Chronic Intermittent Ethanol Exposure Dysregulates Nucleus Basalis Magnocellularis Afferents in the Basolateral Amygdala

Abstract Nucleus basalis magnocellularis (NBM) cholinergic projections to the basolateral amygdala (BLA) regulate the acquisition and consolidation of fear-like and anxiety-like behaviors. However, it is unclear whether the alterations in the NBM-BLA circuit promote negative affect during ethanol withdrawal (WD). Therefore, we performed ex vivo whole-cell patch-clamp electrophysiology in both the NBM and the BLA of male Sprague Dawley rats following 10 d of chronic intermittent ethanol (CIE) exposure and 24 h of WD. We found that CIE exposure and withdrawal enhanced the neuronal excitability of NBM putative “cholinergic” neurons. We subsequently used optogenetics to directly manipulate NBM terminal activity within the BLA and measure cholinergic modulation of glutamatergic afferents and BLA pyramidal neurons. Our findings indicate that CIE and withdrawal upregulate NBM cholinergic facilitation of glutamate release via activation of presynaptic nicotinic acetylcholine receptors (AChRs). Ethanol withdrawal-induced increases in NBM terminal activity also enhance BLA pyramidal neuron firing. Collectively, our results provide a novel characterization of the NBM-BLA circuit and suggest that CIE-dependent modifications to NBM afferents enhance BLA pyramidal neuron activity during ethanol withdrawal.


Introduction
Alcohol use disorder (AUD) is a chronic brain disease characterized by cycles of excessive drinking, abstinence, and relapse. Chronic alcohol consumption causes structural and functional changes in stress and reward pathways that promote negative affect during abstinence (Koob and Volkow, 2016). Relapse is commonly motivated by a desire to alleviate the anxiety and craving associated with alcohol withdrawal (WD; Sliedrecht et al., 2019). Since anxiety during abstinence strongly correlates with the risk of relapse (Sliedrecht et al., 2019), understanding the neurophysiology of withdrawal-induced negative affect is essential to mitigate the incidence of relapse.
The basolateral amygdala (BLA) is a node of affective processing that regulates behavioral responses to emotionally salient stimuli. BLA pyramidal neurons receive sensory and cognitive-related information via glutamatergic afferents entering the region through distinct input pathways. One example is the stria terminalis (ST), a pathway that contains glutamatergic afferents from the medial prefrontal cortex and various thalamic nuclei (Ray and Price, 1992). Glutamate release from the ST drives BLA pyramidal neuron excitability and ultimately the activity of downstream efferent reward and aversion circuits. Chronic ethanol exposure and withdrawal increase glutamatergic signaling in the BLA, facilitate BLA pyramidal neuron excitability and ultimately the expression of anxiety-related and reward-related behaviors (Morales et al., 2018;McGinnis et al., 2020a,b;Price and McCool, 2022). The present study suggests that dysregulation of nucleus basalis magnocellularis (NBM) afferents potentiate BLA pyramidal neuron activity during withdrawal.
The NBM is part of the basal forebrain cholinergic system that projects throughout the cortex, hippocampus, and amygdala. Notably, the NBM is the primary source of cholinergic innervation in the BLA (Carlsen et al., 1985;Aitta-Aho et al., 2018); and, the projections are among the most dense cholinergic inputs in the CNS (Ben-Ari et al., 1977). Acetylcholine (ACh) released by NBM cholinergic projections binds to nicotinic and muscarinic ACh receptors (n/mAChRs) expressed by ST afferents, local GABAergic interneurons, and BLA pyramidal neurons (Zhu et al., 2005;Pidoplichko et al., 2013;Lee and Kim, 2019). NBM cholinergic input is thus poised to modulate BLA output by facilitating neurotransmitter release and fast synaptic transmission via ionotropic nAChRs and metabotropic mAChRs. NBM projections may also contain GABAergic and glutamatergic components (Gritti et al., 2006). One-third of NBM-BLA ChAT 1 terminals express the vesicular glutamate transporter 3 (vGLUT3; Nickerson Poulin et al., 2006), which suggests a potential for the corelease of acetylcholine and glutamate from these synapses (Higley et al., 2011;Nelson et al., 2014). Additionally, ;10% of NBM projections to the BLA are GABAergic (Mascagni and McDonald, 2009) and appear to synapse directly onto both BLA pyramidal neurons and local GABAergic interneurons (McDonald et al., 2011). However, the physiological significance of these NBM GABAergic projections and the potential for glutamate co-release from cholinergic terminals within BLA remains unclear.
Behavioral and electrophysiology studies illustrate that the NBM-BLA circuit facilitates learning by controlling the consolidation of salient cues. For example, NBM cholinergic neurons fire rapidly in response to appetitive or aversive stimuli (Zhang et al., 2004;Hangya et al., 2015;Rajebhosale et al., 2021) and cause transient and scaled increases in BLA acetylcholine levels (Mark et al., 1996;Jing et al., 2020;Liu et al., 2021). Optogenetic stimulation of ChAT 1 terminal fields in the BLA enhances cue-reward learning (Crouse et al., 2020) and prolongs fear extinction (Jiang et al., 2016;Kellis et al., 2020). These preclinical data suggest that acetylcholine strengthens the acquisition of fear-associated and reward-associated memories by enhancing BLA pyramidal neuron activity. Electrophysiology studies support these findings and show that endogenous acetylcholine release increases in vivo and ex vivo neuron firing (Jiang et al., 2016), the signal-to-noise ratio (Unal et al., 2015), and elicits long-term potentiation (LTP; Jiang et al., 2016) in BLA pyramidal neurons. However, no studies have examined withdrawal-induced alterations in the NBM-BLA circuit and the subsequent effects on glutamatergic signaling in the BLA.
Previous findings suggest that chronic intermittent ethanol (CIE) exposure, a model that produces dependence and enhances anxiety-like behaviors in rodent models (Morales et al., 2018), robustly elevates BLA acetylcholine levels. Therefore, we hypothesized that these findings reflect a potentiation of NBM cholinergic projections to the BLA during withdrawal. In the present study, we use optogenetics and electrophysiology to understand how CIE exposure and withdrawal alter NBM cholinergic neuron excitability and measure the neurophysiological outcomes in the BLA. Our results indicate that CIE exposure and withdrawal enhances the excitability of NBM "cholinergic" neurons and increases NBM cholinergic neurotransmission in the BLA.

Animals
Adolescent male Sprague Dawley rats were purchased from Envigo and arrived at the Wake Forest University Medical School between postnatal day (P) 35 and P40 (;110-125 g). Lights in the animal facility were configured to a reverse 12/12 h light/dark cycle where lights automatically turned off between 9 A.M. and 9 P.M. Rodents were given ad libitum access to standard rat chow (LabDiet) and water. Animal welfare was monitored daily by laboratory staff and the Animal Resource Program (ARP). All experimental procedures were preapproved by the Institutional Animal Care and Use Committee (IACUC) at Wake Forest University School of Medicine.
Rodents were given 2-ml warmed sterile saline (0.9% sodium chloride injection; Hospira) and an anti-inflammatory drug (1 mg/kg Meloxicam). After the incision was sutured and secured with skin adhesive (Skin Affix), rodents regained consciousness on a warm water blanket before returning to their home cage (single-housed). Laboratory staff monitored rodent health daily for one week following surgery until suture removal. Rodents were then pair-housed and recovered for four weeks to allow opsin expression at NBM terminals before experimentation. Injection sites were confirmed by collecting coronal slices of the NBM and visualizing EYFP (green) or tdTomato (red) using fluorescence microscopy postmortem. Rats with unintended viral placement were excluded from the study.

Chronic intermittent ethanol vapor exposure
Pair-housed rodents remained in their home cages and were placed within Plexiglas vapor inhalation chambers (Triad Plastics) for the duration of the chronic intermittent ethanol (CIE) exposure paradigm. Ethanol was vaporized and continuously pumped into the inhalation chamber throughout the light cycle (12 h/d; 9 P.M. to 9:00 A.M.) for 10 consecutive days to produce repeated cycles of exposure and acute withdrawal. Rodents began the CIE chambers between postnatal days P66 and P71 and ended the CIE chambers between postnatal days P76 and P81. Laboratory personnel monitored animal health, and food pellets and water were discarded and replenished daily. Age-matched controls (denoted AIR in figures) were similarly housed but exposed to ambient air only. Blood samples were collected via tail snip twice throughout the exposure to measure blood ethanol concentrations (BECs) using a commercially available alcohol dehydrogenase/NADH (nicotinamide adenine dinucleotide plus hydrogen) enzymatic assay kit (Carolina Liquid Chemistries). The average BECs during the exposure was 229.3 6 3.8 mg/dl (n = 102 rats). Rodents with average BECs outside the 150-275 mg/dl target range were excluded from the study. All whole-cell patchclamp electrophysiology studies were completed following 24 h of withdrawal, between postnatal days P77 and P82.

Whole-cell patch-clamp electrophysiology
Coronal slices were moved into a submersion-type recording chamber for whole-cell patch-clamp electrophysiology experiments that were continuously perfused with oxygenated, room temperature (;25°C) standard aCSF at a rate of 2 ml/min using a peristaltic pump (MasterFlex). For NBM and BLA current-clamp recordings, recording electrodes were filled with a potassium gluconate intracellular solution that contained: 145 mM K-gluconate, 10 mM EGTA, 10 mM HEPES, 5 mM NaCl, 1 mM MgCl 2 ·6H2O, 2 mM Mg-ATP, and 0.1 mM Na-GTP. The pH was adjusted to ;7.2 À 7.3 with KOH. The osmolarity of this solution was ;285 Osm/l. For all other recordings, glass electrodes were filled with a cesium gluconate intracellular solution that contained: 145 mM CsOH, 10 mM EGTA, 5 mM NaCl, 1 mM MgCl2·6H 2 O, 10 mM HEPES, 4 mM Mg-ATP, 0.4 mM Na-GTP, 0.4 mM QX314, and 1 mM CaCl 2 ·2H 2 O. The pH was adjusted to ;7.2-7.3 with D-gluconic acid, and the osmolarity was adjusted to ;285 Osm/l. Data were acquired at 5 kHz and low-pass filtered at 2 kHz via an Axopatch 700B Amplifier and pClamp10.7 software (Molecular Devices).
Based on published studies, we used exclusion/inclusion criteria to distinguish NBM "cholinergic" neurons from "noncholinergic" neurons. "Cholinergic" neurons must exhibit the following characteristics: (1) slow and regular firing patterns; (2) afterhyperpolarization (AHP) duration following the first action potential must be !100 ms; and, (3) no rebound burst firing after a hyperpolarizing current step. This latter characteristic is a property of GABAergic neurons (Griffith and Matthews, 1986;Sotty et al., 2003). Although the NBM is an amorphous brain region, we targeted the NBM in coronal slices by placing the recording electrode just dorsal to the intersection between the lateral ventricle and ventrolateral striatum (NBM: AP À1.5 mm, ML 62.5 mm, DV 7.2 mm). For these experiments, there were no pharmacological synaptic blockers added to the external aCSF. NBM neurons were injected with 600-ms current beginning at -100 pA and increasing by 25 pA every 20 s up to 1300 pA. These recordings measured the AHP duration of the first action potential, the resting membrane potential (RMP), and the number of action potentials elicited from application of each current step. Additionally, we used a current ramp protocol beginning at -100 pA and increasing to 1300 pA over 600 ms. The ramp protocol was repeated for five trials, where the first action potential was used to measure action potential properties like the action potential half-width and the peak amplitude (Price and McCool, 2022).
For BLA pyramidal neuron electrophysiology recordings, putative BLA principal/pyramidal neurons were distinguished from local GABAergic interneurons based on their characteristic low access resistance (,25 MV) and high membrane capacitance (.100 pF). We recorded neurons within the anterior/posterior subdivisions of the BLA as previous research indicates this region receives the densest NBM cholinergic input (Woolf and Butcher, 1982;Carlsen et al., 1985;Lee and Kim, 2019). Glutamatergic EPSCs were recorded from BLA neurons at a holding potential of À65 mV and were pharmacologically isolated with GABAA antagonist picrotoxin (100 mM) and APV (NMDA receptor antagonist, 50 mM). In separate recordings, GABAergic IPSCs were recorded at a holding potential of À10 mV and were pharmacologically isolated by adding glutamate receptor antagonists DNQX (AMPA/kainite receptor antagonist, 20 mM) and APV (50 mM) to the external aCSF. In addition, the nonselective mAChR antagonist atropine (Zwart and Vijverberg, 1997) was added to the external aCSF (500 nM) in a subset of recordings to isolate nAChR-dependent effects of NBM terminal activation/inhibition. Under the appropriate conditions, presynaptic nAChR-dependent modulation of glutamate release can serve as a proxy for circuit-level dysregulation of cholinergic NBM-BLA projections following chronic ethanol (Sizer et al., 2021).
Glutamatergic EPSCs were electrically evoked with platinum/iridium concentric bipolar stimulating electrodes (FHC Inc.) with an inner pole of 12.5 mm. The stimulating electrode was placed several hundred microns away from the basolateral nucleus within the stria terminalis (ST) fiber tract, just dorsal to the central amygdala and medial to the lateral/basolateral nuclei, to activate cortical and thalamic glutamatergic afferents. Modest electrical stimulation of the ST was normalized across all recordings to elicit synaptic currents with amplitudes of ;100 pA to ensure these represent monosynaptic glutamatergic responses. Polysynaptic glutamatergic responses, likely because of an erroneous spread of electrical stimulation and subsequent activation of "local" circuits, were rarely encountered and excluded from data analysis. Direct activation of antidromic action currents within BLA principal neurons was never observed.
Research Article: New Research (Bejing Viasho Technology Co, Ltd.). Halorhodopsin (NpHR; inhibitory opsin) and ChrimsonR (red-shifted excitatory opsin) were both activated using a 589-nm, 100-mW DPSS yellow laser (Bejing Viasho Technology Co, Ltd.). Lasers were connected to a fiber optic cable (Thorlabs); the naked end of the cable was placed just above the stria terminalis on the medial side of the BLA to activate/inhibit neurotransmitter release from NBM terminals. The 473/589-nm laser output was measured across a range of laser intensities with a photodiode (Switchable Gain Detector 320-1100 nm; Thorlabs) and an oscilloscope (Tektronix TDS1002B) to measure light intensity as voltage (V out in mV). Laser intensity was expressed as input power (mW), calculated using the following formula: V out = (R)(l )(Transimpedence Gain)(Scale Factor)(Input Power)(1000), where R l ð Þ is defined as the responsivity of the photodiode.
To provide a measure of release probability, two electrical stimulations of equal intensity was delivered at an interstimulus interval of 50 ms. At short interstimulus intervals, the ratio between the first and second EPSC response amplitudes serves as a proxy for presynaptic neurotransmitter release probability (Andreasen and Hablitz, 1994;Dobrunz and Stevens, 1997). The paired-pulse ratio (PPR) was calculated using EPSC amplitudes as follows: [(Peak 2 amplitude)/(Peak 1 amplitude)]. For stria terminalis glutamate PPR recordings, the Peak 1 and Peak 2 amplitudes are distinct and were measured from the baseline preceding the first EPSC, as previously reported (Morales et al., 2018). In a subset of experiments (Fig. 3), we used optogenetics to activate/inhibit NBM terminal release with a yellow or blue laser (10 Hz, 5 ms) delivered 1 ms before electrical stimulation of stria terminalis inputs.

Immunohistochemistry
Four weeks following stereotaxic microinjection of channelrhodopsin/eYFP into the NBM, rats (n = 16) underwent transcardial perfusions of PBS and 4% formalin for tissue fixation. Whole brains were extracted and stored in a 4% formalin solution at 4°C overnight. Brains were rinsed with a phosphate-buffered saline (0.12 M Na 2 HPO 4 , 0.18 M NaH 2 PO 4 , and 0.12 M NaCl) and placed in phosphate-buffered saline containing 30% sucrose solution for long-term storage at 4°C. This sucrose solution was replaced once a week to prevent bacterial growth. Coronal slices (50 mm thick) containing the NBM and the BLA were prepared using a Leica VT1200/S vibrating microtome. Unstained brain slices were immersed in 0.5 ml of cryoprotectant at À20°C for long-term storage. Fixed coronal slices were rinsed several times with 0.01 M PBS (Fisher Bioreagents) containing 0.3% Triton X-100 (PBS-Tx solution) and transferred into a blocking solution consisting of 5% normal donkey serum in PBS-Tx for 2 h. After rinsing with PBS-Tx, brain slices were incubated overnight with blocking solution (0.5 ml/well) containing chicken anti-GFP primary antibody (1:1000 dilution; Aves Labs; RRID: AB_10000240). Following three consecutive PBS-Tx rinses, brain slices were incubated in blocking solution containing Alexa Fluor 488 donkey anti-chicken secondary antibody (1:250 dilution; Jackson ImmunoResearch Labs; RRID: AB_2340375) for 1.5 h. Slices were repeatedly rinsed with PBS and mounted onto slides using ProLong Gold Antifade Mounting Media with DAPI (ThermoFisher). Coverslips were fixed to the slides and allowed to dry at 4°C overnight before imaging with a confocal microscope.

Statistical analysis
Statistical analyses were completed using Prism 8 (GraphPad Software). Data were analyzed with repeatedmeasures two-way ANOVA, repeated-measures mixedeffects analysis, unpaired t tests, and Bonferroni's post hoc tests depending on the experiment (see Table 1). A value of p , 0.05 was considered statistically significant, and statistical significance was denoted in the figures as follows: *p , 0.05, **p , 0.01, ***p , 0.001, and ****p , 0.0001. Graphs are represented as mean 6 SEM.
Several studies have shown that CIE exposure and withdrawal increase the glutamate release probability (decrease the paired-pulse ratio) of electrically-evoked stria terminalis afferents (Christian et al., 2013;Morales et al., 2018). In fact, this effect may be due in part to tonic activation of presynaptic a 7 nicotinic acetylcholine receptors  Research Article: New Research during withdrawal (Sizer et al., 2021). Therefore, we hypothesized that upregulation of NBM terminal activity helps produce pathologic glutamate release by activating presynaptic nicotinic acetylcholine receptors (nAChRs) expressed by stria terminalis synapses. We used optogenetics to activate channelrhodopsin or chrimson-expressing NBM terminals before electrical stimulation of stria terminalis paired-pulse ratios (PPR) to understand the effect of NBM terminal activity on glutamate release probability. The recordings concluded with a mecamylamine drug wash (MEC; nonselective nAChR antagonist; 100 mM) to determine whether NBM terminal stimulation activated presynaptic nicotinic receptors at stria terminalis synapses. The effect of NBM terminal activation and effects of MEC perfusion was the same regardless of gating by Channelrhodopsin or ChrimsonR. We therefore collapsed the AIR and CIE data across these excitatory opsins for Figure  Laser stimulation of NBM terminals increases the stria terminalis glutamate release probability in AIR neurons, with no effect in CIE neurons (repeated-measures two-way ANOVA; main effect of opto 1 MEC, F (2,28) = 9.935, p = 0.0002; significant opto 1 MEC X CIE interaction F (2,28) = 4.738, p = 0.012; Fig. 3B i ). Bonferroni's multiple comparison test illustrates significant differences in AIR baseline versus opto (t = 4.112, p = 0.0004), but no significant differences in CIE baseline versus opto (t = 0.2855, p . 0.999). Pharmacological upregulation of synaptic acetylcholine levels with physostigmine (0.5 mM; acetylcholinesterase antagonist) cause similar alterations in stria terminalis release probability (opto AIR: n = 20 cells from 9 pairs of rats, opto CIE: n = 11 cells from 5 pairs of rats, physostigmine AIR: n = 23 cells from 12 pairs of rats, physostigmine CIE: n = 21 cells from 9 pairs of rats; two-way ANOVA; significant main effect of CIE F (1,71) = 19.74, p , 0.0001; no significant effect  Based on these findings, we hypothesized that laser inhibition of halorhodopsin-expressing NBM terminals before electrical stimulation of stria terminalis afferents would reverse elevated glutamate release during withdrawal. Planned unpaired t test show significant differences in AIR and CIE paired-pulse ratios at baseline (AIR: 1.3 6 0.1, n = 12 cells from 6 pairs of rats; CIE: 1.0 6 0.1, n = 14 cells from 6 pairs of rats; planned unpaired t test, t (24) = 2.429, p = 0.012; Fig. 3F k ). Inhibition of NBM terminal activity with increasing 589-nm laser intensity (expressed as input power, mW) significantly increases the PPR in CIE neurons, with no effect in AIR neurons (repeated-measures mixed-effects analysis; laser X CIE interaction F (4,78) = 2.687, p = 0.037 Fig. 3F  Collectively, these findings show that inhibiting NBM terminal activity restores glutamate release at stria terminalis afferents during withdrawal to levels that are similar to AIR controls.
We recorded from BLA pyramidal neurons and measured optically-evoked EPSCs from channelrhodopsin-expressing NBM terminals to understand whether CIE exposure and withdrawal upregulates NBM terminal activity in the BLA. We measured the EPSC amplitude across a range of laser intensities (expressed as input power, mW) and found that CIE exposure significantly increases the EPSC amplitude relative to AIR controls (AIR: n = 22 cells from 7 pairs of rats; CIE: n = 20 cells from 7 pairs of rats; repeated-measures mixed-effects analysis; main effect of laser intensity F (5,192) = 41.91, p , 0.0001; trending main effect of CIE F (1,40) = 4.041, p = 0.051; laser X CIE interaction F (5,192) = 4.674, p = 0.0005; Fig. 5G r ). Bonferroni's post hoc tests show significant differences between AIR and CIE EPSCs at the two highest laser intensities (input power: 0.50 mW, t = 2.921, p = 0.023; Input power: 0.91mW; t = 3.588, p = 0.002). Withdrawal-induced increases in EPSC amplitude were not because of significant differences in access resistance between AIR and CIE neurons (data not shown; AIR: 11.1 6 0.6MV, n = 22 cells from n = 7 pairs of rats; CIE: 12.6 6 1.1MV, n = 20 cells from 7 pairs of rats; unpaired t test, t (18) = 1.18, p = 0.123).

Upregulation of glutamatergic and cholinergic NBM terminals increases BLA pyramidal neuron firing during withdrawal
Next, we sought to understand whether withdrawal-induced potentiation of NBM cholinergic terminal activity modulates BLA pyramidal neuron firing. For these current-clamp recordings, we included picrotoxin in the aCSF to block the influence of NBM GABAergic release. Pilot studies indicated that atropine significantly reduced BLA pyramidal neuron firing independently of AIR or CIE exposure, so it was excluded from the aCSF (data not shown). We injected a depolarizing current into AIR-exposed and CIE-exposed BLA pyramidal neurons to reach a membrane potential of À48 mV and measured action potential firing frequency (Hz) for a 30-s baseline (laser OFF baseline ), during 60-s laser inhibition of NBM terminals (laser ON), and a 30-s recovery period (laser OFF recovery ). There was no significant difference in depolarizing current injected in AIR and CIE neurons to maintain -48-mV membrane potential (AIR: 152.7 6 12.6 pA, n = 9 cells from four pairs of rats; CIE: 125.0 6 22.0 pA, n = 6 cells from 3 pairs of rats; unpaired t test, t (13) = 1.174, p = 0.131; Fig. 6B s ). Also, the resting membrane potential (RMP) was not significantly different between AIR-exposed and CIE-exposed neurons or before and after the application of depolarizing current (repeated-measures two-way ANOVA; no main effect of CIE exposure F (1,10) = 0.2158, p = 0.652; no main effect of depolarizing current F (1,10) = 3.598, p = 0.087; no CIE X depolarizing current interaction F (1,10) = 1.022, p = 0.336; Fig. 6C t ). CIE exposure and withdrawal significantly elevates BLA pyramidal neuron firing frequency at baseline (AIR: 0.3 6 0.2, n = 9 cells from four pairs of rats; CIE: 1.0 6 0.1, n = 6 cells from 3 pairs of rats; planned unpaired t test, t (13) = 2.679, p = 0.0091; Fig. 6D u ). Laser inhibition of  NBM terminals reverses enhanced BLA pyramidal neuron firing in CIE neurons, with no effect in AIR neurons (repeated-measures mixed-effects ANOVA; main effect of laser F (2,26) = 8.865, p = 0.001; trending main effect of CIE F (1,13) = 4.557, p = 0.052; laser X CIE interaction F (2,26) = 3.778, p = 0.036; Fig. 6D u ). Bonferroni's multiple comparison test CIE: OFF baseline versus ON (t = 3.971, p = 0.001) and ON versus OFF recovery (t = 3.725, p = 0.003), AIR: OFF baseline versus ON (t = 0.7082, p . 0.99) and ON versus OFF recovery (t = 1.380, p = 0.540). These data suggest that NBM terminal activity directly contributes to withdrawaldependent increases in BLA pyramidal neuron firing.

Discussion
CIE exposure and withdrawal upregulate acetylcholine levels in the BLA (Sizer et al., 2021). Since it is well established that the BLA receives dense cholinergic input from Figure 5. CIE exposure and withdrawal potentiates monosynaptic glutamatergic and nicotinic EPSCs elicited from NBM terminals. A, Schematic illustrating placement of 473-nm laser for optical activation of NBM terminals while recording from BLA pyramidal neurons. B, TTX (1 mM) perfusion occludes EPSC and is fully recovered with 4-AP (1 mM) perfusion in AIR neurons (n = 5 cells; repeated-measures one-way ANOVA; p = 0.050), Bonferroni's multiple comparisons test (baseline vs TTX, *p = 0.049 and baseline vs 4-AP, p = 0.99). C, Representative traces of NBM EPSCs following TTX and 4-AP perfusion. D, No significant differences in EPSC latency in AIR (n = 20) and CIE (n = 22) neurons (unpaired t test, p = 0.282) EPSC latency and TTX 1 4-AP data indicate that the optically evoked NBM EPSCs are monosynaptic. E, DNQX (AMPA/kainate receptor antagonist; 20 mM) and mecamylamine (MEC; nonselective nicotinic receptor antagonist; 100 mM) ablate optically evoked EPSC in AIR (n = 7) and CIE neurons (n = 5). Mixed-effects analysis, main effect of CIE exposure (*p = 0.016), main effect of DNQX 1 MEC (p = 0.0002), CIE X DNQX 1 MEC interaction (p = 0.0303). Bonferroni's multiple comparisons test (CIE baseline vs DNQX, **p = 0.0047 and DNQX vs MEC, *p = 0.0395). F, Representative traces of AIR and CIE EPSCs following DNQX and MEC perfusion. G, Optical excitation of NBM terminals at increasing laser intensities [measured as input power (mW)] in AIR (n = 22) and CIE (n = 20) neurons. Mixed-effects analysis showed main effect of laser (p , 0.0001), trending main effect of exposure (p = 0.051), and laser X exposure interaction (p = 0.0005). Bonferroni's post hoc tests show significant differences in AIR and CIE EPSCs at the two highest laser intensities (*p = 0.023 and **p , 0.002). H, Representative traces of EPSCs in AIR and CIE neurons. Scale bars: y-axis 20 pA and x-axis 20 ms for all traces in figure. the NBM, we hypothesized that dysregulation of the NBM-BLA circuit may be responsible. Therefore, the present study sought to understand whether alterations in NBM cholinergic neurons and their projections increase glutamatergic signaling in the BLA during ethanol withdrawal. Our data suggest that NBM "cholinergic" neurons respond to synaptic or intrinsic changes following CIE exposure and withdrawal that enhance neuronal excitability. We subsequently used optogenetics to specifically activate or inhibit NBM terminals within the BLA and measure the impact on both stria terminalis glutamate release and BLA pyramidal neuron firing. We found that CIE exposure and withdrawal potentiates NBM cholinergic neurotransmission in the BLA, facilitating tonic activation of nicotinic receptors which helps drive glutamate release from stria terminalis synapses. Upregulation of NBM terminal activity also increases BLA pyramidal neuron firing. Additionally, we found that CIE exposure and withdrawal attenuate NBM GABAergic neurotransmission in the BLA. These findings suggest that withdrawal-induced alterations in NBM afferents may shift the relative impact of cholinergic and GABAergic neurotransmission and contribute to BLA pyramidal neuron excitability. Notably, this is the first study to characterize ethanol-induced alterations of NBM-BLA projections at glutamatergic circuits in the BLA.
Our initial experiments recorded NBM "cholinergic" and "noncholinergic" neurons in AIR-exposed and CIE-exposed rodents. Consistent with previous reports, we show that the firing frequency of ethanol naive "cholinergic" neurons plateaus at ;20 Hz (Hedrick and Waters, 2010). Although previous work suggests acute ethanol modulates medial septum/diagonal band cholinergic neuron firing (Ericson et al., 2010), our work appears to be the first study to directly measure the stimulatory effects of chronic ethanol exposure and withdrawal on NBM "cholinergic" neuron excitability. We should note that the variability across these measures may reflect the presence of subpopulations of NBM "cholinergic" neurons. For example, basal forebrain ChAT 1 cholinergic neurons are characterized as early and late firing neurons that both express distinct firing patterns and may regulate different facets of attention and arousal (Unal et al., 2012). We found that CIE exposure and withdrawal increased the resting membrane potential, increased the action potential peak amplitude, and decreased the half-width of NBM "cholinergic" neurons ( Fig. 7). Similar to the effects of ethanol withdrawal, auditory fear conditioning also robustly increases the firing frequency and decreases the halfwidth of NBM ChAT 1 neurons (Rajebhosale et al., 2021). Although alterations in the resting membrane potential, peak amplitude, and half-width may be responsible for increases in NBM "cholinergic" neuron excitability, the specific ion channels and signaling pathways underlying these effects warrant further investigation.
However, since synaptic transmission was intact in these experiments, it is impossible to discern whether the alterations in NBM cholinergic neuron action potential properties reflect synaptic influence or intrinsic mechanisms that promote excitability. The extensive list of afferents or circuits which modulate basal forebrain cholinergic neuron excitability makes it difficult to speculate on potential synaptic mechanisms. For example, the NBM receives serotonergic (Gasbarri et al., 1999), noradrenergic, and histaminergic input from the dorsal raphe, locus coeruleus, and hypothalamus, respectively . BLA-projecting NBM cholinergic neurons also receive input from the striatum [caudate putamen and nucleus accumbens (NAc)] and other basal forebrain regions (Gielow and Zaborszky, 2017). There are three known GABAergic projections from the amygdala to the NBM; these include inputs arising from a small population of intrinsic BLA GABAergic interneurons (McDonald et al., 2012), GABAergic projections from the intercalated cell masses (ICMs; Pare and Smith, 1994;Likhtik et al., 2008), and central amygdala GABAergic projections (Gielow and Zaborszky, 2017). Notably, chronic ethanol reduces GABA release from BLA lateral paracapsular GABAergic cells (Diaz et Figure 6. NBM terminal activity upregulates BLA pyramidal neuron firing following CIE exposure and withdrawal. A, Schematic illustrating placement of 589-nm laser for optical inhibition of NBM terminals expressing halorhodopsin while recording from BLA pyramidal neurons. BLA pyramidal neurons were injected with depolarizing current to a membrane potential of -48 mV. Action potential firing frequency was recorded for 30-s baseline (laser OFF), during 60-s laser inhibition of NBM terminals (laser ON), and 30-s recovery (laser ON). B, Depolarizing current injected to reach -48-mV membrane potential is not significantly different in AIR (n = 9 cells) and CIE neurons (n = 6 cells), unpaired t test (p = 0.131). C, Resting membrane potential (RMP) was not significantly different between AIR-exposed and CIE-exposed neurons, or before and after application of depolarizing current during recording (repeated-measures two-way ANOVA, see Results). D, Inhibition of NBM terminals with 589-nm laser (10 Hz, 60 s) reverses increases in BLA pyramidal neuron firing in CIE neurons (n = 6), with no effect in AIR neurons (n = 9). Two-way repeated-measures ANOVA, main effect of laser (p = 0.001), trending effect of exposure (p = 0.052), and laser X exposure interaction (p = 0.036). Bonferroni's post hoc tests show significant differences between laser OFF (baseline) and laser ON (**p = 0.001) and laser ON versus laser OFF (recovery; **p = 0.003). E, Representative traces of AIR neurons and CIE neurons. Scale bars: y-axis 20 mV and x-axis20 s for all traces in figure. al., 2011), close relatives of the ICMs. Glutamatergic projections to the NBM also arise from several regions, including the BLA (Arieli et al., 2020), medial PFC (Zaborszky et al., 1997), hypothalamus, and thalamus (Carnes et al., 1990). Dysregulation of any of these neuromodulatory, GABAergic, or glutamatergic afferents could promote NBM cholinergic neuron excitability during ethanol withdrawal. Our study appears to be the first to measure putative NBM cholinergic neuron excitability following CIE exposure/withdrawal and lays the foundation for future work to dissect the specific mechanisms involved.
In order to manipulate NBM terminal activity and measure the neurophysiological outcomes in the BLA, we targeted the NBM through precise stereotaxic microinjection of excitatory and inhibitory AAV viral constructs and found robust GFP expression at both the injection site and the NBM terminals in the BLA (AE AE Power et al., 2003). Since the NBM-BLA circuit contains cholinergic and GABAergic projections, we used picrotoxin and atropine to interrogate the downstream effects of NBM cholinergic neurotransmission on nAChR-dependent modulation of glutamatergic neurotransmission in the BLA (Jiang and Role, 2008). BLA neurons express both heteromeric and homomeric nicotinic acetylcholine receptors containing different permutations of a 2-4,7 and b 2,4 subunits (Zhu et al., 2005). These BLA nicotinic receptors control the expression of anxiety-like and depressive-like behaviors (Mineur et al., 2016;Wise et al., 2020) and regulate fear extinction (Jiang et al., 2016). Importantly, activity of these presynaptic nAChRs can be used as a proxy measure for ethanol-induced alterations in cholinergic signaling at distinct synapses (Sizer et al., 2021). For example, CIE exposure and withdrawal upregulate glutamate release from stria terminalis afferents (Christian et al., 2013;Morales et al., 2018;McGinnis et al., 2020a;Price and McCool, 2022); and this process is driven by the tonic activation of presynaptic a 7 nAChRs expressed on these glutamate terminals (Sizer et al., 2021). The present study expands on these findings and uses optogenetics to manipulate NBM terminal activity. In AIR controls, optogenetic stimulation of NBM terminals causes nAChR-dependent facilitation of release from stria terminalis inputs. These findings are consistent with recent reports using transgenic rodent models (Tryon et al., 2021) and confirm that optical stimulation of channelrhodopsin-expressing NBM terminals releases acetylcholine. Pharmacological upregulation of synaptic acetylcholine levels produces similar alterations in stria terminalis glutamate release. Since CIE exposure/withdrawal upregulate acetylcholine levels leading to activation of presynaptic nicotinic receptors, optical stimulation of NBM terminals did not further increase glutamate release in this treatment group. However, optical inhibition of these terminals reversed the effect of CIE on stria terminalis glutamate release. Collectively, these data suggest that withdrawal-induced potentiation presynaptic nAChR activity and facilitation of glutamate release relies almost exclusively on NBM cholinergic projections to the BLA (Fig. 7).
Notably, this is also the first study to measure the functional significance of NBM-BLA GABAergic projections. Reports estimate that ;10% of the NBM-to-BLA projection consists of GABAergic input which may regulate BLA output by inhibiting either "local" GABAergic interneurons or BLA pyramidal neurons (McDonald et al., 2011). The TTX 1 4-AP and IPSC latency data strongly suggest that optically-evoked GABAergic responses from NBM terminals are monosynaptic. We cannot however specifically rule out GABA co-release from other NBM projections. Subpopulations of basal forebrain cholinergic neurons express both ChAT and GAD67 mRNA (Sotty et al., 2003); and NBM cholinergic projections co-release GABA in the cortex (Saunders et al., 2015). However, co-localization GABAergic and cholinergic markers in basal forebrain represents ;1% of the total neuron population (Sotty et al., 2003) suggesting that GABA/acetylcholine co-release Figure 7. Schematic representation of NBM afferents in AIR control rodents (left) and CIE and withdrawal rodents (right). CIE and withdrawal potentiate NBM cholinergic neuron excitability and increases NBM cholinergic terminal activity in the BLA. The upregulation of NBM acetylcholine release in the BLA tonically activates presynaptic nAChRs and facilitates glutamate release from stria terminalis afferents. CIE also decreases release from NBM GABAergic terminals in the BLA. The effects together appear to enhance BLA pyramidal neuron excitability. may be a relatively rare phenomena. It should be noted that the IPSC amplitude optically evoked from NBM inputs onto BLA principal neurons was significantly larger than the amplitude of NBM excitatory EPSCs. This is consistent with anatomic studies showing that the relative sparse NBM GABAergic terminal fields in the BLA nonetheless contain densely packed synaptic vesicles, numerous mitochondria, multiple synaptic release sites, and large synaptic contact areas (McDonald et al., 2011). Therefore, NBM GABAergic projections may be poised to exert considerable regulatory impact on BLA neurophysiology. Our data also indicate that CIE exposure/ withdrawal attenuate NBM GABAergic IPSCs at BLA pyramidal neuron synapses which may reflect the decreased function of these GABAergic projections (Fig. 7). One possible mechanism is that increased cholinergic neurotransmission in the BLA may reduce NBM GABA release via tonic activation of presynaptic M2 muscarinic receptors (Marchi et al., 1990;Salgado et al., 2007;Muller et al., 2016;Fajardo-Serrano et al., 2017). Collectively, our findings illustrate that CIE exposure and withdrawal produce opposing effects on NBM cholinergic and GABAergic neurotransmission onto BLA pyramidal neurons.
Since populations of NBM-BLA cholinergic projection neurons co-express glutamatergic synaptic markers, we found we could optically-evoke EPSCs from NBM terminals in the BLA. The TTX 1 4-AP and EPSC latency data demonstrate that these optical synaptic responses likely arise from monosynaptic inputs onto BLA pyramidal neurons. Optically-evoked EPSCs expressed short latencies that were not significantly different between AIR and CIE neurons, yet CIE neurons showed significant increases in the EPSC amplitude across a range of laser intensities. Thus, ethanol withdrawal upregulates direct, monosynaptic NBM excitatory neurotransmission onto BLA principal neurons. Surprisingly, we also found that these opticallyevoked EPSCs contain both nicotinic and glutamatergic components. It is unclear whether this reflects acetylcholine/glutamate co-release or optical activation of distinct NBM glutamatergic projections. Approximately one-third of NBM-BLA ChAT 1 terminals express VGLUT3, suggesting these neurons potentially co-release glutamate (Higley et al., 2011;Nelson et al., 2014). While the substantia innominata-medial PFC circuit contains glutamate-specific projections (Henny and Jones, 2008), NBM glutamatergic projections have not been documented within the NBM-BLA circuit. Regardless, NBM terminals in the BLA express three distinct axonal morphologies which may indicate the presence of a small population of NBM glutamatergic projections (McDonald et al., 2011).
Finally, we used optogenetics to measure the effect of manipulating NBM terminals on BLA pyramidal neuron excitability in AIR-exposed and CIE-exposed groups. Our results suggest that CIE/WD upregulates NBM projections which help promote enhancements in BLA pyramidal neuron firing. This is consistent with recent publications showing that optical stimulation of cholinergic terminals produces sustained excitation of BLA pyramidal neurons (Jiang et al., 2016;Aitta-Aho et al., 2018). While we included atropine in the external aCSF during our synaptic studies (Figs. 3, 5), we did not include atropine in the excitability experiments (Fig. 6) because it alone significantly attenuated BLA pyramidal neuron firing in pilot studies (data not shown). In support of this, muscarinic antagonists occlude long-term potentiation (LTP) in the BLA (Watanabe et al., 1995). Muscarinic receptors could influence BLA pyramidal neuron firing through several mechanisms, including inhibition of the slow-afterhyperpolarization (Washburn and Moises, 1992;Womble and Moises, 1993;JM Power and Sah, 2008) via IP 3 -gated calcium stores Sah, 2007, 2008). These findings suggest that muscarinic receptors play an essential role in regulating BLA pyramidal neuron firing. More broadly, the BLA pyramidal neuron excitability and stria terminalis release studies show that potentiation of NBM cholinergic activity during withdrawal may promote BLA pyramidal neuron excitability through both postsynaptic activation of muscarinic receptors and presynaptic facilitation of glutamate release via nicotinic receptors.
Since male rodents express more robust CIE-dependent changes in BLA glutamate neurotransmission compared with females (Morales et al., 2018;Price and McCool, 2022), we chose to characterize withdrawal-induced dysregulation of the NBM-BLA circuit in male Sprague Dawley rats. One important future direction is clearly to measure NBM-BLA circuit following CIE exposure and withdrawal in females. The NBM is a sexually dimorphic region with respect to the number of ChAT 1 NBM neurons, ChAT activity, magnitude of acetylcholine release, and density/conductance of presynaptic and postsynaptic cholinergic receptors. These ultimately translate to differential stress reactivity, arousal, and attention between males and females (Mitsushima et al., 2003;Masuda et al., 2005;Takase et al., 2009). Approximately 50-80% of ChAT 1 neurons in the NBM are immunoreactive for the estrogen receptor GPR30 (Gibbs, 1996;Hammond et al., 2011), and estrogen is protective against cholinergic deficits following chronic stress (Bangasser et al., 2019). Because our studies indicate that NBM cholinergic neurotransmission tightly controls the facilitation of stria terminalis glutamate release during withdrawal (Sizer et al., 2021;Tryon et al., 2021), sex differences within the NBM-BLA circuit may promote disparities related to the effects of CIE exposure at these synapses. Males require fewer days of CIE exposure to express significantly increased stria terminalis glutamate release (Morales et al., 2018). This suggests that BLA neurophysiology may be more vulnerable to CIE exposure in males relative to females. It is therefore reasonable to expect that males could express heightened sensitivity to CIE at NBM cholinergic synapses as well, which may control the differential, sexspecific downstream effects on glutamatergic neurotransmission in the BLA and ultimately related behaviors.
The NBM-BLA circuit strengthens the consolidation of salient cues, a process essential for survival across species. In the presence of an acute stressor, cholinergic NBM neurons fire rapidly (Zhang et al., 2004;Hangya et al., 2015). A variety of acute stressors stimulate NBM cholinergic neuron activity, including mild food shock, predator odor, and exposure to novel/stressful environments Rajebhosale et al., 2021;Mineur et al., 2022). NMB cholinergic neuron firing releases acetylcholine within the BLA and modulates attention, arousal, memory storage, reward processing, and threat-encoding (Mark et al., 1996;Jing et al., 2020;Liu et al., 2021). This acetylcholine also strengthens the acquisition of fear-/reward-associated memories, causes deficits in fear extinction (Jiang et al., 2016;Kellis et al., 2020), and increases anxiety-like behavior (Mineur et al., 2018). Importantly, chronic stress causes a persistent elevation in cholinergic signaling that impairs memory encoding, decreases cognitive flexibility, and drives impulsivity and novelty-seeking behaviors (Yilmazer-Hanke et al., 2016;Mineur and Picciotto, 2021). The current work shows that chronic ethanol exposure appears to produce similar persistent changes to NBM cholinergic function. This may contribute to both mood disorders and vulnerability to alcohol (Yilmazer-Hanke et al., 2016).
To conclude, CIE exposure and withdrawal cause neuroadaptations in emotional centers of the brain that facilitate ethanol withdrawal-induced anxiety and relapse. The present study suggests enhancements in NBM "cholinergic" neuron excitability and the downstream effects on BLA neurophysiology are involved in mediating these behaviors. In particular, we show that withdrawal-induced alterations in NBM terminal activity contribute to BLA pyramidal neuron hyperexcitability via three mechanisms. First, ethanol withdrawal strengthens NBM cholinergic neurotransmission and produces tonic activation of nicotinic receptors on stria terminalis glutamatergic synapses, resulting in a "pathologic" elevation of glutamate release. Second, ethanol withdrawal attenuates monosynaptic NBM GABAergic neurotransmission onto BLA pyramidal neurons. Finally, CIE-dependent upregulation of NBM cholinergic terminals helps drive BLA pyramidal neuron excitability during withdrawal. This study thus provides a novel characterization of the NBM-BLA circuit in AIR-exposed and CIE-exposed rodents. Collectively, our results implicate the NBM-BLA circuit in disrupting BLA neurophysiology and potentially mediating anxiety-like behavior during ethanol withdrawal.