INTRODUCTION

Increasing evidence supports the view that persistent alterations in gene expression and excitatory synaptic plasticity in nucleus accumbens (NAc), a main component of the mesolimbic dopamine reward circuit, have an important role in mediating aspects of drug addiction (Kauer and Malenka, 2007; Nestler, 2001). The major neuronal type in the NAc is the GABAergic medium spiny neuron (MSN), which comprises 95% of all neurons in this brain region (Gerfen, 1988; Graybiel, 2005; Shi and Rayport, 1994). As the activity of these GABAergic MSNs is regulated not only by dopaminergic and glutamatergic inputs (Gerfen and Surmeier, 2011) but also by local GABAergic interneurons and collateral projections between neighboring MSNs (Stuber et al, 2012), it is likely that altered GABAergic function in the NAc has an important role in the neural and behavioral plasticity induced by drugs of abuse (Xi and Stein, 2002).

Brain-derived neurotrophic factor-tyrosine receptor kinase B (BDNF-TrkB) signaling in NAc is also important in the actions of drugs of abuse (Graham et al, 2007; Graham et al, 2009; Horger et al, 1999; Pickens et al, 2011; Russo et al, 2009). Interestingly, BDNF-TrkB signaling modulates the function of GABAA receptors (GABAARs) in other brain regions. For example, it increases cell-surface GABAAR expression in the hippocampus (Jovanovic et al, 2004) and promotes construction of inhibitory GABAergic networks in the hippocampus and cerebellum (Huang et al, 1999; Rico et al, 2002). In addition, BDNF elevates cellular GABA content in cultured striatal neurons and in the striatum of neonatal rats (Mizuno et al, 1994). These observations raise the possibility that altered GABA function in NAc might link BDNF-TrkB signaling to its potent effects on behavioral phenotypes associated with drug addiction.

Recent work demonstrates that cell type-specific manipulations of BDNF-TrkB signaling, and downstream effects on neuronal activity, in the two subpopulations of NAc MSNs, those expressing dopamine D1 (Drd1a) vs D2 (Drd2) receptors, produce opposite behavioral responses to cocaine (Lobo et al, 2010). Although a growing body of evidence demonstrates some shared and some distinct molecular, cellular, and circuit mechanisms between psychostimulant and opiate action (Albertson et al, 2006; Badiani et al, 2011; Chang et al, 1998; Koo et al, 2012; Nestler, 2005), the contribution of D1-type vs D2-type NAc MSNs to morphine reward remains unknown.

In the present study, we show that cell type-specific optogenetic stimulation of D1-type vs D2-type NAc MSNs produces opposite effects on morphine reward, as shown previously for cocaine. In addition, selective deletion of TrkB from D1-type MSNs enhances morphine reward, whereas TrkB deletion from D2-type MSNs is without effect. D1-TrkB KO (knockout of TrkB from D1-type MSNs) mice also display decreased GABAAR subunit expression and reduced spontaneous inhibitory postsynaptic currents (sIPSCs) in D1-type MSNs. In contrast, morphine reward and sIPSCs are not significantly altered in D2-TrkB KO mice. We demonstrate further that chronic morphine exposure decreases mRNA expression of TrkB and GABAAR subunits as well as sIPSCs in D1-type MSNs. Together, our findings support a scheme whereby a reduction in inhibitory GABAAR responses, mediated by morphine-induced reduction in BDNF-TrkB signaling in D1-type MSNs, enhances morphine reward.

MATERIALS AND METHODS

Mice

Drd1a (D1)-Cre (line FK150) and Drd2 (D2)-Cre (line ER44) bacterial artificial chromosome (BAC) transgenic mice on a C57BL/6J background were obtained from N. Heintz and P. Greengard (Rockefeller), C. Gerfen (NIMH), and NINDS/GENSAT (www.gensat.org) (Gong et al, 2007; Gong et al, 2003). D1-tdTomato (TMT) mice (line 5) on a C57BL/6J background were obtained from N. Calakos (Duke; Shuen et al, 2008). D1-TrkB KO and D2-TrkB KO mice on a C57BL/6J background were generated by crossing D1-Cre or D2-Cre mice to floxed TrkB (flTrkB) mice (Luikart et al, 2003). Two- to three-month-old male mice were used in all experiments. Mice were fed ad libitum and maintained at 2225 °C on a 12-h light/dark cycle. All experiments were performed in accordance with the Animal Care and Use Committee of the Icahn School of Medicine at Mount Sinai.

Stereotaxic Surgery, Virus and Drug Infusion, and Optogenetic Stimulation

Thirty-three gauge needles were used to unilaterally infuse 0.5 μl of adeno-associated virus (AAV) vectors (ie, DIO (double loxP-flanked inverted)-AAV-ChR2 (channel rhodopsin)-EYFP (enhanced yellow fluorescent protein) or DIO-AAV-EYFP) into the NAc of the right hemisphere (AP +1.5; ML +1.5; DV −4.4 from bregma; 10° angle) at a rate of 0.1 μl/min. Three weeks later, 20-gauge guide cannulae (4 mm length) were implanted into the right hemisphere (AP +1.5; ML +1.3; DV –3.9; 0° angle). Mice were given 1-week recovery before conditioned place preference (CPP) experiments. For in vivo optical stimulation, a modified 200-μm core optic fiber (Thor labs, with 50 μm of the stripped optic fiber exposed below the cannula) were attached through an FC/PC adaptor to a 473-nm blue laser diode (BCL-473-050-M, Crystal Lasers), and light pulses were generated through a stimulator (33220A, Agilent; Koo et al, 2012; Lobo et al, 2010). Ten Hertz blue light pulses (100 ms square width) were delivered for four 3-min periods (one period consists of 3 min on and 5 min off) during the 30-min conditioning. Optic fiber light intensity ranged from 15.9 to 27.8 mW/mm2 at the tip and similarly through experimental groups. For bicuculline (Tocris) infusions, bilateral guide cannulae (26-gauge, 5 mm length) were implanted above the NAc (AP +1.5; ML +0.75; DV –3.9; 0° angle). Bicuculline (5 ng per 0.5 μl/side) was delivered through an injector needle at a rate of 0.1 μl/min. After the 5-min infusion, injector needles remained in place for 45 min until each morphine CPP training session started.

Morphine CPP

An unbiased CPP paradigm was used as previously described (Koo et al, 2012; Lobo et al, 2010). Mice were evaluated for morphine or morphine/blue light CPP in a three-chamber CPP box using the Med Associates software. The box consisted of a smaller middle chamber (12 (W) × 15 (L) × 33 (H) cm) and two conditioning larger chambers (24.5 (W) × 15 (L) × 33 (H) cm) with different contextual clues including a gray verses stripe chamber, different floor mesh, and different lighting. On day 1, mice were placed in the CPP box for 20 min to ensure no chamber bias. For standard morphine CPP, mice were conditioned to one chamber for 45 min in the morning (saline) and to the opposite chamber in the afternoon (15 mg/kg morphine, sc or ip) over 3 days. The dose was chosen based on a dose–response analysis of morphine CPP in a previous study (Koo et al, 2012). On day 5, the CPP test day, mice were allowed to freely explore all three chambers for 20 min. For morphine/light CPP, optic fibers were secured to the cannula before saline or morphine injections. Mice were conditioned to saline/no light and morphine/blue light for 30 min over 2 days. We used a subthreshold morphine dose (5 mg/kg) where our hypothesis was to test for an enhancement of CPP, and higher doses (10 mg/kg) where our hypothesis was to test for an attenuation of CPP. For all experiments, CPP scores represent time spent in the paired−time spent in the unpaired chamber.

Immunohistochemistry

Immunoreactivity of EYFP and Cre was detected as described previously (Koo et al, 2012). Thirty-five-micrometer free-floating brain sections were incubated in 1 : 2000 of rabbit polyclonal anti-GFP (A11122, Invitrogen) and/or 1 : 1000 mouse anti-Cre recombinase (MAB3120, Millipore) in block solution overnight at 4 °C. The next day, sections were incubated in 1 : 500 of donkey anti-rabbit Cy2 (Immuno Research) and 1 : 500 of donkey anti-mouse Cy3 in PBS for 1 h. All sections were mounted with antifade solution and subsequently imaged on a LSM 710 confocal microscope.

RNA Extraction and qRT-PCR

For RNA isolation, 14-gauge bilateral NAc punches were processed according to the published protocol (Koo et al, 2012). Primers were designed to amplify regions of 100–250 base pairs located within the gene (Table 1). SYBR Green qRT-PCR was run in triplicate and analyzed using the ΔΔCt method as previously described (Livak and Schmittgen, 2001) with GAPDH as a normalization control. GAPDH mRNA expression was not altered among the groups in all experiments.

Table 1 Primers for Real-Time PCR

Fluorescence Activated Cell Sorting (FACS)

Twelve-gauge bilateral NAc punches from TMT mice were dissociated to single cells using papain (Worthington). A pure population of D1-tdTomato+ MSNs was obtained via FACS according to a previously published protocol with minimal modifications (Lobo et al, 2006). RNA was extracted using a PicoPure RNA Isolation Kit (Life Technologies) including a DNase I digestion according to the manufacturer’s protocol. RNA quality and relative concentration were confirmed on the Picochip using Agilent 2100 Bioanalyzer. qRT-PCR was run as described above with 18S rRNA as a normalization control. 18s rRNA mRNA expression was not altered among the groups.

Electrophysiological Recordings

Whole-cell voltage-clamp recordings were obtained from EYFP+ MSNs in the NAc shell of acute brain slices from D1-Cre, D2-Cre, D1-Cre-flTrkB, and D2-Cre-flTrkB mice that had been stereotaxically injected into the NAc with DIO-AAV-EYFP, or from tdTomato+ MSNs in the NAc shell of acute brain slices from TMT mice as previously described (Lobo et al, 2010). Briefly, brain slices were maintained in the holding chamber for 1 h at 37 °C. sIPSCs or sEPSCs were recorded in voltage clamp mode (Bessel filtered at 4 kHz) either in the presence of the AMPA/NMDA antagonist (kynurenic acid; 2 mM) or GABAAR antagonist (picrotoxin; 100 μM), respectively, at 34 °C. Cells were held at −70 mV. Voltage clamp recordings were performed using the amplifier Multiclamp 700B and data acquisition was realized with pClamp 10. Series resistance was monitored during experiments. Responses were obtained from two to six neurons from each animal.

Statistical Analysis

Student’s t-tests were used for the analysis of experiments with two experimental groups. One-way ANOVAs were used for the analysis of three or more groups, followed by Fisher’s PLSD post hoc tests, when appropriate. For CPP, electrophysiology, and qRT-PCR data using D1/D2-Cre and/or D1/D2-Cre-flTrkB mice, two-way ANOVAs were used followed by Fisher’s PLSD post hoc tests, as appropriate. Main and interaction effects were considered significant at p<0.05. All data are expressed as mean±SEM.

RESULTS

To examine the effect of cell type-specific optical stimulation of NAc neurons on the rewarding effects of morphine, we injected DIO-AAV-ChR2-EYFP into the NAc of D1-Cre or D2-Cre BAC transgenic mice, in which Cre recombinase expression is driven by the D1 or D2 promoter and their regulatory elements (Gong et al, 2007) (Figure 1a–c). The activation of D1-type MSNs by use of established stimulating protocols (Chandra et al, 2013; Lobo et al, 2010) enhanced CPP to a subthreshold dose of morphine (5 mg/kg), with opposite effects induced by activation of D2-type MSNs with a higher morphine dose (10 mg/kg; Figure 1d). The ability of light activation of D1-type MSNs to increase morphine CPP was non-additive with the effect of morphine itself: such light activation did not enhance morphine CPP to maximal morphine doses (15 mg/kg; data not shown). The opposite effects of optogenetic stimulation of D1-type vs D2-type MSNs on morphine reward are similar to those seen for cocaine reward (Lobo et al, 2010). We then assessed the functional role of BDNF-TrkB signaling in D1- and D2-type MSNs in morphine reward, using D1-Cre- or D2-Cre-flTrkB mice, where TrkB is selectively deleted from each MSN subtype (Lobo et al, 2010). Selective deletion of TrkB from D1-type MSNs (D1-TrkB KO) in NAc and dorsal striatum increased morphine CPP, whereas D2-TrkB KO mice showed no effect (Figure 1e).

Figure 1
figure 1

Cell type-specific activation, or selective deletion of tyrosine receptor kinase B (TrkB), from D1- or D2-type medium spiny neurons (MSNs) alters morphine conditioned place preference (CPP). (a) Schematic coronal section from nucleus accumbens (NAc) with an inset depicting localized DIO (double loxP-flanked inverted)-AAV (adeno-associated virus)-mediated ChR2-EYFP (channel rhodopsin-enhanced yellow fluorescent protein; green) and Cre (red) in NAc. Scale bar, 20 μm. (b) Schematic diagram depicting the experimental procedures for optical stimulation of D1- or D2-type MSNs in morphine CPP. (c) DIO-AAV-ChR2-EYFP carries an inverted version of ChR2 fused to the fluorescent marker EYFP. In the presence of Cre, ChR2-EYFP is inverted into the sense direction and expressed from the EF-1α (elongation factor 1α) promoter. (d) Activation of D1-type MSNs in D1-Cre DIO-AAV-ChR2-EYFP mice enhances morphine reward compared with D1-Cre DIO-AAV-EYFP controls (5 mg/kg, subcutaneous (sc)). In contrast, activating D2-type MSNs in D2-Cre DIO-AAV-ChR2-EYFP mice decreases morphine reward compared with D2-Cre DIO-AAV-EYFP controls (10 mg/kg, sc; two-way ANOVA, stimulation × cell-type effect: F1,24=11.09, p<0.01; stimulation effect: F1,24=0.2051, p=n.s.; cell-type effect: F1,24=2.551, p=n.s.; n=5–9). *p<0.05, compared with each control group in the Fisher’s PLSD post hoc test. (e) Loss of TrkB from D1-type MSNs in D1-TrkB KO (knockout of TrkB from D1-type MSNs) mice enhances morphine reward (15 mg/kg, sc), whereas the loss of TrkB from D2-type MSNs in D2-TrkB KO mice has no effect (15 mg/kg, sc; two-way ANOVA, genotype × cell-type effect: F1,29=1.896, p=n.s.; genotype effect: F1,29=4.386, p<0.05; cell-type effect: F1,29=3.608, p=0.068; n=7–10).

PowerPoint slide

To gain insight into a possible functional relationship between deletion of BDNF-TrkB signaling and optogenetic activation in the MSN subtypes, we measured mRNA expression levels of GABAAR subunits in NAc of D1- and D2-TrkB KO mice. D1-TrkB KO reduces mRNA levels of several GABAAR subunits including Gabra4, Gabra5, Gabrb1, Gabrg1, and Gabrg3, but the Gabra2 subunit was increased (Figure 2a). In contrast, we found a reduction only in Gabra4, with an increase in Gabra2, Gabrg2, and Gabrg3, in the NAc of D2-TrkB KO mice (Figure 2a). To determine whether altered GABAAR subunit expression reflects functional changes in MSNs, we investigated the sIPSCs and the sEPSCs of these MSN subtypes in NAc shell after the cell type-specific deletion of TrkB. Consistent with GABAAR expression patterns, there was a significant reduction in the amplitude of sIPSCs of D1-TrkB KO mice, but not in D2-TrkB KO mice, compared with control mice (Figure 2b and c). No alterations in the frequency of sIPSCs were found either in D1-TrkB KO or in D2-TrkB KO mice (Figure 2b and d). There were also no alterations in the amplitude or frequency of sEPSCs either in D1-TrkB KO or in D2-TrkB KO mice (Figure 2e and f).

Figure 2
figure 2

Effects of cell type-specific tyrosine receptor kinase B (TrkB) deletion on GABAA receptor (GABAAR) subunit gene expression and spontaneous inhibitory postsynaptic currents (sIPSCs) in nucleus accumbens (NAc). (a) Selective deletion of tyrosine receptor kinase B (TrkB) from D1- or D2-type medium spiny neurons (MSNs) alters mRNA expression of GABAAR subunits. Notably, knockout of TrkB from D1-type MSNs (D1-TrkB KO) reduces mRNA expression of many GABAAR subunits including Gabra4, Gabra5, Gabrb1, Gabrg1, and Gabrg3 (two-way ANOVA, genotype × cell-type effect: F23,329=4.318, p<0.001; genotype effect: F1,329=4.030, p<0.05; cell-type effect: F23,329=4.318, p<0.001, n=6–9). τp<0.1, *p<0.05, **p<0.01, ***p<0.001, compared with each control; #p<0.05, ##p<0.01, ###p<0.001, compared with D1-Cre-flTrkB (floxed TrkB) mice in the Fisher’s PLSD post hoc test. (b) Raw sIPSC traces recorded from D1- and D2-type MSNs of D1-/D2-TrkB KO and D1-/D2-Cre control mice that were injected with DIO-AAV-EYFP (double loxP-flanked inverted-adeno-associated virus-enhanced yellow fluorescent protein) into the NAc for visualization of D1- and D2-type MSNs. (c) D1-TrkB KOs exhibit decreased amplitude of sIPSCs compared with control mice. In contrast, there was no change in the amplitude of sIPSCs in D2-TrkB KO mice (two-way ANOVA, genotype × cell-type effect: F1,28=0.4384, p=n.s.; genotype effect: F1,28=4.661, p<0.05; cell-type effect: F1,28=0.2496, p=n.s.; n=4–13 animals per group). *p<0.05, compared with each control group in the Fisher’s PLSD post hoc test. (d) There was a significant main effect of cell type, but not a TrkB KO effect on the frequency of sIPSCs (genotype × cell-type effect: F1,28=0.6903, p=n.s.; genotype effect: F1,28=2.561, p=n.s.; cell-type effect: F1,28=35.13, p<0.001). (e and f) There were no changes in the amplitude (e, genotype × cell-type effect: F1,21=2.637, p=n.s.; genotype effect: F1,21=0.4618, p=n.s.; cell-type effect: F1,21=1.203, p=n.s., n=5–8 animals per group) or in the frequency (f, genotype × cell-type effect: F1,21=0.5902, p=n.s.; genotype effect: F1,21=2.113, p=n.s.; cell-type effect: F1,21=0.1181, p=n.s.) of sEPSCs in D1- or D2-TrkB KO mice.

PowerPoint slide

We next investigated the effects of chronic exposure to morphine on GABAAR subunit expression in whole NAc extracts. We found that repeated intraperitoneal (ip) injections of morphine (20 mg/kg, 7 days) decrease the mRNA expression of several GABAAR subunits—Gabra2 and Gabrd—in the NAc of C57BL/6J wild-type (WT) mice (Figure 3a). We then examined the role of GABAA signaling in NAc in morphine reward by bilaterally infusing bicuculline (5 ng/side), a competitive GABAAR antagonist, into NAc before each training session of morphine CPP with a maximal dose of morphine (15 mg/kg). Figure 3b shows that the inhibition of GABAA signaling selectively in NAc dramatically enhances morphine reward.

Figure 3
figure 3

Effects of chronic morphine on GABAA receptor (GABAAR) subunit gene expression and spontaneous inhibitory postsynaptic currents (sIPSCs) in D1-type medium spiny neurons (MSNs). (a) Chronic morphine (20 mg/kg, ip, 7 days) downregulates mRNA expression of certain GABAAR subunits in whole nucleus accumbens (NAc) extracts, such as Gabra2 and Gabrd (one-way ANOVA, F23,228=3.037, p<0.001, n=8–12). **p<0.01, ***p<0.001, compared with each control in the Fisher’s PLSD post hoc test. (b) Intra-NAc infusion of bicuculline (10 ng), a competitive GABAAR antagonist, dramatically blocks morphine reward (15 mg/kg, ip). Student t-test, *p<0.05, n=9. (c) Typical scatterplots for fluorescence activated cell sorting (FACS) from wild-type (WT) and D1-tdTomato mice. The data from WT mice were used to gate the tdTomato (TMT+) signals. (d) Purified D1-type TMT+ cells display enrichment of prodynorphin (Pdyn) mRNA compared with TMT-negative controls (TMT−). t-test, **p<0.01, n=4–5. (e) Chronic morphine (implanted sc with one pellet, 2 days) reduces mRNA expression of several GABAAR subunits including Gabra2, Gabra3, Gabra4, Gabra5, and Gabrb1 in purified D1-type MSNs, compared with controls (implanted sc with sham pellet; one-way ANOVA, F17,68=2.713, p<0.01, n=4–5). τp<0.1, *p<0.05, **p<0.01, compared with each control in the Fisher’s PLSD post hoc test. (f) Chronic morphine also robustly reduces TrkB (tyrosine receptor kinase B) mRNA expression in purified D1-type MSNs. t-test, **p<0.01, n=4–5. (g) Raw sIPSC traces recorded from TMT+ D1-type MSNs in drug naive vs morphine-treated mice. (h and i) Chronic morphine significantly decreases amplitude (h), but not frequency (i), of sIPSCs in D1-type MSNs. t-test, **p<0.01, n=6 animals per group.

PowerPoint slide

Finally, we studied the effects of chronic morphine on inhibitory GABAA signaling in D1-type MSNs selectively, given the direct connection between the loss of BDNF-TrkB signaling and reduction of inhibitory GABAA signaling in D1-type MSNs with respect to morphine reward. D1-type MSNs were isolated and purified using FACS from NAc of TMT mice (Figure 3c). The FACS-purified D1-type TMT+ cells display enrichment of prodynorphin (Pdyn, a marker of D1-type neurons in NAc), compared with TMT-controls (Figure 3d). Consistent with GABAAR expression patterns in whole NAc extracts from WT mice, chronic morphine reduces mRNA expression levels of several GABAAR subunits including Gabra2, Gabra3, Gabra4, Gabra5, and Gabrb1 in purified D1-type MSNs (Figure 3e). Gabra1, Gabrb2, and Gabrg2 were not detected in purified D1-type MSNs. In addition, chronic morphine robustly reduced full-length TrkB mRNA expression in the D1-type MSNs (Figure 3f). Consistent with GABAAR expression patterns in D1-type MSNs, the amplitude, but not the frequency, of sIPSCs in D1-type MSNs was significantly reduced by chronic morphine (Figure 3g–i).

DISCUSSION

It has been hypothesized that opioid receptor-mediated inhibition of NAc GABAergic MSNs is critical for opiate reward (Xi and Stein, 2002). Consistent with this hypothesis, self- or local administration of morphine or heroin into the NAc significantly suppresses the spontaneous firing of NAc MSNs (Hakan and Henriksen, 1989; Lee et al, 1999). In the present study, we find that chronic morphine administration, or its resulting suppression of BDNF-TrkB signaling in D1-type MSNs, decreases expression of several GABAAR subunits in this neuronal cell type of the NAc and that the concomitant reduction of inhibitory GABAergic tone in NAc D1-type MSNs promotes morphine reward.

Our observation that chronic morphine suppresses GABAAR subunit expression in NAc is opposite to prior observations with psychostimulants. For example, we showed recently that chronic cocaine increases the expression of several GABAAR subunits, including Gabra1, Gabra2, and Gabrb2 in NAc and that GABAAR antagonism blocks behavioral plasticity to cocaine (Kennedy et al, 2013), findings consistent with earlier studies (Dixon et al, 2010; Heiman et al, 2008). In contrast, we show here that chronic morphine decreases levels of several GABAAR subunits, including Gabra2 and Gabrd, in whole NAc and that GABAAR antagonism in this region robustly enhances morphine reward. Such morphine-induced downregulation of GABAAR subunits is even more apparent when D1-type MSNs are examined selectively (see below). Interestingly, mRNA expression levels of Gabra2, which encodes the GABAAR-α2 subunit and is implicated in reward learning and behavioral sensitization as well as human cocaine addiction (Dixon et al, 2010), is regulated oppositely in the NAc by morphine vs cocaine. A previous study showed that GABAAR-α2 subunit is not directly associated with cocaine reward as evidenced by no alterations in cocaine CPP and conditioned reinforcement in Gabra2-deficient mice (Dixon et al, 2010). Furthermore, we observed that D1-TrkB KO enhances Gabra2 levels, as opposed to reduced expression of many other GABAAR subunits and to the consequent reduction of sIPSCs, suggesting that the Gabra2 induction by reduced TrkB in morphine-treated D1-type MSNs might cancel out the effect of Gabra2 reduction by morphine. Rather, Gabra4, another GABAAR subunit of high abundance in NAc (Schwarzer et al, 2001; Wisden et al, 1992) that is decreased by morphine in NAc, particularly in D1-type MSNs, as well as by D1-TrkB KO or by itself in the present study, appears to be critical for morphine reward. A very recent report shows that Gabra4 deletion in D1-type MSNs increases cocaine CPP, but the deletion in D2-type MSNs does not (Maguire et al, 2014). These data suggest a GABAAR subunit-specific and cell type-specific role of GABAAR subunits for drug reward effects.

A role for BDNF-TrkB signaling in NAc in enhancing the behavioral effects of cocaine is well established (Graham et al, 2007; Graham et al, 2009; Horger et al, 1999). By contrast, less is known about the influence of BDNF-TrkB signaling in NAc on morphine action. Intra-NAc infusion of BDNF or genetic deletion of TrkB from all NAc neurons has no effect on morphine reward (Koo et al, 2012). This is consistent with results reported here because TrkB signaling predominates in D2-type MSNs (Lobo et al, 2010) and we show enhancement of morphine reward upon TrkB KO from D1-type MSNs, with no effect observed upon TrkB KO from D2-type MSNs. We also showed recently that chronic morphine administration decreases BDNF expression in the VTA, which would be expected to decrease BDNF signaling from the VTA to the NAc (Koo et al, 2012). Such reduced signaling, in turn, based on results reported here, would be expected to reduce GABAergic activity in D1-type MSNs only, with a resulting increase in morphine reward. This is in parallel with our recent demonstration that D1-type NAc neurons mediate morphine reward: D1 receptor antagonism, but not D2 receptor antagonism, in NAc blocks the facilitation of morphine reward induced by optogenetically activating dopaminergic inputs to NAc (Koo et al, 2012).

In addition, our FACS and electrophysiological studies show that chronic morphine reduces TrkB expression and GABAergic activity selectively in D1-type MSNs in NAc. These findings support the hypothesis that chronic exposure to morphine directly decreases BDNF-TrkB signaling in D1-type MSNs and that the morphine-induced suppression of GABAergic activity in D1-type MSNs occurs through this morphine-induced reduction of BDNF-TrkB signaling, as D1-TrkB KO by itself (ie, without morphine) reduces GABAA signaling in this NAc neuronal subtype. Nonetheless, it is possible that reduced BDNF signaling in D1-type MSNs of the NAc might enhance the activity of these neurons and morphine reward through GABAAR-independent mechanisms. Likewise, future studies are needed to understand the complex circuit mechanisms by which reduced activity of D1-type MSNs promote morphine reward.

Together, our data indicate that enhanced morphine reward in D1-TrkB KOs is attributed at least in part to the suppression of GABAergic inhibition, and that this accounts for the similar behavioral responses observed upon optogenetic activation of D1-type MSNs. The effect of morphine-induced suppression of BDNF-TrkB signaling in D1-type MSNs on morphine reward suggests a double contrast between morphine and cocaine action, as cocaine-induced increases in BDNF-TrkB signaling in D2-type MSNs promotes cocaine reward (Lobo et al, 2010). Whereas chronic exposure to psychostimulants or opiates is known to produce many common molecular and cellular actions in NAc MSNs (Nestler, 2005), there are also clear differences as evidenced by the current findings on BDNF-TrkB signaling. Elucidation of the complicated BDNF-TrkB signaling mechanisms underlying alterations in drug reward by the two main subtypes of NAc MSNs in response to psychostimulants vs opiates will help identify and characterize new therapeutic targets for drug addiction.

FUNDING AND DISCLOSURE

This study was supported by grants from the National Institute on Drug Abuse (E.J.N.). The authors declare no conflict of interest.