Opioids differentially modulate two synapses important for pain processing in the amygdala

Abstract

Background and Purpose

Pain is a subjective experience involving sensory discriminative and emotionally aversive components. Consistent with its role in pain processing and emotions, the amygdala modulates the aversive component of pain. The laterocapsular region of the central nucleus of the amygdala (CeLC) receives nociceptive information from the parabrachial nucleus (PB) and polymodal, including nociceptive, inputs from the basolateral nucleus of the amygdala (BLA). Opioids are strong analgesics and reduce both the sensory discriminative and the affective component of pain. However, it is unknown whether opioids regulate activity at the two nociceptive inputs to the amygdala.

Experimental Approach

Using whole‐cell electrophysiology, optogenetics, and immunohistochemistry, we investigated whether opioids inhibit the rat PB–CeLC and BLA–CeLC synapses.

Key Results

Opioids inhibited glutamate release at the PB–CeLC and BLA–CeLC synapses. Opioid inhibition is via the μ‐receptor at the PB–CeLC synapse, while at the BLA–CeLC synapse, inhibition is via μ‐receptors in all neurons and via δ‐receptors and κ‐receptors in a subset of neurons.

Conclusions and Implications

Agonists of μ‐receptors inhibited two of the synaptic inputs carrying nociceptive information into the laterocapsular amygdala. Therefore, μ‐receptor agonists, such as morphine, will inhibit glutamate release from PB and BLA in the CeLC, and this could serve as a mechanism through which opioids reduce the affective component of pain and pain‐induced associative learning. The lower than expected regulation of BLA synaptic outputs by δ‐receptors does not support the proposal that opioid receptor subtypes segregate into subnuclei of brain regions.

What is already known

• The amygdala is important for associative learning and the affective component of pain.

• Opioids reduce the intensity and negative affective component of pain.

What this study adds

• μ‐receptor activation reduces information delivery from the parabrachial nucleus and basolateral amygdala to CeLC neurons.

• δ‐receptor and κ‐receptor activation reduces information delivery from the basolateral amygdala to some CeLC neurons.

What is the clinical significance

• μ‐receptor agonists may act in the amygdala to reduce the negative affective component of pain.

• δ‐receptors in the amygdala may not be a useful target to treat acute pain.

Abbreviations

AAV

adeno‐associated virus

aCSF

artificial CSF

BLA

basolateral nucleus of the amygdala

CeA

central nucleus of the amygdala

CeLC

laterocapsular region of the central nucleus of the amygdala

ChR2

channel rhodopsin‐2

DAMGO

[d‐Ala², NMe‐Phe⁴, Gly‐ol⁵]‐enkephalin

DL‐APV

dl‐2‐amino‐5‐phosphonopentanoic acid

eEPSC

evoked EPSC

EYFP

enhanced yellow fluorescent protein

ICI

ICI‐174, 864

Met‐Enk

methionine‐enkephalin

Nor‐BNI

nor‐binaltorphimine

PB

parabrachial nucleus

PBel

external lateral parabrachial nucleus

PPR

paired pulse ratio

1. INTRODUCTION

Pain is a complex sensory experience encompassing both affective and sensory discriminative components (Rainville, Duncan, Price, Carrier, & Bushnell, 1997). We begin to associate the experience of pain with the environment in which it occurs, and ultimately, this causes sufferers to limit activities that have become associated with pain, such as exercise or work (Gureje, Von Korff, Simon, & Gater, 1998). This association is formed in the laterocapsular region of the central nucleus of the amygdala (CeLC; Sato et al., 2015; Han, Soleiman, Soden, Zweifel, & Palmiter, 2015), also known as the nociceptive amygdala (Fu et al., 2008). A convergence of nociceptive and other sensory information allows the CeLC to form associations between the affective component of pain and the environment in which it occurs. The CeLC receives a combination of “purely” nociceptive information, via synaptic inputs from the parabrachial nucleus (PB) and polymodal sensory information via synaptic inputs from the basolateral nucleus of the amygdala (BLA; Sah, Faber, Lopez De Armentia, & Power, 2003) and other brain regions such as the thalamus (Moga, Weis, & Moore, 1995) and cortex (Mcdonald & Mascagni, 1997). In animals, this association can be modelled using conditioned place aversion, where animals avoid the environment in which they experienced a painful stimulus (Zhang, Zhang, Hu, & Xu, 2011). The central nucleus of the amygdala (CeA; including the CeLC) is required for formalin (Tanimoto, Nakagawa, Yamauchi, Minami, & Satoh, 2003) and neuropathic (Pedersen, Scheel‐Kruger, & Blackburn‐Munro, 2007) pain‐induced conditioned place aversion. Apart from a role in pain‐induced aversion, the CeLC may also be important for the negative affective aspects of pain via its projection to the substantia innominata dorsalis (Bourgeais, Gauriau, & Bernard, 2001). Thus, the CeLC and the strength of its synaptic inputs delivering nociceptive or polymodal information could influence pain‐induced aversion and the negative emotional experience of pain.

“Purely” nociceptive information is delivered to the CeLC by a synaptic input from the external lateral PB (PBel) that relays nociceptive information received from the spinal cord (Bernard, Alden, & Besson, 1993). This synaptic pathway is required for the affective component of pain as inactivation of the PBel attenuates affective pain behaviours, such as escape following foot shock (Han et al., 2015). Acute noxious heat (Kissiwaa & Bagley, 2018) or chronic inflammatory (Adedoyin, Vicini, & Neale, 2010; Fu et al., 2008; Fu & Neugebauer, 2008; Han, Li, & Neugebauer, 2005), neuropathic (Ikeda, Takahashi, Inoue, & Kato, 2007), or acid‐induced pain (Cheng et al., 2011) potentiates delivery of this nociceptive information to the CeLC. This synaptic potentiation correlates with increased hyperalgesia (Adedoyin et al., 2010; Fu et al., 2008; Ikeda et al., 2007; Kissiwaa & Bagley, 2018), and “in vivo” inhibition of this synapse inhibits both the synaptic potentiation and hyperalgesia (Fu et al., 2008). This suggests that the strength of the synaptic input from the PB to the CeLC influences the experience of pain and pain‐induced aversion. The polymodal sensory information delivered to the CeLC will also regulate pain‐induced aversion and the affective component of pain. One of the sources of the polymodal sensory information, including nociceptive information, is the BLA (Sah et al., 2003). Like the CeLC, the BLA is important for the affective and associative learning aspects of pain. Pharmacological inactivation of the BLA attenuates pain‐induced affective behaviours, such as vocalization (Ji et al., 2010), and lesions of the BLA attenuate pain‐induced conditioned place aversion (Tanimoto et al., 2003). Very recently, it has been shown that a subpopulation of BLA pyramidal neurons is activated by specific nociceptive stimuli and that inactivation of the relevant neurons reduces the aversive response to the stimuli (Corder et al., 2019). Although it was not determined whether these neurons make synaptic connections in the CeLC (Corder et al., 2019), BLA neurons that code for negative valence do make synaptic connections in the CeA (Beyeler et al., 2016). Additionally, the BLA–CeLC synapse is potentiated in pain conditions with ongoing peripheral injury such as neuropathic (Ikeda et al., 2007) and arthritic pain (Neugebauer, Li, Bird, Bhave, & Gereau, 2003). Therefore, it is possible to predict that regulation of these two convergent synaptic inputs onto CeLC neurons would alter pain‐induced associative learning and the affective component of pain.

Widely used opioids, such as morphine, activate the https://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=319, which reduces both pain intensity and the affective component of pain in humans (Kupers, Konings, Adriaensen, & Gybels, 1991; Price, Von der Gruen, Miller, Rafii, & Price, 1985) and experimental animals (Lagraize, Borzan, Peng, & Fuchs, 2006; Oliveras et al., 1986; Zhang et al., 2013). Opioids act in the amygdala to produce these effects as intra‐amygdala injections of opioids reduce pain‐induced conditioned place aversion (LaGraize et al., 2006; Zhang et al., 2013). Although opioids act at several sites in the amygdala (Chieng, Christie, & Osborne, 2006; Winters et al., 2017; Zhu & Pan, 2004), including directly inhibiting the excitability of half of CeLC neurons (Chieng et al., 2006), it is unknown whether opioids inhibit delivery of information via the BLA–CeLC and PB–CeLC synapses. The neurons in the PBel, which relay the nociceptive information from the spinal cord to the CeLC, express high levels of μ‐receptors (Chamberlin, Mansour, Watson, & Saper, 1999) and much lower levels of https://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=318s (Mansour et al., 1994; Unterwald, Knapp, & Zukin, 1991) and very few https://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=317s (Arvidsson et al., 1995; Mansour et al., 1994). Consistent with this expression profile, lateral PB neurons are only directly inhibited by activation of μ‐receptors (Christie & North, 1988). In contrast, opioid receptor immunohistochemistry and mRNA levels in the BLA are overwhelmingly dominated by the very high levels of δ‐receptors (Le Merrer, Becker, Befort, & Kieffer, 2009; Mansour et al., 1994; Poulin, Chevalier, Laforest, & Drolet, 2006; Wang et al., 2018), while there are moderate levels of κ‐receptors (Unterwald et al., 1991) and low levels of μ‐receptors (Ding, Kaneko, Nomura, & Mizuno, 1996; Mansour et al., 1994; Wang et al., 2018). This high expression of δ‐receptors in the BLA and the high expression of δ‐receptors in the CeA (Wang et al., 2018) has resulted in the proposal that there is segregation of opioid receptors in distinct subnuclei of the amygdala and therefore distinct functions in amygdala circuits modulating pain affect (Wang et al., 2018). This profile of opioid receptor expression suggests that the “purely” nociceptive information coming from the PBeL will be inhibited by μ‐receptor agonists, such as morphine, but the polymodal information from the BLA will enter the CeLC unchecked.

In this study, we used whole‐cell electrophysiology, optogenetics, and immunohistochemistry to determine whether opioid agonists inhibit delivery of nociceptive or polymodal information to the rat nociceptive amygdala. We found that opioid agonists inhibited glutamate release at the PB–CeLC and BLA–CeLC synapses via μ‐receptors at the PB–CeLC synapse and via μ‐receptors and additionally via δ‐receptors and κ‐receptors in some cells at the BLA–CeLC synapse. These findings suggest that μ‐receptor agonists inhibit both the “purely” nociceptive and polymodal information entering the CeLC and through this could reduce pain‐induced associative learning and negative affect. Further, the balance of μ‐receptor and δ‐receptor regulation of BLA–CeLC synapses does not support the proposal that opioid receptors have distinct functions in these amygdala nuclei.

2. METHODS

2.1. Animals

All animal care and experimental procedures complied with the guidelines and were approved by the University of Sydney Animal Ethics Committee. Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny, Browne, Cuthill, Emerson,, & Altman, 2010) and with the recommendations made by the British Journal of Pharmacology. Male Sprague–Dawley rats (3–7 weeks old) were sourced from Animal Resources Centre, Perth Australia. Rats were housed in a temperature‐controlled environment with a 12‐hr light/dark cycle and were provided with food and water ad libitum.

2.2. Stereotaxic surgeries

Anaesthesia was induced in rats (age 3–7 weeks) with 5% isoflurane and maintained with 2.5% isoflurane. Once deeply anaesthetized, rats were placed in the stereotaxic apparatus (model 942, Kopf Instruments, Tujunca, CA) and head shaved to reveal skin surface. Prior to the incision, a subcutaneous injection of caprofen (5 mg·kg⁻¹; Cenvet, NSW, Australia) and 0.5% bupivacaine (Cenvet) injected under the surface of the incision site was given. After the incision, two small holes were drilled above the lateral PB (anteroposterior −8.5 to 9.2 and mediolateral ±2.5 from bregma and dorsoventral −6.0 mm from skull surface; Paxinos & Watson, 1986). Bilateral injections were performed using glass pipettes (Drummond Scientific, Broomall, PA) filled with 300 or 500 nl of adeno‐associated virus (AAV)5–HSYN–HChR2(H134R)–EYFP (UNC GTC Vector core; AAV titre: 5.5 × 10¹² virus molecules·ml⁻¹) at a rate of 120 or 200 nl·min⁻¹ respectively (Nanoject 2000, Drummond Scientific) over 2.5 min. The volume of AAV injected did not change the level of expression of virally driven proteins. After the injection, the pipette was left in place for 5 min to allow diffusion of viral solution before being slowly withdrawn. Bone wax (Coherent Scientific) was used to seal the skull opening. After the surgery, rats were given 0.3 ml of 300 mg·ml⁻¹ solution of procaine penicillin (Benacillin, Cenvet) and 0.3 ml of 100 mg·ml⁻¹ solution of cephazolin (Hospira, Cenvet). Rats were monitored daily, and postoperative procedures including weight and infection management were performed for the remainder of the experiment.

2.3. Preparation of brain slices

2.3.1. Naïve animals

Rats (4–16 weeks old) were anaesthetized with isoflurane and decapitated, and their brains were removed into ice‐cold artificial CSF (aCSF) solution containing (in mM) 125 NaCl, 2.5 KCl, 1.25 NaH₂PO₄·2H₂O, 2.5 MgCl₂, 0.5 CaCl₂, 25 NaHCO₃, and 11 glucose. Coronal slices (280 μm) containing the amygdala were obtained using the Leica VT 1200s vibratome (Leica Biosystems, Nußloch, Germany). Slices were transferred to a submerged chamber containing 34°C aCSF equilibrated to pH 7.4 with 95% 0₂ and 5% CO₂ for at least 1 hr.

2.3.2. Surgery animals

Three to twelve weeks after surgeries, rats were anaesthetized with isoflurane and decapitated, and their brains were removed into ice‐cold aCSF saturated with carbogen (95% O₂/5% CO₂). Coronal slices (280 μm) containing the amygdala and PB were obtained using the Leica VT 1200s vibratome. Slices were hemisected and initially incubated in an NMDG‐HEPES recovery solution containing (in mM) 93 NMDG chloride, 2.5 KCl, 1.2 NaH₂PO₄, 30 NaHCO₃, 20 HEPES, 25 d‐glucose, 5 sodium ascorbate, 2 thiourea, 3 sodium pyruvate, 10 MgCl₂, and 0.5 CaCl₂, pH 7.3, 300–310 mOsm·L⁻¹ heated at 34°C and saturated with carbogen for 10 min. Slices were then transferred to a submerged chamber containing aCSF equilibrated to pH 7.4 with carbogen at 34°C for a further 10 min and left at room temperature for at least 1 hr.

2.4. Electrophysiology

Slices were transferred to a recording chamber and superfused continuously at 2.5 ml·min⁻¹ with aCSF containing (in mM) 125 NaCl, 2.5 KCl, 1.25 NaH₂PO₄·2H₂O, 1 MgCl₂, 2 CaCl₂, 25 NaHCO₃, and 11 d‐glucose saturated with carbogen. The temperature was maintained between 33°C and 34°C using an inline heater and monitored using a thermistor. Slices were visualized using an Olympus BX51 microscope equipped with 40× water immersion objective and Dodt gradient contrast optics and epifluorescence illumination. Whole‐cell patch‐clamp recordings were made from neurons in the CeLC and PBel. For recordings in CeLC, patch electrodes (2–4 MΩ) were filled with internal solution containing (in mM) 140 CsCl, 5 HEPES, 10 EGTA, 2 CaCl₂, 2 Mg₂ATP, and 0.3 NaGTP (pH 7.3, osmolarity 280–285 mOsm). For recordings in PB, patch electrodes (2–4 MΩ) were filled with internal solution containing (in mM) 135 K‐gluconate, 10 HEPES, 0.5 EGTA, 8 NaCl, 2 Mg₂ATP, and 0.3 NaGTP (pH 7.3, osmolarity 280–285 mOsm). Neurons were voltage or current clamped using a patch‐clamp amplifier (Multi‐clamp 700B, Axon Instruments, Foster City, CA). Current signals were low‐pass filtered at 5 kHz and sampled at 10 kHz. Series resistance (≤12 MΩ) was compensated by 60% and continuously monitored throughout the experiment. Data were discarded if series resistance fluctuated by more than 20% during recording. Recordings were not corrected for liquid junction potentials. EPSCs were electrically and light evoked. Electrically evoked EPSCs were evoked by two consecutive stimuli (inter‐stimulus interval of 50 ms) of identical strength via concentric bipolar stimulating electrodes (rate, 0.05 Hz; stimuli, 2–99 V, 100 μs, FHC, Bowdoin, ME, USA). EPSCs and action potentials were evoked in ChR2‐expressing neurons and axons using a 473‐nm DPSS laser with a built in optic fibre (Ikecool, CA, USA). Laser output was 3–6 amps at the recording site. Light pulses were delivered to the slice through a 40× immersion objective lens and controlled by computer generated transistor–transistor logics. Data were acquired and analysed using Axograph software (Molecular Devices).

2.5. Immunohistochemistry

Following electrophysiology experiments, slices were fixed overnight in 4% paraformaldehyde in PBS at 4°C. The following day, sections were washed in 0.1 M of PBS. Sections were incubated in a 10% normal goat serum (NGS)/0.5% BSA/0.3% Triton X‐100 in PBS (wt/vol) for an hour at room temperature and then washed using 0.1 M of PBS. This was followed by an overnight incubation of sections in rabbit antibody to GFP (IgG, 1:1,000, Molecular Probes, Cat No. A6455, RRID:AB_221570) primary antibody in 1% NGS/1% Triton X‐100 in PBS at 4°C. GFP primary antibody was washed off the following day with 0.1 M of PBS. Sections were then incubated for 2 hr in Alexa Fluor 488 goat anti‐rabbit (IgG, 1:1,000, Invitrogen Cat No. A‐11008, RRID:AB_143165) secondary antibody in 1% NGS/1% Triton X‐100 in PBS at room temperature. Antibodies were not reused. Sections were washed with 0.1 M of PBS and mounted onto glass slides and coverslipped with Fluoromont‐G (ProSciTech, Queensland, Australia). Sections were imaged with Zeiss LSM510 Meta Confocal Microscope (Carl Zeiss, Germany). The paper complies with the goals and practicalities of immunoblotting and immunohistochemistry for submission to the British Journal of Pharmacology (Alexander et al., 2018). The Immuno‐related procedures used comply with the recommendations made by the British Journal of Pharmacology.

2.6. Data and statistical analysis

The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology. Results are expressed as mean ± SEM. Statistical significance was determined by Student's two‐tailed t test (paired or unpaired where appropriate) and two‐tailed Wilcoxon matched pairs signed‐rank test. The total number of assays (n) for each experiment were above or equal to 5 except for the μ‐receptor modulation of AAV5‐injected animals, due to the complexity of obtaining quality recordings in AAV5‐injected animals. A result was considered statistically significant if P < .05. GraphPad Prism (Version 7, GraphPad, San Diego, CA, USA) was used for all statistical analysis. Synaptic responses were deemed to be opioid‐sensitive if they were reduced by ≥20%.

2.7. Materials

https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=4051 was from Sigma‐Aldrich (St. Louis, MO, USA). https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1614 (Met‐Enk) was from Bachem AG (Bubendorf, Switzerland). https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1636 (ICI), https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1647, and https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1615 were purchased from Tocris Bioscience (Bristol, UK). https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1635 was from Cayman Chemicals (Michigan, USA). dl‐2‐Amino‐5‐phosphonopentanoic acid (DL‐APV), https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=4264, https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1655 and https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1642 (Nor‐BNI) were purchased from Abcam (Cambridge, UK). Picrotoxin was added directly to aCSF, whilst stock solutions were made for all other drugs. Stock solutions of drugs were made in distilled water except for U‐69593, which was made in DMSO. Stock solutions were diluted to working concentrations in aCSF immediately before use and applied by gravity driven superfusion. Opioid receptor agonists were applied by superfusion, and antagonist sensitivity was tested by addition of opioid receptor antagonists to the superfusate in the ongoing presence of the agonist. While these experiments were not performed blind, the order of agonist and antagonist application was routinely alternated.

2.8. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander et al., 2017).

3. RESULTS

3.1. PB–CeLC synapse is inhibited by μ‐receptor activation

To stimulate the parabrachial synaptic inputs to the CeLC, we placed stimulating electrodes dorsomedial to the CeA (Figure 1a). We then made whole‐cell patch‐clamp recordings from CeLC neurons and recorded the evoked EPSC (eEPSC) at −70 mV. We delivered two stimuli with a 50‐ms interval and isolated the eEPSC by inclusion of the https://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=72 antagonist picrotoxin (100 μM) in the superfusate. Superfusion of the endogenous opioid Met‐Enk (10 μM) inhibited the first eEPSC amplitude by 53.3 ± 3.7% (n = 6 neurons from six animals, Figure 1b). Met‐Enk has similar affinity for the μ‐receptor and the δ‐receptor (Raynor et al., 1994). To determine whether Met‐Enk inhibits glutamate release at the PB–CeLC synapse through μ‐receptors or δ‐receptors, we tested whether the Met‐Enk inhibition was reversed by the addition to the superfusate of the selective μ‐receptor antagonist CTAP (1 μM) and/or the selective δ‐receptor antagonist ICI (1 μM). In a subpopulation of neurons, both the μ‐receptor and δ‐receptor antagonists were required to completely reverse the Met‐Enk inhibition of the PB–CeLC eEPSC amplitude (four of six neurons, Figure 1b). CTAP alone completely reversed the Met‐Enk inhibition in the other neurons (two of six neurons, Figure 1b). We also determined whether activation of the κ‐receptors regulates the PB–CeLC synapse. Superfusion of the selective κ‐receptor agonist U‐69593 (300 nM) did not inhibit the eEPSC amplitude at this synapse (−2.4 ± 2.8% inhibition of eEPSC amplitude, n = 5 cells from three animals, Figure 1c). The δ‐receptor modulation of PB–CeLC synapse is surprising, and therefore, we wondered whether the finding was because of stimulation of non‐PB δ‐receptor‐positive fibres. Electrical stimulation of the PB–CeLC synapse is relatively selective and reproducible because of the visibility of the fibre tracts (Bernard et al., 1993) but it is still possible to recruit non‐PB fibres. To investigate whether this δ‐receptor modulation at the PB–CeLC synapse was due to lack of selectivity of electrical stimulation, we used optogenetics to selectively activate the synapse. We injected AAV5–HYSN–ChR2(H134R)–EYFP into the PBel (Figure 1d). Three weeks after virus injection, this produced cell body EYFP expression in the outer PBel and terminal EYFP labelling in surrounding areas (Figure 1e). Neurons with EYFP expression in the cell body responded directly to light activation. Blue light pulses of 2‐ or 10‐ms duration produced time‐locked inward currents in 8/8 PBel neurons (from four animals) voltage clamped at −70 mV (current amplitude: 327.5 ± 77.6 pA, Figure 1f). In neurons with inward currents above 200 pA in voltage clamp (5/8), blue light pulses in current clamp resulted in action potentials (Figure 2f). Neurons without ChR2/EFYP expression did not respond to blue light activation.

Figure 1.

Optogenetics can be used to selectively activate the parabrachial nucleus (PB)–laterocapsular region of the central nucleus of the amygdala (CeLC) synapse. (a) Schematic diagram of stimulation and recording site. Stimulating electrodes were placed dorsomedial to the central nucleus of the amygdala (CeA) to stimulate parabrachial fibres. The response of the CeLC neurons to this stimulation was recorded. (b) Example traces of evoked EPSCs (eEPSCs) of a CeLC neuron in control and the reduction of eEPSC amplitude in methionine‐enkephalin (Met‐Enk; 10 μM). Paired scatter plot of the reversal of Met‐Enk inhibition of eEPSC amplitude by sequential application of CTAP (1 μM) and ICI (1 μM) in individual neurons. Order of antagonist superfusion was routinely alternated. (c) Example traces of eEPSCs of a CeLC neuron in control and in the presence of U‐69593 (U69, 300 nM). Scatter plot of the lack of inhibition of eEPSC amplitude by U‐69593. Each dot represents an individual neuron. (d) Coronal schematic at bregma −9.68 indicating bilateral injections of AAV5 to the external lateral parabrachial nucleus (PBel). Adapted from the brain atlas by Paxinos and Watson. (e) Low magnification (left) and higher magnification (right) confocal images of ChR2/EYFP expression in the PB. Injection was targeted at the PBel (red outline), where most ChR2/EYFP‐expressing cell bodies were found (white arrows). Scale bars: 100 μm. Images were taken 3 weeks after injection. (f) Light flashes (blue bars) of 10‐ms duration evoked inward currents (top trace, holding potential −70 mV) and action potentials (bottom trace) in ChR2/EYFP‐expressing cells in the PBel. (g) Confocal image of EYFP/ChR2 terminal expression in the CeLC. Scale bar: 100 μm. (h) Nine to twelve weeks after viral injections, light pulses of 0.1‐ or 0.5‐ms duration were directly illuminated over ChR2/EYFP‐expressing presynaptic boutons in the CeLC. The response of CeLC neurons to this stimulation was recorded. (i) Example traces of light (blue lines; 0.1‐ms duration) evoked EPSCs of a CeLC neuron in control (left) and after the application of NBQX (10 μM) and APV (100 μM). (j) Data (means ± SEM) of NBQX and APV inhibition of light‐activated EPSCs in CeLC neurons. *P<.05, two‐tailed paired Student's t test. Numbers above error bars indicate number of neurons. Stimulus artefacts have been removed from example traces using electrical stimulation

Figure 2.

Activation of μ‐receptors inhibits presynaptic glutamate release at the parabrachial nucleus–laterocapsular region of the central nucleus of the amygdala (CeLC) synapse. (a) Example traces of light‐activated evoked EPSCs (eEPSCs) of a CeLC neuron in control and DAMGO (1 μM) inhibition of eEPSCs amplitude and reversal by the selective antagonist CTAP (1 μM). (b) Paired scatter plot of the reversal of DAMGO inhibition of eEPSC amplitude by subsequent application of CTAP (1 μM) in individual neurons. Each dot represents an individual neuron. Square is mean, and error bars are SEM. (c) Example traces of light‐activated eEPSCs of a CeLC neuron in control and the reduction of eEPSCs amplitude in DAMGO and scaled to the peak amplitude of eEPSC1 of control to demonstrate the increase in paired pulse ratio. (d) Before‐and‐after scatter plot of DAMGO induced increase in paired pulse ratio. Statistical significance was tested using a two‐tailed paired Student's t test, t = 3.191, df = 5. Each dot represents a neuron. Square is mean, and error bars are SEM. (e) Example traces of a CeLC neuron, with and without deltorphin II (300 nM). Light‐activated eEPSCs are not affected by deltorphin II. (f) Scatter plot of the lack of inhibition of eEPSC amplitude by deltorphin II. Each dot represents an individual neuron. Square is mean, and error bars are SEM

Nine to twelve weeks after virus injections, there was intense expression of ChR2/EYFP terminals in the CeLC (Figure 1g). To test whether these PBeL terminals made synaptic connections with CeLC neurons, we recorded from CeLC neurons with soma surrounded by ChR2/EYFP boutons. Picrotoxin (100 μM) was perfused onto slices to block GABA_A receptor‐mediated spontaneous activity, and neurons were voltage clamped at −70 mV. We delivered two consecutive light pulses of 0.1‐ or 0.5‐ms duration with an inter‐stimulus interval of 50 ms directly over these presynaptic EYFP‐expressing boutons (Figure 1h), and this elicited eEPSCs in 27/34 CeLC neurons (eEPSC amplitude: 209.8 ± 40.4 pA, n = 27, Figure 1i). Light‐evoked EPSCs were glutamatergic as they were abolished by application of the glutamate receptor antagonists NBQX (10 μM) and APV (100 μM; Figure 1i,j).

To determine the opioid sensitivity of the synaptic inputs from the PBeL, we applied selective opioid agonists. The selective μ‐receptor agonist DAMGO (1 μM) significantly reduced the first eEPSC amplitude (63.9 ± 5.7% inhibition of baseline eEPSC amplitude, n = 6 neurons from four animals, Figure 2a,b). The DAMGO inhibition was reversed by the selective μ‐receptor antagonist CTAP (1 μM; 90.6 ± 5.6% reversal of DAMGO inhibition; Figure 2a,b). DAMGO significantly increased the paired pulse ratio (PPR; control: 0.54 ± 0.09 vs. DAMGO: 1.05 ± 0.24, Figure 2c,d). This suggests that DAMGO reduces the eEPSC amplitude through acting pre‐synaptically to reduce glutamate release probability. In contrast, the δ‐receptor agonist deltorphin II (300 nM) had no effect on the first eEPSC amplitude (10.4 ± 6.7% inhibition of baseline eEPSC, n = 4 neurons from four animals, Figure 2e,f) or PPR (control: 0.35 ± 0.07 vs. deltorphin II: 0.38 ± 0.09, n = 4 neurons from four animals) Therefore, when ChR2 expression in PB neurons is used to selectively activate PB–CeLC synapses, only μ‐receptor activation inhibits glutamate release.

3.2. The BLA–CeLC synapse is modulated by all three opioid receptors

To stimulate the BLA synaptic inputs to the CeLC, we placed stimulating electrodes in the BLA (Figure 3a). Application of Met‐Enk (10 μM) inhibited the eEPSC amplitude by 49.7 ± 5.8% (n = 6 neurons from five animals). Met‐Enk also significantly increased the PPR (control: 1.17 ± 0.10 vs. Met‐Enk: 1.72 ± 0.19; Figure 3b,c). Thus, like the PB–CeLC synapse, this suggests Met‐Enk reduces eEPSC amplitude at the BLA–CeLC synapse by reducing glutamate release probability. To determine whether Met‐Enk inhibits glutamate release at the BLA–CeLC synapse through μ‐receptors or δ‐receptors, we used selective antagonists and found that the receptor responsible varied between neurons. In half of the neurons, both the μ‐receptor and δ‐receptor antagonists CTAP (1 μM) and ICI (1 μM) were needed to reverse the Met‐Enk inhibition of eEPSC amplitude (three of six neurons; Figure 3d,f). In the other half of the neurons, the inhibition was entirely reversed by the μ‐receptor antagonist CTAP (three of six neurons; Figure 3e,f).

Figure 3.

The basolateral nucleus of the amygdala (BLA)–laterocapsular region of the central nucleus of the amygdala (CeLC) synapse is modulated by all three opioid receptors. (a) Schematic diagram of stimulation and recording site. Stimulating electrodes were placed centrally in the BLA to stimulate BLA fibres. The response of the CeLC neurons to this stimulation was recorded. (b) Example traces of evoked EPSCs (eEPSCs) of a CeLC neuron in control and the reduction of eEPSC amplitude in Met‐Enk (10 μM) and normalized to the peak amplitude of eEPSC1 of control. (c) Before‐and‐after scatter plot of the Met‐Enk induced increase in paired pulse ratio. Each dot represents an individual neuron. Square is mean, and error bars are SEM. Statistical significance was tested using a two‐tailed Wilcoxon matched pairs signed‐rank test. (d) Example traces of eEPSC of a CeLC neuron and Met‐Enk inhibition of eEPSC amplitude and reversal by sequential application of selective antagonists. CTAP (1 μM) and ICI (1 μM) are both needed to reverse the Met‐Enk inhibition of eEPSC amplitude. (e) Example traces of eEPSC of another CeLC neuron and the Met‐Enk inhibition of eEPSC amplitude and reversal by sequential application of selective antagonists. CTAP completely reverses the Met‐Enk inhibition of eEPSC amplitude. (f) Paired scatter plot of the reversal of Met‐Enk inhibition of eEPSC amplitude by CTAP and ICI. Order of antagonist superfusion was routinely alternated. Each point represents an individual neuron. (g) Example traces of eEPSC of a CeLC neuron and the DAMGO (1uM), deltorphin II (300 nM), and U‐69593 (300 nM) inhibition of eEPSC amplitude. (h) Paired scatter plot of the inhibition of BLA–CeLC synapse by selective opioid receptor agonists and the inhibition by subsequent addition of either CTAP (1 μM), ICI (1 μM), or Nor‐BNI (100 nM). Synaptic responses were deemed to be opioid sensitive if the amplitude was reduced by ≥20%. (i) Before‐and‐after scatter plot of the selective opioid agonists on paired pulse ratio of the opioid‐sensitive neurons. Statistical significance for the effect of DAMGO was tested using a two‐tailed paired Student's t test, t = 4.736, df = 7, n = 8. Each dot represents a neuron. Stimulus artefacts have been removed from example traces

We also examined whether selective opioid receptor agonists inhibit the BLA–CeLC synapse. Superfusion of the μ‐receptor agonist DAMGO (1 μM) inhibited eEPSC amplitude by 35.7 ± 4.7% in all CeLC neurons (n = 10 neurons from four animals, Figure 3g,h) and increased the PPR (control: 1.27 ± 0.11 vs. DAMGO: 1.56 ± 0.13; Figure 3i,g). The inhibition was reversed by the μ‐receptor antagonist CTAP (95.6 ± 2.1% reversal of DAMGO inhibition, n = 10, CTAP, 1 μM, Figure 3h). Superfusion of the DOR agonist deltorphin II (300 nM) inhibited eEPSC amplitude by 38.4 ± 8.2% in a subpopulation of neurons (four of 12 neurons from four animals, Figure 3g,h), and the PPR in these neurons was 1.48 ± 0.13 under control conditions and 1.91 ± 0.47 during application of deltorphin II (n = 4, Figure 3i). The inhibition was reversed by the δ‐receptor antagonist ICI (99.9 ± 5.3% reversal of deltorphin II inhibition, n = 4, Figure 3h). Deltorphin II had no effect on eEPSC amplitude in the other neurons (2.09 ± 2.09% inhibition of eEPSC amplitude, eight of 12 neurons from four animals, Figure 3h). Superfusion of the κ‐receptor agonist U‐69593 (300 nM) inhibited eEPSC amplitude by 38.7 ± 6.5% in a subpopulation of neurons (four of eight neurons, Figure 3g,h), and the PPR in these neurons was 0.99 ± 0.07 under control conditions and 1.17 ± 0.14 during application of U‐69593 (n = 4, Figure 3i). The inhibition was reversed by the κ‐receptor antagonist Nor‐BNI (80.5 ± 1.04% reversal of inhibition by U‐69593, n = 4, Nor‐BNI, 100 nM, Figure 3h). U‐69593 had no effect on eEPSC amplitude in the other neurons (8.72 ± 2.11% inhibition of eEPSC amplitude, four of eight neurons from four animals, Figure 3h). Therefore, opioids inhibit the amplitude of the eEPSC via μ‐receptors and additionally via δ‐receptors and κ‐receptors in a subset of cells at the BLA–CeLC synapse.

4. DISCUSSION

The present study investigated opioid regulation of the PB–CeLC and BLA–CeLC synapses. We found that opioid agonists inhibited both synapses through a presynaptic mechanism; however, the receptors responsible differ between the two synapses. Opioid inhibition at the PB–CeLC synapse is through μ‐receptors, whereas at the BLA–CeLC synapse, opioid inhibition is via μ‐receptors and additionally via δ‐receptors and κ‐receptors in some cells (Figure 4).

Figure 4.

Opioids differentially regulate activity at the parabrachial nucleus (PB)–laterocapsular region of the central nucleus of the amygdala (CeLC) and basolateral nucleus of the amygdala (BLA)–CeLC synapse. Opioids inhibit glutamatergic transmission at the PB–CeLC and BLA–CeLC synapse. The μ‐opioid receptors (MOR) upon activation by an opioid agonist inhibits the release of glutamate and thus inhibits activity at the PB–CeLC synapse. The μ‐opioid receptors alone inhibits glutamate release from BLA terminals on to all CeLC neurons. Activation of δ‐opioid receptors (DOR) and κ‐opioid receptors (KOR) also inhibits the BLA–CeLC synaptic response in a subpopulation of neurons. AMPA receptor, AMPAR; NMDA receptor, NMDAR; opioid receptor, OR

When an injury or noxious stimulus occurs, the PBel relays nociceptive information received from the spinal cord to the CeLC (Bernard et al., 1993; Bernard, Huang, & Besson, 1992). Clinically used opioid agonists activate μ‐receptors. In this study, we showed that when μ‐receptors are activated, delivery of nociceptive information from the PB is reduced by 60% and polymodal sensory information, including complex nociceptive information coming from the BLA (Sah et al., 2003), was also reduced by 50%. In addition, the excitability of a third of CeLC is directly inhibited by activation of μ‐receptors (Chieng et al., 2006). Therefore, when an injury occurs or there is activation of the SC–PB–CeLC nociceptive pathway, a μ‐receptor agonist is likely, through these multiple actions, to substantially reduce the activation of CeLC neurons. Given that the PB–CeLC pathway and the central nucleus of the amygdala, including the CeLC, are necessary for conditioned learning about noxious stimuli (Pedersen et al., 2007; Tanimoto et al., 2003) and the aversive component of pain (Han et al., 2015; Sato et al., 2015), it is possible that activation of μ‐receptors within the CeLC or its synaptic inputs contributes to opioid‐mediated inhibition of the affective component of pain (LaGraize et al., 2006; Zhang et al., 2013). Apart from the actions of clinically used drugs, these μ‐receptors could also be activated by endogenous release of enkephalin or β‐endorphin, both of which activate μ‐receptors and are expressed in the CeLC (Le Merrer et al., 2009; Poulin et al., 2006).

Delivery of polymodal information, including complex nociceptive information, from the BLA was inhibited by μ‐receptors in all CeLC neurons and by δ‐receptors and κ‐receptors in 38% and 50% of CeLC neurons respectively. These experiments do not allow us to determine whether there is co‐expression of the opioid receptors on individual BLA neurons. The dominance of μ‐receptor over δ‐receptor inhibition at the BLA–CeLC synapse is surprising in the light of the low expression of μ‐receptors and very high expression of δ‐receptors in the BLA using light‐level immunohistochemistry (Ding et al., 1996; Poulin et al., 2006; Wang et al., 2018). Although there have been some concerns about the specificity of opioid receptor antibodies, in particular those for δ‐receptors, the localization of high levels of δ‐receptors and low levels of μ‐receptors in the BLA is also found in δ‐receptors–GFP and μ‐receptor–mCherry mutant mice, which do not rely on opioid receptor antibodies (Wang et al., 2018). BLA pyramidal neurons do express μ‐receptors, when studied using immuno‐electron microscopy (Zhang, Muller, & Mcdonald, 2015), and activation of μ‐receptors, but not of δ‐receptors, directly inhibits some lateral amygdala pyramidal neurons (Faber & Sah, 2004; Sugita & North, 1993). One possibility is that some of the receptors labelled by immunohistochemistry are not available for ligand activation. Consistent with this, a measure of functional opioid receptors, selective opioid agonist radioligand binding, shows very high levels of μ‐receptors and moderate levels of δ‐receptors in the BLA (Mansour, Khachaturian, Lewis, Akil, & Watson, 1987; Smith, Ratnaseelan, & Veenema, 2018; Xia & Haddad, 2001). Therefore, the radioligand binding (Mansour et al., 1987; Smith et al., 2018; Xia & Haddad, 2001), the direct inhibition of pyramidal neurons (Faber & Sah, 2004; Sugita & North, 1993), our current findings, and our previous work on the BLA‐intercalated neurons synapse (Winters et al., 2017) suggest that functionally, the μ‐receptors are as important, or more so, than δ‐receptors for regulating the synaptic output of BLA pyramidal neurons. Although the inhibition mediated by δ‐receptors was lower than predicted by the immunohistochemistry (Le Merrer et al., 2009; Poulin et al., 2006; Wang et al., 2018), it is actually somewhat unusual to observe inhibition by δ‐receptors, at all, in naïve animals. In naïve animals, a large proportion of the δ‐receptors is localized intracellularly in the cytoplasm (Arvidsson et al., 1995; Cahill et al., 2001; Cheng, Liu‐Chen, & Pickel, 1997), which has been suggested to explain why some areas, such as the PAG, have relatively high expression of δ‐receptors but no δ‐receptor‐mediated activity (Chieng & Christie, 1994; Connor, Schuller, Pintar, & Christie, 1999; Hack, Bagley, Chieng, & Christie, 2005). Activity of δ‐receptors is increased when chronic pain and chronic opioid treatment cause trafficking of these receptors to the plasma membrane (Cahill et al., 2001; Cahill, Morinville, Hoffert, O'donnell, & Beaudet, 2003; Chieng & Christie, 2009). Therefore, the high levels of δ‐receptor immunohistochemistry in the BLA may signify high potential for regulation by δ‐receptors of the BLA outputs to the CeLC under some conditions. Consistent with this, activation of δ‐receptors has minimal effects in acute measures of pain (Cahill et al., 2001; Cahill et al., 2003) but effectively reduces measures of pain in a variety of chronic pain states (Cahill et al., 2003; Mika, Przewlocki, & Przewlocka, 2001; Pradhan, Smith, Zyuzin, & Charles, 2014). These measures of pain include reductions in mechanical allodynia and thermal hyperalgesia (Nadal, Banos, Kieffer, & Maldonado, 2006), pain‐induced conditioned place aversion, and aversive emotional state (Pradhan et al., 2014). It is possible that the inhibition by δ‐receptors of the BLA–CeLC synapse contributes to this. Unlike μ‐receptors and δ‐receptors, the κ‐receptor is not a useful analgesic target in the CNS due to κ‐receptor‐induced dysphoria and aversion (Land et al., 2008; Pfeiffer, Brantl, Herz, & Emrich, 1986). In fact, κ‐receptor activation in the BLA increases measures of aversion (Knoll et al., 2011) Given this, if the κ‐receptors were activated by endogenously released dynorphin that is expressed in the CeLC (Le Merrer et al., 2009), the functional consequences remain unclear.

In conclusion, the use of optogenetics allowed selective activation of the spinal–PB neural pathway delivering nociceptive information into the amygdala. This selective activation of the nociceptive input and local stimulation of polymodal sensory inputs to the nociceptive amygdala allowed us to determine that both inputs to the CeLC are strongly inhibited by activation of μ‐receptors. These two inputs provide CeLC neurons with information to form the association between a painful stimulus and the environment. CeLC neurons project to areas such as the substantia innomimata dorsalis (Bourgeais et al., 2001) relay nociceptive information to areas that encode pain affect. Thus, in the presence of the μ‐receptor agonist morphine, the inhibition of CeLC synaptic inputs could serve as a mechanism through which opioids reduce the affective component of pain and pain‐induced associative learning. Further, while expression of opioid receptors may be segregated into distinct amygdala subnuclei (Le Merrer et al., 2009; Mansour et al., 1994; Poulin et al., 2006; Wang et al., 2018), these findings indicate that functionally, this segregation is not upheld. Care therefore must be taken when attributing distinct functions to opioid receptors within amygdala circuits modulating pain affect, on the basis of their expression profile alone. This is particularly important given the identification of “pain” affect responsive neurons in the BLA (Corder et al., 2019) that could drive development of new analgesics that target δ‐receptors to inhibit the synaptic outputs of these neurons.

AUTHOR CONTRIBUTIONS

This study was conceived by E.E.B. and designed by E.E.B. and S.A.K. Brain microinjections were performed by S.A.K. and B.L.W. Electrophysiology experiments and immunohistochemistry assays and confocal microscopy were conducted by S.A.K. Electrophysiology experiments were conducted by S.D.P. This manuscript was written and edited by S.A.K. and E.E.B.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

DECLARATION OF TRANSPARENCY AND SCIENTIFIC RIGOUR

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for https://bpspubs.onlinelibrary.wiley.com/doi/full/10.1111/bph.14207, https://bpspubs.onlinelibrary.wiley.com/doi/full/10.1111/bph.14208, and https://bpspubs.onlinelibrary.wiley.com/doi/full/10.1111/bph.14206 and as recommended by funding agencies, publishers, and other organizations engaged with supporting research.

Kissiwaa SA, Patel SD, Winters BL, Bagley EE. Opioids differentially modulate two synapses important for pain processing in the amygdala. Br J Pharmacol. 2020;177:420–431. 10.1111/bph.14877