Nicotinic and opioid receptor regulation of striatal dopamine D2-receptor mediated transmission

In addition to dopamine neuron firing, cholinergic interneurons (ChIs) regulate dopamine release in the striatum via presynaptic nicotinic receptors (nAChRs) on dopamine axon terminals. Synchronous activity of ChIs is necessary to evoke dopamine release through this pathway. The frequency-dependence of disynaptic nicotinic modulation has led to the hypothesis that nAChRs act as a high-pass filter in the dopaminergic microcircuit. Here, we used optogenetics to selectively stimulate either ChIs or dopamine terminals directly in the striatum. To measure the functional consequence of dopamine release, D2-receptor synaptic activity was assessed via virally overexpressed potassium channels (GIRK2) in medium spiny neurons (MSNs). We found that nicotinic-mediated dopamine release was blunted at higher frequencies because nAChRs exhibit prolonged desensitization after a single pulse of synchronous ChI activity. However, when dopamine neurons alone were stimulated, nAChRs had no effect at any frequency. We further assessed how opioid receptors modulate these two mechanisms of release. Bath application of the κ opioid receptor agonist U69593 decreased D2-receptor activation through both pathways, whereas the μ opioid receptor agonist DAMGO decreased D2-receptor activity only as a result of cholinergic-mediated dopamine release. Thus the release of dopamine can be independently modulated when driven by either dopamine neurons or cholinergic interneurons.

Electrophysiology. Whole-cell recordings were performed using either Axopatch 200 A or Axopatch 200B amplifiers (Molecular Devices). Membrane potentials were not corrected for liquid junction potentials. Patch pipettes (1.5-2 MΩ) were made from borosciliate glass capillary tubes (World Precision Instruments). Patch pipettes for MSNs contained (in mM): 115 K-methylsulfate, 20 NaCl, 1.5 MgCl 2 , 10 HEPES(K), 10 BAPTA-tetrapotassium. Patch pipettes for ChIs contained (in mM): 135 D-gluconate(K), 10 HEPES(K), 0.1 CaCl 2 , 2 MgCl 2 , 0.1 EGTA. All internal patch solution contained: 1 mg/mL ATP, 0.1 mg/mL GTP, and 1.5 mg/ mL phosphocreatine (pH 7.35, 275 mOsm). Membrane potential was held at − 60 mV during voltage clamp recordings. Cells were discarded if their series resistance exceeded 15 MΩ, as series resistance was not compensated. ChIs were identified by the presence of an h-current when stepped to − 90 mV. Data was acquired with Axograph X (Axograph Scientific) at 5 kHz. Voltage clamp recordings were filtered to 2 kHz. Electrical stimulation (0.7 ms) was applied with a monopolar extracellular stimulating electrode filled with ACSF (World Precision Instruments). Optogenetic stimulation was evoked with 2 ms widefield illumination (470 nm) using a custom made LED (470 nm Rebel LED Star Saber, Luxeon Star). All drugs were applied via bath application unless otherwise specified. All recordings were made in the dorsal striatum.

Statistics and analysis.
Data are shown as mean ± SEM. Statistical significance was assessed using either Wilcoxon matched-pairs signed rank test, Mann-Whitney test, or one-way ANOVA where appropriate (InStat 3.0, Graphpad).

Activation of D2 receptors in MSNs.
To examine D2-receptor activation in response to nicotinic receptor activity, we overexpressed G-protein coupled inwardly rectifying K + (GIRK2) channels in striatal MSNs 17 . Injection of an adeno-associated virus (AAV) encoding GIRK2 and a soluble td-Tomato fluorophore resulted in expression of both proteins in MSNs 17,18 . After allowing three weeks for expression, coronal striatal slices were obtained from AAV-injected mice. Viral expression was observed in both direct and indirect pathway MSNs 17,18 , but was restricted from ChIs likely as a result of poor efficiency of transfection of ChIs using a synapsin promoter 18 . It was not examined whether other GABAergic interneurons expressed GIRK following AAV injection. All experiments in the present study were performed in the presence of antagonists for GABA A , GABA B AMPA, NMDA, dopamine D1, and muscarinic receptors. Td-Tomato + MSNs were voltage-clamped at a holding potential Scientific RepoRts | 6:37834 | DOI: 10.1038/srep37834 of − 60 mV. As about half of striatal MSNs belong to the D2-receptor expressing indirect pathway 19 , D2-receptor mediated GIRK2 currents were observed in roughly half of GIRK2 + MSNs 17 .
To separate cholinergic-mediated dopamine release from direct dopamine terminal driven release, we employed an optogenetic approach. We injected a double-floxed virus encoding the light activated cation channel channelrhodopsin-2 (ChR2, AAV.DIO.ChR2.eYFP) into either the substantia nigra (SNc) of DAT-Cre mice or into the striatum of ChAT-IRES-Cre mice. Photostimulation of either dopamine terminals (DAT-Cre:ChR2) or ChIs (ChAT-Cre:ChR2) with a single flash of blue light (470 nm, 2 ms) was sufficient to evoke D2-IPSCs in indirect pathway GIRK2 + MSNs ( Fig. 1C and D). In both cases, IPSCs were eliminated by the D2-receptor antagonist sulpiride (400 nM; n = 4; p < 0.05, U = 4, Mann-Whitney) ( We next examined whether a single action potential in a ChI was sufficient to evoke D2-receptor activation. To test this, we made paired recordings of ChIs (current clamp) and GIRK2 + -expressing MSNs (voltage clamp). ChIs were hyperpolarized to just below threshold during these recordings to prevent tonic firing. To increase the resolvable amplitude of paired IPSCs, dopamine transporters were inhibited with the dopamine transport blocker cocaine (10 μ M). While electrical stimulation evoked robust D2-IPSCs, single action potentials in ChIs failed to elicit D2-IPSCs ( Fig. 1F and G). Single action potentials also did not evoke an IPSC when recorded in the presence of ambenonium (100 nM), which inhibits acetylcholinesterase thereby increasing the concentration of striatal ACh (4.7 ± 2 pA, n = 5). This confirms that multiple ChIs activated together are required to engage nicotinic receptor mediated dopamine release 11 .

Frequency dependence of cholinergic-induced dopamine release.
Past work has demonstrated that nicotinic receptors provide stronger modulation of dopamine release during low frequency electrical stimulation 12 . To examine this through the activation of D2-receptors, we compared the effect of a single electrical stimulus to bursts of stimuli (5 at 40 Hz). Bursts evoked larger amplitude D2-IPSCs than single stimuli (burst 125.8 ± 5% of single stimulation amplitude, n = 14, p < 0.001, W = 91, Wilcoxon) ( Fig. 2A and I) 21 . In addition, DHβ E (1 μM) inhibited electrically evoked D2-IPSCs less when driven by bursts than single stimuli (single: 39.2 ± 4% inhibition, n = 4; burst: 13 ± 4% inhibition, n = 6; p < 0.01, Mann-Whitney) ( Fig. 2B and C). This indicates that ChI-mediated dopamine transmission contributed less during high frequency stimulation and is consistent with the high-pass filter model describing nicotinic modulation of dopamine release that has been seen before with electrical stimulation 12 .
To determine the kinetics of nAChR desensitization, we next examined the time course of recovery of D2-IPSCs. Paired pulse experiments revealed that ChAT-Cre:ChR2 evoked IPSCs took significantly longer to recover than DAT-Cre:ChR2 evoked IPSCs (Fig. 3). As D2-IPSCs evoked by direct dopamine terminal stimulation recovered faster than D2-IPSCs evoked by ChI-stimulation, (Fig. 3B), the long recovery of ChAT-Cre:ChR2 evoked D2-IPSCs was likely not due to intrinsic properties of dopamine release. We found that ChAT-Cre:ChR2 IPSCs evoked at a 1 second interpulse interval were almost completely abolished, while DAT-Cre:ChR2 IPSCs only decreased by half their amplitude (1 sec PPR: DAT 0.49 ± 0.02, ChAT 0.09 ± 0.01, n = 6 for both groups, p < 0.01, U = 0, Mann-Whitney). This difference in PPR suggests that nicotinic receptors on SNc terminals likely exhibit a slow recovery from desensitization in response to synchronous ACh release 26 . Together, these data indicate that nicotinic receptors on dopamine terminals positively modulate D2-receptor transmission during synchronous ChI firing yet desensitize at high frequencies. Opioid modulation of striatal D2-IPSCs in MSNs. Mu, kappa and delta opioid receptors are expressed in the striatum where they regulate transmission through multiple circuits [27][28][29] . Although mu and delta opioid agonists induce reward behavior 30 , kappa opioid agonists are aversive in animal models 31,32 . Mu and delta receptors are both expressed postsynaptically on MSNs. Mu opioid receptors (MORs) are present on ChIs 29,33-35 while kappa opioid receptors (KORs) are on dopamine terminals 36,37 . Both mu and kappa receptors have been shown to modulate striatal dopamine release 38 . To examine how these two opioid receptors regulate D2-IPSCs in MSNs, we applied agonists of either MORs or KORs while stimulating either dopamine terminals or ChIs. As MORs are located on CHIs, we found that bath application of the MOR antagonist DAMGO (1 μ M) did not alter DAT-Cre:ChR2 evoked D2-IPSCs (2.3 ± 5% inhibition, n = 5, p > 0.05, W = 7, Wilcoxon) ( Fig. 4A) but strongly decreased the amplitude of D2-IPSCs evoked by ChI stimulation (42.4 ± 5% inhibition, n = 6, W = 21 p < 0.05, Wilcoxon) (Fig. 4B). During these experiments, we also observed an outward current of 157 pA in one of six MSNs resulting from MOR activation on the MSN itself. The inhibition of ChAT-Cre:ChR2 evoked D2-IPSCs by DAMGO was reversed by the non-selective opioid receptor antagonist naloxone (1 μ M, n = 6; p < 0.05 vs control, W = 17, Wilcoxon). This indicates that MORs on ChIs selectively modulate dopamine transmission release through the nicotinic pathway 38 . In contrast, the KOR agonist U69593 inhibited D2-receptor activation from both pathways (42.2 ± 4% inhibition of DAT-Cre:ChR2 evoked IPSCs, n = 6; 36.1 ± 6% inhibition of ChAT-Cre:ChR2 evoked IPSCs, n = 8; W = 21, p < 0.05 for both groups) ( Fig. 5A and B). Thus, KORs equally modulate dopamine release when either dopamine terminals or ChIs are stimulated.

Discussion
As dopamine release in the striatum is critical for motivated behaviors and action selection, modulation of this release is an important regulator of striatal circuitry. Independent of dopamine neuron impulse activity, nicotinic receptors on striatal dopamine terminals also facilitate dopamine release 5,11 . Using overexpressed GIRK2 channels in MSNs, we found that nicotinic receptors can modulate synaptic release of dopamine and subsequent D2-receptor activation. Direct comparison of D2-receptor activation in response to dopamine release driven through these two pathways revealed differential regulation of dopamine transmission.
Dopamine neurons typically fire at tonic low frequencies, and their firing switches to a bursting pattern in response to rewards and their cues [1][2][3] . Several studies have previously examined the role of nicotinic receptors as a high-pass filter of dopamine release at different firing frequencies using electrical stimulation [11][12][13] . This work has shown that nicotinic receptors strongly facilitate the release of dopamine at low stimulation frequencies (single pulse or 5 Hz) but have less of an effect 12 or even depress dopamine release at higher frequencies (between 20 and 100 Hz) 13 . This nicotinic-mediated alteration of dopaminergic release probability was thought to result in contrast enhancement of dopamine release at different firing rates. However, dopamine neurons and ChIs show opposing firing patterns in response to salient stimuli 16 and striatal dopamine release induces a pause in ChI tonic activity 22,23 . Thus, it is unlikely that dopamine neurons and ChIs would be simultaneously and synchronously active in vivo. We found that, similar to the electrochemical measurements of Zhang and Sulzer (2004), nicotinic receptors more strongly modulated dopamine release at low frequency electrical stimulation. However, by segregating these two pathways using optogenetics, we found that nicotinic receptors only act as high pass filters when both pathways are activated simultaneously. Selective stimulation of only dopamine terminals by activating ChR2 expressed in SNc axons of DAT-Cre mice revealed that D2-IPSCs exhibited no change in nicotinic receptor-mediated facilitation at any frequency examined. Thus, under conditions when dopamine neurons and ChIs are not synchronously active, nicotinic receptors failed to modulate the release of dopamine that underlies the activation of D2-receptors.
High-affinity α 4β 2 nicotinic receptors on midbrain dopamine neurons rapidly desensitize in response to low concentrations of ACh or nicotine [39][40][41][42] . As such, it is thought that these receptors on dopamine terminals readily desensitize in response to the concentrations of ACh that are released when multiple ChIs are simultaneously active 6,39,43 . The reduced contribution of this di-synaptic pathway during high frequency cholinergic stimulation enables nicotinic receptors to only respond to the first synchronous stimulus of a burst 5,11 . Thus, trains of optogenetic ChI stimuli failed to evoke further dopamine facilitation through nicotinic receptors when compared to single stimulation.
As DAT-Cre:ChR2 evoked D2-IPSCs facilitated during high frequency optogenetic stimulation and recovered from paired-pulse depression at a faster rate than ChAT-Cre:ChR2 evoked IPSCs, our results reveal a ChI-mediated mechanism of depression rather than dopamine vesicle depletion or D2-receptor desensitization. Similar to the observed recovery of ChAT-Cre:ChR2 evoked IPSCs, human α 4β 2 nAChRs expressed in heterologous systems can take ~20 seconds to recover their activity to normal levels in response to ACh and nicotine 26 . Further, ACh released onto MSNs at muscarinic synapses recovers within 10 seconds 18 , further suggesting that cholinergic vesicle depletion is unlikely to be responsible for the observed paired pulse depression of ChAT-Cre:ChR2 evoked IPSCs.
ChIs are tonically active both in vivo and in the absence of synaptic inputs 14 , largely due to the presence of hyperpolarization-activated cyclic nucleotide gated (HCN) channels and small-conductance potassium (SK) channels 44,45 . This has led to the idea that tonic ChI firing provides the striatum with a tone of ACh that activates postsynaptic receptors. We found that ACh tone in the slice provided little modulation of D2-receptor activation as DAT-Cre:ChR2 evoked IPSCs showed little change in amplitude in response to nicotinic receptor antagonists. Importantly, tonic ChI activity does not release sufficient amounts of ACh to desensitize nicotinic receptors, as optogenetic stimulation of ChIs could still evoke robust dopamine release. The lack of nicotinic modulation of DAT-Cre:ChR2 evoked IPSCs may result from super-physiological release of dopamine in response to light activated ChR2 conductance in the axon terminal. ChR2 is a light activated cation channel, stimulation of which causes an increased influx of Ca 2+ in axon terminals compared with AP triggered release that has been shown to alter probability of release at other synapses 46 . As a result, one possibility may be that the effect of nicotinic receptors on dopamine terminals may have been occluded by ChR2-mediated release.
ACh tone may not affect dopamine release because acetylcholinesterase (AChE), the degradative enzyme of ACh, is enriched in the striatum and highly efficient, effectively terminating ACh signaling before it can activate striatal nicotinic receptors 6 . Thus, in agreement with previous electrochemical studies 11 , D2-IPSCs evoked from single action potentials in ChIs were unresolvable, whereas synchronous activity of ChIs is required to evoke D2-IPSCs. As AChE is integral for phasic ACh transmission at muscarinic synapses 18 , it likely degrades ACh rapidly such that tonic ChI firing is insufficient to drive dopamine release through nicotinic receptors. Thus, AChE may have a major role in titrating nicotinic receptor activity in the striatum, as its activity decreases striatal ACh while its inhibition desensitizes nicotinic receptors.
Within the striatum, mu, kappa and delta opioid receptors are widespread inhibitory modulators of striatal circuitry (Fig. 6). Enkephalin is released from indirect pathway MSNs and activates delta receptors largely on indirect pathway MSNs 4,27 . In addition, striatal enkephalin can activate striatal MORs on MSNs 27 and ChIs 47 . Dynorphin released primarily from direct pathway MSNs activates KORs to drive either reward or aversion in different striatal compartments 48 . Electrochemical studies examining the effect of opioid receptors on striatal dopamine release have found that although KORs regulate dopamine release throughout the striatum, MOR regulation of dopamine release was spatially segregated to specific areas 38 . We similarly saw that KORs decreased dopamine release from both dopamine terminals and through ChI-mediated release. The areas of MOR regulation that others have observed could potentially be delineated by the striatal patch and matrix, as MORs are selectively expressed in the patch 19,49 as is acetylcholinesterase 50 . ChIs are thought to reside either in the matrix or on the interface of striatal patch and matrix compartments 51 . However, they have far-reaching axons that cross the patch/matrix boundaries 52 . It is possible that we reliably observe MOR-mediated inhibition of ChAT-Cre:ChR2 evoked D2-IPSCs because convergent ChIs activated during optogenetic stimulation extend their axons from MOR-expressing patches.
Together, this work suggests that the role of the burst-firing pattern of ChIs common following reward stimuli is to synchronize firing, allowing for transient, phasic dopamine release, rather than to alter gain of dopamine release in the striatum. This parallels in vivo microdialysis studies that have shown that nicotinic receptor antagonism specifically in the striatum causes a small decrease in dopamine concentration over time 9 . Based on our results, this change in striatal dopamine levels may be differentially gated through opioid receptors. These results further provide insight into the regulation of dopamine release through two different mechanisms.