage (HC) for 5 min. An NO (5 cm cylindrical bottle cap from a 1L Pyrex laboratory bottle) or a container holding NF (bacon-flavored food pellets covered with a fine mesh; catalog #F3580, BioServ) was placed in the center of the arena in a counterbalanced fashion. Mice were allowed to explore for 30 min. Mice were then taken out of the arena and placed back in their HC for 5 min. Subsequently, the alternate stimulus was placed in the center of the arena, and mice were allowed to explore for another 30 min. The elevated zero maze (EZM) was customized for mice and was purchased from Med Associates (64 cm diameter, 65 cm height). The EZM purchased came with hinged covers for the closed compartments, allowing closing the compartments completely or exposing only the top of the closed compartment. We used two separate conditions for our studies. In one condition, the EZM was illuminated by bright light (161 lux), and the closed compartments were fully enclosed. In the second condition, mice experienced the EZM under dim conditions (49 lux) with the tops of the closed compartments exposed. Mice were exposed to both conditions in a counterbalanced fashion and were exposed to each condition 7 d apart. We predicted that mice would modulate their behavior, specifically their time spent and entries into the open compartments, based on which condition they were exposed to. Each EZM session was 5 min long. In the sucrose preference test, mice were individually housed for these experiments. Mice were acclimated to the two-bottle home cage apparatus, in which both bottles contained water, for 3 d. On the day 4, one bottle contained 4% or 1% sucrose (different cohorts of animals) and one contained water. The side where the sucrose water was placed was counterbalanced. Sucrose and water bottles were weighed for 3 d, alternating sides each day, and the results were averaged. For the light-dark box, a separate cohort of mice that had no prior experience with other behavioral chambers was used for these experiments. Mice were placed in the dark compartment of the light-dark box with bright overhead illumination (161 lux; catalog #10-000-355, Stoelting). Activity and time allocation in the light or dark compartment was assessed over a 5 min duration of time. Mice were subsequently placed back in their HC for 5 min. Mice were then reintroduced to the testing chamber (starting in the dark side) with a novel palatable food item (bacon-flavored pellets) placed in the bright side. Activity and time allocation in the light or dark compartment was assessed over a 5 min duration of time.
Stressor exposure: repeated forced swim stress
On day 1, male or female mice were placed in a 4L bucket containing 30 ± 1°C water for 15 min. On days 2–7, mice were re-exposed to the 30 ± 1°C water for 5 min each day at different times of day. Male mice were killed for fast scan cyclic voltammetry (FSCV) experiments 24–72 h after the last swim exposure. chronic restraint stress: In the chronic restraint stress procedure, male mice were confined to a mouse restrainer (50 ml conical tube with air holes) for a period of 30 min every day at 10:00 A.M. for 2 weeks. For both paradigms, control mice were handled for the same number of days as the corresponding stressor exposure.
Statistics
G*Power (version 3.1) software was used to conduct power analyses based on known effect sizes from Shin et al. (2015) and prior behavioral studies from the Wess laboratory using the M5 KO mice. Statistical analysis was performed in Prism (GraphPad) and Microsoft Excel. Two-tailed unpaired t tests, two-tailed paired t tests, one-way ANOVAs, or two-way repeated measures ANOVAs were used when appropriate and are stated in the results. One-way and two-way ANOVAs were followed up with Dunnett's, Tukey's, or Sidak-corrected t test comparisons. All data are presented as mean ± SEM. Results were considered significant at an alpha of 0.05; asterisk (*) denotes t tests or post hoc t tests following one-way or two-way ANOVAs, number symbol (#) denotes significant interaction or second post hoc t test, and an ampersand (&) denotes two-way ANOVA, main effect, δ = trend 0.100 > p > 0.05; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.
Results
Chrm5 is ubiquitously expressed on Th+ neurons in the VTA
The expression of Chrm5 mRNA that encodes the M5 receptor protein was assessed. According to the Allen brain atlas in situ hybridization data, Chrm5 has a markedly limited expression compared with the other four cloned muscarinic acetylcholine receptors Chrm1-4 (https://alleninstitute.org; Fig. 1a; Lein et al., 2007). One region of concentrated Chrm5 mRNA expression is the ventral midbrain (Fig. 1a). To assess Chrm5 mRNA expression specifically on dopamine neurons, we used a custom RNAscope probe targeted to the first 750 bp of the Chrm5 gene where the neomycin-resistant cassette was inserted to generate the M5 KO mouse (see above, Materials and Methods). Using this probe, we confirmed that Chrm5 mRNAs are localized to dopamine neurons in the VTA and substantia nigra pars compacta by multiplexing with a probe targeted to Th mRNAs. Using an automated methodology, we determined that Chrm5 is localized to 84% of Th+ neurons in VTA (n = 9 mice; Fig. 1b,c). Given the ubiquity and density of Chrm5 expression on VTA neurons and the undetectable expression of Chrm5 in the NAc in the Allen brain atlas, we initially concluded that the vast majority of Chrm5 in the NAc is expressed on dopamine terminals. However, given the increased sensitivity of RNAscope in situ hybridization, we also assessed the localization of Chrm5 to NAc cells. Analysis of Chrm5 localization to NAc cells demonstrated that only 11 ± 1% of NAc cells expressed mRNA for the M5 receptor (n = 12 mice), confirming our speculation that the predominant source of M5 receptor function in the NAc is at the dopamine terminal. Comparing males and females, we found a significant difference in Chrm5 mRNA puncta in putative dopamine neurons that was isolated using a Th mask (with nucleus excluded) in ImageJ software (males, 593 ± 40; females, 756 ± 42 particles, t test, t(7) = 2.499, p = 0.041, n = 6,3 mice, respectively). We speculated that the higher expression of Chrm5 mRNA in females may manifest functional differences, which motivated the subsequent analysis. We began by confirming and validating previous reports that M5 receptor activation potentiates dopamine transmission in the NAc.
Muscarinic M5 receptors are localized to midbrain dopamine neurons. a, Composite images of brainwide mRNA expression data for Chrm1-5 (M1–M5) based on in situ hybridization data collected by the Allen Institute and available from https://alleninstitute.org. b, c, Low-power (b) and high-power (c) fluorescent images of Th (green) and Chrm5 (red) mRNA expressed in the VTA. Quantification of Chrm5/Th coexpression in the VTA of male and female M5 WT mice, 22,177 Th+ cells counted from nine animals (6 males, 3 females).
M5-dependent potentiation of dopamine transmission
The reciprocal interaction between cholinergic and dopaminergic systems within the striatum is complex. Electrical stimulation within the NAc core produces a composite dopamine transient, a summation of direct dopamine fiber stimulation and indirect dopamine release. Indirect dopamine sources arise from stimulation of cholinergic interneurons and acetylcholine release that mediate dopamine transmission via activation of β2-containing nACh-Rs on dopamine varicosities (Fig. 2a). In the presence of DhβE, electrically evoked dopamine transients are derived solely from excitation of dopamine fibers and therefore are not affected by activity of M2/M4 autoreceptors. Under these conditions, the nonselective agonist OXO-M potentiates the peak amplitude of the dopamine transient in an M5-dependent fashion (Fig. 2b; Shin et al., 2015). In the current study, we replicated and extended the findings of Shin et al. (2015) by including M5 HET in our analysis and assessing the temporal kinetics of this response across several OXO-M concentrations in both male and female mice.
Muscarinic M5 receptor activation potentiates dopamine transmission in the NAc core. a, Schematic diagram depicting cholinergic and dopaminergic interactions that are engaged during the voltammetry paradigm using electrical stimulation within the NAc. b, Representative current times time and 3D FSCV color plots for electrically evoked dopamine transients recorded in M5 WT (a) and M5 KO (b) mice before and following OXO-M (10 μm) in slices maintained in DhβE (1 μm). c, Time course of normalized peak amplitude of electrically evoked dopamine transients recorded from ex vivo NAc slices taken from M5 WT, HET, and KO male mice in which OXO-M (0.5 μm), followed by scopolamine, is bath applied to the slice. d, Average OXO-M (0.5 μm) effect (average of last 5 time points) for M5 WT, HET, and KO. e, Time course of the normalized peak amplitude of electrically evoked dopamine transients recorded from ex vivo NAc slices taken from M5 WT, HET, and KO male mice in which OXO-M (10 μm), followed by scopolamine, was bath applied to the slice. f, Average OXO-M (10 μm) effect (average of last 5 time points) for M5 WT, HET, and KO. g, Average OXO-M effect for M5 WT, HET, and KO mice in the first (early, left) and last (late, right) 4 min/2 time points of the OXO-M bath application. h, Time course of the normalized peak amplitude (to DhβE baseline) of electrically evoked dopamine transients from M5 WT in which scopolamine was bath applied to the slice and maintained, followed by OXO-M (10 μm). i, Average scopolamine effect or SCOP+OXO-M (10 μm) effect (average of the last 5 time points) relative to DhβE baseline (black and gray bars) or average SCOP+OXO-M (10 μm) effect relative to scopolamine baseline (white bar). j, Representative current times time electrically evoked dopamine transients recorded in M5 WT before and following OXO-M (0.5 μm) application (black and blue lines, respectively. Top transients were evoked by 1 p stimulation, bottom traces were evoked by 5 p, 25 Hz stimulation. k, Peak dopamine transient amplitude activity-dependence curves (normalized to 1 p baseline) at baseline and following OXO-M (0.5 μm) application. l, OXO-M (0.5 μm) effect as a percentage of baseline for each stimulation parameter. m, Time course of the normalized peak amplitude of electrically evoked dopamine transients before and following bath application of DhβE (1 μm) in NAc slices taken from M5 WT and KO mice. n, Time course of the normalized peak amplitude of electrically evoked dopamine transients before and following bath application of OXO-M (10 μm) in NAc slices taken from M5 WT and KO mice with no DhβE onboard. Asterisk (*) outside bars denote post hoc t test, asterisk within bars denote one-sample t test of actual mean versus 100%.
OXO-M (0.5 µm) produced a significant and sustained increase in the peak amplitude of the electrically evoked dopamine transient in the presence of DhβE in male and female M5 WT mice (123.4 ± 7.7% of baseline, one-sample t test vs 100%, p = 0.01, n = 13 slices, Fig. 2c,d). The potentiation was fully reversed by the nonselective muscarinic antagonist scopolamine (1 µm), confirming that the effect is muscarinic receptor dependent. In M5 KO mice, OXO-M (0.5 µm) had no effect on the peak dopamine amplitude, confirming that the OXO-M-induced potentiation is because of activation of M5 receptors (94.1 ± 5.2% of baseline, one-sample t test vs 100%, p = 0.29, n = 11 slices; Fig. 2c,d). OXO-M (0.5 µm) had a slight effect on peak dopamine amplitude in M5 HET mice, indicating a predicted intermediate gene dose effect, although this was not significant (108.9 ± 5.4% of baseline, one-sample t test vs 100%, p = 0.13, n = 13 slices; two-way repeated measures ANOVA M5 WT, HET, KO, time times genotype interaction, F(48,814) = 2.564, p < 0.0001, Fig. 2c; one-way ANOVA, F(2,34) = 5.118, p = 0.01; post hoc Tukey's t test, WT vs KO, p = 0.0078, Fig. 2d).
In agreement with previous findings, OXO-M (10 µm) also produced a significant potentiation of the peak amplitude of the electrically evoked dopamine transient in the presence of DhβE in male WT mice. The potentiation was also fully reversed by the nonselective muscarinic antagonist scopolamine (1 µm). Interestingly, we found that this concentration of OXO-M produced a rapid increase in dopamine transmission to 126 ± 4% of baseline within the first 5 min. However, this effect was transient and significantly attenuated over time to a steady-state amplitude of 111 ± 5% of baseline. Thus, there was a difference in the early versus late phase of the OXO-M effect in male WT mice (paired t test, t(40) = 2.433, p = 0.0195, n = 21, slices; Fig. 2e–g). Interestingly, OXO-M-induced potentiation of dopamine transmission at this concentration was not different between M5 HET and WT mice (both 126 ± 4% of baseline in early phase). However, unlike WT mice, M5 HET mice displayed no significant attenuation of the OXO-M response over time (M5 HET, early phase, 126 ± 4% of baseline; late phase, 123 ± 8% of baseline; paired t test, t(12) = 0.5061, p = 0.6219, n = 13 slices; Fig. 2e–g). Again, we confirmed that this OXO-M-induced potentiation of dopamine transmission required M5 receptor activation using M5 KO littermates. Indeed, the OXO-M-mediated potentiation was absent in M5 KO mice. However, this higher concentration of OXO-M produced a small but significant inhibition of the peak dopamine transient during OXO-M application. The inhibition was evident in the early phase and remained stable through the late phase, confirming that OXO-M potentiation of dopamine transmission was dependent on M5 receptors under these conditions (one-way ANOVA, M5 WT, HET, KO early phase, F(2,44) = 34.52, p < 0.0001; post hoc Tukey's t tests, p < 0.0001, late phase, F(2,45) = 8.623, p = 0.007, post hoc Tukey's t tests, M5 KO vs M5 WT, p = 0.0135; M5 KO vs M5 HET, p = 0.0005, n = 13–21 slices; Fig. 2e–g).
The previous set of experiments used scopolamine at the end of the experiment to confirm that the effects were dependent on muscarinic receptors. We also wanted to confirm that scopolamine could completely prevent the OXO-M-induced potentiation of dopamine transmission at both early and late time points. Further, we wondered whether scopolamine preapplication would reveal any endogenous acetylcholine tone. After normalizing the dopamine peak amplitude to the pre-scopolamine baseline, we found a slight nonsignificant decrement in the dopamine transient peak amplitude with no further effect following OXO-M application (one-sample t test, scopolamine (SCOP) vs 100%, p = 0.1908, n = 7; SCOP vs OXO-M, paired t test, p = 0.09; Fig. 2h,i). These data indicate that although there may be a small modulatory acetylcholine tone, it is either not particularly robust or not well detected by these methods. OXO-M time points normalized to the SCOP baseline also demonstrated that scopolamine pretreatment completely blocked the OXO-M-induced potentiation of dopamine transmission (one-sample t test, SCOP plus OXO-M vs 100%, p = 0.1755, n = 7 slices).
Single-pulse stimulation is thought to recapitulate tonic dopamine transmission, whereas stimulation trains are thought to emulate phasic or burst firing by dopamine neurons. Work from the Cragg laboratory and others has demonstrated that both nicotinic and muscarinic receptor modulation are stimulation dependent. Blockade of nicotine receptors reduces the amplitude of single-pulse stimulated dopamine transients yet potentiates high-frequency stimulation of dopamine release (Rice and Cragg, 2004; Threlfell et al., 2010). A similar effect was seen with OXO-M application under conditions that favor activation of M2/4 autoreceptors (no DhβE present). Here, we show that in the presence of DhβE (1 µm), OXO-M (0.5 µm) robustly potentiates single-pulse stimulation but does not significantly potentiate dopamine transients evoked by stimulation trains (5 p/5 Hz, 5 p/10 Hz, 5 p/25 Hz, 5 p/100 Hz; one-sample t tests vs 100; 1 p, p = 0.0474, 5 p/5 Hz–100 Hz, p > 0.05, n = 7 slices; Fig. 2k,l). These data suggest that M5 receptor activation may play a greater role in modulating dopamine tone compared to phasic dopamine.
We wanted to test whether M5 deletion had an impact on the nACh-R-mediated component of the dopamine transient, which is sensitive to the nicotinic blocker DhβE. DhβE (1 µm) produced a 56% reduction in the peak dopamine transient in males (Fig. 2m; n =13 slices) and a 52% reduction in females (n = 31 slices; Fig. 2m). There was no significant difference in the effect of DhβE between male and female WT mice (two-way repeated measures ANOVA, main effect of sex and sex times time interaction, all p values > 0.05). Interestingly, in knock-out mice for the muscarinic M5 receptor (M5 KO), dopamine transients were slightly less sensitive to DhβE (46 ± 1% reduction from baseline in male M5 KO compared with 56 ± 2 in male WT mice; t test, t(18) = 3.198, p = 0.0050, n = 7–13 slices; Fig. 2m). This was a small quantitative difference; qualitatively, the contribution of nACh-R activation to the electrically evoked composite dopamine transient was similar.
Finally, we assessed the function of the muscarinic autoreceptors M2/M4, which are expressed in cholinergic interneurons and inhibit acetylcholine transmission. Again, we used the nonselective agonist OXO-M (10 µm) but in the absence of DhβE in this experiment. We found that the OXO-M effect was nearly identical between M5 WT and M5 KO mice (M5 WT, 59 ± 4; M5 KO, 60 ± 6% reduction from baseline, t test, t(7) = 0.1590, p = 0.8782, n = 4–5 slices). From this, we concluded that the cholinergic mechanisms responsible for triggering dopamine release (involving β2-nACh-Rs) and for mediating negative feedback on acetylcholine release (involving M2/M4 receptors) are for the most part intact in M5 KO mice.
OXO-M concentration response profile in male and female mice
Because we were intrigued by the time course of the OXO-M effect on M5-dependent potentiation in control mice, we probed this effect further by examining a range of OXO-M concentrations in both males and females. In all cases, we used scopolamine (1 µm) to reverse the effects of OXO-M and confirm the involvement of muscarinic receptors (Fig. 3). Increasing concentrations of OXO-M from 0.1 to 25 µm (0.1, 0.5, 2.5,10. 25 µm) did not increase the size of potentiation of the peak dopamine transient but significantly sped up the time-to-max across concentrations (Fig. 3a–g, n = 5–20 slices; max OXO-M effect, two-way ANOVA, F(4,71) = 0.2917, p = 0.8824; time-to-max, two-way ANOVA, main effect of concentration, F(4,71) = 20.84, p < 0.0001; Fig. 3g,h). There was also a significant difference in the late-phase attenuation of the response so that at 25 µm, the response rapidly returned to baseline levels. This was demonstrated by comparing the early- and late-phase concentration-response curves for males and females (Fig. 3f; male, two-way ANOVA, concentration times time interaction, F(4,40) = 7.347, p = 0.0002, n = 5-20 slices; female, two-way ANOVA, concentration times time interaction, F(4,32) = 12.66, p < 0.0001, n = 7–8 slices) and calculating the percentage attenuation as early to late phase (percentage baseline, Fig. 3i). Males and females had very similar responses to OXO-M across concentrations. One exception was at the 0.5 µm concentration, where we did see a significant difference in the overall time course (Fig. 3b) as well as a trend in the time-to-peak (post hoc t test, p = 0.0579). However, we concluded that generally there were only slight differences in the OXO-M response between males and females. There were no significant differences in the ability of scopolamine to reverse the potentiating effects of OXO-M (Fig. 3j).
M5-dependent potentiation of dopamine transmission displays unique temporal kinetics. a–e, Time course of normalized peak amplitude of electrically evoked dopamine transients recorded from ex vivo NAc slices taken from male (gray) and female (teal) M5 WT mice in which OXO-M (0.1, 0.5, 2.5, 10, 25 μm) followed by scopolamine (1 μm) is bath applied to the slice. f, OXO-M concentration response curve for the early (orange) and late (green) stages of OXO-M bath application for males (left) and females (right). g, Peak OXO-M effect at 0.1, 0.5, 2.5, 5, 10, and 25 μm OXO-M in males and females. h, Time-to-peak for 0.1, 0.5, 2.5, 5, 10, and 25 μm OXO-M in males and females. i, Attenuation (percentage baseline, late phase - percentage baseline, early phase) observed during OXO-M bath application (0.1, 0.5, 2.5,10, 25 μm) for male and female mice. j, Scopolamine reversal (percentage baseline) observed during OXO-M plus scopolamine bath application. k, Time course of normalized peak amplitude of electrically evoked dopamine transients before and following OXO-M (0.1 μm) in slices maintained in vehicle or M5 PAM VU 0365114 (5 μm). l, Time course of normalized peak amplitude of electrically evoked dopamine transients before and following OXO-M (0.5 μm) in slices maintained in vehicle or M5 NAM ML375 (10 μm). Asterisk (*) denotes post hoc t tests, number symbol (#) denotes significant interaction or main effect.
Based on our in-depth characterization of the OXO-M response profile, we were curious how co-application of either the commercially available positive allosteric modulator (PAM) VU 0365114 (5 µm) or the negative allosteric modulator (NAM) ML375 (10 µm) had on the effect of OXO-M on dopamine transmission. We coapplied VU 0365114 with OXO-M (0.01 µm) and compared this to our previous OXO-M (0.01 µm) effects. There was a significant drug by time interaction in which application of VU 0365114 appeared to quicken the time-to-max of the OXO-M effect (two-way repeated measures ANOVA, drug times time interaction, F(21,147) = 2.595, p = 0.0005, n = 4–5 slices; Fig. 3k). In contrast, the time course of ML375 coapplication with OXO-M (0.5 µm) was similar to the time course of the DMSO vehicle control, but there was a significant main effect of ML375 (drug main effect, F(1,9) = 5.457, p = 0.0443, n = 5–6 slices; Fig. 3l) because of the attenuation of the OXO-M effect.
Repeated stressor exposure disrupts M5 potentiation of dopamine transmission
We previously showed that repeated exposure to stress disrupts CRF-mediated potentiation of dopamine transmission in the NAc core (Lemos et al., 2012). Here, we further assessed whether repeated stress similarly disrupted M5 modulation of dopamine transmission. We exposed mice to either 7 d of repeated forced swim stress (FSS) or 10 d of repeated restraint stress (CRS). In both cases, stress reduced the maximal potentiation induced by 0.5 µm OXO-M compared with stress-naive mice (Fig. 4a–c; stress naive, 137 ± 7; CRS, 117 ± 5; FSS, 100 ± 8% baseline; one-way ANOVA, F(2,28) = 7.325, p = 0.0028; Dunnett's t test, control vs CRS, p = 0.0427, vs FSS, p = 0.0016, n = 7−13 slices).
Repeated stressor exposure disrupts M5 receptor function. a, Schematic demonstrating stress paradigms used for b and c. b, Time course of normalized peak amplitude of electrically evoked dopamine transients recorded from ex vivo NAc slices taken from stress-naive control male mice and male mice exposed to repeated FSS for 7 d or CRS for 10 d in which OXO-M (0.5 μm), followed by scopolamine, was bath applied to the slice. c, Mean OXO-M (0.5 μm) response in control, forced swim stress, and mice exposed to chronic restraint stress. d, Representative images of fluorescent in situ hybridization studies multiplexing Chrm5 and Th mRNAs in a stress-naive control (left) and FSS-exposed mouse (right). e, Mean percentage Chrm5/Th coexpression in the VTA of control and FSS-exposed mice. f, Mean total Chrm5 puncta in the VTA of control and FSS-exposed mice. g, Mean total Th+ cells in the VTA of control and FSS-exposed mice. Asterisk (*) denotes post-hoc t-test.
To test whether the functional disruption was because of downregulation of Chrm5 mRNA, we performed RNAscope in situ hybridization on tissue containing the VTA acquired from stress-exposed (FSS, 7 d) or stress-naive mice. Tissue was generated, processed, and analyzed in parallel. We found no differences in Chrm5/Th mRNA coexpression, total Chrm5 puncta within the VTA, or total number of Th+ neurons between stress-exposed and stress-naive mice (unpaired t tests, all p values > 0.05, n = 4–5 mice; Fig. 4d–g). These data demonstrate that the stress-induced M5 functional disruption is not because of downregulation of mRNAs encoding the M5 receptor.
M5 deletion manifests disparate behavioral deficits in the open field in male and female mice
Based on our finding that stress disrupts M5 signaling, we wondered how genetic constitutive M5 deletion would affect dopamine-dependent behaviors including locomotion, exploratory behaviors, and hedonic behaviors (Düzel et al., 2010). We first assessed the behavior of M5 WT, HET, and KO male and female littermates in the novel open field task performed under dim lighting conditions (49 lux). We assessed the time spent in the center of a novel open field over a 60 min period and found that M5 deletion significantly reduced the time in center during the first 30 min only in male mice. This was not observed in female M5 KO mice, which were similar to WT and HET mice. In both male and female M5 WT mice, exploration of the center of the open field was always higher in the first 30 min than in the last 30 min, as expected for exploratory behavior that attenuates as time passes (Fig. 5a–d). However, in both male and female M5 KOs this habituation in exploration of the center was disrupted. M5 HET mice had an intermediate phenotype, where males showed behavior similar to control M5 WT, whereas M5 HET females showed disruption in habituation, similar to M5 KO (Fig. 5a–d; males, M5 WT, HET, KO, two-way ANOVA, time times genotype interaction, F(2,61) = 5.342, p = 0.0073, asterisk (*) denotes post hoc Sidak's t test, first 30 min for WT vs HET, p = 0.8945; WT vs KO, p = 0.0299; HET vs KO, p = 0.0477; number symbol (#) denotes post hoc t test of 0–30 vs 30–60 min, WT, p = 0.0208; HET, p < 0.0001; KO, p = 0.9964, n = 14–25 mice; females, M5 WT, HET, KO, two-way ANOVA, main effect of time, F(1,40) = 18.88, p = 0.0001; post hoc Sidak's t test, first 30 min for WT vs HET, p = 0.6723; WT vs KO, p = 0.6466; HET vs KO, p > 0.9999; post hoc t test of 0–30 vs 30–60 min, WT = p = 0.0064, HET = p = 0.1270, KO = p = 0.0894, n = 13–16 mice). Under these conditions, females spent more time in the center compartment compared with males; however, a significant sex by genotype interaction was not present (two-way ANOVA, main effect of sex, F(1,101) = 6.925, p = 0.0098, sex times genotype interaction, F(2,101) = 0.6531, p = 0.5226). In addition to time exploring the center, we also analyzed rearing behavior in M5 WT and KO male and female littermates during the first 10 min of the open field as rearing has been shown to be a secondary measure of exploration, indicative of vigilance and dependent on dopamine transmission (Vallone et al., 2002; Sturman et al., 2018). Female M5 KO mice had a significant reduction in rearing behavior compared with WT littermates, whereas there was no difference in the number of unsupported rears between M5 WT and KO male mice (male, WT, 7 ± 1; KO, 8 ± 2 rears; t test, t = 0.0356, p = 0.7103, n = 14 mice each; female, WT, 14 ± 2; KO, 9 ± 1 rears; t test, t = 2.338, p = 0.0265, n = 15–16 mice; Fig. 5e,f). To confirm that these deficits in exploratory behavior were not because of a general deficit in spontaneous locomotion, we analyzed horizontal ambulatory behavior. Generally, females showed enhanced locomotion in the open field compared with males (two-way ANOVA, main effect of sex, F(1,108) = 12.14, p = 0.007). Importantly, we found no difference in spontaneous horizontal locomotion across genotypes for neither males nor females (Fig. 5g–j; males, one-way ANOVA on total time, F(2,66) = 1.252, p = 0.2925, n = 19–31 mice; females, one-way ANOVA on total time, F(2,42) = 0.6094, p = 0.5484, n = 14–16 mice). Together, these behavioral findings point to a deficit in exploratory behavior in mice lacking muscarinic M5 receptors. We also found sex-specific differences in the impact of M5 deletion on exploration, indicating that M5 function regulates different aspects of these behaviors.
M5 deletion produces disparate behavioral deficits in the open field in males and females. a, b, Representative motor tracks extracted from video monitoring of the mice in the open field for male (a) and female (b) M5 WT (left), M5 HET (middle), and M5 KO (right) mice. c, Time in the center of a novel open field in the first and second 30 min of a 60 min open field session for male M5 WT, HET, and KO mice. Asterisk (*) denotes post hoc Sidak's t test comparing M5 WT/HET versus M5 KO at 30 or 60 min; number symbol (#) denotes post hoc Sidak's t test comparing 30 versus 60 min for M5 WT, HET, or KO. d, Time in the center of a novel open field in the first and second 30 min of a 60 min open field session for female M5 WT, HET, and KO mice. e, Total number of unsupported rears in the first 10 min of the novel open field session for male M5 WT and KO mice. f, Total number of unsupported rears in the first 10 min of the novel open field session for female M5 WT and KO mice. Asterisk (*) denotes unpaired t test. g, Time course of spontaneous locomotor activity during a 60 min exposure to the novel open field for male M5 WT, HET, and KO mice. h, Mean total motor activity in a 60 min novel open field session for male M5 WT, HET, and KO mice. i, Time course of spontaneous locomotor activity during a 60 min exposure to the novel open field for female M5 WT, HET, and KO mice. j, Mean total motor activity in a 60 min novel open field session for female M5 WT, HET, and KO mice.
On the day following open field testing, we used the same open field apparatus to assay the response to novel stimuli. Mice were habituated to the OF for 60 min under dim conditions. We investigated the time spent exploring both a novel object (1L bottle cap) or a novel palatable food (bacon pellets). The novel food was contained in a dish and covered by a mesh so that mice could smell the food but not consume it. The NO and NF exposure was counterbalanced across two sessions. We analyzed the exploratory response to the first stimulus compared with the second stimulus exposure regardless of stimulus type. In this case, there was a significant main effect of both genotype and time, indicating that male M5 KO mice have a deficit in exploratory behavior, particularly during the first stimulus presentation. In contrast to our findings in males, females had no significant differences in exploratory behavior (Fig. 5c,e; males, two-way ANOVA, main effect of time, F(1,32) = 28.44, p < 0.0001; main effect of genotype, F(1,32) = 4.627, p = 0.0391; post hoc t test of M5 WT vs KO for Stimulus 1, p = 0.0259, n = 17 mice each group; females, two-way ANOVA, main effect of time, F(1,22) = 1.713, p = 0.204; main effect of genotype, F(1,26) = 0.4313, p = 0.5181, n = 11–13 mice; Fig. 6a,b,d). We had hypothesized that mice would increase their exploration time toward the NF compared with the NO in WT but not KO mice. Although there was a trend for this pattern in males, overall, there were no significant differences in exploration of the two stimuli in either WT or KO male or female mice. There was, however, an overall main effect of genotype on exploration in males only (males, two-way ANOVA, stimuli times genotype interaction, F(1,32) = 0.6007, p = 0.4440; main effect of genotype, F(1,32) = 4.627, p = 0.0391, n = 17 mice each group; females, two-way ANOVA, stimuli times genotype interaction, F(1,26) = 0.6583, p = 0.4245, main effect of genotype, F(1,26) = 0.9796, p = 0.3314, n = 13–15 mice; Fig. 6e,f).
M5 deletion reduces exploration of novel stimuli selectively in male mice. a, Schematic diagram of the novel stimulus paradigm. b, Time in the center during a 30 min session for Stimulus 1 (presented first) or Stimulus 2 (presented second) regardless of the stimulus type placed in the center of an open field for male M5 WT or M5 KO mice. Asterisk (*) denotes Sidak's post hoc t test comparing M5 WT versus M5 KO for Stimulus 1; number symbol (#) denotes main effect of Stimulus 1 versus Stimulus 2. c, Time in the center during a 30 min session in which either an NO or NF was placed in the center of an open field for male M5 WT or M5 KO mice. Ampersand (&) denotes main effect of genotype. d, Time in the center during a 30 min session for Stimulus 1 (presented first) or Stimulus 2 (presented second) regardless of stimulus type placed in the center of an open field for female M5 WT or M5 KO mice. e, Time in the center during a 30 min session in which either an NO or NF was placed in the center of an open field for female M5 WT or M5 KO mice. f, Sucrose preference (4%) for male or female M5 WT or KO mice (left). Sucrose preference (1%) for M5 WT or KO pooled male and female mice (right). Asterisk (*) denotes unpaired t test. Sucrose preference equals sucrose volume consumed/sucrose plus water volume consumed times 100. g, Percentage time immobile in the forced swim test over a period of 7 d for M5 WT, HET, and KO male and female mice. On day 1, a 15 min session was broken down into 5 min intervals. On days 2–7, there were 5 min sessions.
Because M5 KO mice, males in particular, display a reduction of exploration of appetitive stimuli, we wondered whether these same mice displayed anhedonia. Given our previous data, we hypothesized that M5 deletion should reduce sucrose preference. To ensure we had a robust preference in M5 WT mice, we used a relatively high sucrose concentration of 4%. We found that neither male nor female M5 KO mice displayed a deficit in sucrose preference (using 4% sucrose solution) compared with M5 WT littermate controls (Fig. 6f; two-way ANOVA, all comparisons, p > 0.05, n = 7–14 mice). We then used a more typical 1% concentration of sucrose to determine whether there were any M5-dependent anhedonic effects. In contrast, with 1% sucrose, M5 KO mice showed a very slight but nevertheless significant increase in sucrose preference (M5 WT, 65.8 ± 1.7%; M5 KO, 69.9 ± 0.6%, unpaired t test, p = 0.0360, n = 6–7 mice, pooled male and female, Fig. 6f). Together, we conclude that the reduction in exploration is not because of general anhedonia.
An alternative hypothesis is that the novel stimuli act as acute stressors that may trigger a combination of active investigation and passive avoidance. A change in this behavior may be because of a general deficit in stress-induced active versus passive coping strategies. To test this, we assessed the percentage immobile time in the forced swim test over a period of 7 d. The forced swim test reveals changes in active (i.e., swimming, climbing) versus passive (i.e., floating) stress coping strategies. There were no overall differences in percentage immobile time between male and female mice (two-way repeated measures ANOVA, sex times time interaction, F(8,232) = 1.210, p = 0.2940, n = 15–16 mice). There were also no differences in percentage immobile time by genotype (two-way repeated measures ANOVA, genotype times time interaction, males, F(16,104) = 0.3696, p = 0.9866, n = 5–6 mice; females, F(16,96) = 0.9931, p = 0.4708, n = 5 mice per group, Fig. 6g). Thus, we concluded that the difference in novelty exploration is not driven by a general alteration in active versus passive stress coping strategies.
M5 deletion disrupts normal behavioral adaptations to changing environmental conditions
Time allocation in the center of an open field is sensitive to anti-anxiety drugs and thus is considered a test of anxiety-like behavior (Prut and Belzung, 2003). We used the elevated zero maze as a secondary measure of anxiety-like behavior. Performance in the EZM and the more commonly used elevated plus maze (EPM) is sensitive to changes in environmental conditions, especially light (Walf and Frye, 2007). Mice were exposed to the EZM under two different, counterbalanced conditions, dim light, where the closed arms were not covered, and bright light, where the closed arms were covered (Fig. 7a). It has been shown that modulating the luminescence as well as conditions of the closed arms (i.e., transparent vs opaque walls) has an impact on time spent in the open arms. Generally, there was no overall effect on time in the open arms by genotype. However, M5 WT male mice spent significantly more time in the open arms under the dim conditions as did M5 HET. However, M5 KO mice did not significantly change their behavioral response to EZM between conditions (Fig. 7b; two-way ANOVA, main effect of lighting, F(1,100) = 16.48, p < 0.0001; main effect of genotype, F(2,100) = 3.521, p = 0.0333; post hoc t test, bright vs dim, M5 WT, p = 0.0391; M5 HET, p = 0.0008; M5 KO, p = 0.5964, n = 13–25 mice). Although females showed the same qualitative pattern, the effects were not statistically significant for any comparison. M5 WT females showed a trend in difference between dim and bright conditions that was absent in M5 HET and KO mice. However, overall, the quantitative differences were not significant (Fig. 7c; two-way ANOVA, main effect of lighting, F(1,47) = 6.705, p = 0.0128; main effect of genotype, F(2,47) = 0.8458, p = 0.4356; post hoc t test, bright vs dim, M5 WT, p = 0.0852; M5 HET, p = 0.8487; M5 KO, p = 0.3037, n = 15–18 mice). However, if we compared the percentage time in the open arms in the dim condition to 50% (no preference), we found that female M5 WT mice showed no difference from 50% (p = 0.2196), compared with M5 HET and M5 KO (all p values = 0.0004 and 0.0053, respectively) suggesting a subtle deficit in exploration in female M5 KOs under these conditions. Thus, although the deficits in exploration of riskier compartments in novel environments tend to be context specific, the deficit in normal adaptations in behavioral output persist across contexts. Furthermore, this deficit in exploratory behavior appears to be sex dependent.
M5 deletion produces deficits in behavioral adaptation in response to changes in environmental conditions in the elevated zero maze. a, Representative overhead images of bright, covered (left), and dim no cover (right) elevated zero maze conditions. b, Percentage time in the open arms for bright and dim conditions for male M5 WT, HET, and KO littermates. c, Percentage time in the open arms for bright and dim conditions for female M5 WT, HET, and KO littermates. Shaded regions highlight the relative difference from 50% (no preference) for each group under each condition. Asterisks (*) post-hoc t-test. Delta denotes trend.
M5 deletion reveals dissociation between generalized anxiety-like behavior and selective deficit in novelty exploration
The discrepancy in phenotype between two classic measures of anxiety-like behavior led us to use yet a third behavioral measure that combines elements of both the OF and EZM—the light-dark box. Importantly, this was done in an entirely different cohort of mice that did not have prior experience with other behavioral tests. Following the standard 5 min test, we reintroduced the mice to the chamber with the NF stimuli in the light side. Interestingly, there was no difference in the percentage time spent in the light side between M5 WT, HET, and KO for either male or female mice (or between all males and all females) in the standard 5 min test, similar to the EZM (two-way ANOVA, F(2,55) = 0.9393, p = 0.3971, n = 9–13 mice; Fig. 8a,b). Novel palatable food placed in the bright light without much habituation was generally aversive across all mice. However, this manipulation revealed a similar aversion to novelty exploration specifically in male mice as shown in the OF (two-way ANOVA, sex times genotype interaction, F(2,51) = 3.325, p = 0.0439; post hoc Sidak's t test, male M5 WT vs KO, p = 0.007; male M5 WT vs HET, p = 0.0591, n = 8−13 mice; Fig. 8a,c). Although there was no main effect of sex (two-way ANOVA, main effect of sex, F(1,51) = 0.2201, p = 0.6410), it is possible that the lack of effect in females is because of a floor effect. However, it should be noted that the percentage of light side exploration for male M5 KO mice was nearly half (5.6%) that of the average female time (10.8%), indicating that a general floor effect was not present for the female mice. Together, these data reveal a behavioral dissociation between general anxiety-like behavior and specific deficits in novelty exploration.
M5 deletion produces a sex-dependent deficit in novel food exploration in light/dark box. a, Representative overhead images of a brightly lit light-dark box without (left) and with (right) a novel palatable food stimulus. b, Percentage time in the light compartment for male and female M5 WT, HET, and KO littermates in the absence of the NF stimulus. c, Percentage time in the light compartment for male and female M5 WT, HET, and KO littermates in the presence of the NF stimulus. Asterisks (*) post-hoc t-test. Delta denotes trend.
Discussion
The M5 receptor is an interesting therapeutic target protein because its overall expression is restricted compared with the M1–M4 receptors, potentially reducing the likelihood of unwanted side effects. In addition, the crystal structure of the M5 receptor has recently been solved, enabling pharmacological targeting of this protein (Vuckovic et al., 2019). Indeed, negative and positive allosteric modulators targeted toward the M5 receptor are actively being developed (Bridges et al., 2009, 2010; Gentry et al., 2010a,b, 2012, 2013, 2014). Manipulations of M5 receptor function have indicated that it may play a role in drug seeking and taking, as well as sensorimotor gating (Basile et al., 2002; Thomsen et al., 2005, 2007; Gunter et al., 2018; Gould et al., 2019). However, much remains unknown with respect to the more basic characterization of the cellular and behavioral functioning of the M5 receptor, with little attention paid to potential sex differences. In this study, we quantified M5 expression in VTA dopamine neurons in males and females and performed a full functional characterization of M5 potentiation of dopamine transmission in both males and females. Although there may be some more nuanced differences, overall, we found that both M5 expression and M5 potentiation of dopamine transmission in the NAc were similar in male and female mice. Surprisingly, despite similar cellular function, M5 deletion manifests disparate behavioral deficits in males and females. In females, M5 deletion primarily impairs rearing behavior, a dopamine-dependent motor behavior that is both a form of exploration as well as vigilance. In contrast, M5 deletion in males primarily affected exploratory behavior, particularly of riskier or more exposed areas of our testing chambers or toward novel stimuli. In addition, M5 receptors appear to be important for adapting behaviors in response to new information, a behavioral function that has been attributed to cholinergic interneurons in the striatum. Although this study provides a foundation for understanding M5 function at the cellular, systems, and behavioral levels, there are many questions that still exist.
Unique features of M5 modulation of dopamine transmission in the NAc
There was a concentration-dependent gene-dose effect on OXO-M potentiation of dopamine transmission. At a relatively lower concentration, there was a predictable intermediate effect of OXO-M in M5 HET mice. In contrast, at a higher concentration, there was a similar maximal effect between M5 WT and HET mice, and rather the secondary inhibition was disrupted. There may be spare receptors on dopamine terminals within the NAc core, which have been shown to exist for several GPCRs including muscarinic receptors (Buchwald, 2019). However, it appears that both M5 alleles are necessary to trigger the secondary attenuation of the response as a difference in the early versus the late phase was only detected in M5 WT mice and not M5 HET mice. The mechanism that mediates this concentration-dependent attenuation is still unclear. It could be that the receptor is desensitized and internalized at higher concentrations. Alternatively, it is possible that M2/M4 heteroreceptors, not present on cholinergic interneurons, are recruited at these higher concentrations. Last, it is also possible that M5 receptor activity on NAc neurons triggers the release of a known inhibitor of dopamine transmission such as an endocannabinoid or dynorphin.
We had expected that increasing concentrations of OXO-M would increase the maximal response as is often the case for other GPCRs. Notably, the majority of research on GPCR regulation of dopamine transmission at the terminals within the NAc has focused on inhibitory modulation of dopamine transmission via Gi/o-coupled receptors (i.e., D2, KOR, M2/4) with perhaps the exception of 5-HT2 and mGlu1 receptors (Bruton et al., 1999; Alex et al., 2005; Britt and McGehee, 2008; Shin et al., 2017). In the case of the M5 receptor, increasing concentrations of OXO-M shortened the time to achieve what appears to be a ceiling of maximal potentiation of dopamine release. Increasing concentrations hastened equilibrium within the slice preparation along with receptor binding. We speculate that the coupling effectors that lead to potentiation, likely via Gq-induced release of calcium stores, limit the maximal effect. Furthermore, it is possible that there is a narrow dynamic range of Gq or Gs modulation compared with Gi/o. This may also explain why M5 potentiation is more effective using single-pulse stimulation as opposed to high-frequency stimulation. We do note that for some reason M5 potentiation was larger for single pulse when it was part of the train paradigm. It is possible that trains are triggering some LTP-like mechanism that we do not fully understand. Collectively, further investigation of these cellular mechanisms is required to fully understand GPCR-mediated modulation of dopamine terminals, yet remains challenging to probe with our current tools. Understanding the concentration response profile of the M5 receptor is key to the pharmacological development of agonists, antagonists, and allosteric modulators. For example, as we show, a commercially available PAM does not necessarily increase the maximal response as one might predict based on other GPCRs but rather shortens the time to achieve the maximal response. Moving forward, this concentration response profile will aid in the assessment of pharmacological efficacy.
Our previous studies demonstrate that repeated stress ablates CRF-mediated potentiation of dopamine transmission at terminals within the NAc core (Lemos et al., 2012). Like CRF, our test using an intermediate concentration of OXO-M showed that repeated stress disrupted M5-dependent potentiation. We went on to show that this disruption is not because of a stress-induced downregulation of M5-expressing dopamine neurons, total M5 transcripts in the VTA, or total number of dopamine neurons. However, a caveat to this conclusion is that we are only able to assess mRNA levels because of poor M5 antibody quality. It is possible that stressor exposure does reduce M5 receptor total protein, protein insertion into the membrane, and M5 receptor coupling. It is also possible that repeated stress elevates ambient levels of acetylcholine to either desensitize and internalize the receptor or recruit inhibitory mechanisms similar to how high concentrations of OXO-M might evoke the secondary attenuation. These will be interesting questions to address in future studies. Our observation that stress disrupts M5 function led us to examine how M5 deletion may lead to anxiety- or depression-related phenotypes.
The role of M5 receptors in mediating exploratory behaviors
First, we acknowledge the caveat to this work in that we are using a constitutive M5 knock-out mouse as opposed to a cell-type-specific transgenic and/or viral strategy. Presently, there are no commercially available tools to address this caveat. Our laboratory is developing these tools, and we intend to follow up this study in the future using these novel techniques. This study provides a road map for initial validation of subsequent modern techniques.
With this caveat in mind, we explored the consequence of M5 deletion on dopamine-dependent motor, exploratory, and hedonic behaviors. Spontaneous ambulatory behavior was intact. This is consistent with what has been previously reported (Yamada et al., 2001). Yet, we found that in females, M5 deletion disrupted rearing behavior, which is a dopamine-dependent form of vertical inspection and vigilance displayed in novel environments. In contrast, M5 deletion manifested differently in males. Males displayed disruption in exploration of the center of a novel open field as well as disruption in the exploration of novel appetitive stimuli (presented in two different contexts, open field and light/dark box). For both males and females, these disruptions in exploratory behaviors were not accompanied by anhedonia assayed using the sucrose preference test (for either 1 or 4% sucrose), despite evidence that both scopolamine and reduced cholinergic interneuron firing promotes a reduction in sucrose seeking and preference (Warner-Schmidt et al., 2012; Addy et al., 2015; Navarria et al., 2015; Cheng et al., 2019). These data indicate that cholinergic modulation of hedonia is dependent on activation of M1, 2, 3, or 4. Both novelty and risk preference have been linked with increased drug taking and seeking (Belin et al., 2011; Wingo et al., 2016). Therefore, a disruption in novelty exploration and risk aversion occurring in the same genotype in which decreases in drug seeking and taking are frequently observed is consistent with the previous literature.
Our findings are not entirely consistent with a previous study in rats done by the Addy laboratory in which rats received intra-VTA infusion of physostigmine with or without coadministration of the M5 negative allosteric modulator (NAM) ML375 (VU6000181; Nunes et al., 2019). The investigators found that intra-VTA physostigmine, a treatment that elevates acetylcholine levels in the VTA, produced anhedonia and anxiety-like behavior in both males and females assayed with the sucrose preference test, EPM, and forced swim test. Coinfusion of the M5 NAM prevented these effects only in males. Interestingly, the M5 NAM on its own had no effect, indicating that this effect only occurs under conditions of elevated acetylcholine tone within the VTA. The sex-dependent component is consistent with the sexual dimorphism we have observed. However, this study suggests that M5 function in the VTA produces aversion and is anxiogenic. There may be several reasons for the inconsistencies between our findings and those of Nunes et al. (2019), including species differences, developmental compensation because of constitutive M5 deletion, regional specificity of the Nunes study manipulations, and potential nonselectivity of the ML375 compound. We believe that some of these inconsistencies may be reconciled with the development of cell-specific transgenic and viral techniques.
The role of M5 receptors in behavioral adaptation
In the open field test, control mice showed a reduction in center exploration of the novel open field over time. However, this habituation was not present in mice with constitutive M5 deletion. Likewise, in the EZM, mice adjusted their time spent in the open compartments based on the lighting conditions and the level of enclosure of the closed compartments. This was disrupted in both male and female M5 KO mice. Work from the Cain lab has demonstrated that M5 KO mice have a deficit in prepulse inhibition (Thomsen et al., 2007). Although prepulse inhibition is considered a sensorimotor gating assay and an animal model for features of schizophrenia, at a broader level it relates to the ability to integrate information and adjust the behavioral response. Together, it is possible that M5 receptors play a broader role in behavioral updating in response to changes in environmental contingencies. Indeed, it has been suggested that striatal cholinergic interneurons are important for behavioral updating and flexibility (Bradfield et al., 2013; Apicella, 2017). It is plausible that M5 is a downstream effector of cholinergic interneuron responses to environmental changes.
Conclusion
There are several open questions remaining that are beyond the scope of this initial study, which we hope are actively pursued by other laboratories in addition to our own. Considering both the therapeutic potential and the interesting biological function of the M5 receptor, we hope this study brings M5 to the forefront of GPCR and neuroscience research.
Footnotes
This work was supported by National Institute of Mental Health Grants MH109627 (J.C.L.), and MH122749 (J.C.L.), Department of Neuroscience and Medical Discovery Team on Addiction Start-Up Funds (J.C.L.), and National Institute on Alcohol Abuse and Alcoholism Grant AA000421 (V.A.A.). We thank Dr. Jürgen Wess at the National Institute of Diabetes and Digestive and Kidney Diseases for providing the M5 KO mouse line, Ms. Rachel Dick for help with cryosectioning used for RNAscope techniques, Dr. Anna Lee for the use of the MATLAB-based RNAscope analysis pipeline, Dr. Sade Spencer and the Spencer lab for performing pilot Western blot experiments, and Drs. Erin Calipari, Mark Thomas, Michael Bruchas, and Larry Zweifel, and Ms. Elizabeth Souter for input and comments on this work.
The authors declare no competing financial interests.
- Correspondence should be addressed to Julia C. Lemos at jlemos{at}umn.edu
