Activation of lateral hypothalamic group III mGluRs suppresses cocaine-seeking following abstinence and normalizes drug-associated increases in excitatory drive to orexin/hypocretin cells

Abstract

The perifornical/lateral hypothalamic area (LHA) orexin (hypocretin) system is involved in drug-seeking behavior elicited by drug-associated stimuli. Cocaine exposure is associated with presynaptic plasticity at LHA orexin cells such that excitatory input to orexin cells is enhanced acutely and into withdrawal. These changes may augment orexin cell reactivity to drug-related cues during abstinence and contribute to relapse-like behavior. Studies in hypothalamic slices from drug-naïve animals indicate that agonism of group III metabotropic glutamate receptors (mGluRs) reduces presynaptic glutamate release onto orexin cells. Therefore, we focused on the group III mGluR system as a potential target to reduce orexin cell excitability in-vivo. First, we verified that group III mGluRs regulate orexin cell activity in behaving animals by showing that intra-LHA infusions of the selective agonist L-(+)-2-Amino-4-phosphonobutyric acid (L-AP4) reduces c-fos expression in orexin cells following 24h food deprivation. Next, we extended these findings to show that intra-LHA L-AP4 infusions reduced discriminative stimulus-driven cocaine-seeking following withdrawal. Importantly, L-AP4 had no effect on lever pressing for sucrose pellets or general motoric behavior. Finally, using whole-cell patch-clamp recordings from identified orexin cells in orexin-GFP transgenic mice, we show that enhanced presynaptic drive to orexin cells persists for up to 14d into withdrawal and that this plasticity can be normalized by L-AP4. Together, these data indicate that activation of group III mGluRs in LHA reduces orexin cell activity in vivo and may be an effective strategy to suppress cocaine-seeking behavior following withdrawal. These effects are likely mediated, at least in part, by normalization of presynaptic plasticity at orexin cells that occurs as a result of cocaine exposure.

1. INTRODUCTION

Orexin (hypocretin) cells in the perifornical/lateral hypothalamic area (LHA) play a key role in motivated behavior (Brodnik et al., 2018; Harris et al., 2005; James et al., 2017b; James et al., 2018b; Mahler et al., 2014; Sakurai, 2014). For example, the orexin system is particularly important for driving drug-seeking behavior elicited by environmental cues associated with drug use (for review, see (James et al., 2017b)). Indeed, orexin cells are recruited by discriminative and contextual stimuli previously paired with drug (Dayas et al., 2008; Martin-Fardon et al., 2018; Moorman et al., 2016), and the magnitude of orexin cell activation is strongly correlated with drug-seeking behavior (Harris et al., 2005; Moorman et al., 2016; Richardson and Aston-Jones, 2012). Furthermore, orexin receptor antagonists, particularly those acting at the orexin-1 receptor, are highly effective at blocking stimulus-driven drug-seeking for various drugs of abuse (Martin-Fardon and Weiss, 2014; Moorman et al., 2017; Plaza-Zabala et al., 2013; Smith et al., 2009; Smith et al., 2010), but these compounds have minimal effects on drug-seeking elicited by drug priming injections (Zhou et al., 2012). Based on these data, the orexin system is a potential therapeutic target to ameliorate craving in response to drug-associated stimuli during abstinence (James et al., 2012; Yeoh et al., 2014a).

Excitatory synapses onto orexin cells rapidly rewire in response to environmental and physiological challenges, including food restriction, sleep deprivation and drug exposure/withdrawal (Horvath and Gao, 2005; Rao et al., 2013; Yeoh et al., 2012). We and others have reported that cocaine exposure enhances excitatory drive to LHA orexin cells and that these changes persist into acute withdrawal (Rao et al., 2013; Yeoh et al., 2012). Such changes may drive enhanced orexin output in response to drug-associated cues during abstinence and contribute to relapse-like behavior. Thus, suppression of excitatory drive to LHA orexin circuits during withdrawal might represent a strategy to reduce relapse risk (Yeoh et al., 2012). To this end, previous work has shown that metabotropic glutamate receptors (mGluRs) in the hypothalamus maintain tonic presynaptic inhibition at excitatory (and inhibitory) synapses (Kuzmiski and Bains, 2010; Kuzmiski et al., 2009). In particular, the group III mGluR subtype has been shown to gate presynaptic glutamate release onto orexin cells, such that activation of these receptors with the selective agonist L-(+)-2-Amino-4-phosphonobutyric acid (L-AP4) reduces presynaptic glutamate release onto orexin cells in vitro (Acuna-Goycolea et al., 2004). Currently, it is unclear whether agonism of group III mGluRs reduces orexin cell activity in vivo, or whether this approach could be used to suppress enhanced excitatory input to orexin cells following cocaine.

Here we sought to examine the effect of infusing L-AP4 directly into LHA in freely behaving rats, both in terms of impacts on orexin cell activity, and drug-seeking behavior. First, we evoked c-fos expression in orexin cells using food deprivation, a stimulus known to enhance excitatory drive to orexin cells (Horvath and Gao, 2005), and show that Fos induction is suppressed by intra-LHA L-AP4 infusions. In rats with a history of cocaine self-administration, we show that intra-LHA L-AP4 injections reduces drug-seeking elicited by discriminative stimuli following 14d withdrawal. Next, we sought to verify LAP-4’s mechanism of action by carrying out whole-cell patch clamp recordings in orexin-GFP mice with a history of cocaine exposure. We extend upon previous findings to show that the 14d withdrawal time point is associated with enhanced measures of presynaptic and postsynaptic plasticity to orexin cells, and that L-AP4 selectively normalizes withdrawal-induced increases in excitatory drive. Together, our findings identify the group III mGluR system as a potential mechanism through which drug-induced plasticity to orexin cells can be normalized, leading to reduced drug-seeking elicited by drug-associated stimuli during abstinence.

2. MATERIALS & METHODS

2.1. Experiment 1: Behavioral studies in rats

2.1.1. Animals

Rats weighing 200–250g upon arrival (Charles River, USA) were single-housed on a reversed 12hr light/dark cycle (lights off at 0800) with food and water available ad libitum. All procedures were approved by the Institutional Animal Care and Use Committee of Rutgers University. All experiments were carried out in the active period.

2.1.2. Surgery for intravenous catheter implantation

Rats were handled daily for 1wk before undergoing intravenous (i.v) catheter surgery as described previously (James et al., 2016). Briefly, animals were anaesthetized with isoflurane (1–3% with a flow rate of 2L/min) and received the analgesic rimadyl (5mg/kg, s.c.). Chronic indwelling catheters were inserted into the right jugular vein and exited the body via a port between the scapulae. Rats received prophylactic i.v. cefazolin (10mg) and heparin (10U) daily starting 1 day after surgery and continuing throughout self-administration training.

2.1.3. Surgery for LH-directed cannulae

Animals were placed in a stereotaxic frame, anaesthetized as above, and then underwent stereotaxic surgery for guide cannulae (23G; PlasticsOne) implantation 2mm above the fornix (AP:−2.65; ML:±1.90; DV:−7.25). Cannulae were affixed to the skull with four stainless steel screws and dental cement, and were occluded with 26G gauge bilateral steel stylets (PlasticsOne).

2.1.4. Food deprivation and c-fos expression experiments

A subset of rats (n=5) previously used for sucrose self-administration and locomotor testing (described below) were used in this experiment. These rats were maintained on an ad libitum diet for all other tests; here they were subjected to 24h food deprivation, which is known to enhance mEPSCs frequency and induce c-fos expression in orexin cells (Horvath and Gao, 2005; Mieda et al., 2004). Rats received unilateral microinjections of L-AP4 (100nmol) and aCSF into opposite hemispheres (hemispheres counterbalanced across animals). This dose of L-AP4 was selected based on previous studies indicating that injections into the striatum blocks cocaine hyperlocomotion without affecting general motor behavior (Mao and Wang, 2000). Ninety minutes later, animals were anesthetized (ketamine/xylazine; 56.5/8.7mg/kg, respectively, i.p.) and then transcardially perfused with saline followed by 4% paraformaldehyde (PFA). Brains were removed and post-fixed overnight, and then transferred to a 20% sucrose solution for storage until they were sectioned into 40um sections on a cryostat.

Tissue from LHA was processed for Fos-protein and orexin-A using procedures outlined elsewhere (James et al., 2018b; Yeoh et al., 2012). Briefly, tissue underwent three washes in PBST-azide (phosphate buffer saline + 0.03% Triton X-100 + azide), before being incubated overnight at room temperature in primary antibodies goat anti-orexin-A (1:500; Santa-Cruz Biotechnology, catalog number SC-8070; validation details at JCN Antibody database AB_653610) and rabbit anti-Fos (1:1000; Synaptic Systems, catalog number 226003; validation details at JCN Antibody database AB_2231974) in 2% normal donkey serum. The next day, the tissue was washed three times in PBST and incubated for 2h in appropriate secondary antibodies coupled to Alexa-Fluor 488/594-conjugated donkey anti-goat/rabbit (Jackson Immunoresearch Laboratories). Sections were then rinsed in PBS, mounted onto glass slides and cover slipped using Fluoroshield mounting medium with DAPI (Abcam). Hypothalamic sections were imaged in tiles with a 16× objective on a Zeiss AxioZoom V16 microscope and then stitched together using Zen Imaging software. Cell counts were also performed using the Zen Imaging software. The number of cells immunoreactive for orexin-A, as well as the number of orexin-A cells that also expressed Fos-protein, were quantified in three sections taken through the main rostrocaudal extent of the hypothalamic orexin cell region in each rat by an experimenter blind to hemispheric treatment conditions. Counts of orexin+, Fos+/orexin+ and Fos+/orexin-cells were carried out in a 300×300μm region immediately ventral to the injector tip in each hemisphere. The percentage of orexin+ immunoreactive cells that co-expressed Fos-protein, as well as the number of Fos-expressing non-orexin cells, were averaged across each hemisphere for each subject, and then compared across L-AP4 versus aCSF microinjected hemispheres across all subjects.

Cocaine self-administration training

A separate group of rats (n=16) were trained to self-administer cocaine in standard Med-Associates operant chambers. Each chamber was equipped with two retractable levers with white lights above it, a red house-light located at the top of the chamber wall opposing the levers and a tone generator. A syringe pump located outside of the sound-attenuating chamber delivered cocaine intravenously. Data acquisition and behavioral testing equipment were controlled by MED-PC IV program (Med Associates, USA). Seven days after surgery, animals were trained to respond for i.v. cocaine hydrochloride infusions. Animals were first trained on a fixed ratio 1 (FR1) schedule of reinforcement for 2hr/day, 6–7days/week. During this period, responding on the active (right) lever resulted in a 3.6s infusion of cocaine (0.2mg/50ul infusion) via the i.v. catheter, which was paired with the activation of white cue light above the active lever. A 20s timeout period followed each infusion delivery, which was signaled by the house light turning off, during which additional presses on the active lever were of no consequence. Responses on the inactive (left) lever were recorded but had no scheduled consequences at any time. Rats underwent daily 2hr cocaine self-administration sessions until they met a criterion of >10 cocaine infusions/day over three consecutive days. At this time, rats were transitioned to daily 2hr randomized conditioning sessions for cocaine or saline infusions (FR1), in the presence of distinct discriminative stimuli (‘first round’ of conditioning). We used a discriminative stimulus paradigm, as this form of stimulus is associated with robust activation of orexin cells (Dayas et al., 2008; Martin-Fardon et al., 2018; Moorman et al., 2016). For cocaine (DS⁺), this involved black-striped wallpaper on the chamber door and a vanilla odor placed in a receptacle under the active lever. The time out period following an active lever response was signaled with the illumination of house light. For saline (DS⁻), chambers were equipped with black-spotted wallpaper on the chamber door and a lemon odor. Time out periods were signaled by a constant 78dB 2.9kHz tone. After 16d of conditioning (8DS⁺/8DS⁻ sessions, randomized order), animals underwent 13d of abstinence in their home cage where they were left undisturbed except for weekly bedding changes.

2.1.5. Test of drug-seeking after abstinence

After 13d of abstinence, animals were returned to the operant chambers, presented with the DS⁻cues, and tested for drug-seeking behavior for 30min; active lever responses resulted in CS⁺ delivery but not a saline infusion. Immediately following testing, rats were gently restrained and injectors were lowered into the bilateral cannulae (28G, PlasticsOne) and then removed; no injection was made. The following day (14d following final conditioning session), animals underwent testing of drug-seeking following abstinence under DS⁺ conditions for 30min. 10min prior to testing, animals received a bilateral microinfusion of L-AP4 (100nmol; (Mao and Wang, 2000)) or vehicle (aCSF) into the orexin field (0.5ul over 1min). Injectors were left in place for an additional 1min to limit upward diffusion of the injectate. Rats were then placed in the self-administration chambers under DS⁺ conditions and tested for drug-seeking behavior for 30 min; lever presses resulted in CS⁺ delivery but no cocaine. The following day, animals were returned to self-administration conditioning training (8 days DS⁺, 8 days DS⁻; ‘second round’ of conditioning), and then underwent a second round of abstinence. After 13d abstinence, rats were tested under DS⁻ conditions, as before. The following day, rats were again tested for drug-seeking in the DS⁺ context, but prior to testing, received the opposite treatment (L-AP4 or aCSF) to their first test. The order in which animals received the two different treatments was counterbalanced across all animals in the study.

2.1.6. Sucrose pellet self-administration

Drug-naïve animals (n=7) were trained to self-administer sucrose pellets (45mg, Test Diet) on an FR1 schedule during 2h sessions (James et al., 2018a; McGlinchey et al., 2016). As with cocaine training, each rewarded active lever response was paired with a light and tone cue, followed by a 20s timeout period. Training continued until the number of active lever presses and pellets earned was stable (<20% variability) over 3 consecutive days. The effect of intra-LH L-AP4 or aCSF infusions on sucrose self-administration was tested under a FR1 schedule for 2h, using injection procedures identical to those used for cocaine (described above). Test days were separated by at least 2d of further FR1 training (<20% variability relative to pre-test baseline).

2.1.7. Locomotor assay

The same rats tested for sucrose self-administration (n=7) were also assessed for general locomotor activity. Procedures are published elsewhere (James et al., 2018a; James et al., 2018b). Briefly, rats were placed in locomotor chambers (clear acrylic, 42cm × 42cm × 30cm) equipped with SuperFlex monitors (Omintech Electronics Inc, Columbus, OH) containing a 16 × 16 infrared light beam array for the x/y axis (horizontal activity) and 16 × 16 infrared light beams for the z axis (vertical activity). Activity was recorded by Fusion SuperFlex software. Animals were habituated to the locomotor boxes for 2h/day for at least 3d, and until the average total distance traveled by each rat was within a range of ±25% of the mean of those days. On test days, total distance traveled, as well as horizontal and vertical beam breaks were recorded over a 2h session. Between test days, animals were again habituated to the boxes until stable behavior was again observed over 3 consecutive days. All animals received L-AP4 and aCSF as described above in a counterbalanced order.

2.1.8. Verification of injector placement

For cocaine rats, injector placement was verified by deeply anesthetizing the animals with isoflurane and then lowering injectors into the cannulae. Rats were then decapitated and brains were flash-frozen in 2-methylbutane and stored at −80°C. Brains were sectioned into 40μm sections on a cryostat, slide mounted, and visually inspected for injection tracts under a 40× microscope. All other experiments (sucrose, locomotor and Fos studies) utilized the same animals; injection sites were verified during sectioning and confirmed during cell quantification.

2.1.10. Statistical analyses

For the food deprivation experiment, the percentage of orexin-positive cells co-expressing Fos-protein was compared across treatment groups using a paired t-test analysis (with the aCSF-injected hemisphere acting as the control for each animal). In cocaine self-administering rats, self-administration values for the final DS⁺ and DS⁻ sessions were compared using paired samples t-tests. Responses during DS⁻ drug-seeking tests sessions were calculated as the average of both sessions and compared to DS⁺ drug-seeking values following L-AP4 or vehicle treatment using repeated measures ANOVA. Two animals were excluded from the drug-seeking analyses as they failed to exhibit robust DS⁺/induced-behavior following vehicle treatment. For these two animals, responding on the DS⁺ test was lower than their responding on the DS⁻ test and was more than 3SDs below the overall group average. Analysis of behavioral data (including DS⁺/DS⁻ training data) excluded these animals, as well as those ultimately found to have inaccurate microinjection sites (outlined below).

2.2. Experiment 2: Electrophysiological experiments in mice

2.2.1. Animals

All experimental procedures were approved by the University of Newcastle Animal Care and Ethics Committee and performed in accordance with the NSW Animal Research Act, Australia. Breeding pairs of orexin-GFP mice DBF1 (F1 of DBA2XB6J, (Sakurai et al., 2005) were transferred from Kanazawa University and backcrossed onto C57BL/6J background for at least nine generations at the Animal Resources Centre, NSW, Australia. Mice aged 4–6 weeks were housed 2–4/cage in temperature- and humidity-controlled conditions on a reverse 12hr daylight cycle (lights off 0700) with ad libitum access to food and water. All experiments were carried out during the animals’ dark (active) phase.

2.2.2. Confirmation of GFP-orexin cells in the LHA

This mouse line has been well validated and we observed GFP labeling exclusively in the LHA medial and lateral to the fornix as per previous studies (see also Figure 5C; (Yamanaka et al., 2003a; Yamanaka et al., 2003b).

Figure 5. Cocaine induces locomotor sensitization & distribution of GFP-orexin cells in the LHA.

A-B, repeated cocaine injections produced a significant increase in locomotor activity (distance travelled) as compared to saline-treated animals (*P<0.05, relative to baseline). Locomotor activity in response to cocaine was significantly higher in the final 3d of cocaine exposure compared to the first 3d of cocaine exposure (*P<0.05; Fig 1, left panel). C, GFP expression is contained exclusively to orexin cells (blue). Scale bar 100μM.

2.2.3. Drugs

Cocaine Hydrochloride (GlaxoSmithKline,Australia) was dissolved in sterile physiological saline: (15mg/ml) for intraperitoneal (i.p.) injection. Picrotoxin and 6-cyano-7-nitroquinoxaline-2,3-dione, CNQX, were purchased from Sigma-Aldrich (MO,USA). L-(+)-2-Amino-4-phosphonobutyric acid (L-AP4) was purchased from Tocris Bioscience (Bristol,UK).

2.2.4. Behavioral Paradigm

Prior to treatment, animals were conditioned to daily handling and subjected to single-daily sham injections, after which they were placed in an enclosed arena (50cm × 50cm) for 30 min. Animals were then randomly allocated into two treatment groups, either saline (n=10) or cocaine (15mg/kg; i.p. n=8) for seven consecutive days (Yeoh et al., 2014b; Yeoh et al., 2012). Immediately following injections, animals were placed in the same enclosed arenas (30mins) before they were returned to their home cage. In a subset of mice (n=3 saline v.s. n=3 cocaine), locomotor activity (total distance traveled) was assessed using Ethovision software. Brain slice electrophysiology experiments were undertaken 14 days after the last cocaine/saline injection session.

2.2.5. Slice Preparation for electrophysiology experiments

Mice were deeply anaesthesized with ketamine (1ml/kg, i.p) and decapitated. Brains were rapidly removed and immersed in ice-cold oxygenated (95%O₂, 5%CO₂) sucrose-substituted artificial cerebrospinal fluid (S-ACSF, containing in mM: 236.2 sucrose, 25NaHCO₃, 13.6glucose, 2.5KCl, 2.5CaCl₂, 1NaH₂PO₄, 1MgCl₂). Coronal slices (250μm) containing the LHA were obtained using a vibrating microtome (Campden, England), then transferred to an oxygenated storage chamber containing ACSF (119.4mM NaCl substituted for sucrose) and incubated for 1hr at room temperature prior to recording.

2.2.6. Electrophysiology Recording

Slices were transferred to a recording chamber and continually superfused with oxygenated ACSF, maintained between 32–34°C by a temperature controller (Warner Instruments, USA). Whole-cell recordings were made using Multiclamp 700B amplifier (Molecular Devices,CA), digitized at 10kHz, via an ITC-18 computer interface (Instrutech,NY) and recorded onto a Macintosh computer running Axograph X software. All recordings were restricted to the LHA brain region spanning between Bregma −1.06 mm and −1.70 mm (Paxinos et al., 2001). After obtaining whole-cell recording configuration, series resistance and input resistance were calculated based on the response to a −5mV voltage step from a holding potential of −70mV. These values were monitored at the beginning and end of each recording and data were rejected if values changed by more than 20%. A bipolar stimulating electrode was positioned immediately medial and dorsal to the PF/LHA to stimulate excitatory inputs (0.1ms pulse duration, 1.2× threshold stimulus intensity). Evoked EPSCs (eEPSCs) were recorded with an internal solution containing (in mM): 120Cesium methanesulfonate, 20CsCl, 10HEPES, 4Mg-ATP, 0.3Na₃-GTP, 0.2EGTA, 10NA-phosphocreatine, 5QX-314 (pH7.3 with CsOH). These recordings assessed both the paired-pulse ratio and AMPA:NMDA ratio of eEPSCs. Cells were voltage clamped at −70mV, in the presence of picrotoxin (100μM, Yamanaka et al, 2003) to record the paired-pulse ratio (50ms interstimulus interval) of AMPAR-mediated EPSCs. NMDAR-mediated EPSCs were recorded at a holding potential of +40mV with the addition of CNQX (10μM) to abolish AMPA-receptor mediated currents. Miniature excitatory postsynaptic currents (mEPSCs) were recorded in voltage clamp at a holding potential of −70mV with series resistance of <25MΩ, in the presence of tetrodotoxin (1μM) and picrotoxin (100μM). In addition, L-AP4 (50μM) was bath applied to determine the effects of GPIII mGluR activation. Recording pipettes (6–9MΩ) for mEPSC recordings contained an internal solution containing (in mM): 135KCH₃SO₄, 6NaCl, 2MgCl₂, 10HEPES, 0.1EGTA, 2MgATP, 0.3NaGTP (pH7.3 with KOH) and 0.1% Neurobiotin.

2.2.7. Data Analysis

Peak amplitude, rise-time and decay time constant were evaluated for eEPSCs using semi-automated procedures in the Axograph X software. Paired-pulse ratio (PPR) was calculated by dividing the mean peak amplitude of the second eEPSC by the mean peak amplitude of the first. AMPA:NMDA ratios were calculated by dividing the mean peak amplitude of evoked AMPA EPSCs (recorded at −70mV) by the mean peak amplitude of evoked NMDA EPSCs (recorded in CNQX at +40mV). mEPSCs were detected and captured using a sliding template method (Clements and Bekkers, 1997), along with a minimum amplitude threshold criteria of 10pA. Captured mEPSCs were individually inspected and excluded from the analysis if they included overlapping events or had an unstable baseline before the rise or during the decay phase of the mEPSC. Analyses were performed on averaged mEPSCs, generated by aligning the rising phase of all accepted events. Peak amplitude, rise-time (calculated over 10–90% of peak amplitude) and decay time constant (calculated over 20–80% of the decay phase) were obtained using semi-automated procedures. Average mEPSC frequency was obtained by dividing the number of captured events by the analysis duration in seconds.

2.2.8. Statistical analyses

For sensitization behavior, data were grouped into ‘baseline’ (last 3d of habituation), ‘early’ (first 3d of cocaine/saline treatment) and ‘late’ (last 3d of cocaine/saline treatment), and compared using repeated-measures ANOVA. All electrophysiological data were compared using either Student’s t-tests or one-way ANOVA.

3. RESULTS

3.1. Effects of L-AP4 on orexin cell activity and drug-seeking behavior in-vivo.

3.1.1. Microinfusions of L-AP4 into the LHA reduced the percentage of Fos-positive LH and Fos-positive orexin neurons in food-deprived animals

To test whether intra-LHA injections of L-AP4 can alter orexin cell activity in vivo, we infused L-AP4 and aCSF unilaterally (opposite hemispheres, counterbalanced) following 24h food deprivation. The total number of orexin-immunoreactive cells quantified was similar across aCSF- and L-AP4-treated hemispheres (P>0.05, Figure 1A). Consistent with previous studies (Lutter et al., 2008; Mieda et al., 2004), approximately 75% of orexin-immunoreactive cells co-expressed Fos following food deprivation in the aCSF control group. This percentage was significantly lower (~50%) in the L-AP4-treated group (t₄=9.774, P=0.0006, paired-samples t-test; Figure 1B). We also observed an L-AP4-induced suppression of Fos in non-orexin neurons in LHA (t₄=2.974, P=0.041; Figure 1C), although to a slightly lesser degree (~30%). Representative images of Fos-expression following aCSF or L-AP4 microinfusions are presented in Figure 1D. Together, these data indicate that activation of group III mGluRs with L-AP4 suppressed food restriction-induced cell activity in LHA, including in orexin-expressing neurons.

Figure 1. Intra-LHA microinfusions of L-AP4 attenuate orexin cell activity.

Activation of Fos-expressing orexin neurons was quantified in food-deprived rats following intra-LH infusions of L-AP4 into one hemisphere, and aCSF into the opposite hemisphere. A, The total number of orexin-immunoreactive cells quantified in L-AP4-injected hemispheres was identical to that of the aCSF-injected hemispheres. B, L-AP4 microinfusions into LHA significantly reduced the number of orexin-immunoreactive cells that co-expressed Fos (***P<.001). C, L-AP4 microinfusions also decreased Fos expression in non-orexin LHA neurons. D, Representative photomicrographs from the same animal showing Fos (green) in orexin (red) cells following intra-LHA injections of aCSF (top panel) and L-AP4 (bottom panel) in opposing hemispheres. Scale bar = 25μm. White arrows indicate neurons immunoreactive for orexin and Fos. Yellow arrows indicate cells immunoreactive for orexin only.

3.1.2. Microinfusions of L-AP4 into LHA reduced cocaine-seeking following 14d abstinence

Animals learned to discriminate between DS⁺ and DS⁻ stimuli during the first round of conditioning, as evidenced by a significantly greater number of active lever presses (t₁₀=2.330, P=0.0421) and infusions (t₁₀=4.648, P=0.009) on the final DS⁺ session compared DS⁻ session (Figure 2A). Inactive lever responding was significantly higher on the final DS⁻ sessions compared to the DS⁺ sessions (t₁₀=4.901, P=0.0006; Figure 2A). On the second round of self-administration training, the number of active lever presses and infusions earned during DS⁺ and DS⁻ training sessions was almost identical to those in the first round of training (P’s>0.05, data not shown). While inactive lever responding was similar during DS⁺ training sessions (P>0.05, data not shown), there was a significant decrease in the number of inactive lever responses during the DS⁻ sessions in the second round of training compared to the first (t₁₀=3.295, P=0.0093; data not shown). Importantly, performance during the second round of training did not differ according to whether animals were treated with aCSF or L-AP4 on the first reinstatement test (described below; P’s>0.05; data not shown).

Figure 2. Intra-LHA microinfusions of L-AP4 attenuate cocaine seeking behavior elicited by drug-associated discriminative stimuli following 14d abstinence.

A, Rats were trained to self-administer cocaine (DS⁺) or saline (DS⁻) in the presence of discriminative stimuli. Rats learned to discriminate between DS⁺/DS⁻ stimuli, as indicated by higher active lever responses and infusions during the final DS⁺ session, relative to the final DS⁻ session of the first round of conditioning (*P<0.05; **P<0.01). Behavior during the second round of conditioning was indistinguishable from the first (data not shown). B, Following 14d withdrawal, testing (30min) under DS⁺ conditions was associated with significantly higher levels of active lever responding compared to testing under DS⁻ conditions in aCSF-treated rats (^^P<0.01). Active lever responding in the DS⁺ test was significantly attenuated following intra-LH microinfusions of L-AP4 (*P<0.05). C, Comparison of cumulative lever responding across the DS⁺ test indicates that lever pressing was significantly attenuated by L-AP4 across the entire duration of the test session (****P<.0001).

To test the effect of group III agonism on drug-seeking behavior following 14d abstinence, we infused L-AP4 or aCSF directly into LHA immediately prior to testing animals for drug-seeking in the presence of DS⁺ stimuli. Following aCSF microinfusions, re-exposure to DS⁺ stimuli was associated with significantly greater levels of drug-seeking (active lever responding) compared to testing under DS⁻ conditions (F_2,32=8.54, P=0.0093; Holm-Sidak post-hoc comparison, P=0.0088; Figure 2B). The DS⁺-elicited drug-seeking was significantly attenuated by intra-LHA infusions of L-AP4 (P=0.0190, Holm-Sidak post-hoc test; Figure 2B). This LAP-4-induced suppression of active lever responding was evident across the entire duration of the test session (main effect of ‘treatment’, F_1,120=35.94, P<0.0001, two-way ANOVA; Figure 2C). In contrast, L-AP4 had no effect on inactive lever responding during the DS⁺ test (P>0.05, data not shown). Analyses of DS⁺-elicited drug-seeking on the first versus second round of reinstatement testing revealed a trend towards higher average active lever presses in the first round (39 vs 27), although this did not reach significance (t₉=0.90, P=0.3904).

3.1.3. Microinfusions of L-AP4 into LHA had no effect on lever pressing for sucrose or general locomotor activity

To rule out any general effects of intra-LHA L-AP4 infusions on lever pressing behavior, we tested the effect of L-AP4 on a low-effort (FR1) responding for sucrose pellets. Animals exhibited a high level of active lever responding (M=248 active lever presses/2h session), with over half of this responding occurring within the first 30min of the session; thus this paradigm served as an effective control task to test for any impairments in lever pressing behavior induced by L-AP4 microinfusions. On test, infusions of either aCSF or L-AP4 did not affect the number of responses made on the active lever during the 30min test session (P>0.05; Figure 3A,B), nor the number of sucrose pellets earned during the session (P>0.05; Figure 3C).

Figure 3. LHA microinfusions of L-AP4 have no effect on high-rate responding for sucrose pellets or general locomotor activity.

A, In animals trained to self-administer sucrose pellets on an FR1 schedule, L-AP4 had no effect on the total number of active lever responses made in the 30min test session. B, Active lever responding for sucrose was similar between L-AP4-and aCSF-treated animals at all time points during the 30min test. C, L-AP4 did not affect the total number of sucrose pellets earned during the 30min test. D, On a 30min test of general locomotor activity, intra-LHA L-AP4 microinfusions had no effect on the distance traveled in the locomotor test in any of the 5min epochs. E, Similarly, L-AP4 did not affect the total distance traveled across the entire duration of the test. F, L-AP4 did not affect horizontal activity. G, L-AP4 did not affect vertical activity. H, Consistent with evidence that orexin signaling has anxiogenic properties, intra-LHA microinfusions of L-AP4 significantly increased the amount of time spent in the centre square of the open field during the locomotor test. *P<0.05.

On a test of exploration in a novel open field environment, there was no effect of treatment on total distance travelled, (Time × Treatment interaction, P>0.05; Figure 3D,E), horizontal activity (P>0.05; Figure 3F) or vertical activity (P>0.05; Figure 3G) during locomotor testing, indicating that L-AP4 infusions did not affect general motoric activity. Interestingly, L-AP4 treatment significantly increased time spent in the centre square of the open field during locomotor testing (t₆=2.96, P<0.05; Figure 3H), indicating that L-AP4 may have reduced anxiety-like behavior during this test.

3.1.4. Histology

For drug-seeking experiments, microinjections (L-AP4 and vehicle) made into the LHA ranged from 2.64 mm to 3.12 mm posterior to bregma (Figure 4, green dots). Three animals were excluded from drug-seeking analyses based on misplaced injection sites. In one case, both injectors were ventral to the orexin cell field (pictured in Figure 4; pink dots). This animal exhibited much higher reinstatement behavior following L-AP4 compared to the group average of animals that received accurate injections of L-AP4 (active lever presses=27, compared to a group L-AP4 average 16±3 SEM). Another animal had one accurate injection site in one hemisphere, but the other was located ventral to the orexin cell field (pictured in Figure 4; pink dots). In another, both injectors were located in the very rostral extent of the orexin cell field (−2.28mm; not pictured) rather than the center of the orexin cell field as in all other cases. For Fos and locomotor control experiments, all cannula placements were accurate; microinjections ranged from 2.64mm to 2.92mm posterior to bregma (Figure 4, blue dots).

Figure 4. Localization of LHA-directed L-AP4 microinjections.

Schematics of the hypothalamus showing accurately placed infusion sites for drug-seeking (green) and Fos/locomotor control experiments (blue). Misplaced infusions from the drug-seeking experiment that were excluded from behavioral analyses are depicted in pink.

3.2. Effects of i.p. cocaine on plasticity in LHA orexin cells and impact of GPIII mGluR activation

3.2.1. Cocaine produced locomotor sensitization

Orexin-GFP mice that were exposed to cocaine exhibited a significant increase in locomotor activity compared to baseline (Figure 5A). Locomotor activity in response to cocaine was significantly higher in the final 3d of cocaine exposure (t(4)=6.976, P<0.05) compared to the first 3d of cocaine exposure (t(4)=3.539, P<0.05; Figure 5A), indicating a significant sensitization of locomotor activity (main effect of ‘Treatment’, F_{2, 4}=24.33, P=0.0058). In contrast, saline injections did not affect locomotor activity across testing days (Figure 5B).

3.2.2. Cocaine-induced changes to LHA-orexin circuits persist for 2 weeks post-cocaine

To determine whether cocaine-induced plasticity in orexin cell circuits persisted into withdrawal, we recorded orexin cells in hypothalamic slices taken from animals that had undergone 14d home cage abstinence following 7d of cocaine injections. Cocaine-exposed animals displayed a paired-pulse depression compared to controls (t₂₃=2.598, P<0.05, cocaine vs saline, cell yield/animal = 5.5 vs 3.2, range = 0.26–1.38 vs 0.66–1.41 respectively, Figure 6A) and a significant increase in the AMPA:NMDA ratio in orexin cells from 14d withdrawn animals was also observed (t₁₃=−2.731, P<0.05, cocaine vs saline, cell yield/animal = 3.0 vs 1.8, range = 1.93–5.96 vs 1.44–3.54 respectively, Figure 6B).

Figure 6. Cocaine-induced changes in paired-pulse ratio, AMPA:NMDA ratio, and mEPSCs frequency on D14 of cocaine withdrawal.

A, bar graph compares group data indicating that recordings from cocaine-exposed mice exhibited paired-pulse depression (*P<0.05). Inset shows representative traces of eEPSCs recorded in orexin cells from saline- and cocaine-exposed animals. B, bar graph compares group data indicating that recordings from cocaine-exposed mice exhibited an increase in AMPA:NMDA ratio (*P<0.05). Inset shows representative overlaid AMPA-mediated (recorded at −70mV) and NMDA-mediated (recorded at +40mV, in 10μM CNQX) eEPSCs. C, traces show representative mEPSCs recordings of orexin neurons from animals exposed to either saline (top) or cocaine (bottom). D, Bar graphs compare the mEPSCs frequency recorded in orexin cells from saline-exposed animals, saline + bath application of L-AP4, cocaine-exposed animals, and cocaine + bath application of L-AP4 at withdrawal day 14. Cocaine exposure significantly increased the mEPSCs frequency compared to saline control (**P<0.01), and application of L-AP4 reduces the mEPSCs frequency back to control level. E, bar graphs compare the mEPSCs amplitude between saline-, saline + LAP4, cocaine-, and cocaine + L-AP4. L-AP4 did not affect mEPSC amplitude in both treatment groups. The data presented are mean +/− SEM.

We also recorded mEPSCs properties in LHA cells 14d after final cocaine exposure (Figure 6C). mEPSCs frequency was significantly increased after 14d cocaine abstinence compared to saline (t₂₆=−3.796, P<0.01, cocaine vs saline, cell yield/animal = 2.8 vs 3.5, range = 8.0–49.38 vs 4.77–33.48 respectively, Figure 6D), whereas amplitude (P>0.05, Figure 6E) and decay kinetics (P>0.05) were unchanged. Bath application of L-AP4 significantly reduced mEPSCs frequency in 14d cocaine withdrawn animals (t₁₉=3.041, P<0.01, pre L-AP4 vs post L-AP4 treatment, cell yield/animal = 2.8 vs 2.3, range = 8.0–49.38 vs 14.98–20.87 respectively, Figure 6D), but had no effect on mEPSCs frequency in saline animals (P>0.05). L-AP4 had no effect on amplitude (P>0.05, Figure 6E) or decay kinetics in both groups (P>0.05).

4. DISCUSSION

We show that microinfusions of the group III mGluR agonist L-AP4 into the LHA orexin field suppressed food restriction-induced cell activity in this region, including in orexin cells. This treatment also reduced cocaine-seeking behavior following 14d withdrawal. Importantly, intra-LH L-AP4 had no apparent effects on general locomotor activity or self-administration of sucrose pellets. We confirm that L-AP4 acts via a presynaptic mechanism by showing that withdrawal-induced increases in mEPSC frequency are reversed by L-AP4 in recordings of identified orexin cells in mice. Together, these findings indicate that agonism of group III mGluRs can suppress presynaptic plasticity onto orexin cells following cocaine withdrawal, and identify the LHA group III mGluR system as a possible target to normalize orexin cell activity following cocaine.

4.1. Agonism of LHA group III mGluRs reduces orexin cell activity and reduces cocaine-seeking behavior

Agonism of group III mGluRs with L-AP4 has previously been shown to inhibit synaptic input onto orexin cells in slice (Acuna-Goycolea et al., 2004). Here, we extend these findings to show that infusions of L-AP4 directly into the LHA orexin cell field reduced activity of orexin cells following 24h food deprivation. We chose this stimulus based on previous studies that demonstrated food deprivation increases presynaptic input onto orexin cells, as measured by mEPSCs frequency and the number of vGlut2+ve inputs in close apposition to orexin cells, as well as increased Fos expression in orexin cells (Horvath and Gao, 2005; Mieda et al., 2004). We also observed a significant decrease in the activity of non-orexin cells following L-AP4 injections, although to a lesser extent than in orexin cells. LHA neurons produce several neuropeptides other than orexin that are important for feeding and energy homeostasis, including melanin-concentratin hormone (MCH) and cocaine- and amphetamine-regulated transcript (Bittencourt et al., 1992; Elias et al., 2001; Pissios et al., 2006; Vicentic and Jones, 2007). There are also significant populations of local GABA and glutamatergic neurons that gate the output of the major neuropeptide expressing neurons in LH, as well as other populations that project directly to the VTA (Faget et al., 2016; Sakurai et al., 2005; Sharpe et al., 2017). Indeed, based on the analysis of miniature postsynaptic current, Acuna-Goycolea and colleagues (Acuna-Goycolea et al., 2004) postulated that group III mGluRs are most likely located on both glutamatergic and GABAergic presynaptic axons that innervate orexin cells. Importantly, considering that all cells external to LH were severed before in vitro recording, it is likely that a substantial part of both excitatory presynaptic axons arise locally within the LH. Thus, it is probable that activation of GPIII mGluRs also suppressed activity in local interneuron populations, as well as other non-orexin LH projection neurons, and that this contributed to the reductions in orexin cell activity that we observed here.

Previous studies have shown that orexin cells are activated by cocaine-associated discriminative stimuli (Martin-Fardon et al., 2018) and that their signaling is critical for the expression of drug-seeking behavior elicited by these cues (Martin-Fardon and Weiss, 2014). Here, we show that infusions of L-AP4 directly into the LHA orexin cell field significantly reduced drug-seeking provoked by reintroduction of cocaine-paired discriminative-stimuli following 14d withdrawal. Although we did not directly assess the effect of L-AP4 on orexin cell activity in this test, our findings in food-deprived rats as well as previous electrophysiological evidence (Acuna-Goycolea et al., 2004), indicates that this behavioral effect involved suppression of orexin cell activity elicited by the drug-context. This finding aligns well with previous studies that have reported reduced stimulus-driven cocaine-seeking following abstinence or extinction in orexin knock-out mice (Steiner et al., 2018) or following either systemic or local pretreatment with an orexin-1 receptor antagonist (James et al., 2011; Mahler et al., 2013; Martin-Fardon and Weiss, 2014; Smith et al., 2010). While this is the first study to directly implicate LHA group III mGluRs in drug-seeking, these same receptors have been shown to regulate other forms of drug behavior in caudate. Intra-caudate microinjections of L-AP4 blocked hyperlocomotion induced by an acute systemic injection of cocaine, amphetamine or apomorphine, and this effect was blocked by co-administration of the group III antagonist, a-methyl-4-phosphonophenylglycine (Mao and Wang, 2000). Thus, group III mGluRs may act at various sites to regulate several glutamate-dependent drug behaviors, potentially increasing the utility of compounds targeting these receptors in blocking addiction-like behavior.

Importantly, L-AP4 had no effect on lever pressing for sucrose pellets or spontaneous locomotor activity in an open field. These findings indicate that L-AP4 did not affect animals’ ability to lever press or engage in general exploratory locomotor activity. This is consistent with previous studies that reported that blockade of orexin signaling did not affect low-effort reward seeking or general locomotor activity (Borgland et al., 2009; Espana et al., 2010; James et al., 2018b; Smith et al., 2009). Moreover, a previous study reported that bilateral intra-caudate microinjections of L-AP4 at a similar dose to that used here had no effect on basal locomotor activity (Mao and Wang, 2000). As such, the effects of L-AP4 on cocaine-seeking behavior are likely related to changes in motivation, rather than impaired motor behavior.

It is important to note that our test of sucrose seeking was a low-effort (FR1) task and that rats were not food restricted prior to testing. We and others have previously shown a role for the orexin system in operant responding for sucrose when animals are food restricted (Cason and Aston-Jones, 2014). Such data are consistent with our ‘motivational activation’ hypothesis of orexin system function, whereby the orexin system is preferentially engaged under circumstances where high levels of motivation are required to obtain reward, or when motivation for reward is augmented by external stimuli or stressors (James et al., 2017b; Mahler et al., 2014). Thus, it is likely that activation of group III mGluRs in LHA would reduce responding for sucrose in food restricted rats, or in a stimulus-driven reinstatement task such as the DS⁻/DS⁺ paradigm tested here for cocaine - especially given that craving for sucrose incubates following forced abstinence (Grimm et al., 2005).

It is also noteworthy that our study did not examine the effect of L-AP4 on drug-seeking behavior under DS⁻ conditions following abstinence. A study using the renewal model of reinstatement of cocaine seeking, where rats were trained to seek cocaine in ‘context A’ and were then extinguished in ‘context B’, showed that exposure to both contexts A (ABA) and B (ABB) induced greater Fos expression in orexin neurons compared to no-test controls, and that the magnitude of the Fos response was similar between ABA and ABB groups (Hamlin et al., 2008). This is despite drug-seeking being significantly higher under ABA conditions compared to ABB. One interpretation of these data is that orexin neurons mediate general (not necessarily drug-directed) motivated activity or exploration; although evidence from pharmacological studies and our control experiments (discussed above) would appear to rule out this interpretation. Alternatively, it is possible that context B (or in our case, the DS⁻ environment) maintains some of the context-cueing properties associated with context A (or the DS⁺ environment); thus it will be important that future studies test whether the effects of L-AP4 reported here are also observed under DS⁻ conditions.

Interestingly, intra-LHA microinjections of L-AP4 increased the amount of time rats spent in the centre square of the open field apparatus during locomotor testing, indicative of an anxiolytic effect. This observation aligns with several studies implicating orexin signaling in the expression of anxiety behavior; exogenous orexin administration of orexin peptides (Ida et al., 1999; Ida et al., 2000) or optogenetic stimulation of orexin cells (Bonnavion et al., 2015; Heydendael et al., 2013) promotes stress behaviors, whereas orexin receptor antagonists generally suppress stress- and anxiety-like behaviors (James et al., 2017a; Plaza-Zabala et al., 2010; Staples and Cornish, 2014; Vanderhaven et al., 2015). Some evidence indicates that the medial orexin cell populations (DMH and PF) specifically modulate anxiety and stress-behavior, whereas the lateral population plays a stronger role in regulating reward behavior (see (James et al., 2017b) for review). Here, injections of L-AP4 were made into the bulk of the orexin cell field (immediately dorsal to the fornix) and thus both medial and lateral orexin cell populations were affected, perhaps accounting for the effects of this manipulation on both reward and anxiety behaviors.

Rodents have been shown to exhibit elevated levels of anxiety 24hrs after both non-contingent and contingent cocaine exposure (Aujla et al., 2008; Basso et al., 1999; Harris and Aston-Jones, 1993). Interestingly, Aujla and colleagues (2008) reported that agonism of group 2 mGluRs - which have a similar function to the group 3 mGluRs studied here -reduced defensive burying in animals with escalated cocaine intake, indicating an anxiolytic effect. In our hands, it is plausible that the anxiolytic effects of L-AP4 underlie the attenuation of drug-seeking behavior; further studies are needed to confirm this hypothesis.

4.2. Group III mGluR agonism reverses presynaptic plasticity within LHA orexin circuits following cocaine withdrawal

We also sought to characterize, at the synaptic level, the mechanisms through which L-AP4 acts to suppress orexin cell activity. To do this, we used a well-validated mouse model in which GFP is driven by the human orexin promoter, thus allowing us to record from identified orexin cells in slice. We exposed mice to 7d of experimenter-administered cocaine, a regimen that we have previously shown to produce plasticity within LHA orexin circuit that is identical to that produced by cocaine self-administration (Yeoh et al., 2012). This regimen was associated with a prototypical sensitization of locomotor activity, thought to reflect neuroadaptive changes governing the development of drug-dependence (Cornish and Kalivas, 2001). We have previously reported enhanced excitatory input onto orexin cells 1d after cocaine sensitization or self-administration (Yeoh et al., 2012); here we extend these findings to show a significant paired-pulse depression at excitatory synapses, accompanied by an increase in mEPSC frequency, on orexin-GFP cells following 14d withdrawal from a cocaine sensitization regime. This finding suggests that the increased drive to orexin cells persists well after cessation of drug exposure, potentially contributing to the persistent risk of relapse observed in animals and humans. Notably, we did not observe a change in mEPSCs amplitude. This finding is surprising given our observation using electrically evoked ESPCs that cocaine exposure was associated with a significant increase in AMPA:NMDA ratio. mEPSC analysis effectively samples all excitatory inputs on orexin cells, providing an average of the excitatory drive from all afferent pathways (local and extrinsic). In contrast, electrically evoked EPSCs, assess a subset of excitatory inputs determined by the placement of our stimulating electrode, which was within the LHA (local). Thus, a more restricted LHA-based population of inputs may be potentiated by cocaine and revealed only after local stimulation. Future studies could address this issue using virally-mediated stimulation of specific afferents.

Our findings are largely consistent with a previous study that reported that repeated i.p. injections of cocaine, but not a single injection, increased the AMPA:NMDA ratio and was associated with enhanced mEPSCs amplitude in GFP-orexin cells (Rao et al., 2013). This study also showed that cocaine facilitated the induction of LTP by high frequency stimulation in orexin cells. Interestingly however, this study reported that these measures of plasticity persisted for only 5d; a significantly shorter period of time than that reported here (14d). While the reasons for these differences is unclear, it is possible that the higher dose of cocaine (15 versus 10mg/kg) and the duration of cocaine injections (7d versus 3d) used in the current study contributed to longer lasting effects reported here. Regardless, both studies point to significant rewiring of LHA orexin system following repeated cocaine exposure. It is possible that rewiring of LHA orexin neurons by drugs of abuse augment the extent to which orexin cells are recruited by other salient stimuli, such as stress or cues linked with drug-taking, even after significant periods of abstinence, and thus may contribute to the persistent relapsing nature of addiction.

Consistent with our in vivo data, we show that activation of group III mGluRs with L-AP4 significantly reduced mEPSCs frequency in orexin-GFP cells recorded from cocaine-exposed animals. L-AP4 had no effect on mEPSC amplitude, indicating a selective normalization of presynaptic plasticity in these animals. Interestingly, L-AP4 had no effect on either mEPSC frequency or amplitude in saline-treated animals. This indicates that cocaine produced a change in the expression and/or function of LHA mGluRs, as has been reported in other brain regions under different circumstances (Gordon and Bains, 2003). Future studies will need to explore this possibility by probing the upstream signaling pathways (e.g. PKC) responsible for regulating GPIII mGluR function. Furthermore, while not directly addressed here, preliminary whole cell recordings from our laboratory indicate that L-AP4 might also reduce excitability of non-orexin LHA neurons. As noted above, it will be important to determine the relative contribution of L-AP4-induced normalization of presynaptic input onto orexin versus non-orexin neurons to the behavioral phenomena that we report here.

Our finding that L-AP4 did not affect mEPSCs frequency in saline-treated rats contrasts with those of Acuna-Goycolea et al (2004), who reported a reduction in spontaneous glutamate-and GABA-mediated synaptic currents following bath applications of L-AP4 in drug-naïve animals. Although speculative, it is possible that repeated saline injections in our study resulted in a shift in baseline mEPSC frequency or mGluR function. It should also be highlighted that L-AP4 is thought to be more selective for mGluR4 or 8 receptors at the concentration used in our experiments, indicating that the group III mGluR responsible for these effects is likely through the activation of one or both of these ‘high-affinity’ subtype (Gereau and Swanson, 2008). Further studies are required to delineate the specific subtype involved.

It is important to note that our behavioral experiments were carried out in self-administering rats whereas the electrophysiological experiments were conducted in experimenter-exposed mice. The discriminative stimulus operant drug-seeking model used here has been characterized far more extensively in rat compared to mouse (Dayas et al., 2008; Martin-Fardon et al., 2018; Martin-Fardon and Weiss, 2013, 2014; McGlinchey and Aston-Jones, 2017); thus our behavioral data can be more readily compared with previous studies. In contrast, the orexin-GFP transgenic mouse line has been verified and characterized across several studies, whereas comparable studies in rat are limited (Sakurai et al., 2005; Yamanaka et al., 2003a; Yamanaka et al., 2003b). Therefore, some caution should be taken when considering how our electrophysiological findings relate to the reported behavioral data, and vice versa. However, we note that all published literature to date points to a large degree of overlap in the orexin system between rat and mouse in terms of topography (de Lecea et al., 1998; Sakurai et al., 1998; Sakurai et al., 1999; Stricker-Krongrad et al., 2002), projections and receptor distribution (Ch’ng and Lawrence, 2015; Chen et al., 1999; Lin et al., 2002; Marcus et al., 2001; Peyron et al., 1998; Puskás et al., 2010), electrophysiological properties (Yeoh et al., 2012) and role in reward-seeking behaviors (James et al., 2017b; Schmeichel et al., 2018; Steiner et al., 2018). Moreover, we have previously reported that plasticity at orexin cells occurs similarly between rats that self-administer cocaine and mice that receive experimenter-administered cocaine (as was the case here) on d1 of withdrawal (Yeoh et al., 2012). Nevertheless, future studies should confirm that the plasticity that we observed on withdrawal d14 in mice is similar in rats with a history of self-administration.

4.3. Conclusions

In conclusion, we show that agonism of LHA group III mGluRs in vivo suppresses activity of LHA neurons, including orexin cells, and that this is associated with reduced drug-seeking behaviour following 14d abstinence. We also show that L-AP4 reverses cocaine-associated increases in presynaptic plasticity at orexin cells in vitro. The exact source of increased excitatory drive to LHA circuits remains to be determined, although prior studies have identified the hypothalamus (local populations of excitatory interneurons) and prefrontal cortical as key inputs to LHA (Sakurai et al., 2005; Sharpe et al., 2017); it will be important for future studies to examine these inputs and others using circuit-specific manipulations. Together, our findings highlight group III mGluRs as a potential novel target for pharmacotherapies designed to suppress hyperactivity of orexin cells during abstinence and ameliorate relapse risk.

Highlights:

• Cocaine induces plasticity in the lateral hypothalamic area (LHA)

• Group III mGluRs within the LHA may be a target to suppress this plasticity

• LHA L-AP4 (Group III mGluR agonist) reduced cocaine-seeking after 14d abstinence

• Locomotor behavior/responding for sucrose was unaffected by LHA L-AP4

• L-AP4 suppressed excitatory drive to orexin cells associated with cocaine in slices

Abbreviations:

DMH

dorsomedial hypothalamus

PF

Perifornical hypothalamus

LHA

lateral hypothalamic/perifornical area

L-AP4

L-(+)-2-Amino-4-phosphonobutyric acid

mGluR

metabotropic glutamate receptor