Mu Opioid Receptors in Gamma-Aminobutyric Acidergic Forebrain Neurons Moderate Motivation for Heroin and Palatable Food

Abstract

Background: Mu opioid receptors (MORs) are central to pain control, drug reward, and addictive behaviors, but underlying circuit mechanisms have been poorly explored by genetic approaches. Here we investigate the contribution of MORs expressed in gamma-aminobutyric acidergic forebrain neurons to major biological effects of opiates, and also challenge the canonical disinhibition model of opiate reward.

Methods: We used Dlx5/6-mediated recombination to create conditional Oprm1 mice in gamma-aminobutyric acidergic forebrain neurons. We characterized the genetic deletion by histology, electrophysiology, and microdialysis; probed neuronal activation by c-Fos immunohistochemistry and resting-state functional magnetic resonance imaging; and investigated main behavioral responses to opiates, including motivation to obtain heroin and palatable food.

Results: Mutant mice showed MOR transcript deletion mainly in the striatum. In the ventral tegmental area, local MOR activity was intact, and reduced activity was only observed at the level of striatonigral afferents. Heroin-induced neuronal activation was modified at both sites, and whole-brain functional networks were altered in live animals. Morphine analgesia was not altered, and neither was physical dependence to chronic morphine. In contrast, locomotor effects of heroin were abolished, and heroin-induced catalepsy was increased. Place preference to heroin was not modified, but remarkably, motivation to obtain heroin and palatable food was enhanced in operant self-administration procedures.

Conclusions: Our study reveals dissociable MOR functions across mesocorticolimbic networks. Thus, beyond a well-established role in reward processing, operating at the level of local ventral tegmental area neurons, MORs also moderate motivation for appetitive stimuli within forebrain circuits that drive motivated behaviors.

Keywords: Conditional gene knockout; Dopamine; Motivation; Mu opioid receptor; Opiate; Reward.

Figure 1

Dlx-MOR mice lack MORs mainly in striatal MSNs. (A) Quantification of triple in situ hybridization in mouse striatal sections (details in Supplemental Figure S1) shows major Oprm1 mRNA co-localization with Drd1, Drd2 or both in NAc and DS. (B) Quantitative RT-PCR analysis of Oprm1 mRNA in microdissected brain regions from Ctl, conditional knockout Dlx-MOR and total knockout CMV-MOR mice reveals complete Oprm1 deletion specifically in DS and NAc of Dlx-MOR mice. Transcript levels in mutants are expressed relative to Ctl (dotted line, n=3 to 4, in triplicates) and shown in rostrocaudal order. (C) Quantitative RT-PCR analysis of Oprd1, Oprk1, Penk and pDyn mRNAs in NAc (above) and DS (bellow) shows no change in Dlx-MOR mice (n=3 to 4, in triplicates). (D) Representative [³H]DAMGO (MOR agonist) binding autoradiograms of brain sections from Ctl and Dlx-MOR mice reveals complete receptor deletion at the level of striatum (bregma 1.1) but not in a more posterior section (Hb, bregma −1.46). Color bar, relative densities in fmol/mg tissue with non-specific binding at background level. (E) Quantification of [³H]DAMGO binding in areas with sufficiently high signal (≥60 fmoles/mg tissue) from the forebrain to spinal cord (n=3 to 4) confirms significant MOR decrease in the NAc and VP. MOR density is also lower in DS and VTA, although this is not significant. (F, G) Slice VTA electrophysiology (n=6 to 12). (F) In GABA interneurons (yellow), DAMGO (1μM) decreased evoked GABAA IPSCs (eIPSCs) in slices from Ctl but not Dlx-MOR mice, while N₆-CPA (A1 receptor agonist, 1μM) reduced eIPSCs in the two groups. (G) In dopamine neurons (blue), DAMGO (1μM) decreased GABAA eIPSCs similarly for Ctl and Dlx-MOR, while N₆-CPA had no significant effect. (H) Schematic representation of the MOR deletion within NAc/VTA circuitry of Dlx-MOR mice, as indicated by RNA, protein and electrophysiological analysis. Data are represented as mean + SEM. Blue star, significant genotype effect; one star, p<0.05; three stars, p<0.001. Serpentine, MOR. Amy, Amygdala; DS, dorsal striatum; DRN, dorsal raphe nucleus; EP, endopiriform nucleus; Hb, habenula (MHb, medial); LH, lateral hypothalamus; NAc, nucleus accumbens (NAcc, core; NAcsh, shell); PAG, periaqueductal gray; PFC, prefrontal cortex; RMTg, rostromedial tegmental area; SC, spinal cord (LI/II, lamina I/II); SN, substantia nigra; Th, thalamus (ThCL, central lateral; ThCM, central medial; ThIMD, intermediodorsal); VP, ventral pallidum; VTA, ventral tegmental area.

Figure 2

Neuronal activity is modified in Dlx-MOR mice. (A, B) Heroin-induced neuronal activation. c-Fos immunohistochemistry (IHC) was performed on brain sections from animals perfused 2h after saline or heroin (10 mg/kg, ip) administration. (A) Representative images illustrate c-Fos IHC in the NAcsh, DLS and VTA of Ctl and mutant mice. (B) Quantification of c-Fos IHC, expressed as c-Fos-positive cells/mm² (bilateral, 4–7 sections/animal, n=5 to 6), shows a treatment effect for Ctl mice in all three regions. For Dlx-MOR mice, heroin-induced increase of c-Fos IHC was detected in VTA but not in NAcsh and DLS, indicating strongest modification in the striatum. (C, D, E) Whole-brain functional connectivity in live mice using resting-state functional MRI. High resolution spatial independent component analysis followed by partial correlation in time domain and graph theory identifies functional connectivity (FC) hubs displaying above-mean normalized connectivity strength and diversity, which differ across genotypes. (C) Focus on NAc-VTA FC shows two hubs covering NAc (left) detectable in Ctl mice (normalized strength>0.041 and/or diversity>0.966) but not Dlx-MOR mice (normalized strength<0.047 and/or diversity<0.940) and, conversely, two hubs with anatomical correspondence to VTA (right) detected in Dlx-MOR mice (normalized strength>0.047 and/or diversity>0.940) but not Ctl mice (normalized strength<0.041 and/or diversity<0.966). Thus, NAc nodes lose, while VTA nodes gain their hub status, indicating modified relay function between these two brain nodes. (n=10 to 12). (D, E) Seed-based Granger causality analysis considering 6 MOR-enriched regions within reward circuitry shows distinct information flow in Ctl and Dlx-MOR mice (see full analysis in Supplemental Figure S2). (D) Identification of functional dominant or bi-directionality across NAc, DS and VTA is represented in matrices for Ctl and Dlx-mice. Brain regions on the horizontal axis are sources, and regions on the vertical axis are destination of the information flow and the number of subjects showing a significant directionality (by 100 permutations) is indicated. Mean conditional Granger causality (C-GC) values represent intensity of information flow for each connection (gradient grey scale). The dominant directionality is determined by comparing both number of significant subjects and mean C-GC for each connection (see Methods). (E) Scheme summarizing dominant or bi-directionality in Ctl and Dlx-MOR mice for the 3 seed regions of interest in this study (orange arrows) shows that VTA-DS connectivity is reversed in mutants, consistent with abolished heroin-induced locomotor activation and altered heroin SA, whereas VTA-NAc directional connectivity is unchanged concordant with maintained heroin CPP and heroin-induced DA release in the NAc (see Figure 4). Data are represented as mean + SEM. Black star, significant treatment effect; blue star, significant genotype effect. Two stars, p<0.01; three stars, p<0.001. DS, dorsal striatum; DLS, dorsolateral striatum; Hb, habenula; NAcsh, nucleus accumbens shell; PFC, prefrontal cortex; PAG, periaqueductal gray; VTA, ventral tegmental area. Scale bar = 100 μm.

Figure 3

Dlx-MOR mice show modified motor responses to heroin. Morphine analgesia was investigated in tail immersion [52°C (A)] test, tail flick (B) and hot plate (C) tests. (A, B) Acute thermal nociception in spinal reflex responses is unchanged in mutant mice (cumulative dose-response, ip injections, n=12) (C) Also, the three parameters measured in the hot plate test upon ip morphine administration (2 or 5 mg/kg, n=5 to 9) reveal treatment but no genotype effects. (D) Physical dependence was induced by repeated injections of ascending doses of morphine (10–100 mg/kg, ip, twice daily, 6 days). Scoring of naloxone-precipitated (1 mg/kg, sc) withdrawal signs shows no genotype difference in the expression of morphine physical dependence (n=9 to 15) (global score) (details in Supplemental Figure S3). (E) Dose-dependent heroin-induced locomotor activation in Ctl but not Dlx-MOR mice is shown (mg/kg ip, 2h-session, n=4 to 32). (F) Sensitization to the heroin locomotor effect (10 mg/kg ip) develops in Ctl mice only (5X 2h-sessions, n=12 to 32). (G) Heroin-induced catalepsy, expressed as latency to paw withdrawal in the bar test, develops in both Ctl and Dlx-MOR mice, and is enhanced in Dlx-MOR mice at 10 mg/kg (n=8 to 27). Data are represented as mean + or − SEM. Symbols: (A, B) black stars, significant difference to baseline; (C, D, E, G) black stars, significant difference to saline treatment; (E, G) blue star, significant genotype effect; (F) black stars, significant difference to all other groups; # significant time effect session 1 vs session 5 for Ctl 10. One symbol, p<0.05; two symbols, p<0.01; three symbols, p<0.001.

Figure 4

Dlx-MOR mice show enhanced motivation for heroin and palatable food. (A) CPP to heroin (mg/kg sc, n=4 to 21), established over 6 conditioning sessions, does not differ between Ctl and Dlx-MOR mice (significant session X treatment interaction). Preference is expressed as % time spent in the drug-paired compartment during pre- (left) and post- (right) conditioning sessions. (B) DA microdialysis and concurrent locomotor activation. Heroin (10 mg/kg ip, injected after 120 min, recording during 320 min) increases extracellular DA above baseline in the NAc (Supplemental Figure S4) of Dlx-MOR and Ctl mice, with no genotype effect (n=9 to 13) (left). AUC ([DA] ng/mL, inset) analysis confirms no difference between Dlx-MOR and Ctl. Simultaneous to DA increase, heroin stimulates locomotion in Ctl but not Dlx-MOR mice (right). (C–E) Operant responding to heroin (n=11 to 18). (C) Both Dlx-MOR and Ctl mice acquire and increase heroin SA with decreasing doses (20 successive 1h-daily sessions, FR1 schedule of reinforcement). SA levels (mean number of infusions/session) are higher in the Dlx-MOR group. Heroin SA is significantly higher at 2 doses (dose-response analysis, inset). (D) Dlx-MOR mice achieve a higher breaking point in a 3h-PR session (0.0125 mg/kg/inf). (E) After extinction (1h-daily sessions for 13 days), Dlx-MOR but not Ctl mice reinstate heroin SA (0.006 mg/kg/inf) upon cue presentation (CIR). (F, G) Operant responding to chocolate-flavored pellets (n=20 Ctl, 13 Dlx-MOR). (F) Acquisition and maintenance of active nose-poking for chocolate reinforcement is shown (mean number of nose-pokes during 10 days FR1 and 5 days FR5 schedule of reinforcement, 1h-daily sessions). In both FR1 and FR5 sessions, Dlx-MOR mice show higher operant responding than Ctl mice. (G) Breaking-point for chocolate-flavored pellets is higher in mutant mice (5h PR session). Data are represented as mean + or - SEM. Black star, significant treatment effect; blue star, significant genotype effect. #, significant time effect. One symbol, p<0.05; two symbols, p<0.01; three symbols, p<0.001. AUC, area under the curve; CIR, cue-induced reinstatement; FR, fixed ratio; PR, progressive ratio. See Tables 1 and 2 for statistical analysis of SA experiments.