The orbitofrontal cortex (OFC) integrates information about the environment to guide decision-making. Glutamatergic synaptic transmission mediated through N-methyl-d-aspartate receptors is required for optimal functioning of the OFC. Additionally, abnormal dopamine signaling in this region has been implicated in impulsive behavior and poor cognitive flexibility. Yet, despite the high prevalence of psychostimulants prescribed for attention deficit/hyperactivity disorder, there is little information on how dopamine modulates synaptic transmission in the juvenile or the adult OFC. Using whole-cell patch-clamp recordings in OFC pyramidal neurons, we demonstrated that while dopamine or selective D2-like receptor (D2R) agonists suppress excitatory synaptic transmission of juvenile or adult lateral OFC neurons; in juvenile lateral OFC neurons, higher concentrations of dopamine can target dopamine receptors that couple to a phospholipase C (PLC) signaling pathway to enhance excitatory synaptic transmission. Interfering with the formation of a putative D1R–D2R interaction blocked the potentiation of excitatory synaptic transmission. Furthermore, targeting the putative D1R–D2R complex with a biased agonist, SKF83959, not only enhanced excitatory synaptic transmission in a PLC-dependent manner, but also improved the performance of juvenile rats on a reversal-learning task. Our results demonstrate that dopamine signaling in the lateral OFC differs between juveniles and adults, through potential crosstalk between dopamine receptor subtypes.
The orbitofrontal cortex (OFC) is implicated in cognitive flexibility and goal-directed behavior. OFC neurons hold information about reward expectancies “online” to guide future decision-making, thus estimating the likelihood of specific outcomes to guide future responses (Schoenbaum et al. 2011). Dopaminergic signaling in the OFC has been implicated in various behaviors, including impulsivity, risky decision-making, reversal learning, and responding on progressive ratio schedules of reinforcement (Cetin et al. 2004; Winstanley et al. 2005; Zeeb et al. 2010; Stopper et al. 2014). OFC neurons receive dopaminergic inputs from the midbrain and express both excitatory Gαs-coupled dopamine 1 (D1Rs) and inhibitory Gαi/o-coupled dopamine type 2 receptors (D2Rs; Berger et al. 1976; Simon et al. 2011).
Despite prevalent administration of monoaminergic drugs for the treatment of attention deficit/hyperactivity disorder (ADHD) in children (Brinker et al. 2007; McCarthy et al. 2012), little is known about dopamine signaling in the OFC of prepubescent juvenile rodents. Indeed, very few studies have demonstrated differential effects of monoaminergic drugs on impulsivity, a hallmark symptom of ADHD, in juvenile versus adult rodents (Bizot et al. 2007; Garske et al. 2013) and none have explored the mechanism behind these dopamine-mediated effects in the OFC. The density of dopaminergic innervation to the superficial layers of the prefrontal cortex (PFC) increases during development and adolescence until approximately postnatal day 60 (Lindvall et al. 1974; Berger et al. 1976; Kalsbeek et al. 1988). Parallel increases in D1R and D2R expression across development have also been demonstrated (Tarazi and Baldessarini 2000; Garske et al. 2013). Furthermore, electrophysiology studies in other cortical regions have shown D1R-mediated potentiation of N-methyl-d-aspartate receptors (NMDARs) in an age-dependent manner (Flores-Barrera et al. 2014). This effect was mediated by a late adolescent D1R-mediated acquisition of GluN2B NMDA subunits to the synapse in the medial PFC (mPFC; Flores-Barrera et al. 2014). Therefore, we hypothesize that there are age-dependent differences in dopaminergic signaling in the OFC.
The role of dopamine in its ability to regulate firing activity and excitatory synaptic transmission has been well characterized in the mPFC. Here, dopamine tunes the activity of pyramidal cell networks, and is believed to underlie working memory function (Watanabe et al. 1997; Zahrt et al. 1997; Durstewitz et al. 2000; Paspalas and Goldman-Rakic 2005). NMDAR currents are important for maintaining the activity of cortical neurons involved in an active memory trace (Durstewitz 2009) and are modulated by dopamine (Seamans et al. 2001; Chen et al. 2004; Kruse et al. 2009; Li et al. 2009, 2010). In whole-cell recordings from prepubescent rat brain slices containing mPFC, dopamine D1Rs typically enhance NMDAR responses, whereas D2Rs attenuate NMDAR-mediated responses (Zheng et al. 1999; Seamans et al. 2001; Wang and O'Donnell 2001; Gonzalez-Islas and Hablitz 2003; Chen et al. 2004; Wirkner et al. 2004; Beazely et al. 2006). NMDARs expressed on pyramidal neurons are required for efficient function of the OFC. NMDAR currents drive firing of selective neuronal populations to relevant stimulus–outcome associations. Specificity of firing to salient stimuli may underlie the ability of the OFC to integrate prior experience with a current context, in order to adaptively respond to changing environments (van Wingerden et al. 2012; Brigman et al. 2013). However, it is unknown how dopamine signaling alters excitatory synaptic transmission in the OFC.
Given that mPFC and OFC functions are clearly dissociated (McAlonan and Brown 2003; Buckley et al. 2009; St Onge and Floresco 2010) and exhibit differential responses to administration of psychostimulants (Crombag et al. 2005; Homayoun and Moghaddam 2006), a clearer understanding of how dopamine modulates synaptic transmission in the OFC of juvenile and adult animals is warranted.
Materials and Methods
All animals were juvenile (P21–30) or adult male Wistar rats (P60–70), provided by Charles River, and housed in groups of 2–6. Rats were maintained on a 12 : 12 h light : dark schedule (lights on at 7:00 AM), and given food and water ad libitum, except during behavioral experiments. All experimental protocols were in accordance with the Canadian Council on Animal Care and approved by the University of Calgary Animal Care Committee and the University of British Columbia Animal Care Committee. Anesthesia and analgesia were used to minimize pain and discomfort in animals.
Rats were anesthetized with isoflurane, decapitated, and brains were rapidly extracted into ice-cold sucrose solution containing (in mM): 75 sucrose, 87 NaCl, 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO3, 7 MgCl2, and 0.95 CaCl2. Sagittal brain sections containing lateral, medial, or ventral OFC (Paxinos and Watson 2007) were cut at 300 µm on a vibratome (Leica, Nussloch, Germany). Slices were transferred to 250 mL of aCSF containing (in mM): 126 NaCl, 1.6 KCl, 1.1 NaH2PO4, 1.4 MgCl2, 26 NaHCO3, 11 glucose, and 2.4 CaCl2, and incubated for a minimum of 60 min at 31.4–33°C prior to recording. All solutions were continuously saturated with 95% O2/5% CO2.
Slices were placed in the recording chamber and perfused with aCSF with the addition of picrotoxin (100 µM) to block GABAA receptor-mediated inhibitory postsynaptic currents. Cells were visualized on an upright microscope using “Dodt-type” gradient contrast infrared optics (Dodt et al. 2002). Whole-cell voltage-clamp recordings of pyramidal neurons were made using a Multiclamp 700B amplifier (Molecular Devices, Union City, CA, USA). Recording electrodes (3–5 MΩ) were filled with (in mM): 120 cesium methanesulfonate, 20 HEPES, 0.4 EGTA, 5 TEA-Cl, 2 MgCl2, 2.5 MgATP, 0.5 NaGTP, and 5.4 biocytin (in some experiments), with a pH of 7.2–7.3 and 280–295 mOsm. Pyramidal neuron morphology was confirmed with biocytin labeling in a portion of cells as previously described (Labouèbe et al. 2013; seeWang et al. (2008)].
Agonists SKF38393 (10 µM, Tocris, Ellisville, MO, USA), quinpirole (10 µM; Sigma, St Louis, MO, USA), SKF83959 (1 and 10 µM Tocris), and dopamine (0.1 or 10 µM, dissolved in H2O + 75 µM sodium metabisulfite and prepared daily, Sigma) were bath applied for 5 min after establishing 10 min of stable baseline recordings. Dopamine and D1 agonists were protected from light. In some experiments, the antagonists SCH39166 (1 µM; Tocris) or sulpiride [(−/−) sulpiride: 500 nM; Tocris, or (±) sulpiride: 1 µM; Sigma] were bath applied for 5 min prior to and during application of SKF38393 or quinpirole. Protein kinase A (PKA) inhibitor (PKI; 20 µM; Tocris), phospholipase C (PLC) inhibitor U73122 (1 µM; Tocris), or 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) (10 mM; Tocris) was dissolved in internal solution. D2LIL3-29-2 and a scrambled version of this peptide were used intracellularly at 10 µM. Stock solutions of these peptides were prepared daily in deionized water and stored on ice.
All rats were handled approximately 10 min per day, 3 days before surgery. Rats were anesthetized with isofluorane, injected with subcutaneous ketoprofen (10 mg/kg) and saline (2–3 mL, 0.9%), and secured in the stereotaxic frame. A bilateral 23-gauge cannula (made in-house) was implanted into the OFC and secured with 3–4 jeweler's screws and dental acrylic. Stereotaxic coordinates for adults were (in mm) AP: +3.5 (from Bregma), ML: ±3.2, and DV: −4.3 (from dura). Juvenile placements were AP: +3.3, ML: ±3.1, and DV: −2.5 (from dura). All rats were implanted and then given 1 day of recovery before placing them back with their original cage mates (grouped 4–6 for juveniles and 2 for adults). Note that these empirically derived stereotaxic coordinates do not precisely match those given in the rat brain atlas (Paxinos and Watson 2007), which we used as references for the injection-site images. On average, the rostrocaudal axis appeared to be approximately shifted caudally by about 400 μm in juvenile rats. Rats were cannulated on P19/P60–70 for juveniles/adults and then exposed to the behavioral apparatus after recovery. Upon completion of testing, all rats were transcardially perfused with 4% paraformaldehyde, and 40 µm sections mounted on gelatin-coated slides were stained using cresyl violet to confirm cannula placements in lateral OFC (Fig. 8E). All behavioral data from cannulations in brain regions outside the OFC (juvenile n = 5 and adult n = 1) were not included in the analysis.
All place reversal training was performed in a plus maze using previously established protocols (Ragozzino and Choi 2004; Kim and Ragozzino 2005). The maze was constructed of opaque black acrylic with each of the 4 arms 55 cm long, 10 cm wide, and 15 cm in height. An 8-mL ceramic crucible served as a food well and was secured at the end of each arm. The maze was placed on a table that was 72 cm in height. Extra-maze visual cues were placed throughout the room. For juvenile rats, 2 visual cues were placed adjacent to 1 of 2 arms and were kept consistent throughout the experiment. Each arm location was arbitrarily labeled North, South, East, or West, and remained constant throughout behavioral experiments.
Rats were pre-exposed to Froot Loops™ cereal (Kelloggs, Battle Creek, MI, USA) in their home cage 1–2 days before habituation training and were handled for 10 min per day during this time. Habituation consisted of 3–6 days where each rat learned to consume one-fourth to one-half piece of cereal from each of the 4 arms. This continued until each animal ate from all 4 arms a minimum of 5 times in 15 min or less. The last day of habituation consisted of a “block” phase where a black acrylic block was placed into one of the arms, therefore transforming the maze into a T-shape. The rat was placed in the stem (start) arm and allowed to consume cereal at the other 2 baited “choice” arms. This continued for 7 trials, where one trial consisted of the rat consuming both cereal pieces in each of the 2 arms.
The following day (acquisition) consisted of training the rats on a place-discrimination task. Using the same T-maze set-up, only the North choice arm was baited. The rat was placed in either the East or West start arm and the responses toward the North (baited) arm or South (unbaited) arm were recorded. This comprised one trial. In between trials, the rat was placed on a cage top adjacent to the maze. Start arms were pseudorandomized such that the same start arm was never used for more than 3 consecutive trials. After every fourth trial, the maze was turned 90° and wiped out with a sponge sprayed with 2% Virkon (Dupont, Mississauga, ON, Canada) to minimize the use of intra-maze cues in learning the response strategy. Testing continued until each rat reached a criterion of 10 consecutive correct responses. No treatment was given to any animals on acquisition training.
Test day 2 (reversal) was similar to acquisition except that the reinforced arm was now the South arm instead of the North. Each rat was pseudorandomized into 2 groups, so the mean trials to criterion on the acquisition phase were matched. A block of 4 trials was given to confirm that the rats had retained the previously acquired strategy before the reversal. Rats that did not meet a minimum of 75% correct responses in this block were removed from the study (n = 1 rat). Prior to reversal testing, each animal received an infusion of SKF83959 (20 ng/side) or vehicle (10% DMSO in saline). A 30-gauge stainless steel injector was inserted into each guide, so that it extended 1 mm below the cannula tip. For both groups, total injection volume was 500 nL/side at an infusion rate of 100 nL/min. Injectors were left in place for an additional 2 min, then the animal was placed back in its home cage for an additional 10 min before the start of testing.
Errors committed during the reversal phase were analyzed to assess whether the drug treatment altered perseveration of a previously learned place discrimination or regression to a previous strategy once perseveration had ceased (Ragozzino and Choi 2004; Kim and Ragozzino 2005). Trials were grouped into blocks of 4 and then incorrect trials were counted and analyzed. Perseverative errors consisted of 3 or more (>75%) incorrect trials per block (Ragozzino and Choi 2004; Kim and Ragozzino 2005). Once rats made <2 incorrect trials per block, all incorrect trials were labeled regressive errors, indicating the rat was regressing back to its previously learned strategy, and served as an index to measure how well the rat could maintain its choice of a newly learned strategy.
All electrophysiology data are expressed as mean percent change in baseline levels ± SEM. We found that within our cell populations, we had responders (>5% effect) and non-responders (<5% effect). Therefore, we have used non-parametric statistical methods. In all experiments comparing baseline versus the effect of drug treatment, a Wilcoxon matched-pairs signed rank test was used to measure between an averaged 2-min time point from the end of baseline recording versus an averaged 2-min time point at the maximal effect of the agonists. Example traces were constructed from an average of 12 sweeps (2 min) taken before and after drug application. Stimulus artifacts were removed for clarity. For electrophysiology experiments, “n” refers to the number of cells recorded from at least 3 rats expressed as (N/n = cells/rats). A power analysis indicates that to detect a minimum 10% change in effect with a statistical power level of 0.8, we required a sample size of at least 4 cells per group. For multiple group comparisons with repeated observations, a Friedman test was used. For behavioral data, a two-way ANOVA with planned comparisons using a Bonferroni post hoc test was used to compare effects of age and drug treatment. Significance was set at P < 0.05 for all experiments.
To determine if D1Rs modulated NMDAR-mediated EPSCs in the OFC of juvenile rats, we recorded layer II/III pyramidal cells of 3 major OFC subregions: lateral, ventral, and medial (Fig. 1A–D). Application of the D1R agonist, SKF38393 (5 min, 10 µM; Gonzalez-Islas and Hablitz 2001), to the lateral OFC significantly potentiated evoked NMDAR EPSCs of pyramidal neurons (baseline: 102 ± 1% vs. SKF38393: 123 ± 5%; N/n = 7 cells/6 rats, P = 0.015, sum of signed ranks (W) = 28, pairs = 7; Fig. 1A,B). In contrast, SKF38393 did not significantly potentiate NMDAR EPSCs in ventral (baseline: 98 ± 3% vs. SKF38393: 98 ± 5%; N/n = 8/6, P = 0.94, W = −2, pairs = 8; Fig. 1C) or medial (baseline: 98 ± 2% vs. SKF38393: 94 ± 6%; N/n = 8/6, P = 0.74, W = −6, pairs = 8; Fig. 1D) OFC subregions. Taken together, these data suggest that there is regional selectivity of D1R-mediated potentiation of NMDARs in the juvenile OFC.
To confirm that the potentiation of NMDARs by SKF38393 was due to activation of D1Rs, we bath applied SKF38393 in the presence of a selective D1R antagonist, SCH39166 (1 µM; 10 min; Yanovsky et al. 2011). SKF38393 (10 µM; 5 min) applied with SCH39166 did not potentiate NMDAR currents of juvenile lateral OFC neurons (baseline: 97 ± 0.7%; SKF38398 + SCH39166: 98 ± 1%; N/n = 9/6, P = 0.91, W = −3, pairs = 9; Fig. 1E).
To elucidate the signaling mechanism behind the D1R-mediated increase in NMDAR currents, we blocked PKA signaling with intracellular application of a PKI (20 µM; Lopshire and Nicol 1998). PKI abolished D1R-mediated potentiation of NMDAR currents (baseline: 98 ± 2% vs. SKF38393: 93 ± 3%; N/n = 8/5, P = 0.31, W = −16, pairs = 8; Fig. 1F). To test if intracellular calcium was required for potentiation of NMDARs by SKF38393, we applied a calcium chelator, BAPTA (10 mM), intracellularly. SKF38393 potentiated NMDARs in the presence of BAPTA (baseline: 102 ± 1% vs. SKF38393: 119 ± 5%; N/n = 7/6; P = 0.04, W = 24, pairs = 7; see
D1R mRNA in OFC neurons increases from pre-adolescence through adulthood (Garske et al. 2013). Therefore, we investigated whether 10 µM SKF38393 would further potentiate NMDARs of adult OFC pyramidal neurons in a subregion-selective manner. SKF38393 potentiated NMDAR EPSCs in adult lateral OFC pyramidal neurons (baseline: 98 ± 2% vs. SKF38393: 122 ± 7%; N/n = 9/5, P = 0.019, W = 39, pairs = 9; Fig. 2A) with a similar efficacy to that of juveniles (F1,26 = 0.09; P > 0.05, two-way ANOVA). Furthermore, potentiation of NMDAR EPSCs in adult OFC neurons was also regionally selective, as SKF38393 did not potentiate NMDAR EPSCs in either ventral (baseline: 98 ± 2% vs. SKF38393: 100 ± 4%; N/n = 6/4, P = 0.15, W = 15, pairs = 6; Fig. 2B) or medial OFC (baseline: 97 ± 2% vs. SKF38393: 103 ± 4%; N/n = 9/4, P = 0.16, W = 25, pairs = 9; Fig. 2C).
To assess if there were changes in the reversal potential between juvenile and adult animals, we performed a current–voltage relationship. There was no significant difference in the reversal potential [juveniles: 11 ± 0.8 mV (N/n = 8/3) vs. adults: 10 ± 1 mV (N/n = 8/5); P = 0.47, t = 0.7, df = 14, t-test; seeWang et al. 2008), there was no significant difference in the decay of NMDA EPSCs (τslow) between adults and juveniles [juveniles: 174 ± 6 ms (N/n = 9/5) vs. adults: 189 ± 30 (N/n = 6/3); P = 0.72, t = 0.36, df = 13].
To determine whether D2R activation modulates NMDAR EPSCs in the lateral OFC, we bath applied the D2R agonist, quinpirole (10 µM; 5 min; Kotecha et al. 2002). In lateral OFC pyramidal neurons from either juvenile or adult animals, quinpirole significantly inhibited NMDAR EPSCs (juvenile, baseline: 101 ± 2% vs. quinpirole: 85 ± 4%, N/n = 10/6, P = 0.002, W = −55, pairs = 10, Fig. 3A; adult, baseline: 100 ± 1% vs. quinpirole: 71 ± 5%, N/n = 9/7, P = 0.0039, W = −45, pairs = 9; Fig. 3B). The D2R antagonist, sulpiride (500 nM), blocked the quinpirole-mediated suppression of NMDAR EPSCs of juvenile lateral OFC neurons (baseline: 98 ± 1%; quinpirole + sulpiride: 96 ± 4%, N/n = 9/7, P = 0.41, W = −15, pairs = 9; Fig. 3A) or adult lateral OFC neurons (baseline: 97 ± 2%; quinpirole + sulpiride: 94 ± 7%, N/n = 7/4, P = 0.57, W = −18, pairs = 7; Fig. 3B). Taken together, D1R activation potentiates NMDAR EPSCs, whereas D2R activation inhibits NMDAR EPSCs in lateral OFC pyramidal neurons of both juvenile and adult rats.
We also examined if the effects of quinpirole in the OFC of juveniles or adults were subregion-selective. Quinpirole inhibited NMDAR EPSCs in the ventral OFC of juvenile (baseline: 98 ± 1%; quinpirole: 83 ± 5%, N/n = 8/7, P = 0.039, W = −30, pairs = 8; Fig. 3C) or adult (baseline: 96 ± 2%; quinpirole: 75 ± 6%, N/n = 6/4, P = 0.031, W = −21, pairs = 6; Fig. 3D). Furthermore, quinpirole inhibited NMDAR EPSCs in the medial OFC of juvenile (baseline: 98 ± 0.6%; quinpirole: 81 ± 5%, N/n = 8/7, P = 0.023, W = −32, pairs = 8; Fig. 3E) and adult rats (baseline: 100 ± 2%; quinpirole: 78 ± 6%, N/n = 7/3, P = 0.03, W = −26, pairs = 7; Fig. 3F).
Next, we investigated the effects of D1R or D2R activation on AMPA receptor-mediated EPSCs in juvenile lateral OFC neurons. Application of SKF38393 (10 µM) did not change the amplitude of AMPAR currents (baseline: 101 ± 3% vs. SKF38389: 104 ± 4%; N/n = 8/6, P > 0.99, W = 0, pairs = 8; Fig. 4A) or modulate the paired-pulse ratio (baseline: 1.1 ± 0.05 vs. SKF38393: 1.1 ± 0.07; wash: 1.1 ± 0.04; N/n = 6/5, P = 0.95, Freidman statistic = 0.33; Fig. 4B). Furthermore, quinpirole did not modulate AMPAR EPSCs (baseline: 101 ± 2% vs. SKF38389: 106 ± 4%; N/n = 8/3, P = 0.46, W = 12, pairs = 8; Fig. 4C). Taken together, SKF38393 does not modulate the probability of glutamate release and acute application of D1R or D2R agonists does not modulate AMPAR EPSCs.
To test the possibility of dopamine being co-released with evoked glutamate, we bath applied the D1R agonist with a D2R antagonist in lateral OFC slices from juvenile rats, with the hypothesis that evoked endogenous dopamine may reduce D1R-mediated responses by acting at D2Rs. SKF38393 in the presence of sulpiride significantly potentiated NMDAR EPSCs (baseline: 101 ± 2%, sulpiride: 100 ± 5%, SKF38393 + sulpiride: 115 ± 6%, N/n = 7/5; P = 0.004, Friedman statistic = 10.29), but this was not significantly different from the application of SKF38393 alone (Mann–Whitney, P = 0.39, U = 20). These data suggest that endogenous dopamine is not evoked with afferent stimulation to significantly alter D1R-mediated effects at NMDARs.
Previous studies have demonstrated that D1Rs and D2Rs act cooperatively to potentiate evoked firing in the striatum (Hopf et al. 2003; Seif et al. 2011). Therefore, we tested if coactivation of D1Rs and D2Rs modified NMDAR EPSCs of lateral OFC neurons in either juvenile or adult animals. Interestingly, co-application of quinpirole and SKF38393 potentiated NMDAR EPSCs in lateral OFC pyramidal neurons of juvenile rats (baseline: 199 ± 2% vs. SKF38393 + quinpirole: 134 ± 12%; N/n = 8/7, P = 0.0078, W = 36, pairs = 8; Fig. 5A). In contrast, D1R–D2R cooperation was absent in the lateral OFC of adult rats (baseline: 100 ± 2% vs. SKF38393 + quinpirole: 97 ± 5%; N/n = 9/5, P = 0.91, W = 3, pairs = 9; Fig. 5B).
D1R–D2R cooperativity in the striatum can involve coupling to PLC (Rashid, So, et al. 2007; Seif et al. 2011). Therefore, we tested if coactivation of D1Rs and D2Rs required PLC to potentiate NMDARs in juvenile lateral OFC neurons. Co-application of SKF38393 and quinpirole did not potentiate NMDAR EPSCs in the presence of an intracellularly applied PLC inhibitor, U73122 (1 µM; baseline: 101 ± 1% vs. SKF38393 + quinpirole: 102 ± 4%; N/n = 11/8; P = 0.63, W = 12, pairs = 11; Fig. 5C), suggesting that coactivation of D1Rs and D2Rs couple with PLC to potentiate NMDARs in lateral OFC neurons. Intracellular administration of U73122 did not alter evoked NMDAR EPSCs with coapplication of SKF38393 and quinpirole in adults (baseline: 101 ± 2% vs. SKF38393 + quinpirole: 100 ± 7%; N/n = 6/5, P = 0.68, W = −5, pairs = 6; Fig. 5D).
Next, we tested if D1R–D2R cooperativity in juvenile lateral OFC required intracellular calcium to potentiate NMDARs. Intracellular application of the calcium chelator, BAPTA (10 mM), inhibited D1R–D2R-mediated potentiation of NMDAR EPSCs (baseline: 98 ± 2% vs. SKF38393 + quinpirole: 109 ± 5%; N/n = 9/7, P = 0.20, W = 23, pairs = 9; Fig. 5C), suggesting that increased intracellular calcium concentration is necessary for D1R–D2R-mediated potentiation of NMDAR EPSCs in lateral OFC neurons. Taken together, individually applied D1R or D2R agonists have opposing actions at excitatory synapses of lateral OFC pyramidal neurons. However, when co-applied, D1R and D2R agonists may act cooperatively to potentiate NMDARs in a PLC-dependent manner only in pyramidal neurons of juvenile lateral OFC.
A possible explanation for D1R–D2R-mediated cooperativity requiring PLC activation is via atypical PLC-coupled D1Rs (Yu et al. 1996; Pollack 2004) that may be present in juvenile lateral OFC pyramidal neurons. However, it is unclear why these receptors would be selectively targeted with SKF38393 in the presence of quinpirole as we have demonstrated that SKF38393 alone activates PKA-coupled D1Rs and does not require elevated intracellular calcium. A second possibility is that SKF38393 and quinpirole selectively activate D1R–D2R complexes that couple to PLC (Hasbi et al. 2010; O'Dowd et al. 2012).
To determine if an interaction between D1Rs and D2Rs was required for potentiation of NMDAR EPSCs during co-application of SKF38393 and quinpirole, we hypothesized that disruption of the D1R–D2R complex would inhibit the D1R–D2R cooperative potentiation of NMDARs and the net effect would be similar to that of co-application of SKF38393 and quinpirole in adult lateral OFC. To disrupt the D1R–D2R interaction, we used an interfering peptide containing the sequence for a fragment of the third intracellular loop on the D2L isoform, Met257-Glu271 (D2LIL3-29-2), that does not alter expression of native D2Rs (Pei et al. 2010).
Intracellular application of the interfering peptide abolished potentiation of EPSCs by SKF38393 and quinpirole in juvenile lateral OFC neurons (baseline: 99 ± 3% vs. SKF38393 + quinpirole: 94 ± 4%; N/n = 10/7, P = 0.23, W = −25, pairs = 10; Fig. 5E). Conversely, SKF38393 and quinpirole potentiated NMDAR EPSCs in the presence of intracellular application of a scrambled control peptide (baseline: 101 ± 3% vs. SKF38393 + quinpirole: 133 ± 13% increase from baseline; N/n = 8/5, P = 0.0078, W = 36, pairs = 8; Fig. 5E). D2LIL3-29-2 did not alter D1R- or D2R-mediated responses as SKF38393 potentiated (baseline: 102 ± 2% vs. SKF38393: 120 ± 10%, N/n = 10/6, P = 0.027, W = 43, pairs = 10; see5F), suggesting that the interfering peptide does not alter D1R or D2R monomers in adult lateral OFC neurons. Taken together, D1R–D2R coactivation requires an association of the carboxyl tail of the D1R with the third intracellular loop of the D2L isoform, coupling to PLC and increased intracellular calcium.
Next, we tested if SKF83959, a biased agonist that targets PLC-coupled dopamine receptors (Rashid, O'Dowd, et al. 2007; Rashid, So, et al. 2007), could mimic D1R–D2R cooperativity in juvenile lateral OFC neurons. SKF83959 (1 μM) potentiated NMDAR EPSCs in lateral OFC pyramidal neurons (baseline: 104 ± 1% vs. SKF83959: 126 ± 10%; N/n = 7/6, P = 0.03, W = 26, pairs = 7; Fig. 6A). This effect was blocked when U73122 was included in the pipette (baseline: 101 ± 3% vs. SKF83959 + U73122: 98 ± 3%; N/n = 6/4, P = 0.31, W = −11, pairs = 6; Fig. 6A).
To test if both D1R and D2Rs were required for the effect of SKF83959, we applied D1R or D2R antagonists (Rashid, So, et al. 2007). In the presence of SCH39166, an antagonist of D1Rs, SKF83959 did not potentiate NMDAR EPSCs (baseline: 102 ± 1% vs. SKF83959 + SCH39166: 96 ± 6% increase; N/n = 7/4, P = 0.64, W = −16, pairs = 7; Fig. 6B). Similarly, SKF83959 in the presence of sulpiride did not potentiate NMDAR EPSCs (baseline: 100 ± 2% vs. SKF83959 + sulpiride: 94 ± 3%; N/n = 6/5, P = 0.09, W = −17, pairs = 6; Fig. 6B). In adult lateral OFC neurons, bath application of SKF83959 (1 and 10 μM) did not significantly increase NMDAR EPSCs (1 μM: baseline: 101 ± 1% vs. SKF83959: 101 ± 4%; N/n = 8/4, P = 0.84, W = 4, pairs = 8; 10 μM: baseline: 98 ± 3% vs. SKF83959: 105 ± 10%; N/n = 6/5, P = 0.68, W = 5, pairs = 6; Fig. 6C). Thus, SKF83959 potentiates NMDAR EPSCs and requires activation of both D1R and D2R and PLC only in lateral OFC neurons from juvenile rats.
Next, we hypothesized that dopamine may differentially activate dopamine receptors of juvenile and adult lateral OFC neurons. In lateral OFC pyramidal neurons of both juveniles and adults, dopamine (0.1 μM) suppressed NMDAR EPSCs (juvenile, baseline: 99 ± 2% vs. dopamine: 87 ± 5%, N/n = 8/4, P = 0.039, W = −30, pairs = 8; adult, baseline: 99 ± 2% vs. dopamine: 81 ± 5%, N/n = 6/4; P = 0.030, W = −21, pairs = 6; Fig. 7A). At the time point where the lower concentration of dopamine suppressed NMDAR currents in both adults and juveniles, dopamine (10 µM) suppressed NMDARs only in adult lateral OFC (baseline: 100 ± 2% vs. dopamine: 85 ± 5%, N/n = 7/5; P = 0.04, W = −4, pairs = 7; Fig. 7B), but not in juvenile lateral OFC (baseline: 99 ± 1% vs. dopamine: 100 ± 5%, N/n = 7/5; P = 0.99, W = 0, pairs = 7; Fig. 7B). Interestingly, at a later time point, dopamine (10 µM) significantly potentiated NMDAR currents in juvenile (113 ± 6%, N/n = 7/5) compared with adults (91 ± 5%, N/n = 7/5; Mann–Whitney U, P = 0.02, U = 7; Fig. 7B). There was a significant interaction of the effect size of dopamine concentration with the animal's age (P = 0.03; two-way ANOVA, F1,18 = 4.951; Fig. 7C). A Sidak's multiple comparison tests revealed that 0.1 µM dopamine-mediated inhibition of NMDARs was not significantly different between age groups (P > 0.05, Fig. 7A,C). In contrast, there was a significant difference in effect size of the higher dopamine concentration (10 μM) (P < 0.001; Fig. 7B,C). These data suggest that higher concentrations of dopamine have differential effects on the modulation of NMDAR currents in the lateral OFC of juvenile and adult rats.
To determine if the concentration-dependent effects of dopamine on juvenile lateral OFC pyramidal neurons were due to activation of PLC-coupled dopamine receptors, we applied U73122 via the patch pipette. Dopamine (0.1 μM) in the presence of U73122 transiently suppressed NMDAR responses (baseline: 100 ± 4%; dopamine: 83 ± 5%; wash: 101 ± 9%, N/n = 6/3, P = 0.0081, Friedman statistic = 9; Fig. 7D). Interestingly, in contrast to dopamine (10 μM) alone, dopamine (10 μM) in the presence of U73122 caused a long-lasting inhibition of NMDAR EPSCs (baseline: 101 ± 3; 10 μM dopamine: 72 ± 7%; wash: 78 ± 11%; N/n = 6/3, P = 0.0026, Friedman statistic = 10.18; Fig. 7D). Taken together, these data suggest that high dopamine concentration targets PLC-coupled dopamine receptors only on juvenile lateral OFC neurons.
Our in vitro data point to a novel pathway by which dopamine acts in the lateral OFC of juvenile rats. We hypothesized that activation of this pathway could modulate OFC-dependent behaviors selectively in juvenile rats. We trained both juvenile and adult rats on a reversal-learning task using a place-discrimination/-reversal paradigm (Ragozzino and Choi 2004; Kim and Ragozzino 2005). We chose this task as it can be acquired in a short period for juvenile rats (Bizot et al. 2007). Furthermore, reversal learning is dependent on NMDAR activation in the OFC (Bohn et al. 2003; Calaminus and Hauber 2008; Brigman et al. 2013). Additionally, activation of D1Rs and D2Rs in the rat OFC are necessary for effective reversal learning (Calaminus and Hauber 2008; Winter et al. 2009; Mizoguchi et al. 2010). Because SKF83959 potentiated NMDAR EPSCs via PLC in juveniles, but not in adults, we hypothesized that intralateral OFC administration would have age-dependent effects on reversal learning. Juvenile rats learned the place discrimination at the same rate as adults (n = 15 [juvenile] vs. 16 [adult], P > 0.05, t = 0.9, df = 29, t-test; Fig. 8A). There were no significant differences due to age or drug treatment on the time required to complete each trial (F1,20 = 0.7, Fig. 8B), suggesting that there was no impairment of locomotor activity. Analysis of the reversal phase revealed a main effect of drug treatment (F1,27 = 5.9, P < 0.05) and age (F1,27 = 5.2, P < 0.05, two-way ANOVA; Fig. 8C). Post hoc analyses revealed that vehicle-treated juveniles took significantly more trials to reach criterion on the reversal day compared with adults (juveniles: 77 ± 4 trials, n = 7 vs. adult: 56 ± 7 trials, n = 8; t = 2.7, df = 1, P < 0.05; Fig. 8C). This effect was abolished by SKF83959 in the lateral OFC of juvenile rats (juveniles: 55 ± 5 trials, n = 8 vs. adult: 52 ± 6 trials, n = 8; t = 0.46, df = 1, P > 0.05; Fig. 8C), thus facilitating their ability to reverse to the new place as efficiently as adults (F1,27 = 5.2, P > 0.05).
Errors committed during the reversal phase were divided according to whether drug treatment impaired the ability to shift away from the previously learned strategy (perseverative errors), or an inability to maintain or reliably execute a newly learned response pattern (regressive errors) (Ragozzino and Choi 2004; Kim and Ragozzino 2005; Boulougouris et al. 2007; Butts et al. 2013). Adult rats made the same amount of perseverative errors as juvenile rats (F1,27 = 0.2, P > 0.05; Fig. 8D), and significantly less regressive errors than juveniles in the vehicle group (P < 0.01, F1,27 = 8.3; Fig. 8D). While there was no significant effect of SKF83959 on regressive or perseverative errors in adults, SKF83959 significantly reduced regressive errors in juvenile rats (F1,27 = 5.1, P < 0.05). These data suggest that activation of PLC-coupled dopamine receptors in the lateral OFC of juvenile rats can improve the ability to maintain newly learned place discrimination.
Here, we demonstrate that selective activation of D1Rs or D2Rs had opposing effects on NMDAR EPSCs of lateral OFC neurons from both juveniles and adults. However, in lateral OFC pyramidal neurons from juvenile rats, simultaneous activation of D1Rs and D2Rs cooperatively potentiate NMDARs by a PLC-mediated mechanism. Furthermore, intra-OFC infusion of SKF83959, an agonist that targets PLC-coupled dopamine receptors in juvenile lateral OFC, ameliorates performance on a reversal-learning task only in juvenile rats. Taken together, these data suggest that coactivation of D1Rs and D2Rs in the lateral OFC may provide a novel therapeutic target for the treatment of cognitive flexibility disorders in juveniles.
Dopamine terminals in the mPFC and the OFC form symmetrical contacts onto dendritic shafts and spines (Séguéla et al. 1988) that also receive excitatory input (Carr and Sesack 1996). Here, we provide functional evidence that dopamine can gate excitatory inputs to the OFC. Notably, D1R-mediated potentiation of NMDARs was restricted to the lateral OFC region in both juvenile and adult rats. However, D2R-mediated potentiation occurred in most cells of all OFC regions. Regional differences in dopamine receptor expression might explain why D1R-mediated increases in NMDAR currents are restricted to the lateral OFC. Although regional specificity of D1R or D2R expression in the OFC has been described in primates and humans (Bergson et al. 1995; Negyessy and Goldman-Rakic 2005), subregion differences in receptor expression have been little explored in rodents, even though functional differences in behavior may exist. For example, one study demonstrated that the medial OFC did not play a critical role in conditioned cue-induced reinstatement of cocaine seeking, whereas the lateral OFC was involved in the long-term storage, retrieval, or utilization of stimulus-reward associations (Fuchs et al. 2004).
Activation of D1Rs transiently potentiated NMDAR EPSCs in juvenile or adult rats, consistent with reported effects of SKF38393 in other cortical brain regions (Gonzalez-Islas and Hablitz 2003). The D1R group is composed of the D1 and D5 receptor subtypes that are preferentially coupled to Gαs proteins that stimulate adenylyl cyclase and PKA (Monsma et al. 1990; Zhuang et al. 2000). Consistent with this, SKF38393 potentiation of NMDARs was blocked with intracellular application of a PKI or bath application of SCH39166. Three lines of evidence suggest that D1Rs are expressed postsynaptically in the lateral OFC: 1) SKF38393-mediated potentiation of NMDARs was blocked by a pipette-administered inhibitor of PKA signaling, 2) SKF38393 did not modulate the paired-pulse ratio, a measure that highly correlates with release probability (Lin and Faber 2002), and 3) SKF38393 differentially affected NMDARs from AMPARs, suggesting that increased glutamate release was not the mechanism by which SKF38393 potentiated NMDARs. Similarly, activation of D2Rs inhibited NMDARs, but did not alter AMPARs also consistent with a post-synaptic action. D1R and D2R mRNA are present in layer II/III pyramidal neurons of the mPFC (Santana et al. 2009), and co-localization on cell bodies in mPFC slices has been demonstrated (Vincent et al. 1995; Lee et al. 2004). Our data demonstrate similar functional effects of postsynaptic D1Rs and D2Rs in layer II/III pyramidal neurons of the lateral OFC.
Interestingly, we observed a cooperative potentiation of NMDAR EPSCs when SKF38393 and quinpirole were co-applied. Our experimental results could suggest interaction of a D1R–D2R complex as blocking the interaction between the long form of D2Rs and D1Rs inhibited the cooperative potentiation of NMDARs. Importantly, the interfering peptide used in this experiment does not alter signaling of D1R or D2R monomers (Pei et al. 2010). FRET and co-immunoprecipitation studies have demonstrated heterodimerization of D1Rs and D2Rs in the striatum (So et al. 2005; Dziedzicka-Wasylewska et al. 2006), NAc (Perreault et al. 2011), and mPFC (So et al. 2009; Pei et al. 2010). Furthermore, stimulation of D1R–D2R or D5R–D2R heterodimers activates PLC and increases internal calcium concentrations (Rashid, So, et al. 2007; So et al. 2009).
Another possible mechanism underlying D1R–D2R cooperative potentiation of NMDARs is that Gβγ released from activation of D1Rs or D2Rs directly activates PLC signaling leading to increased NMDAR EPSCs. However, D1R-mediated potentiation of NMDARs in lateral OFC neurons of juvenile rats was completely blocked by intracellular application of a PKI. Furthermore, chelation of intracellular calcium did not inhibit the D1R-mediated potentiation of NMDARs in the lateral OFC. Thus, D1R activation leading to freeing of Gβγ is unlikely to directly activate PLC in this model.
A third possibility is that coactivation of D1R and D2R monomers results in crosstalk between Gs and Gi protein-mediated downstream signaling pathways leading to PLC activation, a mechanism that has been described by others (Carroll et al. 1995; Beaulieu and Gainetdinov 2011; Rebres et al. 2011; Medvedev et al. 2013). In expression systems, a potential mediator of D1R–D2R crosstalk is a cytosolic protein, calcyon, which reportedly increases the affinity state for D1Rs (Lidow et al. 2001; Bergson et al. 2003). D1R activation further mobilizes calcyon trafficking to cell membranes (Lidow et al. 2001). When D2Rs are stimulated, calcyon signaling increases intracellular calcium concentrations (Frégeau et al. 2013). Interestingly, in 2 rat models of ADHD, calcyon mRNA is significantly higher in juvenile versus adult OFC neurons (Heijtz et al. 2007). Future experiments could elucidate whether D1R–D2R cooperative potentiation of NMDAR currents via PLC requires calcyon-mediated signaling.
SKF83959, a putative agonist for Gq-coupled D1R–D2R heteromers (Lee et al. 2004; Rashid, So, et al. 2007; Hasbi et al. 2009), potentiated NMDAR EPSCs in a PLC-dependent manner only in juvenile rats. Some reports have indicated that SKF83959 can interact with other non-dopamine receptors, including σ1 receptors (Guo et al. 2013; Lee et al. 2014) and 5HT2A,B and C receptor subtypes (Chun et al. 2013). Importantly, in our slice preparation, selective D1R or D2R antagonists blocked SKF83959-mediated potentiation of NMDARs of juvenile OFC neurons. Furthermore, SKF83959 modulation of NMDARs was selective for juvenile, but not for adult OFC neurons. Thus, at least in this system, SKF83959 appears to be modulating NMDARs via dopaminergic receptor activation. Because SKF83959 could activate NMDARs selectively in juvenile, but not in adult rats, and NMDARs in the OFC are required for reversal-learning tasks (Bohn et al. 2003; Calaminus and Hauber 2008; Brigman et al. 2013), we decided to use this compound to test if PLC-coupled dopamine receptors were important for cognitive flexibility. Unfortunately, there are no alternative compounds that can target PLC-coupled dopamine receptors with a high degree of selectivity. Thus, even though we demonstrate an improvement in cognitive flexibility selectively in juvenile rodents with intralateral OFC SKF83959, our in vivo data should be interpreted with caution. Because SKF83959 has also been reported to interact with σ1, 5HT2A,B, and/or C receptors, there could be age-dependent differences in these receptors that may also underlie differences in cognitive performance. Notably, methylphenidate, a dopamine/norepinephrine transporter blocker that can also act at 5HT2B receptors (Markowitz et al. 2006), reduced impulsive responding in juvenile, but not in adult rats (Bizot et al. 2007), suggesting that serotonin receptor expression may be different in juvenile versus adult rats.
Dopamine also had differential effects on NMDARs in the lateral OFC of juvenile and adult rats. Higher concentrations of dopamine caused an excitatory effect at NMDARs in the OFC of juvenile rats, yet an inhibitory response in adults. Inhibition of PLC blocked NMDAR EPSC potentiation by 10 μM, but not 1 μM dopamine in juvenile rats. These data suggest that higher concentrations of dopamine target PLC-coupled dopamine receptors in juvenile rats. Interestingly, potentiation of NMDARs in prelimbic mPFC by 10 μM dopamine was insensitive to inhibition of PKA signaling (Zheng et al. 1999). Future experiments should address whether activation of a PLC-coupled mechanism could also underlie dopamine-mediated potentiation of NMDARs observed in other PFC regions (Zheng et al. 1999).
NMDAR activation in the PFC is implicated in the maintenance of “up-states” in activated cortical cell networks (Kroener et al. 2009). However, the amplitude of cortical pyramidal neuron up-states in juvenile mPFC is smaller than those recorded in the adult mPFC (O'Donnell et al. 2002). Electrical stimulation of VTA dopamine neurons is sufficient to drive “up-state” events in mPFC neurons (Lewis and O'Donnell 2000). Furthermore, coactivation of D1R and NMDARs is sufficient to induce up-states in adult, but not in juvenile PFC slices (Tseng and O'Donnell 2005). Thus, it is feasible that activation of D1R–D2R complexes in juvenile OFC is required for sustaining NMDAR-mediated up-states.
In summary, our results demonstrate that D1Rs and D2Rs act cooperatively to potentiate NMDARs in the lateral OFC of juvenile rats. While dopamine inhibits NMDAR EPSCs in adult animals, promoting cooperative signaling between dopamine receptor subtypes in the lateral OFC of juvenile rats potentiates NMDAR EPSCs and may facilitate the maintenance of cortical networks in an up-state. Furthermore, SKF83959 in juvenile rats improved cognitive flexibility during reversal learning. Targeting age-dependent differences in dopamine receptor coupling may provide clinical utility in the treatment of cognitive deficit disorders in children.
This work was supported by a CIHR MOP:104357, a Parkinson's Society of Canada Young Investigator Award, and a Mind Foundation of British Columbia Young Investigator Award to S.L.B.
Conflict of Interest: None declared.
- signal transduction
- attention-deficit/hyperactivity disorder
- decision making
- phospholipase c
- dopamine receptor
- dopamine d2 receptor
- n-methyl-d-aspartate receptors
- reversal learning
- synaptic transmission
- signal pathway
- signal transduction pathways
- adult attention deficit hyperactivity disorder
- orbitofrontal cortex