Abstract
Endogenous adenosine plays a crucial role in maintaining energy homeostasis, and adenosine levels are tightly regulated across neural circuits. In the dorsal medial striatum (DMS), adenosine inhibits neurotransmitter release, but the source and mechanism underlying its accumulation are largely unknown. Opioids also inhibit neurotransmitter release in the DMS and influence adenosine accumulation after prolonged exposure. However, how these two neurotransmitter systems interact acutely is also largely unknown. This study demonstrates that activation of µ opioid receptors, but not δ opioid receptors or κ opioid receptors, inhibits tonic activation of adenosine A1Rs via a cAMP-dependent mechanism in both male and female mice. Further, selectively knocking out µ opioid receptors from thalamic presynaptic terminals and postsynaptic medium spiny neurons (MSNs) revealed that activation of µ opioid receptors on D1R-positive MSNs, but not D2R-positive MSNs, is necessary to inhibit tonic adenosine signaling on presynaptic terminals. Given the role of D1R-positive MSNs in movement and motivated behaviors, these findings reveal a novel mechanism by which these neurons regulate their own synaptic inputs.
SIGNIFICANCE STATEMENT Understanding interactions between neuromodulatory systems within brain circuits is a fundamental question in neuroscience. The present work uncovers a novel role of opioids in acutely inhibiting adenosine accumulation and subsequent adenosine receptor signaling in the striatum by inhibiting the production of cAMP. Adenosine receptor signaling regulates striatal neurotransmitters, including glutamate, GABA, dopamine, and acetylcholine. Furthermore, interactions between adenosine2A receptors and numerous other GPCRs, including D2 dopamine and CB1 cannabinoid receptors, suggest that endogenous adenosine broadly modulates striatal GPCR signaling. Additionally, this work discovered that the source of resting endogenous extracellular adenosine is likely D1, but not D2 receptor-positive medium spiny neurons, suggesting that opioid signaling and manipulation of D1R-expressing medium spiny neuron cAMP activity can broadly affect striatal function and behavior.
Introduction
Opioids such as morphine acutely mediate analgesia, and long-term use leads to dependence and potentially addiction. The thalamus and dorsal medial striatum (DMS) are important for regulating opioid dependence and modulating goal-directed behavior, respectively (Balleine et al., 2007; Bertran-Gonzalez et al., 2013; Zhu et al., 2016). Opioids are known to inhibit both excitatory input to the striatum and local GABA release within the striatal microcircuitry (Atwood et al., 2014; Banghart et al., 2015; Birdsong et al., 2019). Additionally, agonists selective to the adenosine A1Rs also inhibit glutamate release in the striatum (Brundege and Williams, 2002) and influence striatal dynamics. Thus, understanding the role of opioid receptors and A1Rs in modulating excitatory inputs to the striatum, and the potential interaction between these receptors, is important for understanding how multiple neurotransmitter systems influence striatal activity.
Opioids binding to µ opioid receptor (MOR) activate Gi/o heterotrimeric G-proteins to inhibit adenyl cyclase (AC) and consequently decrease cAMP levels (Heijna et al., 1992; Izenwasser et al., 1993). Acutely, this inhibition of cAMP, along with other effectors, ultimately inhibits neuronal activity and neurotransmitter release. Activation of the A1Rs also inhibits AC and, under basal conditions, there is a resting extracellular adenosine tone in the striatum. This resting adenosine tone can tonically activate A1Rs, inhibiting neurotransmitter release (Brundege and Williams, 2002). The fact that adenosine and opioids both act through the same effector systems suggests that these two neurotransmitters can influence each other's signaling. But neither the role of opioids in modulating resting adenosine levels nor the source of this resting adenosine tone is known.
Agonists selective to MOR, but not δ opioid receptor (DOR), potently inhibit glutamate release from medial thalamus onto the striatum (Atwood et al., 2014; Birdsong et al., 2019; Muñoz et al., 2020), but the role of κ opioid receptors (KORs) in this circuit has not been examined. However, there is evidence that MORs, DORs, and KORs are widely expressed throughout the striatum (Mansour et al., 1994; Nestler and Carlezon, 2006; Muschamp and Carlezon, 2013; Al-Hasani et al., 2015; Banghart et al., 2015; Massaly et al., 2019) and have been shown to inhibit neurotransmitter release in the striatum to varying degrees in a synapse-specific manner (Atwood et al., 2014; Tejeda et al., 2017; Birdsong et al., 2019). Therefore, although all three subtypes of opioid receptors are present in the striatum, they potentially modulate the activity of the striatum and interact with A1R signaling uniquely.
The present study examines the functional interaction between opioid receptors and adenosine signaling, the mechanism underlying extracellular adenosine accumulation, and the source of adenosine release in the striatum using a combination of brain slice electrophysiology, pharmacology, optogenetics, and genetic manipulation of MOR expression in mice. Optically induced EPSCs (oEPSCs) were recorded in striatal medium spiny neurons (MSNs) following optical excitation of channelrhodopsin-expressing medial thalamic axon terminals in the dorsomedial striatum. The facilitation of oEPSC amplitude by the A1R antagonist 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) was used to measure tonic A1R activation and as a proxy for extracellular adenosine accumulation. The results show that (1) morphine inhibits tonic adenosine accumulation by inhibiting cAMP, (2) this inhibition is specific to MOR agonists and not DOR or KOR agonists, and (3) MOR regulation of dorsomedial striatal adenosine levels requires MOR expression on D1R-positive MSNs.
Materials and Methods
Animals
Male and female C57BL/6J mice (8-10 weeks old) were bred in house and were housed under a 12 h-light/dark cycle. Food and water were available ad libitum. Mice with exons 2 and 3 of the oprm1 gene flanked by the LoxP cassette (FloxedMor; Oprm1fl/fl; JAX stock #030074), with a genetic background of 75%:25% of C57BL/6J were provided by Brigitte L. Kieffer. Vglut2-cre mice (Slc17a6tm2(cre)Lowl; JAX stock #016963) were purchased from The Jackson Laboratory. The 2 mice were crossed to generate FloxedMor-Vglut2-cre mice that lack MORs in presynaptic terminals. A2A-cre mice (Tg(Adora2a-cre)KG139Gsat; MMRC stock #036158-UCD) were provided by Tianyi Mao and were crossed with FloxedMor mice to generate FloxedMor-A2A-cre mice lacking MORs in D2-positive MSNs. D1-cre mice (Tg(Drd1-cre)EY262Gsat; MMRC stock #030989-UCD) were provided by Christopher Ford and were crossed with FloxedMOR mice to generate FloxedMOR-D1-cre mice lacking MORs in D1-positive MSNs.
All animal experiments were conducted in accordance with the National Institutes of Health guidelines and with approval from the Institutional Animal Care and Use Committee of the Oregon Health & Science University.
Viral injection
Mice were anesthetized with isoflurane and placed in a stereotaxic frame (Kopf Instruments, with custom modifications) for microinjections of recombinant adeno-associated virus (AAV2-syn-CsChR-GFP) to express channelrhodopsin (Klapoetke et al., 2014). A glass pipette filled with 40 nl of virus was injected into the medial thalamus (Nanoject II, Drummond Scientific; Kopf Model 1900 Stereotaxic Frame, David Kopf Instruments). Injection coordinates for medial thalamus are in millimeters for ML, AP from bregma, and DV from the top of the skull directly over the target area: ML, ±0.55; AP, −1.2; DV, −3.4. Electrophysiology experiments were done 2-3 weeks after viral injections.
Drugs
Morphine sulfate was obtained from the National Institute on Drug Abuse. Naloxone was from Abcam. Dizocilpine maleate (MK801) was from Abcam or Tocris Bioscience. 8-Cyclopentyl-1,3-dipropylxanthine DPCPX was obtained from Tocris Bioscience or Sigma-Aldrich. CGS21680, SKF81297, mecamylamine, CGP 55845, and MPEP were from Tocris Bioscience. Scopolamine, adenosine, [Met5] enkephalin (ME), bestatin, thiorphan, and Ro-20-1724 were from Sigma-Aldrich. Picrotoxin was from Hello Bio. ME, morphine, adenosine, naloxone, MPEP, scopolamine, and mecamylamine were dissolved in water, diluted in ACSF, and applied by bath superperfusion. Bath perfusion of ME was with bestatin (10 μm) and thiorphan (1 μm) to limit breakdown of ME. Picrotoxin was directly dissolved in ACSF. DPCPX, CGS21680, SCH58261, SKF81297, and Ro-230853 were dissolved in DMSO, diluted in ACSF, and applied during incubation and by bath superperfusion.
Tissue preparation
Acute brain slice preparation was performed as previously described (Birdsong et al., 2019). Briefly, mice were deeply anesthetized and killed using isoflurane. Brains were removed, blocked, and mounted in a vibratome chamber (VT 1200S; Leica Microsystems). Coronal slices (242 μm) were prepared in warm (34°C) ACSF containing the following (in mm): 126 NaCl, 2.5 KCl, 1.2 MgCl2, 2.6 CaCl2, 1.2 NaH2PO4, 21.4 NaHCO3, and 11 D-glucose with +MK-801 (10 μm). Slices were allowed to recover in warm ACSF containing +MK-801 (10 μm) for at least 30 min and then stored in glass vials at room temperature with oxygenated (95% O2/5% CO2) ACSF until used.
Brain slice electrophysiology
Slices were hemisected and then transferred to the recording chamber, which was continuously superfused with 34°C carbogenated ACSF at 1.5-2 ml/min with the following (in µm): 0.2 GABAB-receptor antagonist CGP 55845, 10 GABAA-receptor antagonist picrotoxin, 1 nAChR antagonist mecamylamine, 0.1 muscarinic AChR antagonist scopolamine, and 0.3 mGluR5 antagonist MPEP. Whole-cell recordings from MSNs in the DMS were obtained by Sweta Adhikary with an Axopatch 200A amplifier (Molecular Devices) in voltage-clamp mode, holding potential (Vhold = −75 mV). Data were filtered at 10 kHz and collected at 20 kHz with AxographX. For data collected by Elizabeth Jaeckel, recordings were obtained using a Multiclamp 700B amplifier (Molecular Devices), holding potential (Vhold = −70 mV). Data were filtered at 10 kHz and collected at 20 kHz with MATLAB-based Wavesurfer software. Recording pipettes (Sutter Instruments) with a resistance of 2.8-3.5 mΩ were filled with an internal solution of the following (in mm): 110 potassium gluconate, 10 KCl, 15 NaCl, 1.5 MgCl2, 10 HEPES, 1 EGTA, 2 Na2ATP, 0.3 Na2GTP, 7.8 phosphocreatine, pH 7.35-7.40 (∼280 mOsm). Only recordings in which the series resistance remained <18 mΩ or changed by <20% throughout the experiment were included. A TTL-controlled LED driver and 470 nm LED (Thorlabs) was used to illuminate the slice through the microscope objective directly over the recorded cell with ∼1 mW of power for 0.5 or 1 ms.
Electrophysiology data analysis
Data were analyzed in Axograph. Peak current amplitude was calculated relative to mean current during 50 ms baseline before the stimulus. Statistical analysis was performed using GraphPad Prism 8 software (GraphPad Software). For the time course and summary data, baseline oEPSCs were normalized to oEPSC amplitudes 3-4 min before baseline (prebaseline condition, not shown). All other conditions were normalized to oEPSC amplitudes 3-4 min before drug application. Summary data were presented as the averages of 6-10 trials beginning 3-4 min after drug application after steady state was achieved. For all conditions, mice were used to obtain at least five technical replicates per group; if more than six could be analyzed, all were included. No difference in adenosine tone was observed between male and female mice across all genotypes; therefore, all data were pooled (p = 0.27, unpaired t test). Values are presented as mean ± SEM. Statistical analysis was performed on normalized data. Statistical comparisons were made with paired ratio t test, one-way repeated-measures ANOVA, or one-way ANOVA, followed by multiple comparison-adjusted Tukey's post hoc tests. For all experiments, p < 0.05 was used to describe statistical significance.
Results
Thalamo-striatal glutamate release is sensitive to both opioid and adenosine agonists
AAV2 encoding channelrhodopsin was microinjected into the medial thalamus, and whole-cell voltage-clamp recordings were made from MSNs in the DMS (Fig. 1A). Striatal MSNs were identified by their hyperpolarized resting membrane potential, low input resistance, and a long delay to the initial spike (Kreitzer, 2009). Glutamate release was evoked by optical stimulation with 470 nm light, and AMPAR-mediated EPSCs (oEPSCs) were pharmacologically isolated and recorded as previously described (Birdsong et al., 2019). After a stable baseline of oEPSCs was established, clinically relevant partial agonist morphine (1 μm) was superperfused, followed by antagonist naloxone (1 μm) (Fig. 1C,E). The inhibition by morphine was determined by averaging the oEPSC 3-5 min after drug perfusion and normalizing the response to the average of oEPCS 3-5 min before drug perfusion. Morphine decreased the amplitude of the oEPSCs, and this inhibition was reversed by naloxone (Fig. 1C,E,F; morphine: 0.80 ± 0.05 fraction of baseline, p = 0.0002; naloxone: 0.98 ± 0.01 fraction of baseline, p = 0.002, n = 8 cells from 4 male and 1 female mice, F(2,14) = 17.29, one-way repeated-measures ANOVA, Tukey's multiple comparisons test). In separate cells, an A1R agonist cyclopentyladenosine (CPA, 1 μm) was superperfused, followed by antagonist DPCPX, 200 nm) (Fig. 1D,E). CPA decreased the amplitude of the oEPSCs, and this inhibition was reversed by DPCPX (Fig. 1D,E,G; CPA 0.37 ± 0.04 fraction of baseline, p < 0.0001; DPCPX: 1.3 ± 0.08 fraction of baseline, p < 0.0001, n = 7 cells from 3 male and 3 female mice, F(2,12) = 87.95, one-way repeated-measures ANOVA, Tukey's multiple comparisons test). Furthermore, CPA significantly increased the paired-pulse ratio (PPR) (see Fig. 6A,B), consistent with presynaptic inhibition of neurotransmitter release on A1R activation (1.68 ± 0.30 times larger than baseline PPR, ratio paired t test, p = 0.023, n = 6 cells from 3 male and 3 female mice). Additionally, DPCPX caused a significant over-reversal of the amplitude of the oEPSCs (Fig. 1D,E,G), suggesting tonic inhibition of glutamate release by activation of A1Rs, that was blocked by DPCPX. Thus, glutamate release in the thalamo-striatal synapses is regulated by both MORs and A1 receptors, and there is an additional tonic activation of A1Rs by endogenous adenosine.
Activation of both µ opioid receptor and adenosine A1 receptor leads to inhibition of thalamo-striatal oEPSCs. A, An acute mouse brain slice example of overlaid brightfield and epifluorescent images showing the viral injection site (Mthal; left) and the axonal projections (Striatum; right). B, Schematic showing the locations of both A1Rs and MORs in the thalamo-striatal synapse. C, Representative oEPSCs evoked by 470 nm light (black label), inhibition of oEPSC amplitude by morphine (1 μm; pink label), and reversal by naloxone (1 μm; gray label). D, Representative oEPSCs evoked by 470 nm light (black label), inhibition of oEPSC amplitude by CPA (1 μm; orange label), and over-reversal by DPCPX (200 nm; blue label). E, Plot of the time course of normalized oEPSC amplitude for cells treated with morphine, followed by naloxone (dark circles; n = 8 cells, 4 mice), and for cells treated with CPA, followed by DPCPX (clear circles; n = 7 cells, 6 mice). F, Mean summary data of normalized oEPSC amplitude in baseline condition, after morphine perfusion, followed by naloxone (morphine: 0.80 ± 0.05 fraction of baseline, p = 0.0002; naloxone: 0.98 ± 0.01 fraction of baseline, p = 0.002, n = 8 cells, 4 mice, F(2,14) = 17.29, one-way repeated-measures ANOVA, Tukey's multiple comparisons test). G, Mean summary data of normalized oEPSC amplitude in baseline condition, after CPA perfusion, followed by DPCPX (CPA 0.37 ± 0.04 fraction of baseline, p < 0.0001; DPCPX: 1.3 ± 0.08 fraction of baseline, p < 0.0001, n = 7 cells, 6 mice, F(2,12) = 87.95, one-way repeated-measures ANOVA, Tukey's multiple comparisons test). Line and error bars represent mean ± SEM. *Statistical significance.
MOR regulation of tonic adenosine A1 receptor activation
Since both MORs and A1Rs are coupled to Gi/o G-proteins and both are present in thalamic terminals, functional interaction between the two receptors in regulating glutamate release was examined. oEPSCs were evoked as described above and DPCPX (200 nm) was superperfused to measure the effect of tonic A1R activation. DPCPX increased oEPSC amplitude (Fig. 2A,C,D; DPCPX: 1.3 ± 0.06 fraction of baseline, p = 0.0003, n = 10 cells from 5 male and 2 female mice, t(9) = 5.752, ratio paired t test). Additionally, DPCPX decreased the PPR (see Fig. 6C,D), consistent with a presynaptic facilitation of neurotransmitter release on A1R antagonism (0.81 ± 0.06 times smaller than baseline PPR, ratio paired t test, p = 0.03, n = 10 cells from 5 male and 3 female mice), although this does not rule out additional sites of action of A1R signaling. In separate cells, morphine (1 μm) was superperfused, followed by DPCPX. Morphine reduced the amplitude of oEPSCs (Fig. 2B,C,E) as expected. However, in the presence of morphine, DPCPX did not increase oEPSC amplitude (Fig. 2B,C,E; morphine 0.78 ± 0.03 fraction of baseline, p = 0.0011; DPCPX: 0.77 ± 0.05 fraction of baseline p = 0.04, and 0.99 ± 0.06 fraction of morphine, p = 0.84, n = 6 cells from 2 male and 2 female mice, F(2,10) = 14.00, one-way repeated-measures ANOVA, Tukey's multiple comparisons test), suggesting that morphine inhibited the tonic activation of A1Rs. To determine whether morphine inhibited the tonic A1R activation through MOR activation or a nonspecific morphine effect, global MOR KO mice were used. Similar to WT mice, DPCPX increased the oEPSC amplitude in slices from these mice (Fig. 2F,H,I; DPCPX: 1.37 ± 0.05 fraction of baseline, p = 0.0001, n = 8 cells from 3 male and 2 female mice, t(7) = 8.273, ratio paired t test). In contrast to WT mice, morphine did not reduce the amplitude of oEPSCs (Fig. 2G,H,J: morphine: 1.0 ± 0.05 fraction of baseline, p = 0.9863, n = 6 cells from 1 male and 2 female mice) in slices from MOR KO mice. Further, in the presence of morphine, DPCPX increased oEPSC amplitude (Fig. 2G,H,J; DPCPX: 1.4 ± 0.07 fraction of baseline, p = 0.0006, 1.3 ± 0.09 fraction morphine, p = 0.0008, n = 6 cells from 1 male and 2 female mice, F(2,12) = 17.46, one-way repeated-measures ANOVA, Tukey's multiple comparisons test). There was no difference in the increase in oEPSC amplitude between control slices and slices in morphine, suggesting that MORs are required for morphine to modulate tonic adenosine levels. Therefore, morphine inhibits the tonic activation of A1Rs in the thalamo-striatal circuit by activating MORs.
Morphine inhibits adenosine tone in the thalamo-striatal synapse by activating MORs. A, Representative traces of oEPSCs evoked by 470 nm light (black label) and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label). B, Representative traces of oEPSCs evoked by 470 nm light (black label), inhibition of oEPSC amplitude by morphine (1 μm; pink label), and lack of facilitation by DPCPX (200 nm; blue label). C, Plot of the time course of normalized oEPSC amplitude for cells superperfused with DPCPX (dark circles; n = 10 cells, 7 mice), and for cells superperfused with morphine and then DPCPX (clear circle; n = 6 cells, 4 mice). D, Mean summary data of normalized oEPSC amplitude in control and after DPCPX (1.3 ± 0.06 fraction of baseline, p = 0.0003, n = 10 cells, 7 mice, t(9) = 5.752, ratio paired t test). E, Mean summary data of normalized oEPSC amplitude in control, after morphine superperfusion, and after DPCPX superperfusion. Morphine significantly inhibited oEPSC amplitude (morphine 0.78 ± 0.03 fraction of baseline, p = 0.0011; DPCPX: 0.77 ± 0.05 fraction of baseline, p = 0.04, and 0.99 ± 0.06 fraction of morphine, p = 0.84, n = 6 cells, 4 mice, F(2,10) = 14.00, one-way repeated-measures ANOVA, Tukey's multiple comparisons test). F, Representative traces of oEPSCs evoked by 470 nm light (black label) and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label), in slices from global MOR KO mice. G, Representative traces of oEPSCs evoked by 470 nm light (black label), lack of inhibition by morphine (1 μm; pink label), and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label), in slices from global MOR KO mice. H, Plot of the time course of normalized oEPSC amplitude for cells superperfused with DPCPX (dark circles; n = 8 cells, 5 mice), and for cells superperfused with morphine and then DPCPX (clear circle; n = 6 cells, 3 mice). I, Mean summary data of normalized oEPSC amplitude in control and after DPCPX (1.37 ± 0.05 fraction of baseline, p = 0.0001, n = 8 cells, 5 mice, t(7) = 8.273, ratio paired t test). J, Mean summary data of normalized oEPSC amplitude in control, after morphine superperfusion, and after DPCPX superperfusion. Morphine did not inhibit oEPSC amplitude (morphine: 1.0 ± 0.05 fraction of baseline, p = 0.9863), and there was facilitation by DPCPX in the presence of morphine (1.4 ± 0.07 fraction of baseline, p = 0.0006, 1.3 ± 0.09 fraction morphine, p = 0.0008, n = 6 cells, 3 mice, F(2,12) = 17.46, one-way repeated-measures ANOVA, Tukey's multiple comparisons test). Line and error bars represent mean ± SEM. *Statistical significance.
Tonic endogenous activation of A1Rs is regulated by cAMP levels
MOR is a Gi/o-coupled GPCR that can inhibit adenylyl cyclase, so it is possible that morphine decreased tonic adenosine levels by preventing cAMP production and its subsequent metabolism to adenosine. Therefore, the role of cAMP metabolism on A1R-mediated inhibition of glutamate release was examined. Slices were pretreated with phosphodiesterase (PDE) inhibitor, Ro-20-1724 (400 μm) for at least an hour to block metabolism of cAMP. Ro-20-1724 (400 μm) was also in the perfusate throughout the course of the experiment. In the presence of Ro-20-1724, unlike in control slices, DPCPX (200 nm) did not cause an increase in oEPSC amplitude (Fig. 3A–E; DPCPX 1.47 ± 0.13, p = 0.03, n = 6 cells from 2 male and 2 female mice, in control; Fig. 3B,C,E; DPCPX 0.96 ± 0.1 fraction of baseline, p = 0.789, n = 6 cells from 3 male and 1 female mice, in Ro-201724, F(3,17) = 13.51, one-way repeated-measures ANOVA, Tukey's multiple comparisons test), suggesting that inhibiting the metabolism of cAMP, and thus the conversion of cAMP to adenosine, blocked the tonic activation of A1Rs. Exogenous application of adenosine (100 μm) in either the presence or absence of Ro-20-1724 decreased the oEPSC amplitude, which was reversed by a washout (Fig. 3A–E; adenosine 0.52 ± 0.08 fraction of baseline, p = 0.0001; washout: 0.87 ± 0.06 of baseline, p = 0.0001, n = 6 cells from 2 male and 2 female mice in Ro-201724; adenosine 0.44 fraction ± 0.07 of baseline, p = 0.0094; washout: 1.0 ± 0.05 of baseline, p = 0.999, n = 6 cells from 3 male and 1 female mice, in control, F(3,16) = 36.72, one-way repeated-measures ANOVA, Tukey's multiple comparisons test), demonstrating that Ro-20-1724 was not directly antagonizing the ability of adenosine to inhibit glutamate release via A1Rs.
Morphine inhibits adenosine signaling via a cAMP-dependent mechanism. A, Representative traces of oEPSCs evoked by 470 nm light (black label), inhibition of oEPSC amplitude by adenosine (100 μm; yellow label), washout of adenosine (gray label), and facilitation of oEPSC by DPCPX (200 nm; blue label) in naive conditions. B, Representative traces of oEPSCs evoked by 470 nm light (black label), inhibition of oEPSC amplitude by adenosine (100 μm; yellow label), washout of adenosine (gray label), and lack of facilitation of oEPSC by DPCPX (200 nm; blue label) in slices preincubated in Ro-20-1724. C, Plot of the time course of normalized oEPSC amplitude for cells superperfused with adenosine followed by washout and then DPCPX in naive slices (dark circles, n = 6 cells, 4 mice) and in slices preincubated in Ro-20-1724 (n = 6 cells, 4 mice). D, Mean summary data of normalized oEPSC amplitude for naive slices in baseline condition, after adenosine superperfusion, followed by a washout and then DPCPX. Adenosine significantly reduced oEPSC amplitude in naive slices, and DPCPX significantly facilitated oEPSC in these slices (adenosine 0.44 ± 0.07 fraction of baseline, p = 0.0094; washout: 1.0 ± 0.05 of baseline, p = 0.999, n = 6 cells, 4 mice, F(3,17) = 13.51, repeated-measures ANOVA, Tukey's multiple comparisons test). E, Mean summary data of normalized oEPSC amplitude for slices incubated in Ro-20-1724 in baseline condition, after adenosine superperfusion, followed by a washout and then DPCPX. Adenosine significantly reduced oEPSC amplitude in these slices (adenosine 0.52 ± 0.08 fraction of baseline, p = 0.0001; washout: 0.87 ± 0.06 of baseline, p = 0.0001, n = 6 cells, 4 mice, F(3,17) = 13.51, repeated-measures ANOVA, Tukey's multiple comparisons test), but DPCPX did not significantly facilitate oEPSCs in these slices (DPCPX: 0.96 ± 0.10 fraction of baseline, p = 0.8755, repeated-measures ANOVA, Tukey's multiple comparisons test). F, Representative traces of oEPSCs evoked by 470 nm light (black label) and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label), in control slices. G, Representative traces of oEPSCs evoked by 470 nm light (black label) and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label), in slices preincubated in CGS21680. H, Representative traces of oEPSCs evoked by 470 nm light (black label) and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label), in slices preincubated in SKF81290. I, Plot of the time course of normalized oEPSC amplitude for cells superperfused with DPCPX in control condition (dark circles; n = 7 cells, 5 mice), for cells preincubated in SKF81290 (clear circles; n = 7 cells, 6 mice), and for cells preincubated in CGS21680 (gray circles; n = 6 cells, 4 mice). J, Mean summary data of normalized oEPSC amplitude in control, in slices preincubated in SKF81297, and CGS21680. The increase in amplitude induced by DPCPX was significantly higher in slices treated with SKF89217 (DPCPX 1.6 ± 0.11 fraction of baseline, p = 0.003) and in slices treated with CGS21680 (p < 0.001, F(5,32) = 32.24, one-way ANOVA, Tukey's multiple comparisons test) compared with control slices. Line and error bars represent mean ± SEM. *Statistical significance.
In order to examine whether tonic A1R signaling could be increased by increasing cAMP concentration, Gs-coupled GPCRs in both D1R- and D2R-positive MSNs were pharmacologically activated. Slices were preincubated in D1R-specific agonist SKF89217 (1 μm) for at least an hour, with the drug in the perfusate throughout the course of the experiment. DPCPX (200 nm) caused an increase in oEPSC amplitude (Fig. 3F,H–J; DPCPX 1.6 ± 0.11 fraction of baseline, n = 7 cells from 3 males and 2 female mice). The increase in amplitude induced by DPCPX was significantly higher in slices treated with SKF89217 (p = 0.003, F(5,32) = 32.24, one-way ANOVA, Tukey's multiple comparisons test) compared with control slices. Next, slices were incubated in A2AR agonist CGS21680 (1 μm) for at least an hour, with the drug in the perfusate throughout the course of the experiment. A2ARs colocalize with D2R-positive MSNs only (Fink et al., 1992; Bogenpohl et al., 2012; Severino et al., 2020). DPCPX (200 nm) increased oEPSC amplitude (Fig. 3F,G,I,J; DPCPX 1.76 ± 0.10 fraction of baseline, n = 6 cells from 3 male and 1 female mice). The increase in amplitude induced by DPCPX was also significantly higher in slices treated with CGS21680 (p < 0.0001, F(5,32) = 32.24, one-way ANOVA, Tukey's multiple comparisons test) compared with control slices. There was no difference in oEPSC amplitude after DPCPX superperfusion between slices incubated in SKF89217 and CGS21680 (one-way ANOVA, Tukey's multiple comparisons test). From these experiments, it appears that basal A1R-mediated inhibition of glutamate signaling was affected by manipulating cAMP concentration, consistent with cAMP production and degradation regulating extracellular adenosine accumulation and signaling. Activation of MORs by morphine is known to inhibit cAMP accumulation, which could consequently decrease extracellular adenosine levels preventing A1R activation by adenosine.
Inhibition of cAMP by activation of MORs is reversible
The time dependence of cAMP inhibition by MOR activation was examined next. ME (1 μm) was used instead of morphine, as ME washes from brain slices. oEPSCs were induced as previously described, and ME (1 μm) was superperfused. Like morphine, ME inhibited oEPSCs and DPCPX failed to facilitate oEPSCs in the presence of ME (Fig. 4A–C; ME 0.67 ± 0.03 fraction of baseline, p = 0.0002; DPCPX 0.56 ± 0.03 fraction of baseline, p < 0.0001; DPCPX 0.84 ± 0.05 fraction of ME, p = 0.2867, n = 6 cells from 2 male and 2 female mice, F(3,15) = 78.77, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). ME washed out of the slices in ∼5 min. Following ME washout, there was an over-reversal of oEPSC in the presence of DPCPX (DPCPX 1.37 ± 0.07 fraction of baseline, p < 0.0001, repeated-measures one-way ANOVA, Tukey's multiple comparisons test), which reached steady state in ∼7 min, suggesting that inhibition of cAMP, and therefore inhibition of tonic adenosine levels, may reverse within this 7 min time window.
Inhibition of adenosine signaling by opioids is reversible. A, Representative traces of oEPSCs evoked by 470 nm light (black label), inhibition of oEPSC amplitude by ME (1 μm; pink label), lack of facilitation by DPCPX (200 nm; blue label), and an over-reversal of oEPSC after ME washout (gray label). B, Plot of the time course of normalized oEPSC amplitude for cells superperfused with ME, followed by DPCPX in the presence of ME, and then a washout of ME, but not DPCPX (n = 6 cells, 4 mice). C, Mean summary data of normalized oEPSC amplitude in baseline condition, after ME superperfusion, followed by DPCPX, and a washout of ME, but not DPCPX (ME 0.67 ± 0.03 fraction of baseline, p = 0.0002; DPCPX 0.56 ± 0.03 fraction of baseline, p < 0.0001; DPCPX 0.84 ± 0.05 fraction of ME, p = 0.2867, n = 6 cells, 4 mice, F(3,15) = 78.77, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). Line and error bars represent mean ± SEM. *Statistical significance.
DORs and KORs do not regulate tonic activation of adenosine A1 receptors
To determine whether the loss of A1R tone by opioids was specific to MOR signaling, the effect of DOR and KOR activation on tonic activation of A1Rs was examined. oEPSCs were evoked as described above, and the DOR-selective agonist deltorphin (300 nm) was superperfused, followed by DPCPX (200 nm). Unlike morphine, deltorphin did not reduce the amplitude of oEPSCs; and in the presence of deltorphin, DPCPX increased oEPSC amplitude (Fig. 5A,C,D; deltorphin 1.0 ± 0.04 fraction of baseline, p = 0.90; DPCPX: 1.42 ± 0.07 fraction of baseline, p = 0.0002, and 1.40 ± fraction of deltorphin, p = 0.0004, n = 6 cells from 2 male and 1 female mice, F(2,10) = 24.60, repeated-measures one-way ANOVA, Tukey's multiple comparisons test), suggesting that DOR activation did not affect the tonic activation of A1Rs. Next, in separate cells, the KOR-selective agonist U69593 (1 μm) was superperfused, followed by DPCPX (200 nm). Similar to deltorphin, U69593 did not inhibit oEPSC; and in the presence of U69593, DPCPX increased oEPSC amplitude (Fig. 5B,C,E; U69593 1.02 ± 0.06 fraction of baseline p = 0.9670; DPCPX: 1.6 ± 0.09 fraction of baseline, p = 0.0051, and 1.6 ± 0.13 fraction of U69593, p = 0.0035, n = 6 cells from 2 male and 2 female mice, F(2,10) = 12.24, repeated-measures one-way ANOVA, Tukey's multiple comparisons test), suggesting that KOR activation, like DOR activation, did not inhibit glutamate release from thalamic terminals or affect the tonic activation of A1Rs. Hence, both the direct inhibition of glutamate release from thalamic afferents and the inhibition of tonic activation of A1Rs seems to be agonist specific, both inhibited only by MOR agonists and not DOR or KOR agonists.
DORs and KORs do not mediate inhibition of adenosine signaling. A, Representative traces of oEPSCs evoked by 470 nm light (black label), lack of inhibition of oEPSC amplitude by deltorphin (300 nm; pink label), and facilitation by DPCPX (200 nm; blue label). B, Representative traces of oEPSCs evoked by 470 nm light (black label), lack of inhibition of oEPSC amplitude by U69593 (1 μm; pink label), and facilitation by DPCPX (200 nm; blue label). C, Plot of the time course of normalized oEPSC amplitude for cells superperfused with deltorphin (black circles), followed by DPCPX (n = 6 cells, 3 mice), and for cells superperfused with U69 (clear circles), followed by DPCPX (n = 6 cells, 4 mice). D, Mean summary data of normalized oEPSC amplitude in baseline condition, after deltorphin superperfusion, followed by DPCPX (deltorphin: 1.0 ± 0.04 fraction of baseline, p = 0.90; DPCPX: 1.42 ± 0.07 fraction of baseline, p = 0.0002, and 1.40 ± fraction of deltorphin, p = 0.0004, n = 6 cells, 3 mice, F(2,10) = 24.60, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). E, Mean summary data of normalized oEPSC amplitude in baseline condition, after U69593 superperfusion, followed by DPCPX (U69593: 1.02 ± 0.06 fraction of baseline, p = 0.9670; DPCPX: 1.6 ± 0.09 fraction of baseline, p = 0.0051, and 1.6 ± 0.13 fraction of U69593, p = 0.0035, n = 6 cells, 4 mice, F(2,10) = 12.24, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). Line and error bars represent mean ± SEM. *Statistical significance.
Presynaptic effects of MOR agonists in the thalamo-striatal circuit
Because the inhibition of tonic adenosine release by opioids was selectively mediated by MORs and MORs are expressed in the thalamic terminals and both D1R-positive and D2R-positive MSNs, the location of acute action of MOR agonist on A1R signaling was investigated. PPR changes induced by CPA and DPCPX suggested that A1Rs are expressed on thalamic presynaptic terminals in striatum (Fig. 6A–D). Potential mechanisms by which morphine could disrupt A1R signaling include activation of presynaptic MORs locally decreasing adenosine around thalamic terminals or attenuating A1R signaling within these terminals. To test whether these hypotheses might be true, MOR was knocked out of thalamic neurons. FloxedMOR (Oprm1fl/fl) mice and Vglut2:cre mice were crossed to generate mice lacking MORs from Vglut2-expressing thalamic presynaptic terminals (Oprm1fl/fl, Vglut2-cre +/−) (Vong et al., 2011). FloxedMOR homozygous littermates that did not express Cre were used as controls (Oprm1fl/fl, Vglut2-cre −/−). oEPSCs were evoked as previously described. In control mice, superperfusion of the MOR agonist DAMGO (1 μm) decreased the amplitude of the oEPSCs, and this inhibition was reversed by the antagonist naloxone (1 μm) (Fig. 6E,G,H; DAMGO 0.39 ± 0.05 fraction of baseline, p < 0.0001; naloxone: 0.80 ± 0.06 of baseline, n = 8 cells from 2 male and 2 female mice, F(2,14) = 29.9, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). In FloxedMOR-Vglut2-cre mice lacking MORs in the presynaptic terminals, DAMGO did not inhibit the amplitude of the oEPSCs (Fig. 6F,G,I; DAMGO: 0.99 ± 0.02 fraction of baseline, p = 0.6023; naloxone: 0.92 ± 0.05 fraction of baseline, p = 0.14, n = 7 cells from 4 male and 2 female mice, F(2,12) = 2.08, repeated-measures one-way ANOVA, Tukey's multiple comparisons test), suggesting that opioid action on the thalamo-striatal glutamate release is presynaptic and demonstrating the effectiveness of cre-dependent KO in these animals.
Presynaptic A1Rs and MORs inhibit thalamo-striatal oEPSCs. A, Representative traces of paired oEPSCs evoked by two optical stimuli (1 ms duration, 50 ms interval, 470 nm LED, cyan lines) from medial thalamic axons under baseline conditions (black) and in the presence of CPA (1 μm, orange) demonstrating a decrease in current amplitude. Right, oEPSCs from left normalized to the peak amplitude of the first oEPSC of the pair to demonstrate PPR between baseline and CPA conditions. B, Summary PPR data under baseline conditions and in the presence of CPA (1 μm) as in A, calculated as the peak amplitude of the second oEPSC/peak amplitude of first oEPSC (p = 0.027, n = 6 cells, 6 mice, t(7) = 3.085, ratio paired t test). C, Representative oEPSCs as in A, evoked under baseline conditions (black) and in the presence of DPCPX (200 nm, blue) plotted as absolute amplitude (left) or normalized to the peak of the first oEPSC (right). D, Summary PPR under baseline conditions and in the presence of DPCPX (200 nm) as in B (p = 0.020, n = 10 cells, 8 mice, t(9) = 2.836, ratio paired t test, n = 10 cells, 8 mice). E, Representative traces of oEPSCs evoked by 470 nm light (black label), inhibition of oEPSC amplitude by DAMGO (1 μm; pink label), and reversal by naloxone (1 μm; gray label) in control mice expressing MORs in presynaptic terminals. F, Representative traces of oEPSCs evoked by 470 nm light (black label), lack of inhibition of oEPSC amplitude by DAMGO (1 μm; pink label), and no effect of naloxone (1 μm; gray label) in mice lacking MORs in presynaptic terminals. G, Plot of the time course of normalized oEPSC amplitude for cells superperfused with DAMGO followed by naloxone in control mice (dark circles, n = 8 cells, 4 mice) and in mice lacking MORs in presynaptic terminals (clear circles, n = 7 cells, 6 mice). H, Mean summary data of normalized oEPSC amplitude in for control mice in baseline condition, after DAMGO superperfusion, followed by naloxone (DAMGO: 0.39 ± 0.05 fraction of baseline, p < 0.0001; naloxone: 0.80 ± 0.06 of baseline, n = 8 cells, 4 mice, F(2,14) = 29.9, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). I, Mean summary data of normalized oEPSC amplitude for presynaptic MOR KO mice in baseline condition, after DAMGO perfusion, followed by naloxone (DAMGO: 0.99 ± 0.02 fraction of baseline, p = 0.6023; naloxone: 0.92 ± 0.05 fraction of baseline, p = 0.14, n = 7 cells, 6 mice, F(2,12) = 2.08, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). Line and error bars represent mean ± SEM. *Statistical significance.
Like what was observed in WT mice, DPCPX (200 nm) increased the amplitude of the oEPSCs in the presynaptic MOR KO mice (Fig. 7A,C,D; DPCPX: 1.3 ± 0.07 fraction of baseline, p = 0.0012, n = 8 cells from 2 male and 2 female mice, t(7) = 5.225, ratio paired t test). In separate cells, morphine (1 μm) was superperfused, followed by DPCPX. As expected, morphine did not reduce the amplitude of oEPSCs; however, in the presence of morphine, DPCPX did not increase the amplitude of oEPSCs (Fig. 7B,C,E; morphine 1.0 ± 0.07 fraction of baseline, p = 0.9935; DPCPX: 1.0 ± 0.07 fraction of baseline, p = 0.9119, and 1.0 ± 0.09 fraction of morphine, p = 0.9513, n = 6 cells from 2 male and 2 female mice, F(2,10) = 0.09,141, repeated-measures one-way ANOVA, Tukey's multiple comparisons test), suggesting that morphine still inhibited the tonic activation of A1Rs, even in mice lacking presynaptic MORs. Therefore, although opioids inhibited glutamate release from the thalamic terminals through a presynaptic mechanism and A1Rs are expressed presynaptically, presynaptic MORs did not regulate tonic activation of the A1Rs in this circuit.
Presynaptic MORs do not mediate morphine inhibition of tonic A1R activation. A, Representative traces of oEPSCs evoked by 470 nm light (black label) and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label) in mice lacking MORs in presynaptic terminals. B, Representative traces of oEPSCs evoked by 470 nm light (black label), lack of inhibition of oEPSC amplitude by morphine (1 μm; pink label), and lack of facilitation by DPCPX (200 nm; blue label). C, Plot of the time course of normalized oEPSC amplitude for cells superperfused with DPCPX (dark circles; n = 8 cells, 4 mice), and for cells superperfused with morphine and then DPCPX (clear circle; n = 6 cells, 4 mice). D, Mean summary data of normalized oEPSC amplitude in control and after DPCPX (1.3 ± 0.07 fraction of baseline, p = 0.0012, n = 8 cells, 4 mice, t(7) = 5.225, ratio paired t test). E, Mean summary data of normalized oEPSC amplitude in control, after morphine superperfusion, followed by DPCPX. Morphine did not inhibit oEPSC amplitude (morphine: 1.0 ± 0.07 fraction of baseline, p = 0.9935), and there was no facilitation by DPCPX in the presence of morphine (1.0 ± 0.07 fraction of baseline, p = 0.9119, and 1.0 ± 0.09 fraction of morphine, p = 0.9513, n = 6 cells, 4 mice, F(2,10) = 0.09141, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). Line and error bars represent mean ± SEM. *Statistical significance. F, Representative oEPSCs evoked under baseline conditions (black), in the presence of DPCPX (200 nm, blue), and in the presence of the A2AR antagonist SCH58261 (100 nm) + DPCPX (gray). G, Experiment time course of averaged oESPCs across all cells under baseline conditions (black), in the presence of DPCPX (200 nm, blue bar), and in the presence of the A2AR antagonist SCH58261 (100 nm) + DPCPX (gray bar). H, Summary data plotting average peak amplitude for each cell under baseline, DPCPX, and DPCPX + SCH as in E and F (p = 0.008, F(1.77,12.37) = 7.702 repeated-measures one way ANOVA; p = 0.027 DPCPX vs baseline; p = 0.133 DPCPX + SCH vs baseline Tukey's multiple comparison, n = 8 cells, 5 mice).
If morphine eliminated tonic A1R signaling without causing an acute MOR-mediated presynaptic inhibition of glutamate release, it would be predicted that morphine exposure would result in increased oEPSC amplitudes in the presynaptic MOR KO mice. This increase in the amplitude was not observed. One possible explanation for this lack of facilitation by morphine is that adenosine is acting at multiple sites and a decrease in extracellular adenosine is not functionally similar to perfusing DPCPX and selectively blocking A1Rs. It has been reported that glutamate release from striatal synaptosomes is inhibited by A1R activation and facilitated by A2AR activation (Ciruela et al., 2006), suggesting that adenosine can bidirectionally modulate glutamate release. We investigated this possibility by superfusing slices with the A1R antagonist DPCPX and observed a significant increase in oEPSC amplitude as expected. We then superfused the A2AR antagonist SCH58261 (100 nm) in the continued presence of DPCPX. The addition of SCH58261 reduced the amplitude of the oEPSC such that the oEPSC was not significantly different from baseline (Fig. 7F–H; DPCPX: 1.43 ± 0.13 of baseline, p= 0.027, SCH58621: 1.16 ± 0.09 of baseline, p = 0.133, repeated-measures one-way ANOVA, Tukey's post hoc F(1.768,12.37) = 7.702, n = 8 cells from 2 male and 3 female mice). These results are consistent with extracellular adenosine having multiple effects on glutamate release from thalamic terminals in dorsomedial striatum, although they do not rule out other explanations for a lack of oEPSC facilitation by morphine in Floxed MOR;vGlut2-cre presynaptic MOR KO mice.
MOR-sensitive A1R signaling is regulated by D1 receptor-expressing MSNs, not D2 receptor-expressing MSNs
MORs are expressed in both D1 and D2 receptor-expressing MSNs (Cui et al., 2014; Oude Ophuis et al., 2014), and activation of D1 and A2A receptors, presumably in D1 and D2 receptor expressing MSNs, increased tonic adenosine inhibition of A1Rs (Fig. 3B), suggesting that MSNs are a potential source of extracellular adenosine. Therefore, MORs were selectively knocked out in these cells. FloxedMOR mice and A2A:cre mice were crossed to generate mice lacking MORs from D2R-expressing MSNs (Oprm1fl/fl, A2A-cre +/−) (Gong et al., 2007). oEPSCs were evoked as previously described, and DPCPX increased the amplitude of the oEPSCs (Fig. 8A,C,D; DPCPX: 1.3 ± 0.05 fraction of baseline, p = 0.0049, n = 6 cells from 3 male and 1 female mice, t(5) = 4.787, ratio paired t test). In separate cells, morphine (1 μm) was superperfused, followed by DPCPX. Morphine reduced the amplitude of oEPSCs; and in the presence of morphine, DPCPX did not increase the amplitude of oEPSCs (Fig. 8B,C,E; morphine 0.76 ± 0.03 fraction of baseline, p = 0.0001; DPCPX: 0.75 ± 0.02 fraction of baseline, and 0.99 ± 0.04 fraction of morphine, p = 0.9969, n = 6 cells from 2 male and 2 female mice, F(2,10) = 30.38, repeated-measures one-way ANOVA, Tukey's multiple comparisons test), suggesting that morphine inhibited the tonic activation of A1Rs in mice lacking MORs in D2R-expressing MSNs.
MORs in D1R-expressing MSNs, but not D2R-expressing MSNs, regulate tonic A1R activation. A, Representative traces of oEPSCs evoked by 470 nm light (black label), facilitation of oEPSC amplitude by DPCPX (100 nm; blue label, mice lacking MORs in D2R-expressing MSNs). B, Representative traces of oEPSCs evoked by 470 nm light (black label), inhibition of oEPSC amplitude by morphine (1 μm; pink label), and lack of facilitation by DPCPX (200 nm; blue label) in mice lacking MORs in D2R-expressing MSNs. C, Plot of the time course of normalized oEPSC amplitude for cells superperfused with DPCPX (dark circles; n = 6 cells, 4 mice), and for cells superperfused with morphine and then DPCPX (clear circle; n = 6 cells, 4 mice). D, Mean summary data of normalized oEPSC amplitude in control and after DPCPX (DPCPX: 1.3 ± 0.05 fraction of baseline, p = 0.0049, n = 6 cells, 4 mice, t(5) = 4.787, ratio paired t test). E, Mean summary data of normalized oEPSC amplitude in control, after morphine superperfusion, and after DPCPX superperfusion. Morphine significantly inhibited oEPSC amplitude (morphine: 0.76 ± 0.03 fraction of baseline, p = 0.0001), but there was no facilitation by DPCPX in the presence of morphine (DPCPX: 0.75 ± 0.02 fraction of baseline, and 0.99 ± 0.04 fraction of morphine, p = 0.9969, n = 6 cells, 4 mice, F(2,10) = 30.38, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). F, Representative traces of oEPSCs evoked by 470 nm light (black label) and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label), in slices from mice lacking MORs from D1R-expressing MSNs. G, Representative traces of oEPSCs evoked by 470 nm light (black label), inhibition by morphine (1 μm; pink label), and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label), in slices from mice lacking MORs from D1R-expressing MSNs. H, Plot of the time course of normalized oEPSC amplitude for cells superperfused with DPCPX (dark circles; n = 5 cells, 3 mice), and for cells superperfused with morphine and then DPCPX (clear circle; n = 5 cells, 3 mice). I, Mean summary data of normalized oEPSC amplitude in control and after DPCPX (DPCPX: 1.4 ± 0.09 fraction of baseline, p = 0.006, n = 5 cells, 3 mice, t(4) = 5.253, ratio paired t test). J, Mean summary data of normalized oEPSC amplitude in control, after morphine superperfusion, and after DPCPX superperfusion. Morphine inhibited oEPSC amplitude (morphine 0.72 ± 0.04 fraction of baseline, p = 0.013), and there was facilitation by DPCPX in the presence of morphine (1.14 ± 0.06 fraction of baseline, p = 0.06 and 1.51 ± 0.08 fraction of morphine, p = 0.0001, 5 cells, 3 mice, F(3,12) = 25.36, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). Naloxone caused an over-reversal of oEPSC amplitude (1.30 ± 0.03 fraction of baseline, p = 0.004). Line and error bars represent mean ± SEM. *Statistical significance. K, Representative traces of oEPSCs evoked by 470 nm light (black label) and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label), in slices from mice with a partial MOR KO from D1R-expressing MSNs. L, Representative traces of oEPSCs evoked by 470 nm light (black label), inhibition by morphine (1 μm; pink label), and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label), in slices from mice with a partial MOR knockdown from D1R-expressing MSNs. M, Plot of the time course of normalized oEPSC amplitude for cells superperfused with DPCPX (dark circles; n = 3 cells, 2 mice), and for cells superperfused with morphine and then DPCPX, followed by naloxone (clear circle; n = 3 cells, 2 mice). N, Mean summary data of normalized oEPSC amplitude in control and after DPCPX (1.38 ± 0.22 fraction of baseline, p < 0.001, n = 6 cells, 5 mice, t(5) = 4.466, ratio paired t test). O, Mean summary data of normalized oEPSC amplitude in control, after morphine superperfusion, after DPCPX superperfusion, and after naloxone superperfusion. Morphine inhibited oEPSC amplitude (morphine: 0.70 ± 0.05 fraction of baseline, p = 0.0003), and there was facilitation by DPCPX in the presence of morphine (DPCPX: 0.91 ± 0.06 fraction of baseline, p = 0.39, and 1.34 ± 0.06 fraction of morphine, p = 0.0086, 6 cells, 4 mice, F(3,15) = 27.12, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). Naloxone caused an over-reversal of oEPSC amplitude (1.2 ± 0.05 fraction of baseline, p = 0.0224). Line and error bars represent mean ± SEM. *Statistical significance.
Next, FloxedMOR mice and D1:cre mice were crossed to generate mice lacking MORs from D1R-expressing MSNs (Oprm1fl/fl, D1-cre+/−) (Gong et al., 2007). DPCPX (200 nm) increased the amplitude of the oEPSCs (Fig. 8F,H,I; DPCPX; 1.4 ± 0.09 fraction of baseline, p = 0.006, n = 5 cells from 2 male 1 female mice, t(4) = 5.253, ratio paired t test). In separate cells, morphine (1 μm) reduced the amplitude of oEPSCs and, in the presence of morphine, unlike in WT mice, DPCPX increased the amplitude of oEPSCs (Fig. 8G,H,J; morphine 0.72 ± 0.04 fraction of baseline, p = 0.013; DPCPX: 1.14 ± 0.06 fraction of baseline, p = 0.06 and 1.51 ± 0.08 fraction of morphine, p = 0.0001, 5 cells from 2 male and 1 female mice, F(3,12) = 25.36, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). Next, MOR antagonist naloxone caused an over-reversal of oEPSC compared with baseline (Fig. 8G,H,J; naloxone 1.30 ± 0.03 fraction of baseline, p = 0.004), suggesting that, in mice lacking MOR in D1R-positive MSNs, morphine could no longer inhibit tonic A1R activation. The over-reversal by naloxone suggests that the oEPSC was inhibited by activation of the MORs by morphine, and antagonizing the MORs after antagonizing the A1Rs reveals the presence of adenosine tone in these mice. Surprisingly, mice lacking MORs in only one copy of the D1R gene (FloxedMOR +/−, D1-cre +/−) also showed similar results. In these mice, DPCPX (200 nm) increased the amplitude of the oEPSCs as well (Fig. 8K,M,N; DPCPX; 1.38 ± 0.22 fraction of baseline, p < 0.001, n = 6 cells from 1 male and 3 female mice, t(5) = 4.466, ratio paired t test). In separate cells, morphine (1 μm) reduced the amplitude of oEPSCs; and in the presence of morphine, unlike in WT mice, DPCPX increased the amplitude of oEPSCs (Fig. 8L,M,O; morphine 0.70 ± 0.05 fraction of baseline, p = 0.0003; DPCPX: 0.91 ± 0.06 fraction of baseline, p = 0.39, and 1.34 ± 0.06 fraction of morphine, p = 0.0086, 6 cells from 1 male and 3 female mice, F(3,15) = 27.12, repeated-measures one-way ANOVA, Tukey's multiple comparisons test). Next, MOR antagonist naloxone caused an over-reversal of oEPSC compared with baseline (Fig. 8K,M,N; naloxone 1.2 ± 0.05 fraction of baseline, p = 0.0224), suggesting that a partial deletion of MORs from D1R-expressing MSNs was sufficient to eliminate the inhibition of tonic A1R signaling induced by morphine. Combined, these results are consistent with morphine-mediated inhibition of extracellular adenosine accumulation in the thalamo-striatal circuit, implying that the inhibition of adenosine signaling happens at the D1R-positive MSNs.
A summary of the effects of selective deletion of MOR from various neuronal populations demonstrates that, while the effect of DPCPX was similar in the absence of morphine across all genotypes, only selective KO of MOR in D1R-positive cells resulted in a significant effect of DPCPX in the presence of morphine compared with WT mice in the presence of morphine (Fig. 9; Oprm1fl/fl, D1-cre+/− p = 0.0004). Additionally, there was no statistical difference between Oprm1fl/fl, D1-cre+/− mice and global MOR KO mice in morphine condition (p = 0.3653, unpaired t test, t(10) = 0.9485). Combined, these results demonstrate that morphine's regulation of adenosine signaling in this thalamo-striatal circuit critically requires MORs in D1R-positive MSNs and that, under these experimental conditions, these D1R-positive neurons are likely the major source of extracellular adenosine in dorsomedial striatum.
Summary data comparing DPCPX response in the absence and presence of morphine in mice across all genotypes. Ratio of DPCPX facilitation compared with baseline and in morphine condition in WT mice, mice lacking MORs in presynaptic thalamic terminals, postsynaptic D1R-expressing MSNs, and D2R-expressing genotypes. There was no difference in DPCPX responses across genotypes in baseline condition, but mice lacking MORs in D1R-positive MSNs had a higher facilitation by DPCPX in morphine condition compared with WT mice, mice lacking MORs in presynaptic thalamic terminals, and postsynaptic D2R-expressing MSNs (p = 0.03 compared with WT, p = 0.04 compared with presynaptic MOR KOs, and p = 0.03 compared with MOR KO in D2R-positive MSNs, one-way ANOVA, Tukey's multiple comparisons test). Line and error bars represent mean ± SEM. *Statistical significance.
Morphine regulates A1R signaling in opioid insensitive cortico-striatal circuit
Because the regulation of A1R signaling in thalamo-striatal synapses was from postsynaptic D1R-positive MSNs, it raises the possibility that MOR agonists may affect adenosine signaling in a paracrine manner, affecting A1R activation broadly within the dorsomedial striatum. Cortical inputs also release glutamate onto MSNs in the dorsomedial striatum and inputs from the ACC appear to lack functional MOR (Birdsong et al., 2019). Therefore, MOR regulation of adenosine signaling was investigated in MOR-lacking inputs from the ACC. AAV2 containing channelrhodopsin was microinjected into the ACC, and whole-cell voltage-clamp recordings were made from the MSNs in the DMS (Fig. 10A). oEPSCs were evoked as previously described. DPCPX increased the amplitude of the oEPSCs (Fig. 10C,E,F; DPCPX: 1.42 ± 0.13 fraction of baseline, p = 0.0084, n = 7 cells from 3 male 1 female mice). In separate cells, morphine (1 μm) was superperfused, followed by DPCPX. Morphine did not reduce the amplitude of oEPSCs; and in the presence of morphine, DPCPX did not increase the amplitude of oEPSCs (Fig. 10D,E,G; morphine 1.08 ± 0.14 fraction of baseline, p = 0.9595; DPCPX: 0.99 ± 0.03 of baseline, p = 0.9593, and 0.99 ± 0.03 fraction of morphine, p = 0.9835, n = 6 cells from 3 male and 2 female mice), suggesting that morphine inhibited the tonic activation of A1Rs even in terminals lacking the MORs. This finding suggests that, although opioids do not acutely inhibit glutamate release from cortical terminals, they can indirectly modulate the activity of A1Rs expressed on these terminals and suggest that MOR effects on striatal adenosine signaling are not restricted to thalamo-striatal terminals but broadly disrupt striatal adenosine regulation.
Morphine regulates adenosine release in opioid-insensitive cortico-striatal circuit. A, An acute mouse brain slice example of overlaid brightfield and epifluorescent images showing the viral injection site (ACC; left) and the axonal projections recording site (Striatum; right). B, Schematic showing the locations of MORs and A1Rs in the cortico-striatal synapse. C, Representative traces of oEPSCs evoked by 470 nm light (black label) and facilitation of oEPSC amplitude by DPCPX (200 nm; blue label). D, Representative traces of oEPSCs evoked by 470 nm light (black label), lack of inhibition of oEPSC amplitude by morphine (1 μm; pink label), and lack of facilitation by DPCPX (200 nm; blue label). E, Plot of the time course of normalized oEPSC amplitude for cells superperfused with DPCPX (dark circles; n = 7 cells, 4 mice), and for cells superperfused with morphine and then DPCPX (clear circle; n = 6 cells, 5 mice). F, Mean summary data of normalized oEPSC amplitude in control and after DPCPX (DPCPX: 1.42 ± 0.13 fraction of baseline, p = 0.0084, ratio paired t test). G, Mean summary data of normalized oEPSC amplitude in control, after morphine superperfusion, and after DPCPX superperfusion. Morphine did not inhibit oEPSC amplitude, and there was no facilitation by DPCPX in the presence of morphine (morphine: 1.08 ± 0.14 fraction of baseline, p = 0.9595; DPCPX: 0.99 ± 0.03 of baseline, p = 0.9593, and 0.99 ± 0.03 fraction of morphine, p = 0.9835, repeated-measures ANOVA, Tukey's multiple comparisons test). Line and error bars represent mean ± SEM. *Statistical significance.
Discussion
This study explored how the opioid and adenosine signaling systems interact to inhibit glutamate release in thalamo-striatal and thalamo-cortical circuits. Consistent with previous findings, activation of MORs by morphine inhibited glutamate release from medial thalamic terminals in dorsomedial striatum (Birdsong et al., 2019). Activation of A1Rs also inhibited glutamate release from these terminals, and antagonism of this receptor revealed endogenous adenosine tone that activated the A1Rs. Morphine inhibited this tonic A1R activation through a MOR- and cAMP-mediated mechanism. In mice lacking MORs in D1R-positive MSNs (but not presynaptic or D2R-positive MSN MOR KOs), morphine no longer inhibited the tonic activation of A1Rs. This is consistent with a model in which MOR activation on D1R-positive MSNs inhibits AC, thereby decreasing intracellular cAMP and, ultimately, extracellular adenosine to affect tonic A1R signaling (Fig. 11). Consistent with a postsynaptic effect of morphine on adenosine signaling and a paracrine-like effect of striatal adenosine accumulation, morphine also inhibited A1R signaling in a morphine-insensitive cortico-striatal circuit.
Summary diagram of potential mechanism of morphine-mediated inhibition of tonic adenosine A1R signaling. Tonic A1R activation is able to inhibit glutamate release from MOR-expressing medial thalamic terminals and MOR-lacking ACC terminals in dorsomedial striatum. Morphine activation of MORs on D1-expressing MSNs decreases extracellular adenosine by inhibition of AC, which decreases extracellular striatal adenosine concentration, thereby decreasing A1R activation (and presumably A2AR activation) in a paracrine manner. Created using www.biorender.com.
Interaction between opioids, cAMP, and adenosine
Basal endogenous adenosine concentration increases after withdrawal from chronic morphine treatment (Bonci and Williams 1996; Chieng and Williams, 1998; Matsui et al., 2014). The results of this study suggest that acute morphine exposure had the effect of decreasing adenosine concentration. This is consistent with withdrawal causing a homeostatic increase in adenosine as a result of chronic adenosine deficit. MOR activation is known to decrease AC activity and, consequently, cAMP accumulation. The role of cAMP as a precursor for extracellular adenosine has been previously established in the hippocampus (Brundege et al., 1997; Dunwiddie et al., 1997; Brundege and Dunwiddie, 1998). Therefore, it is not surprising that fluctuations in cAMP concentration might modulate tonic adenosine levels in DMS as well. Previous studies have also shown that cAMP metabolism and transport alter adenosine concentration, and that the regulation of extracellular adenosine depends, in part, on the balance between mechanisms that increase and decrease cAMP concentration (Krupinski et al., 1989; Rosenberg and Dichter, 1989). The transport of adenosine from neurons to the extracellular space has received considerable attention. cAMP can be transported into the extracellular space by probenecid-sensitive transporter (Henderson and Strauss, 1991; Rosenberg and Li, 1995), and can be converted to adenosine through extracellular phosphodiesterase (Rosenberg and Dichter, 1989; Rosenberg et al., 1994). The concentration of cAMP released through this mechanism is sufficient to increase extracellular adenosine concentration (Gereau and Conn 1994; Brundege et al., 1997). Therefore, we hypothesize that the transport of cAMP from D1R-positive MSNs regulates extracellular striatal adenosine accumulation; however, the exact mechanism of cAMP or adenosine release into the extracellular space remains unknown.
In the hippocampus, endogenous adenosine inhibits glutamate release (Dunwiddie and Diao, 1994). Similarly, basal endogenous adenosine affects some striatal synapses. Potentiation of glutamate release by DPCPX in NAc core and of GABA release in rat NAc core and shell have been reported (Brundege and Williams, 2002). This study demonstrated that DPCPX also potentiated glutamate release from both thalamic and cortical terminals in the DMS in mice, and a similar cAMP-dependent mechanism mediates adenosine accumulation.
Opioid selectivity in mediating adenosine release in thalamo-striatal circuit
Consistent with previous findings, activation of MOR, but not DOR, led to inhibition of glutamate release in the thalamo-striatal circuit (Birdsong et al., 2019). This inhibition of glutamate release was lost in FloxedMor-Vglut2-cre mice, corroborating a previous report (Reeves et al., 2021). However, morphine inhibition of A1R signaling remained, suggesting that MOR activation in other striatal cells mediated the effect of morphine on A1R signaling. Despite the fact that DOR and KOR are both Gi/o-coupled GPCRs, neither the activation of DOR nor KOR inhibited tonic adenosine release, suggesting that the MOR uniquely interacts with the adenosine system, perhaps because of enriched expression on D1R-expressing MSNs. However, because cAMP appears to regulate extracellular adenosine, we would predict that other inhibitory GPCRs expressed in D1R-positive MSNs could mimic the effect of morphine, including GABAB receptors or inhibitory DREADDs. Interestingly, A2AR activation increased extracellular adenosine, but DOR agonists did not decrease adenosine tone, despite broad coexpression. This could be because of several factors, including higher cyclase activity in D1- versus A2A-expressing neurons or the subcellular localization of DOR may limit inhibition of A2AR-activated cyclase A2A MSNs. Last, regional heterogeneity may exist in opioid regulation of adenosine tone across medial and lateral dorsal striatum or NAc.
Source of opioid-sensitive adenosine tone
Selectively knocking out the MORs from the presynaptic terminals, D1R-positive MSNs, and D2R-positive MSNs revealed that morphine-induced inhibition of adenosine release was because of morphine's action on MORs in the D1R-positive MSNs, but not D2R-positive MSNs. This finding is consistent with previous work showing that D1R- and D2R-positive MSNs differentially modulate striatal activity (Lobo and Nestler, 2011), and that the somatodendritic region of neurons can release adenosine and retrogradely bind presynaptic A1Rs (Lovatt et al., 2012). Furthermore, MORs in D1R-positive MSNs and D2R-positive MSNs also differentially modulate opioid responses. MOR deletion from D1R-positive MSNs inhibits opioid-induced hyperlocomotion, whereas deletion from D2R-positive MSNs increases opioid-induced hyperlocomotion (Severino et al., 2020). Additionally, MOR expression in D1R-positive MSNs also appears necessary for opioid self-administration and reward (Cui et al., 2014). The broad distribution of adenosine receptors within the striatum makes it difficult to predict the specific behavioral implication of morphine regulation of adenosine effects specifically in thalamo-striatal or cortico-striatal glutamate terminals; however, these findings suggest that the adenosine system as a whole has the potential to be a powerful mediator of acute and chronic morphine effects in the dorsomedial striatum. Thus, a novel role for MORs in regulating adenosine release in the striatum in a cell type-specific manner can have profound implication for opioid dependence and addiction. There is also evidence for astrocytes mediating adenosine release in NAc, although the mechanism behind adenosine release is through increases in Ca2+ activity, and not through an increase in cAMP concentration (Corkrum et al., 2020). The similarities and differences in ways adenosine is regulated to maintain homeostasis in striatal neuron signaling could be a potential new area of study.
Morphine's indirect regulation of opioid-insensitive cortico-striatal circuit
The ACC innervates striatal MSNs, and these glutamatergic afferents are insensitive to acute MOR or DOR opioid actions (Birdsong et al., 2019); however, some cortical inputs to dorsomedial striatum appear modestly sensitive to opioids (Muñoz et al., 2020). Application of morphine in this study failed to significantly inhibit glutamate release from ACC terminals, but in the presence of morphine, A1R signaling was absent. This supports the interpretation that opioid regulation of extracellular adenosine occurs at postsynaptic MSNs and not at presynaptic terminals. This indirect effect on cortico-striatal adenosine signaling emphasizes that opioids modulate neurotransmitter/neurohormone release in the striatum at multiple sites. Recordings were obtained from nonidentified MSNs likely encompassing both D1R- and D2R/A2AR-expressing MSNs. Both tonic adenosine signaling and morphine's inhibition of tonic adenosine signaling were relatively uniform across all cells, suggesting that morphine/adenosine interactions likely do not depend on the identity of the MSN being recorded from. Adenosine from D1R-positive MSNs, therefore, most likely affects adenosine signaling broadly, rather than in a cell-autonomous manner.
While we used A1R effects on thalamo-striatal and cortico-striatal signaling as a readout for morphine regulation of striatal adenosine, morphine may broadly impact striatal adenosine, potentially affecting many striatal neurotransmitter systems and cell types. Adenosine receptor signaling has been suggested to regulate GABA release in striatum and has been linked to changes in dopamine and acetylcholine transmission as well. Gs-coupled adenosine A2ARs are also broadly expressed across D2R-expressing MSNs and would presumably be affected by morphine-induced changes in tonic adenosine concentrations. A2AR activation has been proposed to facilitate glutamate release from striatal synaptosomes, opposing A1R-mediated inhibition of glutamate release (Ciruela et al., 2006). In presynaptic MOR KO recordings and in ACC-striatum recordings, morphine may be expected to increase glutamate release, but this was not observed. It is possible that morphine, by decreasing extracellular adenosine, is preventing tonic activation of both inhibitory A1Rs and facilitating A2ARs that regulate glutamate release in opposing ways, resulting in a loss of adenosine signaling without a net facilitation of cortico-striatal or thalamo-striatal glutamate release.
The present results indicate that morphine inhibits tonic adenosine signaling by activating MORs, and subsequently inhibiting cAMP. This effect of opioid-induced inhibition of adenosine release was specific to MOR and not mediated by DOR or KOR. Selective KO of MORs from presynaptic terminals showed that, although opioids presynaptically inhibit glutamate release, presynaptic MORs do not modulate extracellular adenosine accumulation and adenosine signaling in the thalamo-striatal circuit. Rather, tonic adenosine release was no longer inhibited by morphine when MORs were knocked out from D1R-positive MSNs, but not D2R-positive MSNs or from glutamate terminals. Thus, the endogenous adenosine that tonically activates the A1R comes only from D1R-positive MSNs in the medial thalamus-dorsomedial striatum circuit, and this postsynaptic regulation can affect cortical afferents that are insensitive to acute opioid modulation.
Footnotes
This work was supported by the National Institute on Drug Abuse R01DA042779 to W.T.B., F30 DA051117 to S.A. and T32DA007281 to E.R.J.; the ARCS Foundation S.A. We thank Dr. John T. Williams and Erica Levitt for comments on the manuscript and financial support for this project (National Institute on Drug Abuse R01DA08163 to Dr. John T. Williams); Dr. Brigitte Kieffer for providing Oprm1 fl/fl mice; Dr. Christopher Ford for providing Drd1-cre mice; Dr. Tianyi Mao for providing Adora2a-cre mice; and Katherine Suchland, Dr. James Bunzow, and Dr. Joe Lebowitz for genotyping the transgenic mice.
The authors declare no competing financial interests.
- Correspondence should be addressed to William T. Birdsong at wtbird{at}umich.edu
