Abstract
The lateral habenula (LHb) balances reward and aversion by opposing activation of brain reward nuclei and is involved in the inhibition of responding for cocaine in a model of impulsive behavior. Previously, we reported that the suppression of cocaine seeking was prevented by LHb inactivation or nonselective antagonism of LHb mAChRs. Here, we investigate mAChR subtypes mediating the effects of endogenous acetylcholine in this model of impulsive drug seeking and define cellular mechanisms in which mAChRs alter LHb neuron activity. Using in vitro electrophysiology, we find that LHb neurons are depolarized or hyperpolarized by the cholinergic agonists oxotremorine-M (Oxo-M) and carbachol (CCh), and that mAChRs inhibit synaptic GABA and glutamatergic inputs to these cells similarly in male and female rats. Synaptic effects of CCh were blocked by the M2-mAChR (M2R) antagonist AFDX-116 and not by pirenzepine, an M1-mAChR (M1R) antagonist. Oxo-M-mediated depolarizing currents were also blocked by AFDX-116. Although M2R activation inhibited excitatory and inhibitory inputs to LHb neurons, the effect on excitation was greater, suggesting a shift in excitatory-inhibitory balance toward net inhibition. Activation of VTA inhibitory inputs to LHb neurons, via channelrhodopsin-2 expression, evoked IPSCs that were inhibited by M2Rs. Finally, we measured LHb-dependent operant response inhibition for cocaine and found it impaired by antagonism of M2Rs, and not M1Rs. In summary, we show that a cholinergic signal to LHb and activation of M2Rs are critical to enable inhibition of responding for cocaine, and we define cellular mechanisms through which this may occur.
SIGNIFICANCE STATEMENT The lateral habenula (LHb) is a brain region receiving information from brain areas involved in decision-making, and its output influences motivation, reward, and movement. This interface between thoughts, emotions, and actions is how the LHb permits adaptive behavior, and LHb dysfunction is implicated in psychiatric and drug use disorders. Silencing the LHb impairs control over cocaine seeking in rats, and mAChRs are also implicated. Here, we measured cocaine seeking while blocking different mAChRs and examined mechanisms of mAChR effects on LHb neurons. M2-mAChRs were necessary for control of cocaine seeking, and these receptors altered LHb neuron activity in several ways. Our study reveals that LHb M2-mAChRs represent a potential target for treating substance use disorders.
Introduction
The lateral habenula (LHb) forms part of the habenular complex, a structure at the posterior dorsomedial end of the epithalamus (Sutherland, 1982; Kim and Chang, 2005). Early reports described the LHb as a hub collecting information from limbic forebrain inputs and relaying this to major monoaminergic nuclei (Sutherland, 1982; Christoph et al., 1986). More recently, Matsumoto and Hikosaka (2007) showed that monkey LHb neurons are activated by negative prediction errors triggered by receiving smaller than expected rewards, and are inhibited when expected positive rewards are received (Matsumoto and Hikosaka, 2007). Importantly, this LHb neuron activity opposed that of dopamine (DA) neurons, suggesting that the LHb may counterbalance reward (Matsumoto and Hikosaka, 2007). Later studies showed that the LHb influenced DA neurons indirectly through its synaptic excitation of inhibitory neurons in the rostromedial tegmentum (RMTg) that form GABAergic synapses with DA cells (Jhou et al., 2009; Balcita-Pedicino et al., 2011; Barrot et al., 2012). Through this circuit, the LHb influences motivated behavior by processing rewarding and aversive stimuli to bias the strength and timing of DA system output (Koob and Le Moal, 2008; Stopper et al., 2014).
The LHb is also implicated in human substance use and major depressive disorders (Shabel et al., 2014; Yang et al., 2018), and it has been incorporated into opponent process theory related to reward and motivation (Solomon, 1980; Ettenberg et al., 1999). Consistent with this, drugs having rewarding properties often have delayed aversive effects that promote continued drug seeking to forestall aversion. For example, cocaine is initially rewarding followed by delayed aversion (Ettenberg et al., 1999; Lecca et al., 2017), and LHb neuron activity parallels these phases (Matsumoto and Hikosaka, 2007; Jhou et al., 2013; Neumann et al., 2014). Inactivation of LHb also disrupts reward choice bias and response inhibition for cocaine suggesting a role for the LHb in reward-based decision-making (Stopper et al., 2014) and in control of drug seeking (Zapata et al., 2017).
Most LHb neurons are glutamatergic (Brinschwitz et al., 2010; Aizawa et al., 2012), with few intrinsic GABAergic neurons identified (Webster et al., 2021). Therefore, inhibition of excitatory LHb output may be largely from extrinsic sources (Zhang et al., 2016; Wagner et al., 2017). Inhibitory LHb afferents from the lateral preoptic area (Barker et al., 2017), lateral hypothalamus (LH) (Lecca et al., 2017; Lazaridis et al., 2019), ventral pallidum (Faget et al., 2018), VTA (Stamatakis et al., 2013; Root et al., 2014a), and medial dorsal thalamus (Webster et al., 2020) have all been characterized. These afferents suppress LHb output and are generally rewarding in behavioral assays. In contrast, excitation of the LHb by lateral preoptic area (Barker et al., 2017), entopeduncular nucleus (Li et al., 2021), ventral pallidum (Faget et al., 2018), and LH (Stamatakis et al., 2016; Lazaridis et al., 2019) increase LHb output and is aversive in rodents. Additionally, some LHb inputs (e.g., VTA and entopeduncular nucleus) corelease both GABA and glutamate (Root et al., 2014b, 2018; Shabel et al., 2014).
Understanding the diversity of these afferents and their integration at the neuronal level is important for understanding the LHb's contribution to behavior. In particular, reduced inhibition increases LHb excitation, and LHb hyperactivity is implicated in depression (Shabel et al., 2014) and drug withdrawal (Shabel et al., 2014; Meye et al., 2016; Root et al., 2018). However, the influence of other regulators of LHb neuron activity on behavior is poorly understood. Notably, LHb cholinergic signaling is implicated in inhibiting cocaine seeking in a model of impulsive behavior (Zapata et al., 2017), and there is strong evidence for cholinergic innervation of the LHb from inputs, including the hindbrain pontomesencephalic tegmentum and basal forebrain cholinergic nuclei (Fibiger, 1982; Contestabile and Fonnum, 1983; Woolf and Butcher, 1986). Here, we investigated mechanisms of cholinergic control of LHb neuron activity and used a model of response inhibition to evaluate how this influences cocaine-motivated behavior.
Materials and Methods
Subjects
Male and female Long Evans rats (Charles River Laboratories) aged 4-6 weeks were housed 2-4 same sex animals per cage in an Association for Assessment and Accreditation of Laboratory Animal Care International accredited facility. They were maintained in a temperature- and humidity-controlled environment with ad libitum food and water. Rats used in electrophysiological experiments were housed under standard lighting conditions (lights on 6:00 A.M., off 6:00 P.M.). For behavioral experiments, rats were housed under a reverse 12 h light/dark cycle and experiments were performed during the dark phase. All animal procedures were approved by the Animal Care and Use Committee of the National Institute on Drug Abuse Intramural Research Program and conducted in accordance with the Guide for the care and use of laboratory animals (National Research Council, 2011).
Electrophysiology
Rats were anesthetized with isoflurane and decapitated using a guillotine. The brains were extracted and transferred to ice-cold HEPES-modified cutting solution (in mm as follows: 92 NaCl, 3 KCl, 1.2 NaH2PO4, 30 NaHCO3, 20 HEPES, 25 glucose, 5 ascorbic acid, 10 MgCl2, 0.5 CaCl2). The tissue was blocked with a razor blade and glued to the stage of a vibrating tissue slicer (Leica VT1200S, Leica Biosystems). Coronal slices (280 µm) containing the LHb, and corresponding to ∼3.3 mm to 4 mm posterior to bregma (Paxinos and Watson, 2007), were transferred to a holding chamber containing normal aCSF consisting of the following (in mm): 126 NaCl, 3.0 KCl, 1.5 MgCl2, 2.4 CaCl2, 1.2 NaH2PO4, 11.0 glucose, 26 NaHCO3, saturated with 95% O2/5% CO2, at 35°C for 15-20 min, and then stored at room temperature. A hemisected brain slice was submerged in a recording chamber (∼170 µl volume; Warner Instruments) integrated into a fixed stage of an upright microscope (Olympus BX51WI), and perfused with warm (31°C-33°C) aCSF at 2 ml/min using a peristaltic pump. The aCSF was warmed using an inline solution heater (TC-324B, Warner Instruments). Drugs were prepared as stock solutions in H2O or DMSO and diluted in aCSF to the indicated concentrations. Visualization of LHb neurons was performed using gradient contrast microscopy with infrared illumination. Recordings were performed in the medial LHb (mLHb), a region corresponding to the parvocellular subnucleus and central subnucleus (Geisler et al., 2003). Whole-cell voltage-clamp recordings were performed using a MultiClamp 700B (Molecular Devices), WinLTP software (version 3.0, WinLTP, Ltd.), and an A/D board (PCI-6251, National Instruments). Series resistance was monitored using hyperpolarizing steps (−10 mV, 200 ms), and cells demonstrating >20% change in access resistance were excluded from analyses.
In experiments in which sIPSCs alone were recorded, electrodes (4-6 mΩ) were filled with the following (in mm): 145 KCl, 10 HEPES, 0.2 EGTA, 2 MgCl2, 4 Mg-ATP, 0.3 Na2-GTP, 10 Na2-phosphocreatine, pH 7.2 with KOH. For experiments in which sEPSCs and sIPSCs were recorded in the same cells, or when electrically evoked (eIPSC) or light-evoked IPSCs (oIPSC) were recorded, the electrodes were filled with the following (in mm): 140 K-gluconate, 5 KCl, 10 HEPES, 0.2 EGTA, 2 MgCl2, 4 Mg-ATP, 0.3 Na2-GTP, 10 Na2-phosphocreatine, pH 7.2 with KOH. For neurons clamped at 0 mV, QX-314 (1 mm) was added to the intracellular solution. Except when EPSCs and IPSCs were recorded in the same neurons, IPSCs were pharmacologically isolated using 6,7-dinitroquinoxaline-2,3-dione (DNQX, 10 µM), and EPSCs were isolated in the presence of picrotoxin (100 µm). In some experiments, gabazine (10 µm) or DNQX (10 µm) was bath-applied at the end of recordings to confirm the nature of the synaptic response.
Electrically evoked responses were obtained by positioning the tips of a bipolar stimulating electrode (FHC) on the surface of the brain slice 150-200 µm from the recording electrode within the LHb. Stimulation intensity was adjusted to elicit a response 30%-50% of the peak amplitude. For oIPSCs, a 473 nm diode-pumped solid state laser (OEM Laser BL-473-00200) was used to deliver a single 473 nm blue light pulse (5 ms) that was collimated through a 40× objective using a fiber optic adaptor (IS-OGP; Siskiyou). Evoked responses were obtained every 30s, and spontaneous synaptic events were collected as continuous recordings before and during drug application. Continuous event recordings of spontaneous currents were analyzed offline in 2 min epochs collected before and during drug application. Event detection of spontaneous synaptic currents was performed using WinEDR (University of Strathclyde, Glasgow, UK), using a template-based matching algorithm (Clements and Bekkers, 1997). Typical detection parameters for sEPSCs used rise times of 0.3 ms and decay times of 3 ms, and individual events were confirmed by visual inspection. Excitatory-inhibitory (E-I) ratios were calculated by measuring eEPSCs at −70 mV (inward currents) and eIPSCs at 0 mV (outward currents) within the same cell. The stimulation intensity for these experiments was adjusted to produce an eEPSC that was ∼50% of maximum. The E-I ratio was defined as the proportion of excitatory synaptic current to the total synaptic current recorded in each cell, or E/E + I, using the area under the curve of 6-10 events averaged during the control predrug period and during drug application (Antoine et al., 2019). To analyze E-I ratios independently of electrical stimulation, we also measured sEPSCs (−70 mV holding potential) and sIPSCs (0 mV holding potential) in the same LHb neurons during 30-60 s periods under control conditions and during carbachol (CCh) or CCh + AFDX-116 application. The area of each spontaneous synaptic current (pA*ms) was calculated and all currents summed over the entire time window to obtain the total membrane charge transfer (Q). The E-I ratio was then calculated as a ratio of the total charge transfer associated with the sEPSCs and sIPSCs using the equation [(QsEPSC)/(QsEPSC + QsIPSC)].
Surgery
Virus infusions
Rats were positioned in a stereotaxic frame (Kopf Instruments) and initially anesthetized with isoflurane (4% delivered at a flow rate of ∼1 L/min O2) using a calibrated vaporizer. Thereafter, anesthesia was maintained with 1%-1.5% isoflurane delivered at a flow rate of ∼0.2 L/min O2. Body temperature was maintained at 37°C with a heating pad and a sterile “tips only” technique was used for all surgeries. A 10 µl Hamilton syringe connected to an UltraMicroPump and SYS-Micro4 controller (WPI) was used to deliver 0.7 µl of AAV5-hSyn-hChR2(H134R)-eYFP (University of North Carolina vector core), over 5 min, into the medial VTA (AP: −5.4; ML: ±2.0; DV −8.2; 10° angle). After surgery, incisions were closed with absorbable sutures, and rats received an injection of the nonsteroidal anti-inflammatory meloxicam (1 mg/kg, s.c.) before being returned to their home cage where they were monitored for the next 3 d. Animals were euthanized for in vitro electrophysiology 6-8 weeks following virus infusion.
Self-administration and LHb cannulae implantation
Surgical anesthesia was achieved with equithesin (1% pentobarbital, 2% magnesium sulfate, 4% chloral hydrate, 42% propylene glycol, 11% ethanol, 3 ml/kg, i.p.), diluted 1:3-33% in saline immediately before injection to minimize peritonitis, and anesthetic depth was assessed throughout the procedure. Rats were then stereotaxically implanted with bilateral guide cannulae (C315, Plastics One) aimed 1 mm dorsal to the LHb (coordinates: AP: −3.8, L: ±0.6, V: −3.6 mm relative to bregma). Rats included in Go/NoGo intravenous self-administration (IVSA) studies were also implanted with a polyurethane catheter (Instech) that was inserted 3.5 cm into the right jugular vein. The catheter terminated in a Vascular Access Button (Instech), which was dorsally mounted subcutaneously. Rats received an injection of the nonsteroidal anti-inflammatory meloxicam (1 mg/kg, s.c.) following surgery. Per Animal Care and Use Committee policy, rats were monitored daily for signs of adynamic ileus, and no such signs were observed in any subjects. All animals resumed normal feeding behavior and demonstrated weight gain during a 1 week surgical recovery phase before training in the Go/NoGo task.
Go/NoGo cocaine IVSA task
Operant training took place as described in detail previously (Zapata et al., 2017). Standard rat operant chambers (Med-Associates) were used, and rats were trained to self-administer cocaine (0.75 mg/kg/infusion) for 12 sessions (2 h or 40 infusions per session) on a schedule in which each lever press on the active lever was followed by an injection of cocaine (fixed ratio 1 [FR1] schedule). Following this, rats were trained on a Go/NoGo task consisting of 2-h-long sessions of 6 × 20 min of alternating intervals of cocaine availability (Go) and nonavailability (NoGo), signaled by a house light on during cocaine availability, and house light off when cocaine infusions did not follow lever presses. During Go periods, every fifth response (FR5) was immediately followed by a cocaine infusion, whereas responding during NoGo periods had no consequence. Training progressed until stable discrimination of the Go/NoGo periods was observed (three consecutive sessions in which NoGo responses were <30% of total, 12-14 sessions). After reaching criterion, testing proceeded using a within subject design in which each rat received bilateral LHb infusions of PBS, scopolamine (50 mm), pirenzepine (PZP, 30 mm), or AFDX-116 (30 mm), dissolved in PBS with a 0.5 µl injection volume immediately before starting the Go/NoGo cocaine IVSA session.
Drugs
CCh, oxotremorine-M (Oxo-M), the M4-mAChR-positive allosteric modulator (PAM), VU10010, and AFDX-116 were purchased from Tocris Bioscience. DNQX and picrotoxin were obtained from Hello Bio. PZP, scopolamine, and other reagents were purchased from Sigma-Aldrich. Cocaine hydrochloride was obtained through the National Institute on Drug Abuse Drug Supply Program. TTX was obtained from Alomone Laboratories.
Data analysis
Statistical analyses were conducted using the two-tailed Student's t test and one- or two-way repeated-measures ANOVA (Prism 9, GraphPad Software). An α value of p < 0.05 was considered statistically significant. The Sidak, Tukey, or Dunnett's post hoc tests were performed to assess between group differences.
Results
Effects of cholinergic agonists on LHb neuron membrane currents
We previously reported that the muscarinic agonist Oxo-M produced heterogeneous effects on LHb neuron somatic excitability (Zapata et al., 2017). Here, we measured the effect of the nonselective cholinergic agonist CCh (10 µm) on LHb neuron membrane currents under voltage clamp (holding potential, Vhold = −60 mV). Like our previous Oxo-M data, CCh reversibly evoked outward (inhibitory) currents (n = 16 of 88 cells from 20 rats; 18.18%; Fig. 1A,B) and inward (excitatory) currents (n = 31 of 88, 35.23%; Fig. 1A,B) in neurons located in the mLHb. We additionally observed biphasic currents (brief outward followed by inward currents) in 6 of 88 (7%) of neurons. The remaining 35 cells (39.77%) showed no change in Ihold on CCh application. The plots in Figure 1B show the effect of CCh on each LHb neuron tested, together with symbols indicating the mean and 95% CI for each category. Placement into the “No Change” category occurred when the change in holding current (ΔIhold) was <10 pA, whereas cells in the remaining categories demonstrated ΔIhold >10 pA in either the inward or outward direction. CCh-activated outward currents were associated with a large reduction in input resistance (35.6 ± 10.5% of control), whereas cells displaying inward currents demonstrated a smaller reduction in input resistance (76 ± 6.4% of control). Input resistance was unchanged in cells displaying no change in Ihold (93 ± 4% of control).
Effects of ACh agonists on LHb neuron excitability measured via changes in holding current (ΔIhold). A, Time course showing the diversity of effects of CCh (10 µm, gray bar) on Ihold in all recorded LHb neurons (Vhold = –60 mV, n = 88 cells). CCh application caused inward, outward, or no change in holding currents (mean = solid line, ± SEM = black dots and shading). B, Proportions of cells demonstrating outward, inward, or no ΔIhold with CCh application. Gray lines indicate changes observed in individual neurons. Symbols represent peak mean ± 95% CI ΔIhold. Note the differences in y axis scales. C, Time course showing the concentration-dependent effects of Oxo-M (Oxo, at 10 or 3 µm application indicated by gray bar) on Ihold in all recorded LHb neurons (10 µm, n = 22 cells, 3 µm, n = 16 cells, Vhold = –60 mV). Unlike CCh, Oxo produced inward changes in Ihold in most cells (89%) at either concentration. D, Peak changes in Ihold caused by Oxo (3 and 10 µm) in all recorded LHb neurons (gray lines) and mean ± 95% CI changes for these same cells (symbols). E, Effects of mAChR antagonists on inward currents caused by Oxo (3 µm). Bars represent mean (± SEM) effect of Oxo alone and, in separate groups of LHb neurons, after preincubation with the M1-mAChR antagonist PZP (3 µm), the nonselective mAChR antagonist, scopolamine (Scop, 3 µm), or the M2-mAChR antagonist, AFDX-116 (AFDX, 1 µm). Also shown is the mean difference 95% CI estimation plot (round symbols) of effects of 3 µm Oxo after antagonist preincubation. Both Scop and AFDX significantly reduced the effect of Oxo, whereas PZP had no significant effect (one-way ANOVA, F(3,24) = 10.6, p = 0.0001; p = 0.0002, Tukey's multiple comparisons, *p = 0.011, **p = 0.0029).
In a prior study using Oxo-M under identical experimental conditions, we observed inward currents in ∼50% of LHb neurons (Zapata et al., 2017). However, in the present study, CCh-evoked inward currents were observed in only 35% of cells (Fig. 1A,B). Therefore, we compared the effects of CCh (10 µm) and Oxo-M (10 µm) in the same LHb neurons to determine whether the difference might be explained by distinct pharmacological properties of these agonists (n = 10). We found that Oxo-M produced inward currents in LHb neurons that were insensitive to CCh and this difference in current magnitude was significant (−42.9 ± 11.9 pA and −7.9 ± 4.82 pA, mean ± SEM, respectively: paired t test, t(9) = 3.460, p = 0.0072). This suggests differences in affinity or efficacy between Oxo-M and CCh at mAChRs. Therefore, to study the pharmacology of the inward currents, we used Oxo-M at a concentration (3 µm) that yielded currents similar in amplitude to that of CCh-evoked inward currents (Fig. 1C vs Fig. 1A). The 3 µm concentration of Oxo-M produced inward currents in 16 of 18 (89%) neurons tested (2 showed no change) (Fig. 1C,D). As higher levels of expression of muscarinic M2Rs have been reported in the LHb (Wagner et al., 2016), we next evaluated the effect of the M2R antagonist AFDX-116 (1 µm) on Oxo-M-induced inward currents. We found that these currents were prevented by preincubation of brain slices with AFDX-116 (Fig. 1E; one-way ANOVA, F(3,24) = 10.6, p = 0.0001; p = 0.0002, Tukey's multiple comparisons), or by the nonselective antagonist scopolamine (p = 0.0013, Fig. 1E). In an additional group, the M1R antagonist PZP slightly decreased the effect of Oxo-M, but this was not statistically significant (control Oxo-M inward current vs Oxo-M current in PZP, p = 0.06, Tukey's; Fig. 1E).
Several studies indicate that both PZP and AFDX-116 may also bind to M4-mAChRs, which could explain this discrepancy between antagonists (Bonner, 1989; Hulme et al., 1990; Kashihara et al., 1992; Caulfield and Birdsall, 1998; Dasari and Gulledge, 2011). Therefore, we investigated whether Oxo-M-activated inward currents might involve M4Rs by testing Oxo-M alone (3 µm), or after preincubation with the M4R-PAM VU10010 (3 µm) (Shirey et al., 2008). We found that the effect of Oxo-M was unchanged by preincubation with VU10010 (mean and 95% CI, Oxo-M alone = −22.1 pA, −30.7 pA and −13.5 pA; Oxo-M with VU10010 preincubation = −26.47 pA, −45.8 and −7.85 pA; unpaired t test, t(22) = 0.55, p = 0.588, data not shown). To ensure that the concentration of VU10010 used in this experiment was sufficient to increase M4R function, we measured its effects on the inhibition of Schaffer collateral glutamatergic field excitatory postsynaptic potentials (fEPSPs) in the CA1 region of hippocampal slices, where this has been previously demonstrated (Shirey et al., 2008). We found that the inhibition of hippocampal fEPSPs by CCh (300 nm) was significantly increased by 3 µm VU10010 (mean and 95% CI, CCh alone = −32.04%, −11.5 and −52.6%; CCh with VU10010 preincubation = −59.3%, −32.98 and −82.59%; paired t test = t(4) = 4.40, p = 0.012, n = 5 slices, 3 rats, data not shown). Importantly, the M4-mediated inhibition of hippocampal fEPSPs by CCh (10 µm) was unaffected by the M2R antagonist AFDX-116 (1 µm; mean and 95% CI, CCh alone = −81.44%, −83.76 and −79.12%; CCh + AFDX-116 = −76.65%, −81.32 and −77.98%; paired t test = t(5) = 1.507, p = 0.192, n = 7 slices, 4 rats, data not shown), indicating that M4Rs were not blocked by this antagonist. Therefore, the results of our pharmacological analysis indicate that Oxo-M-activated inward currents in LHb neurons were likely mediated by M2Rs, and not by M1Rs or M4Rs.
The small number of CCh-evoked outward and biphasic currents limited our ability to study the receptors involved in these responses. However, we did find that PZP did not reverse outward currents in 6 cells (not shown).
CCh inhibits GABAergic synaptic input to LHb neurons via M2Rs
To determine whether mAChRs also control synaptic integration in the LHb, we measured the effects of CCh on spontaneously occurring GABAergic IPSCs. Bath application of CCh (10 µm) significantly reduced the frequency of sIPSCs (Fig. 2A,C; n = 13 cells from 4 rats, two-tailed paired t test, t(12) = 4.6, p = 0.0006) but had no effect on the mean amplitudes of these events (Fig. 2B; two-tailed paired t test, t(12) = 0.76, p = 0.46). This effect of CCh on sIPSC frequency was not reversed by PZP (1 µm) (Fig. 2D,E; n = 7 cells from 4 rats; one-way repeated-measures ANOVA, F(2,12) = 7.29, p = 0.0085; CCh, p = 0.005 vs control; CCh + PZP, p = 0.04 vs control, Dunnett's post hoc). However, AFDX-116 reversed the reduction in sIPSC frequency caused by CCh (Fig. 2F,G; one-way repeated-measures ANOVA, F(2,12) = 8.71, p = 0.005; AFDX-116, p = 0.15 vs control, Dunnett's post hoc). IPSCs evoked by local electrical stimulation (eIPSCs) of the brain slice were also significantly inhibited by CCh (Fig. 3A,B; one-way repeated-measures ANOVA, F(2,18) = 13.22, p = 0.003; p = 0.002 vs control, Dunnett's post hoc), and this was reversed by AFDX-116 as well (p = 0.24 vs control, Dunnett's post hoc). The effects of CCh on eIPSCs did not differ by sex (males, n = 9, 44 ± 6% inhibition; females, n = 5, 31 ± 7% inhibition; two-tailed t test, t(12) =1.4, p = 0.18). These data suggest that GABAergic inputs to LHb neurons are equally inhibited by presynaptic M2Rs in male and female rats.
CCh inhibition of GABAergic transmission in LHb neurons via M2-mAChRs. A, Top, Representative traces of sIPSCs under control conditions (Vhold = –70 mV), and during application of CCh (10 µm). Bottom, Representative cumulative sIPSC amplitude distributions (left) and cumulative sIPSC interevent interval distributions (right) in control and CCh periods in the same cell. B, Estimation plot of sIPSC amplitude for all cells under control conditions and during CCh application. The difference between control and CCh periods (CCh-control) in each cell is shown at right along with the mean difference ± 95% CI, and p value. CCh did not alter sIPSC amplitudes (two-tailed paired t test, t(12) = 0.76, p = 0.46). C, Estimation plot of sIPSC frequency in all LHb cells. CCh significantly inhibited sIPSC frequency (two-tailed paired t test, t(12) = 4.6, p = 0.0006). D, Representative sIPSC traces from a single LHb neuron during the control period, during CCh application, and during coapplication of CCh and the M1-mAChR antagonist, PZP (CCh/PZP,1 µm). E, Top, Group effects of CCh and CCh/PZP on sIPSC frequency in LHb neurons. PZP did not significantly reverse the effect of CCh on sIPSC frequency (one-way repeated-measures ANOVA, F(2,12) = 7.29, p = 0.0085; CCh, **p = 0.005 vs control; CCh/PZP, *p = 0.04 vs control, Dunnett's post hoc). Bottom, Mean difference estimation plot of effects of CCh and CCh+PZP on sIPSC frequency for all cells. The difference between control and CCh periods and control and CCh/PZP periods is expressed as the mean difference ± 95% CI. F, Representative sIPSC traces from an LHb neuron during control, CCh (10 µm) application, and during coapplication of CCh and the M2-mAChR antagonist, AFDX-116 (CCh/AFDX, 1 µm). G, Top, Group effects of CCh and CCh/AFDX on sIPSC frequency in LHb neurons. AFDX-116 significantly reversed the effect of CCh on sIPSC frequency (one-way repeated-measures ANOVA, F(2,12) = 8.71, p = 0.005; CCh, **p = 0.009 vs control; CCh/AFDX, p = 0.15 vs control Dunnett's post hoc). Bottom, Mean difference estimation plot of effects of CCh and CCh/AFDX on sIPSC frequency. The difference between control and CCh periods and control and CCh/AFDX periods is expressed as the mean difference ± 95% CI.
CCh inhibits evoked IPSCs through M2Rs in LHb. A, Representative traces of eIPSCs (Vhold = −45 mV, K-Gluconate intracellular solution) showing inhibition during CCh application (10 µm) and reversal by coapplication of the M2R antagonist AFDX-116 (1 µm). B, Mean (± SEM) time course showing the effects of CCh and AFDX-116 on eIPSCs in LHb neurons (n = 10 cells from 6 rats). Significant inhibition was observed following CCh (one-way repeated-measures ANOVA, F(2,18) = 13.22, p = 0.003; p = 0.002 vs control, Dunnett's post hoc), and this was reversed by AFDX-116 (p = 0.24 vs control, Dunnett's post hoc). C, Representative traces of light-evoked oIPSCs recorded from an LHb neuron following infusion of synapsin-driven ChR2 into the VTA. Dashed line indicates the onset of a single, 5 ms duration light pulse (473 nm). D, Mean time course of the effects of CCh and AFDX-116 on VTA to LHb oIPSCs (n = 5 cells from 2 rats). Significant inhibition was observed during CCh (one-way repeated-measures ANOVA, F(2,8) = 16.6, p = 0.001; p = 0.008 vs control, Dunnett's post hoc), and this was reversed by AFDX-116 (p = 0.09 vs control, Dunnett's post hoc).
Inhibition of synaptic GABAergic VTA inputs to LHb by M2Rs
As noted above, the LHb contains relatively few intrinsic sources of GABAergic inhibition (Brinschwitz et al., 2010; Aizawa et al., 2012; Wagner et al., 2017). Therefore, control of LHb neuron excitation likely arises from a diverse array of inhibitory afferents, including those from the VTA (Root et al., 2014b). As VTA neurons respond to both rewarding and aversive stimuli (Root et al., 2020) and express M2R mRNA (Vilaro et al., 1992b), we hypothesized that this pathway would be sensitive to M2R-mediated synaptic control. Rats (n = 5) received VTA injections with AAV encoding for the channelrhodopsin-2 (ChR2) protein, driven by a synapsin promoter (pAAV-hSyn-hChR2(H134R)-EYFP), and light-evoked oIPSCs were recorded in LHb neurons 6-8 weeks later. The oIPSCs evoked by activation of ChR2 in VTA neuron axon terminals in the LHb were significantly inhibited by CCh (10 µm; Fig. 3C,D), and this was reversed by coapplication of AFDX-116 (one-way repeated-measures ANOVA, F(2,8) = 16.6, p = 0.001; CCh, p = 0.008 vs control; CCh + AFDX, p = 0.09 vs control, Dunnett's post hoc). These data suggest that VTA GABAergic inputs to LHb neurons are inhibited by M2R activation.
CCh inhibits glutamatergic synaptic input to LHb neurons via M2Rs
We next examined whether synaptic glutamate input to LHb neurons is modulated by mAChRs by recording sEPSCs. Unlike sIPSCs, CCh (10 µm) significantly decreased sEPSC amplitude (Fig. 4A,C–E; p = 0.0002, t(31) = 4.3, two-tailed paired t test), as well as sEPSC frequency (Fig. 4B,F; p = 0.0002, t(31) = 4.3, two-tailed paired t test; n = 32 neurons from 18 animals), and this was independent of sex (two-way ANOVA; sex × treatment interaction, F(1,32) = 0.3023, p = 0.58). Given that CCh hyperpolarizes some LHb neurons (Fig. 1), the reduction in sEPSC amplitude could reflect decreased somatic excitability of local glutamatergic LHb neurons that provide synaptic inputs to other LHb cells (Kim and Chang, 2005), as described for µ-opioids in the LHb (Margolis and Fields, 2016). Therefore, to limit the influence of somatic excitability changes on glutamate release, we measured CCh effects on mEPSCs following blockade of action potentials by CCh (200 nm). In the presence of TTX, CCh reduced mEPSC frequency (n = 9 cells from 3 rats; control, 3.3 ± 0.39 Hz; TTX, 1.9 ± 0.23 Hz; Fig. 4H; two-tailed paired t test, t(10) = 5.36, p = 0.003) but had no effect on mEPSC amplitude (control, 25.6 ± 4.1 pA; TTX, 24.9 ± 5.4 pA; Fig. 4G; two-tailed paired t test, t(10) = 0.51, p = 0.62). Additionally, noncumulative amplitude histograms indicated that CCh eliminated larger sEPSCs but also strongly reduced sEPSCs across the full range of amplitudes (Fig. 4D). These results suggest that CCh can reduce synaptic glutamate release onto LHb neurons via changes in somatic excitability of local presynaptic neurons, as well as by direct inhibition via effects at axon terminals.
CCh inhibits synaptic glutamate release onto LHb neurons. A, Representative sEPSC traces before and during bath application of CCh (10 µm). B, Representative cumulative distribution histograms of inter-sEPSC interval and (C) amplitude from the same cell shows a reduction in both measures by CCh. D, A noncumulative amplitude histogram from this cell shows that CCh eliminated large sEPSCs and reduced the number of sEPSCs across the full range of amplitudes. Inset, Averaged sEPSCs before and during CCh, as well as the CCh waveform scaled to the peak of the control. No change in sEPSC kinetics was observed. E, Changes in mean sEPSC amplitude for all cells (left, y axis) and estimation plot of the mean difference ± 95% CI in this measure (CCh-control) from control to CCh periods (right, y axis). CCh caused a significant reduction in sEPSC amplitude (two-tailed paired t test, t(31) = 4.3, p = 0.0002). F, Estimation plot of sEPSC frequency for each LHb neuron. A significant reduction in sEPSC frequency was observed following CCh application (two-tailed paired t test, p = 0.0002, t(31) = 4.3, p = 0.0002). G, The amplitude of TTX-resistant (200 nm) action potential-independent mEPSCs was not significantly altered by CCh (two-tailed paired t test, t(10) = 0.51, p = 0.62). H, mEPSC frequency measured in TTX was significantly reduced by CCh (two-tailed paired t test, t(10) = 5.36, p = 0.003). This indicates that CCh reduces sEPSC amplitude by decreasing action potential-dependent glutamate release but also inhibits TTX-resistant release.
Whereas PZP (1 µm) did not affect the inhibition of sEPSC frequency by CCh (Fig. 5A–E; one-way repeated-measures ANOVA, F(2,12) = 6.396, p = 0.01; **p < 0.01, Dunnett's post hoc), AFDX-116 (1 µm) reversed the CCh-mediated inhibition of sEPSC amplitude and frequency (Fig. 5F–J; one-way repeated-measures ANOVA, F(2,14) = 4.48, p = 0.03; CCh, p = 0.02 vs control; CCh + AFDX, p = 0.6 vs control, Dunnett's post hoc; one-way repeated-measures ANOVA, F(2,14) = 8.23, p = 0.005; CCh, p = 0.003 vs control; CCh +AFDX, p = 0.15 vs control, Dunnett's post hoc). Thus, M2Rs, and not M1Rs, inhibit synaptic glutamate release onto LHb neurons.
CCh inhibits glutamate release via M2-mAChRs in the LHb. A, Cumulative sEPSC interval and amplitude (B) distributions from an LHb neuron showing the effect of CCh (10 µm) and PZP (1 µm) added during CCh application. CCh did not alter the amplitude distribution in this neuron but clearly increased the sEPSC interval. C, Representative sEPSC traces from the same cell represented in A and B. D, Mean sEPSC amplitudes for each cell under each condition are shown along with the mean differences (±95% CI) in treatment versus control. Neither CCh nor PZP significantly altered sEPSC amplitude (one-way repeated-measures ANOVA, F(2,12) = 2.351, p = 0.14). E, Effect of CCh and PZP on sEPSC frequency. sEPSC frequency was significantly reduced by CCh, but this was not reversed by PZP (one-way repeated-measures ANOVA, F(2,12) = 6.396, p = 0.01; *p < 0.01, Dunnett's post hoc vs control). F, Cumulative distribution of sEPSC amplitude and sEPSC interval (G). Note the reversal of the effect of CCh by AFDX application. H, Representative sEPSC traces obtained under control conditions and during successive application of CCh and AFDX-116 (AFDX, 1 µm). I, Mean sEPSC amplitudes for each cell. The mean amplitude was significantly reduced (one-way repeated-measures ANOVA, F(2,14) = 4.48, p = 0.03; *p = 0.02 vs control, Dunnett's post hoc), and this was reversed by AFDX (p = 0.6, Dunnett's post hoc). The mean difference (from control) and 95% CI are shown below. J, sEPSC frequency was significantly reduced by CCh (one-way repeated-measures ANOVA, F(2,14) = 8.23, p = 0.005; **p = 0.003 vs control, Dunnett's post hoc), and this was reversed by AFDX (p = 0.15 vs control, Dunnett's post hoc).
M2Rs alter the excitatory-inhibitory balance of the LHb
The preceding data indicate that both excitatory and inhibitory synaptic inputs to LHb neurons are inhibited by M2Rs. As an altered E-I balance in the LHb has been associated with a net change in neuronal activity and with dysfunctional reward valence attribution (Shabel et al., 2014; Meye et al., 2016; Mori et al., 2019), we next determined the effect of M2R activation on LHb E-I balance (Antoine et al., 2019). CCh significantly reduced the E-I ratio measured using eEPSCs and eIPSCs in the same LHb neurons (Fig. 6A,B; n = 15 neurons from 7 animals; paired t test, t(14) = 2.641, p =0.0194). This change in the E-I ratio was also prevented by pretreatment with AFDX-116 (1 µm; 13 neurons from 6 animals; paired t test, t(12) = 0.1294, p = 0.899).
CCh reduces the LHb E-I balance via activation of M2R-mAChRs. A, Signal averages of eEPSCs (inward currents, downward) obtained at –70 mV, and eIPSCs (outward currents, upward) obtained at 0 mV in the same cell using the same electrical stimulus intensity. Traces were obtained before (control, gray) and during CCh (10 µm) application (blue traces). B, E-I ratios for each LHb neuron in control and CCh conditions superimposed on mean E-I ratios (± SEM, bars). Right, Estimation plot showing the difference in E-I ratio for each cell obtained by subtracting the control E-I ratio from the CCh E-I ratio (CCh-Control). Also shown is the mean ± 95% CI of these differences. CCh significantly decreased the E-I ratio (t(14) = 2.641, df = 14, p = .0194, paired t test). C, eIPSC and eEPSC currents obtained during control conditions (gray traces) and during CCh application (red traces) in cells preincubated with the M2R antagonist AFDX-116 (1 µm). D, Control and CCh E-I ratios for each AFDX-pretreated LHb neuron superimposed on mean E-I ratios (± SEM, bars). Right, Estimation plot showing the difference in E-I ratio for each cell obtained by subtracting control E-I ratio from CCh/AFDX E-I ratio (CCh/AFDX-Control), and the mean ± 95% CI of these differences. CCh did not significantly change the E-I ratio in cells treated with AFDX (t(12) = 0.1294, p = 0.899, paired t test). E, sIPSC sweeps obtained at 0 mV holding potential (top) and sEPSCs obtained at −70 mV (bottom) from the same LHb cell before (Control) and during CCh application (10 µm). F, E-I ratios obtained from the sIPSCs and sEPSCs in each cell before and during CCh treatment superimposed on the mean ± SEM E-I ratios (bars). Right, Estimation plot of these E-I ratio differences for each cell (CCh-Control), and the mean ± 95% CI of this difference. CCh significantly reduced the E-I ratio (t(13) = 2.643, p = 0.0203, paired t test). G, Examples of sIPSCs obtained at 0 mV holding potential (top) and sEPSCs obtained at −70 mV (bottom) from the same LHb cell preincubated with AFDX (1 µm). H, E-I ratios obtained from sIPSCs and sEPSCs in each cell before and during CCh treatment in cells preincubated with AFDX (1 µm). This is superimposed on the mean ± SEM E-I ratios (bars). Right, Estimation plot of the differences in these E-I ratios in each cell during control and CCh/AFDX periods (CCh/AFDX-Control), and the mean ± 95% CI of these differences. CCh did not significantly change the E-I ratio in AFDX preincubated cells (t(12) = 0.001432, p = 0.993, paired t test).
As measurement of E-I balance can be affected by differing relative degrees of sensitivity to electrical stimulation, we also measured E-I balance using sEPSCs (−70 mV holding potential) and sIPSCs (0 mV holding potential) in the same LHb neurons under control conditions and during CCh or CCh + AFDX-116 application (Fig. 6E–H). Similar to results with electrically evoked synaptic currents, we found that the E-I ratio of the spontaneous currents was significantly reduced by CCh (Fig. 6E,F; t(13) = 2.643, p = 0.0203, paired t test), and that this was prevented by preincubation of brain slices in 1 µm AFDX-116 (Fig. 6G,H; t(12) = 0.0014, p = 0.993, paired t test). Thus, whereas both excitatory and inhibitory inputs to LHb neurons using these two methods are reduced by M2-mAChR activation, excitatory synaptic inputs are more strongly inhibited, thereby resulting in a net shift in E-I balance toward synaptic inhibition of LHb neurons.
Intra-LHb blockade of M2Rs promotes impulsive cocaine-seeking
We previously reported that blockade of LHb mAChRs by the nonselective antagonist scopolamine impaired operant response inhibition for cocaine in rats trained on a Go/NoGo task (Zapata et al., 2017). This suggests that intact mAChR signaling is necessary to inhibit cocaine seeking in this task. Moreover, our present electrophysiological results show that M2Rs can alter LHb neuron membrane potential and synaptic integration in these cells. Therefore, we next asked whether M2Rs were involved in inhibition of cocaine seeking by comparing the effects of PZP and AFDX-116 in the Go/NoGo paradigm. Rats trained to self-administer cocaine and to withhold responding during signaled drug absence received bilateral infusions of vehicle, scopolamine (50 mm), PZP (30 mm), or AFDX-116 (30 mm) before test sessions. Whereas none of these antagonists significantly altered cocaine seeking during Go intervals (Fig. 7A–C; two-way repeated-measures ANOVA, Go/NoGo vs scopolamine; F(1,8) = 8.264, p =0.02; scopolamine-Go, p =0.71, Sidak's; Go/NoGo vs AFDX; F(1,8) = 8.406, p = 0.02; AFDX-Go, p =0.99, Sidak's; two-way repeated-measures ANOVA, Go/NoGo vs PZP; F(1,8) = 0.3536, p =0.57; PZP-GO, p =0.99, Sidak's), infusion of either scopolamine or AFDX-116 significantly increased responding for cocaine during NoGo periods (Fig. 7A,B, two-way repeated-measures ANOVA, drug × Go/NoGo period interaction, p = 0.0207 and p = 0.0199, respectively, Sidak's). In contrast, PZP did not significantly alter response inhibition for cocaine during NoGo periods (Fig. 7C; two-way repeated-measures ANOVA, Go-NoGo × drug interaction, F(1,8) = 0.3536, p = 0. 569). These data show that M2Rs are critical for response inhibition in this Go/NoGo task.
Blockade of LHb M2-mAChRs impairs inhibition of responses for cocaine in a Go/NoGo cocaine self-administration task. A, Mean (± SEM) number of operant responses emitted during signaled cocaine availability (Go) and nonavailability (NoGo) following infusion of either vehicle (Veh) or the nonselective mAChR antagonist scopolamine (Scop, 50 mm) into the LHb. Also shown are the number of responses observed for each rat in these conditions (circles) and, below, mean ± 95% CI estimation plots of the differences between Veh and drug during Go and NoGo periods. Responding during NoGo periods was significantly increased by infusion of Scop into the LHb (two-way repeated-measures ANOVA, Go/NoGo × drug interaction, F(1,8) = 8.264, p = 0.0207, *p = 0.0219, Sidak's post hoc). However, Scop did not significantly affect Go responding for cocaine (p = 0.71, Sidak's). B, The M2-mAChR antagonist AFDX-116 (AFDX, 30 mm) significantly increased NoGo responding without affecting Go responding (two-way repeated-measures ANOVA, Go-NoGo × drug interaction, F(1,8) = 8.406, p = 0.0199; NoGo responses, **p = 0.0064, Go responses, p = 0.9984, Sidak's post hoc). Below is shown the mean ± 95% CI estimation plot for the effect of AFDX. C, Infusion of the M1-mAChR antagonist PZP into the LHb did not significantly alter Go/NoGo responding for cocaine (two-way repeated-measures ANOVA, Go-NoGo × drug interaction, F(1,8) = 0.3536, p = 0.685).
Discussion
Here we describe robust modulation of neurons in the parvocellular and central subnuclei of the mLHb by mAChRs and show that the central cholinergic system, acting through M2Rs, is critical for the withholding of operant responding for cocaine. We also find that synaptic GABA and glutamate inputs to LHb neurons are inhibited by M2Rs, and with optogenetics we demonstrate that VTA to LHb afferents represent at least one M2R-sensitive inhibitory synaptic pathway. Although we demonstrate that M2Rs can inhibit both glutamate and GABA release onto LHb neurons, we also find that the net effect of M2R activation was to bias the E-I ratio of the LHb toward enhanced inhibition. Based on this, we predict that M2R activation will reduce excitatory LHb output to downstream targets.
In addition to the altered synaptic integration by M2Rs, we found that mAChRs depolarize LHb neurons via activation of these receptors. Consistent with these physiological effects of M2Rs in LHb neurons, our behavioral experiment suggests that endogenous ACh acts via LHb M2Rs to enable suppression of responding for cocaine in a task in which rats had learned to withhold responses when the drug was not available. Although this implicates LHb ACh and M2Rs in response inhibition, our experiments do not permit the identification of the presynaptic or postsynaptic sites of M2R involvement in this model of impulsive behavior. Despite this, response inhibition tasks are used extensively to identify brain circuits underlying impulsive behavior and have translational value in humans (Smith et al., 2014). Therefore, our study suggests that LHb M2Rs could represent a translational target for therapeutics in the treatment of impulsive behavior, particularly as it relates to addiction.
We previously reported that Oxo-M depolarized ∼50% of LHb neurons and hyperpolarized another 10% of these cells (Zapata et al., 2017). Here, we used CCh as an agonist because it exhibits faster pharmacokinetic properties in brain slices. However, we found that CCh initiated inward currents in fewer LHb neurons (31 of 88, 35%), suggesting slightly different pharmacological properties of these agonists. Consistent with this, we observed that some neurons unresponsive to CCh demonstrated robust inward currents with Oxo-M. Therefore, we performed the remainder of the postsynaptic experiments with Oxo-M to characterize mAChRs mediating direct excitation of LHb neurons. We used PZP, considered a relatively selective M1R antagonist (Doods et al., 1987; Buckley et al., 1989; Valuskova et al., 2018), and AFDX-116, an antagonist with high affinity for M2Rs (Buckley et al., 1989; Regenold et al., 1989; Lai et al., 1990; Billard et al., 1995). However, additional evidence shows that PZP binds to other mAChRs, including M4Rs (Bonner, 1989; Hulme et al., 1990; Kashihara et al., 1992; Caulfield and Birdsall, 1998; Dasari and Gulledge, 2011). Similarly, AFDX-116 is reported to have affinity for receptors other than M2Rs, including M4Rs (Buckley et al., 1989; Levey et al., 1991; Kashihara et al., 1992; Vilaro et al., 1992a). Because of this, we also examined whether the M4R PAM, VU10010 (Shirey et al., 2008), would alter the effects of Oxo-M and perhaps reveal the participation of M4Rs in regulating LHb neuron excitability. However, in contrast to its ability to augment the effects of CCh in the hippocampus, where M4-mAChRs are known to inhibit glutamate release (Shirey et al., 2008; Thorn et al., 2017), we found that VU10010 had no effect in the LHb. Therefore, this comparative pharmacology shows that AFDX-116 was more efficacious than PZP at reversing inward currents elicited by Oxo-M, and that the effects of Oxo-M are not potentiated by an M4R PAM. These data, together with the observation that M2Rs are expressed at moderate to high levels in the LHb (Spencer et al., 1986; Wagner et al., 2016), indicate that the M2R is the primary mAChR subtype involved in LHb cholinergic signaling.
In contrast to the heterogeneity of effects on somatic excitability, CCh consistently inhibited glutamatergic and GABAergic synaptic inputs to LHb neurons. These LHb output neurons are largely glutamatergic and receive synaptic glutamate input from both extrinsic sources and local recurrent collaterals (Kim and Chang, 2005; Weiss and Veh, 2011). Moreover, there are few intrinsic sources of inhibitory synaptic input to these cells (Brinschwitz et al., 2010; Aizawa et al., 2012; Wagner et al., 2017; Webster et al., 2021), suggesting that inhibitory control of LHb output may arise largely from extrinsic sources. We found that CCh inhibited glutamatergic synaptic sEPSC frequency and amplitude in LHb neurons. As there is evidence for local connectivity among these glutamatergic cells (Kim and Chang, 2005), and we found that some LHb neurons are directly inhibited by CCh, we reasoned that the CCh inhibition of sEPSC amplitude might reflect somatic hyperpolarization, as described for µ-opioid receptors in the LHb (Margolis and Fields, 2016). Consistent with this, blockade of action potential-dependent glutamate release with TTX eliminated the effect of CCh on sEPSC amplitude, but not frequency, suggesting that the reduction in somatic excitability contributed to sEPSC amplitude changes. Moreover, as the decrease in sEPSC frequency and amplitude by CCh was blocked by AFDX-116 (Hulme et al., 1990), it is likely that M2Rs influence synaptic glutamate release by actions on both somatic excitability and on the axon terminal glutamate release process.
In contrast to its effects on sEPSCs, CCh inhibited sIPSC frequency without changing sIPSC amplitudes. This was also reversed by AFDX-116, but not by PZP, suggesting that M2Rs on inhibitory axon terminals decrease GABA release probability in LHb neurons, and that M1Rs are not involved. Additionally, we found that electrically evoked IPSCs, arising from GABAergic axons from undefined projections, were inhibited by M2Rs, as were those evoked by specific light activation of ChR2 expressed in VTA afferents to LHb. The input from the VTA represents a major inhibitory afferent that can strongly limit LHb neuron output (Shabel et al., 2014; Barker et al., 2017; Faget et al., 2018), and our study indicates that the inhibition of LHb neurons provided by the VTA may be dampened by endogenous ACh acting at M2Rs. Beyond our physiological identification of this M2R modulation, other studies show moderate levels of M2R expression in the LHb, and M2R mRNA in VTA projections to LHb (Vilaro et al., 1992b; Wagner et al., 2016). However, as the majority of VTA neurons that project to the mLHb synaptically corelease GABA and glutamate (Root et al., 2014b); and can encode both reward and aversion (Root et al., 2020), future studies should evaluate M2R involvement on behavioral measures involving this pathway.
Since our study shows that M2Rs inhibit both glutamatergic and GABAergic synaptic input to LHb neurons, one might ask what the net of effect of this dual modulation might be? To address this, we examined CCh effects on E-I balance by measuring electrically evoked and sEPSCs and sIPSCs in the same LHb neurons. Although CCh significantly decreased both GABA and glutamate release, EPSCs were more strongly inhibited, resulting in a net decrease in the E-I balance. This suggests that the overall effect of M2R activation is to increase the relative inhibition of LHb neurons. Within the context of the indirect influence the LHb has on midbrain DA neurons via its output to the RMTg (Jhou et al., 2009), these data would predict that activation of LHb M2Rs might dampen the excitation of the RMTg and hence reduce inhibition of DA neurons to increase their excitability. However, it should also be noted that in vitro measurements of E-I balance occur without consideration of the relative activity and contributions of these pathways during behavior. Therefore, E-I balance measured in vitro provides a somewhat static estimate of the contribution of M2Rs to LHb output, and its contributions to behavior and additional in vivo studies are required to more fully understand the contributions of mAChR-regulated neurotransmitter release to behavior. To this end, our present studies measuring response inhibition of cocaine seeking provides insight into a behavioral role for M2Rs. Here, our results show that the withholding of responses for cocaine is impaired during LHb M2R blockade, suggesting that endogenous ACh and M2Rs are involved in control of the impulse to seek the drug.
We previously found that, unlike blockade of mAChRs, antagonism of LHb ionotropic glutamatergic receptors, β2-subunit containing nicotinic receptors, D1 and D2 DA receptors, and serotonin 5-HT2 and 5-HT3 receptors, did not affect inhibition of responding for cocaine (Zapata et al., 2017). Therefore, these behavioral data imply that the inhibition of glutamate release by M2Rs in LHb, as demonstrated in the present study, and these other neurotransmitter systems are not likely involved in control of impulsive cocaine seeking in this model of response inhibition. This suggests a more important role for M2Rs on inhibitory LHb-projecting neurons.
In conclusion, the results of the present study, and our previous work, suggest that muscarinic cholinergic signaling may have an important influence on LHb function that contributes to the control of impulsive behavior involved in drug seeking, and may also be relevant to substance use, and psychiatric disorders involving impulse control deficits.
Footnotes
This work was supported by the National Institutes of Health and the National Institute on Drug Abuse Intramural Research Program, Project #1ZIADA000457 to C.R.L. E.C.I. was supported by the National Institute on Drug Abuse IRP Scientific Director's Fellowship for Diversity in Research.
The authors declare no competing financial interests.
- Correspondence should be addressed to Carl R. Lupica at clupica{at}mail.nih.gov
