Synapse-specific expression of mu opioid receptor long-term depression in the dorsomedial striatum

The dorsal striatum is a brain region involved in action control, with dorsomedial striatum (DMS) mediating goal-directed actions and dorsolateral striatum (DLS) mediating habitual actions. Presynaptic long-term synaptic depression (LTD) plasticity at glutamatergic inputs to dorsal striatum mediates many dorsal striatum-dependent behaviors and disruption of LTD influences action control. Our previous work identified mu opioid receptors (MORs) as mediators of synapse-specific forms of synaptic depression at a number of different DLS synapses. We demonstrated that anterior insular cortex inputs are the sole inputs that express alcohol-sensitive MOR-mediated LTD (mOP-LTD) in DLS. Here, we explore mOP-LTD in DMS using mouse brain slice electrophysiology. We found that contrary to DLS, DMS mOP-LTD is induced by activation of MORs at inputs from both anterior cingulate and medial prefrontal cortices as well as at basolateral amygdala inputs and striatal cholinergic interneuron synapses on to DMS medium spiny neurons, suggesting that MOR synaptic plasticity in DMS is less synapse-specific than in DLS. Furthermore, only mOP-LTD at cortical inputs was sensitive to alcohol’s deleterious effects. These results suggest that alcohol-induced neuroadaptations are differentially expressed in a synapse-specific manner and could be playing a role in alterations of goal-directed and habitual behaviors.


AIC and OFC inputs do not express mOP-LTD in the DMS. From our previous work, we broadly
showed that cortical inputs express mOP-LTD in the DLS and DMS 5 . We previously probed specific synapses in the DLS to identify which synapses expressed mOP-LTD, but did not explore specific synapses in the DMS. We demonstrated that AIC inputs are the sole input that expresses mOP-LTD in the DLS 5 , however AIC has lower innervation of the DMS, compared with the DLS 19 . As a direct follow-up, we first explored mOP-LTD at AIC inputs to regions of the striatum more medial to our previous recordings 5 . To probe these AIC inputs, we expressed Channelrhodopsin2 (ChR2) in AIC neurons (Fig. 1A), made brain slices containing DMS and activated these AIC inputs using optical stimulation. We activated MORs by bath application of the agonist DAMGO (0.3 μM) for 5 minutes. AIC inputs did not produce any optically-evoked excitatory postsynaptic current (oEPSCs) in the most dorsomedial portion of the area where we typically perform DMS recordings ( Fig. 1B-D). However, when we changed the recording site to slightly more lateral or ventral portions of the dorsal striatum, we were able to evoke oEPSC responses (Fig. 1E). Interestingly, those AIC inputs to those adjacent regions expressed mOP-LTD (84 ± 5%; baseline: −96 ± 48 pA vs post-DAMGO: 84 ± 45 pA; Fig. 1F,G). This is consistent with our findings in the DLS and also indicates that AIC does not send functional projections to the most dorsomedial part of the DMS. Our previous work also demonstrated that mOP-LTD and cannabinoid receptor-mediated LTD can functionally interact and that cannabinoid-LTD is expressed at OFC inputs to DMS 16,25 . Therefore, we predicted that MOR would also mediate LTD at OFC inputs. To our surprise, we found that OFC projections to the DMS did not express mOP-LTD (102 ± 7%; baseline: −190 ± 54 pA vs post-DAMGO: −186 ± 55 pA; Fig. 2B-D). These results indicated that, based on our previous work, two of the most likely sources of mOP-LTD in the DMS did not in fact express this form of plasticity leading us to explore additional sources.

mOP-LTD is expressed at mPFC and ACC inputs to the DMS.
To determine if our previously identified corticostriatal mOP-LTD in DMS occurs at any other well-validated cortical inputs to the DMS, we tested mPFC and ACC inputs (Fig. 3A,B

MORs at BLA and CIN inputs induce mOP-LTD in the DMS.
We were also interested in whether or not other glutamatergic synapses in DMS could express mOP-LTD. Our previous work indicated thalamic inputs do not express mOP-LTD in DMS 5 . BLA is another brain region that sends glutamatergic inputs to the DMS and is a known input that expresses alcohol-mediated plasticity 19,33 . To probe BLA inputs to DMS synapses, we expressed ChR2 in the BLA (Fig. 4A) and we found that DAMGO treatment was able to produce mOP-LTD at these inputs (77 ± 2%; baseline: −153 ± 24 pA vs post-DAMGO: −119 ± 19 pA; Fig. 4B-D). To explore if mOP-LTD is mediated by presynaptic MORs on BLA afferents, we replaced extracellular Ca 2+ with Sr 2+ , and with this change we were able to induce asynchronous synaptic release of glutamate after BLA inputs were optically evoked in the DMS. As expected, measuring the evoked peak, we detected the presence of mOP-LTD from BLA after MOR activation, even in the presence of Sr 2+ (59 ± 4%; baseline: −141 ± 38 pA vs post-DAMGO: −84 ± 24 pA; Fig. 4E-G). We next assessed the properties of the asynchronous events after the evoked peak, measures that can be used to ascertain whether drugs effects are presynaptic or postsynaptic 34 . We found a decrease in the frequency of the events after DAMGO application (baseline: 14 ± 1 Hz vs post-DAMGO: 12 ± 1 Hz; Fig. 4H) without changes in amplitude (baseline: 24 ± 2 pA vs post-DAMGO: 22 ± 1 pA; Fig. 4I), rise time (baseline: 0.9 ± 0.04 ms vs post-DAMGO: 0.9 ± 0.08 ms; Fig. 4J) or decay time (baseline: 3 ± 0.4 ms vs post-DAMGO: 3 ± 0.5 ms; Fig. 4K).
To further characterize the role of presynaptic MORs from BLA, we used a conditional MOR knockout mouse (MOR-flox), co-injected with AAV-Cre and AAV-ChR2 vectors to knock out the expression of MORs specifically in the BLA and to allow for optogenetic stimulation of the BLA inputs to DMS (Fig. 4L). mOP-LTD appeared to be blunted at these BLA inputs (72 ± 11%, Fig. 4M,N) given that DAMGO failed to produce a significant decrease in oEPSC amplitude in these mice (baseline: −53 ± 14 pA vs post-DAMGO: −42 ± 14 pA; Fig. 4O). Altogether, we suggest that presynaptic MORs in the DMS are responsible for mOP-LTD expressed at BLA inputs.
Given that, we previously found that mOP-LTD occurred at CIN inputs to DLS MSNs 5 , we also explored this possibility in the DMS as well. Using ChATCre mice, we infused a cre-recombinase-dependent AAV-ChR2 vector to express ChR2 only in CINs in the DMS. We found that the AAV.DIO.ChR2 vector injected in the DMS specifically infected CINs (Fig. 5A). Functionally, similar to DLS, we found that activation of CINs produces small oEPSCs in MSNs, but these synapses can still express mOP-LTD in the DMS (72 ± 2%; baseline: −107 ± 22 pA vs post-DAMGO: −77 ± 15 pA; Fig. 5B-E). Our previous work indicates that it is likely MORs expressed in CINs themselves that mediate this mOP-LTD 5 . Therefore, mOP-LTD is not only expressed at cortical inputs but also at BLA and CIN inputs on to MSNs in DMS.

MORs on mPFC, ACC and BLA inputs are indispensable for mOP-LTD.
Our data to this point indicated that MOR activation was sufficient to induce mOP-LTD at corticostriatal and BLA-DMS inputs. To The effects of in vivo exposure to ethanol on mOP-LTD are specific to corticostriatal synapses.
We previously reported that mOP-LTD in MSNs within the DMS was not affected by in vivo ethanol exposure 5 , but we just explored those effects using broad electrical stimulation within DMS that would not be selective for any specific input 5 . Therefore, we tested the impact of in vivo ethanol exposure (2.0 g/kg; intraperitoneal, i.p.) on the expression of mOP-LTD at mPFC, ACC, and BLA inputs to the DMS. Mice injected with saline (i.p.) 24 h before harvesting tissue showed normal mOP-LTD at all these DMS synapses (mPFC: 76 ± 8%, Fig . We and others have demonstrated that in vivo ethanol exposure does not alter basal synaptic transmission 5,35 , but to rule out postsynaptic effects of ethanol, we also measured MSN intrinsic excitability. Using current clamp recordings, we found no differences between saline and ethanol injected mice in the frequency of the action potentials ( Supplementary Fig. S1A,B), and in the others membrane parameters evaluated (Supplementary Fig. S1C-G). In sum, the data indicate that a single in vivo ethanol exposure is able to disrupt the induction of mOP-LTD at corticostriatal synapses, but not at BLA inputs to DMS.

Discussion
Our previous work 5 demonstrated the synapse-specificity of mOP-LTD at AIC synapses on to DLS MSNs and its singular disruption by ethanol. Using the strictest anatomical definition of what can be classified as "DMS, " we did not find evidence of functional AIC inputs to this striatal subregion. This aligns with anatomical analyses performed by others 19,36 . Interestingly we found that mOP-LTD does occur at AIC inputs to regions of dorsal striatum that are adjacent to the DMS, closer to ventral striatum (nucleus accumbens) and DLS. In contrast to DLS with its lone mOP-LTD-expressing input, our study demonstrated that mOP-LTD occured at multiple inputs to DMS. MOR activation on specific corticostriatal (ACC and mPFC) and BLA glutamatergic inputs to the DMS all express mOP-LTD. Consistent with our previous work though, only the corticostriatal synapses were sensitive www.nature.com/scientificreports www.nature.com/scientificreports/ to in vivo ethanol exposure. While we cannot completely rule out mOP-LTD at other minor inputs to DMS, we conclude that the major source of mOP-LTD in the DMS is at inputs from mPFC, ACC and BLA, since we demonstrated that the ablating of MOR expression from those areas is sufficient to disrupt mOP-LTD expression in the DMS. The current report together with our previous data demonstrate that MOR-mediated corticostriatal LTD is completely presynaptic using a variety of measures, including paired pulse ratio assessments and measures of spontaneous EPSCs/miniature EPSCs 5,32 . In our previous work and here, we have not found data to suggest that there are target cell type-specific forms of mOP-LTD plasticity, but we acknowledge that we have not specifically tested this possibility here, and this is something that will require further assessment in the future.  www.nature.com/scientificreports www.nature.com/scientificreports/ We previously demonstrated that cannabinoid receptor-mediated LTD occurs at OFC inputs to DMS 16 , but in the present study we found that OFC-DMS synapses did not respond to MOR agonist treatment. These data, in combination with our findings from DLS 5 , suggest that OFC inputs lack expression of presynaptic MORs. This was surprising to us as we had previously found that mOP-LTD and cannabinoid-LTD are mutually occlusive with one another 25 and thought it likely that since OFC inputs to dorsal striatum express cannabinoid-LTD, that we would also find mOP-LTD at these synapses 16 . Future work will need to resolve how mOP-LTD and cannabinoid-LTD interact, if they are indeed expressed at the same synapses or if they display a heterosynaptic relationship.
In addition to identifying mOP-LTD at specific DMS inputs, we also identified the presence of mOP-LTD at CIN inputs on to MSNs within the DMS, consistent with our previous work 5 . CINs are an important component of the network balance of the striatum 37 , and can affect corticostriatal cannabinoid-LTD 38 , thalamic glutamate release 39 , and dopamine tone 40 . In addition to our work, others have reported on the role of MOR in modulating CIN activity 30,41 . Future work will need to explore the specific role of CIN-driven glutamate release and the role of MOR signaling in these neurons.
While many others have reported the effects of MORs on glutamatergic synaptic transmission in dorsal striatum 5,25,28,29,31 , until recently, only one other study has specifically explored regulation of DMS transmission via presynaptic MORs 28 . These investigators found that the only source of MOR-mediated depression of glutamatergic transmission in the DMS occurs at thalamic inputs 28 . Our findings contradict this study, but there are a few possible explanations for the different findings. One possible explanation is the duration of DAMGO exposure. The other study did not report the duration of agonist exposure and representative current traces implied that DAMGO might be having some effect that was also not reversed by the MOR antagonist CTAP. However no time course was provided so it was not possible to evaluate whether MORs were sufficiently engaged to induce LTD. Another likely explanation is methodological. Our recording conditions only block GABA A receptor-mediated transmission, whereas this other study used a cocktail of metabotropic GABA, glutamate, and acetylcholine receptor antagonists in addition to antagonists of GABA and nicotinic acetylcholine ionotropic receptor antagonists. Our study here does not rule out the necessity of other neurotransmitters being co-factors for mOP-LTD and indeed our earlier work suggested that mGluR5 transmission might be an important element of opioid-LTD in DLS, at least for the form mediated by endogenous opioids 25 . Much more work is required to decipher the mechanisms of mOP-LTD including if induction of opioid-LTD by endogenous opioids in DMS utilizes similar mechanisms as in DLS 25 .
We previously reported that mOP-LTD in the DMS was not affected by ethanol exposure 5 . The critical difference here is that the previous study did not explore synapse-specific MOR plasticity like we performed here, but rather broadly probed glutamate transmission. Using local electrical stimulation, it is difficult to ascertain the www.nature.com/scientificreports www.nature.com/scientificreports/ relative contribution of various cortical, thalamic, amygdalar, and cholinergic interneuron inputs to any given MSN. Our previous data and the data presented here show that ethanol has synapse-type-specific effects: ethanol disrupted mOP-LTD at corticostriatal, but not other synapse types. Altogether, these results could explain the lack of ethanol effects on mOP-LTD in the DMS in our previous work, possibly because the ethanol-insensitive MOR-mediated plasticity at non-cortical inputs to DMS MSNs blunted the observation of ethanol's effects on mOP-LTD at corticostriatal synapses, although additional work will need to be performed to validate this hypothesis. Taking these data together with our previous work indicates that there is some mechanism that makes MOR plasticity at corticostriatal synapses sensitive to disruption by ethanol, but MOR plasticity at thalamic, amygdalar, and CIN glutamatergic inputs insensitive. Future work will determine if ethanol needs a particular synaptic environment to affect mOP-LTD and if this is due to different mechanisms of MOR plasticity at these different synapses or if it is due to upstream effects of ethanol on cortex. Others have demonstrated that ethanol is able to alter LTD mediated by other presynaptic G i/o -coupled GPCRs such as cannabinoid-LTD in DLS 15,24,42 and mGluR2-LTD in DLS and DMS of adolescent mice 43 . Ethanol is also known to disrupt other forms of LTD at multiple other synapse types in the brain 6 . It will be of great interest if ethanol has a common mechanism of action at all of these different synapses.
Our data provide some interesting possibilities for how ethanol may influence DMS-mediated goal-directed behavior given that its effects, in relation to mOP-LTD, are limited to cortical afferents. A loss of LTD at After DAMGO application, eGFP-infected brain slices show mOP-LTD, but not brain slices from mice that were infused with AAV-Cre (last 10 minutes, Welch's t-test, P = 0.0326). (F) eEPSC amplitudes were reduced after DAMGO application only in AAV-eGFP-infused mice, and not in AAV-Cre-infected MOR-flox mice, indicating that MORs from mPFC, ACC, and BLA are indispensable for MOR-mediated glutamatergic depression in the DMS (0-10 min baseline v. final 10 min of recording; paired t-test, eGFP: P < 0.0001, t 7 = 11.24, n = 8 from 3 mice; cre: P = 0.8712, t 8 = 0.1674, n = 9 from 3 mice). Data represent mean ± SEM. *P < 0.05, ***P < 0.001.

Scientific RepoRtS |
(2020) 10:7234 | https://doi.org/10.1038/s41598-020-64203-0 www.nature.com/scientificreports www.nature.com/scientificreports/ specific cortical inputs could permit larger synaptic drive from these cortical regions resulting in greater control over striatal activity. For example, one study showed that LTD induction at mPFC synapses within the DMS decreased alcohol-seeking behavior, while LTP induction increased this behavior 22 . A loss of mOP-LTD at these same synapses may be akin to the LTP conditions and may therefore play a role in increased alcohol-seeking behavior in ethanol-dependent animals. Even though mOP-LTD at BLA inputs was insensitive to ethanol exposure, BLA inputs may have a role in ethanol and drug relapse and therefore MORs may modulate these amygdalostriatal-mediated behaviors 44,45 . In the future, it will also be important to not only investigate the origins of MOR-sensitive synapses, but also the afferent targets: D1-or D2-expressing MSNs (direct or indirect pathway respectively) given their different roles in DMS-mediated ethanol-related behaviors 46 .
In conclusion, in combination with our previous work, we have demonstrated that MORs appear to only regulate specific dorsal striatum synapses with each of these different regions expressing specific forms of MOR-mediated plasticity (LTD or short-term depression). In addition, some mOP-LTD-expressing synapses are sensitive to ethanol and others insensitive. Altogether these findings indicate that specific striatal glutamatergic www.nature.com/scientificreports www.nature.com/scientificreports/ synapses express unique complements of signaling processes that result in different types of plasticity being expressed that are differentially sensitive to drugs of abuse. The specific molecular machinery that creates these differences may be a "synaptic fingerprint" of drug-sensitive synapses. In the future, it will be important to further elucidate the synaptic fingerprints of not only MOR signaling at specific striatal synapses, but also other forms of plasticity, to mechanistically understand ethanol's network effects in order to identify novel treatments for alcohol use disorder and drug addiction.

Methods
All experiments were performed similar to our previous studies with some experiment-specific modifications 5,25,47,48 . These methods are described in brief below.
Animals and materials. Animal care and experimental protocols for this study were approved by the Institutional Animal Care and Use Committee at the Indiana University School of Medicine and all guidelines for ethical protocols and care of experimental animals established by the NIH (National Institutes of Health, Maryland, USA) were followed. Male C57BL/6J mice were ordered from the Jackson Laboratory (Bar Harbor, Maine, USA). ChATCre transgenic mice were bred and genotyped in-house (Original stock strain: ChATCre: JAX #006410). Conditional MOR knockout mice (MOR-flox) were a generous gift from Dr. Jennifer Whistler (UC Davis) and have been previously described 5,49 . All transgenic mice used in these studies have been backcrossed to C57BL/6J mice for a minimum of 7 generations. The mice used in these studies were between PND 60-100 at the time of experimentation (with the exception of ChATCre mice that were ~PND 60-120 and MOR-flox AAV-cre-injected mice ~PND 105-126). Animals were group-housed in a standard 12-h light/dark cycle (lights on at 0800) at 50% humidity. Food and water were available ad libitum.

Electrophysiology recordings. Whole-cell recordings of excitatory postsynaptic currents (EPSCs) in
MSNs were carried out at 29-32 °C and aCSF was continuously perfused at a rate of 1-2 ml/min. Recordings were performed in the voltage clamp configuration using a Multiclamp 700B amplifier and a Digidata 1550B (Axon Instruments, Union City, CA). Slices were visualized on an Olympus BX51WI microscope (Olympus Corporation of America). MSNs were identified by their size, membrane resistance, and capacitance. Picrotoxin (50 μM) was added to the aCSF for recordings to isolate excitatory transmission. Patch pipettes were prepared from filament-containing borosilicate micropipettes (World Precision Instruments) using a P-1000 micropipette puller (Sutter Instruments, Novato, CA), having a 2.0-3.5 MΩ resistance. The internal solution contained (in mM): 120 CsMeSO 3 , 5 NaCl, 10 TEA-Cl, 10 HEPES, 5 lidocaine bromide, 1.1 EGTA, 0.3 Na-GTP and 4 Mg-ATP (pH 7.2 and 290-310 mOsm). MSNs were voltage clamped at −60 mV for the duration of the recordings. To detect asynchronous glutamatergic transmission, aCSF Ca 2+ was replaced with Sr 2+ (2 mM). For electrically evoked recordings, a twisted tungsten bipolar stimulating electrode (PlasticsONE, Roanoke, VA) was placed at the border of the overlying corpus callosum. eEPSCs were generated by a DS3 Isolated Current Stimulator (Digitimer, Ft. Lauderdale, FL) every 20 s and stimulus intensity was adjusted to produce stable electrically-evoked EPSCs (eEPSCs) of 200-600 pA in amplitude prior to the initiation of experimental recording. Data were acquired using Clampex 10.3 (Molecular Devices, Sunnyvale, CA). Series resistance was monitored and only cells with a stable series resistance (less than 25 MΩ and that did not change more than 15% during recording) were included for data analysis. Recordings were made 2-7 h after euthanasia.