Orexin neurons in the lateral hypothalamus (LH)/perifornical area (PFA) are known to promote food intake as well as provide excitatory influence on the dopaminergic reward pathway. Dopamine (DA), in turn, inhibits the reward pathway and food intake through its action in the LH/PFA. However, the cellular mechanism by which DA modulates orexin neurons remains largely unknown. Therefore, we examined the effect of DA on the excitatory neurotransmission to orexin neurons. Whole-cell patch-clamp recordings were performed using acute rat hypothalamic slices, and orexin neurons were identified by their electrophysiological and immunohistochemical characteristics. Pharmacologically isolated action potential-independent miniatures EPSCs (mEPSCs) were monitored. Bath application of DA induced a bidirectional effect on the excitatory synaptic transmission dose dependently. A low dose of DA (1 μm) increased mEPSC frequency, which was blocked by the D1-like receptor antagonist SCH 23390, and mimicked by the D1-like receptor agonist SKF 81297. In contrast, higher doses of DA (10–100 μm) decreased mEPSC frequency, which could be blocked with the D2-like receptor antagonist, sulpiride. Quinpirole, the D2-like receptor agonist, also reduced mEPSC frequency. None of these compounds affected the mEPSCs amplitude, suggesting the locus of action was presynaptic. Furthermore, DA (1 μm) induced an increase in the action potential firing, whereas DA (100 μm) hyperpolarized and ceased the firing of orexin neurons, indicating the effect of DA on excitatory synaptic transmission may influence the activity of the postsynaptic cell. In conclusion, our results suggest that D1- and D2-like receptors have opposing effects on the excitatory presynaptic terminals impinging onto orexin neurons.
Why do we eat beyond our energy demand? This question remains conspicuous as the obesity-related health problems continue to mount around the world. A part of the answer might be that food is a natural reward. There is a growing notion that drug addiction and overeating result from similar reinforcing properties of drugs and food that activate the common brain reward circuitry (Kelley and Berridge, 2002; Cota et al., 2005; Volkow and Wise, 2005).
The lateral hypothalamus (LH) has long been viewed as one of the brain areas involved in reward and feeding (Margules and Olds, 1962). One of the primary cellular substrates for these functions may be orexin neurons. Orexins are expressed mostly in the LH and adjacent perifornical area (PFA), originally identified as orexigenic neuropeptides (Sakurai et al., 1998), and thought to coordinate feeding, sleep-wakefulness, and neuroendocrine and autonomic function (Hara et al., 2001; Ferguson and Samson, 2003). In addition, orexin A induces a selective increase in the intake of palatable food suggesting its role in food reinforcement (Clegg et al., 2002; Thorpe et al., 2005). Recent reports suggest that orexins are also involved in reward and drug dependence (Georgescu et al., 2003; Boutrel et al., 2005; Harris et al., 2005; Borgland et al., 2006; Narita et al., 2006).
Dopamine (DA) is another critical player in food reinforcement: genotypes that alter DA reuptake or receptors have a strong influence on food reinforcement and weight gain (Epstein and Leddy, 2006). The DA system intrinsic to the LH/PFA is known to modulate the activity of mesoaccumbens DA projection, as injection of the D2-like receptor antagonist sulpiride into the LH/PFA results in DA release in the nucleus accumbens (NAcc) (Parada et al., 1995). Furthermore, sulpiride, in the LH/PFA, induces a robust conditioned place preference and locomotor activity, by activation of the mesoaccumbens pathway (Morutto and Phillips, 1998).
DA also has an inhibitory influence on food intake through its action in the LH/PFA. DA receptor activation inhibits feeding (Leibowitz and Rossakis, 1979; Leibowitz et al., 1986; Yang et al., 1997), whereas the D2-like receptor antagonist blocks the anorexic effect of DA in the LH/PFA (Leibowitz, 1975; Leibowitz and Rossakis, 1978; Parada et al., 1988). Furthermore, endogenous DA release is associated with food intake (Meguid et al., 1995; Yang and Meguid, 1995), and different energy states (fasting, obesity, anorexia) modulate DA receptor expression and DA release in this area (Fetissov et al., 2000, 2002; Sato et al., 2001). The anorectic DA action may be caused by inhibition of orexin neurons, because peripheral antipsychotic drugs that increase Fos expression in orexin neurons are known to block DA receptors (Fadel et al., 2002).
Modulation of excitatory synaptic inputs is one mechanism by which the firing activity of the postsynaptic orexin neuron can be altered (Li et al., 2002). Orexin neurons exist in the brain area where injection of excitatory transmitter elicits intense feeding (Stanley et al., 1993) and endogenous glutamate release occurs during meal initiation (Rada et al., 2003). Collectively, DA may modulate excitatory transmission to alter the activity of orexin neurons. Here, we report the effect of DA on excitatory transmission onto orexin neurons.
Materials and Methods
All experiments were performed in accordance with the guidelines established by the Canadian Council on Animal Care and were approved by the Memorial University Internal Animal Care Committee. Attention was paid to use only the number of animals necessary to produce reliable results.
Male Sprague Dawley rats (60–100 g) were decapitated under deep halothane anesthesia. The brain was removed and 250-μm-thick coronal slices containing the hypothalamus were generated at 0–2°C in buffer solution composed of the following (in mm): 87 NaCl, 2.5 KCl, 1.25 NaH2PO4, 7 MgCl2, 0.5 CaCl2, 25 NaHCO3, 25 glucose, 30 sucrose, 3 pyruvic acid, 1 ascorbic acid. Slices were incubated at 33–34°C for 45 min and then at room temperature until the recording in artificial CSF (ACSF) composed of the following (in mm): 126 NaCl, 2.5 KCl, 1.2 NaH2PO4, 1.2 MgCl2, 2 CaCl2, 25 NaHCO3, 10 glucose, 1 ascorbic acid. Both solutions were continuously bubbled with a mixture of O2 (95%) and CO2 (5%).
A hemisected slice was transferred into a recording chamber and perfused at 1.5–2 ml/s with ACSF at 33–34°C. Whole-cell patch-clamp recording was performed using a Multiclamp 700B amplifier and pClamp 9.2 software (Molecular Devices, Sunnyvale, CA). The tip resistance of the recording electrode was 3–7 MΩ when filled with the internal recording solution containing the following (in mm): 123 K-gluconate, 2 MgCl2, 8 KCl, 0.2 EGTA, 10 HEPES, 4 Na2-ATP, 0.3 Na-GTP, pH 7.3. With the visual guidance of infrared-differential interference contrast microscope (DM LFSA; Leica Microsystems), neurons adjacent to the fornix with diameter of 10–20 μm were selected. To characterize the electrophysiological features of recorded cells, the cells were injected with a series of hyperpolarizing and depolarizing 300 ms step pulses in current-clamp mode. The remaining experiments were performed in voltage-clamp mode at a holding potential of −80 mV in the presence of picrotoxin (50 μm) and tetrodotoxin (1 μm) to record miniature EPSCs (mEPSCs), or in current-clamp mode to record spontaneous firing activity in the presence of picrotoxin (50 μm). mEPSCs are non-NMDA receptor mediated because they are sensitive to CNQX (data not shown). After completion of electrophysiological recordings, some brain slices were saved for immunohistochemical studies. Membrane currents were filtered at 1 kHz, digitized at 5 kHz and stored for off-line analysis. A 20 mV hyperpolarizing pulse lasting for 100 ms was applied every 20–60 s throughout each experiment, and the steady-state current and decay rate of the capacitance transient were monitored as measures of input resistance and series/access resistance, respectively. Cells that showed significant change in these parameters were excluded from additional analysis.
Miniature EPSCs were detected by using Mini Analysis 6.0 software (Synaptosoft, Decatur, GA). The data are expressed as mean ± SE. Statistical comparisons were performed by using appropriate tests (i.e., Kolmogorov–Smirnov test for testing individual cells, and unpaired or paired Student's t tests for group comparison as appropriate). A value of p < 0.05 was considered significant.
For immunohistochemical identification of the recorded cell, biocytin (1–1.5 mg/ml) was included in the internal solution. Immediately after the recording, slices were placed in 4% paraformaldehyde in 0.1 m PBS overnight at 4°C, then washed and stored in PBS before the addition of primary antibodies. Slices were incubated with anti-orexin A goat polyclonal IgG (1:3000 dilution; Santa Cruz Biotechnology, Santa Cruz, CA) and anti-melanin-concentrating hormone (MCH) rabbit polyclonal IgG (1:2000 dilution; Phoenix Pharmaceuticals, Belmont, CA) for 3 d at 4°C. Slices were then washed and treated for 3 h with a combination of Cy3-conjugated donkey anti-goat antibody, Cy2-conjugated donkey anti-rabbit antibody, and streptavidin-conjugated aminomethylcoumarin acetate (AMCA), all at 1:500 dilution at room temperature. Antibodies were diluted with PBS with 0.05% Triton X. Slices were then washed, mounted, and examined under a fluorescence microscope for detection of orexin A (Cy3), MCH (Cy2) immunoreactivity, and biocytin (AMCA).
All drugs were bath perfused at final concentrations as indicated, by diluting aliquots of 1000× stock in the ACSF immediately before use. DA stock and the solutions included ascorbic acid (1 mm) and were light protected during the recordings to minimize oxidation. The final concentration of DMSO used as a vehicle was 0.1%. SKF 81297, quinpirole, SCH 23390 and sulpiride were purchased from Tocris Bioscience (Ellisville, MO), dopamine, biocytin, picrotoxin from Sigma-Aldrich (St. Louis, MO), and tetrodotoxin from Alomone Labs (Jerusalem, Israel).
In all of the cells that were identified to be orexin A immunopositive (n = 47), we observed a depolarizing sag (Ih current) in response to hyperpolarizing pulses from the resting membrane potential and a rebound depolarization at the current offset (Fig. 1A). In some cells, the rebound was large enough to evoke a spike. In addition, the majority of these cells displayed minimal spike adaptation (76.6%, 36 of 47) and spontaneous firing at rest (89.4%, 42 of 47) (Fig. 1A). The average resting membrane potential for orexin A-immunopositive neurons was −48.6 ± 0.8 mV. In contrast, as shown in Fig. 1C, MCH-immunopositive neurons showed no Ih current (100%, nine of nine) nor rebound depolarization (100%, nine of nine). The average resting membrane potential for MCH neurons was −58.8 ± 1.9 mV, and a majority of them did not fire spontaneously (88.9%, eight of nine). Furthermore, eight of nine cells showed a clear spike adaptation (88.9%). Thus, our result is consistent with that of the previous studies (Eggermann et al., 2003; Gao et al., 2003; Burdakov et al., 2004; Jo et al., 2005). For the electrophysiological study, a total of 39 neurons that displayed the characteristics of orexin neurons (relatively depolarized resting membrane potential, spontaneous firing activity, Ih, rebound depolarization, and minimal spike adaptation) were investigated, of which seven cells were identified as orexin A immunopositive.
DA application induced a reversible change in the frequency of spontaneous excitatory transmission in orexin neurons (Fig. 2). The direction of the effect was dose dependent: 1 μm induced a significant increase (n = 9; p < 0.05) (Fig. 2A1–A3,C), whereas 10 or 100 μm elicited a decrease in mEPSC frequency (10 μm, n = 5, p < 0.05; 100 μm, n = 9, p < 0.01) (Fig. 2B1–B3,C). A concentration of 0.1 μm had no significant effect in four cells tested (p > 0.05) (Fig. 2C). In contrast, there was no significant change in the amplitude of mEPSCs even when there was a change in the frequency, indicating that the locus of DA action is presynaptic [1 μm, n = 5, p > 0.05 (Fig. 2A4); 10–100 μm, n = 6, p > 0.05 (Fig. 2B4)].
It is possible that the bidirectional effect of DA on mEPSCs results from activation of different subtypes of DA receptors. There are two classes of DA receptors, namely D1-like (D1/D5) and D2-like receptors (D2/D3/D4). To test whether these two classes of DA receptors mediate differential effects, DA receptor subtype specific agonists were examined. SKF 81297 (10 μm), the D1-like receptor agonist, significantly increased the frequency of mEPSCs (n = 9, p < 0.05) (Fig. 3A,B1,B2,D) without affecting the amplitude (p > 0.05) (Fig. 3B3). Because 1 μm DA induced an increase in mEPSCs, this effect may be caused by activation of D1-like receptors. In support of this idea, the D1-like receptor antagonist SCH 23390 (10 μm) not only blocked 1 μm DA-induced facilitation of mEPSCs, but also inhibited mEPSCs (n = 5, p < 0.05) (Fig. 3C,D). In addition, the effects of 1 μm DA and SKF 81297 were not significantly different (p > 0.05). In contrast, quinpirole (10–50 μm), the D2-like agonist, reduced the frequency of mEPSC (n = 6, p < 0.05) (Fig. 4A,B1,B2,D) but not the amplitude (p > 0.05) (Fig. 4B3). This inhibitory effect was similar to that of high dose DA, therefore, we tested the effect of the D2-like receptor antagonist on 100 μm DA-induced modulation. In the presence of sulpiride (10 μm), 100 μm DA had no effect on the frequency of mEPSCs (n = 4, p > 0.05) (Fig. 4C,D). Furthermore, the effect of DA 100 μm was not significantly different from that of quinpirole (p > 0.05). Thus, inhibitory effect of high concentration of DA seems to be mediated by D2-like receptors.
Next, we asked the question whether the effects of DA on excitatory transmission could alter the firing activity of orexin neurons. In current-clamp mode, spontaneous action potentials were monitored in the presence of picrotoxin to block the influence of inhibitory synaptic inputs. In this condition, 1 μm DA significantly increased the frequency of action potential firing in all the cells tested (n = 4, p < 0.05) (Fig. 5A1–A3). In contrast, 100 μm DA hyperpolarized and ceased the firing activity of orexin neurons (n = 3, p < 0.05) (Fig. 5B1–B3).
In some of the orexin neurons examined, multiple experiments were performed by sequentially applying compounds having opposite effects (agonists or different doses of DA). Figure 6 depicts examples of these cells. Every cell tested was able to respond in both directions: D1 receptor activation resulted in increase in mEPSC or action potential frequency, whereas D2 receptor activation caused reduction in mEPSC or action potentials. Therefore, both D1 and D2 receptors can modulate the excitatory synapses that impinge on a single orexin neuron.
The present study demonstrates that DA modulates excitatory synaptic transmission in a dose-dependent and reversible manner in orexin neurons. The direction of modulation depends on distinct types of receptors. D1-like receptors mediate the effect of low dose DA (1 μm) that induces facilitation whereas D2-like receptors mediate the effect of higher doses (10–100 μm) that diminish the frequency of spontaneous excitatory transmission. These changes occurred in the absence of any alteration in the amplitude of mEPSCs, indicating that DA affects the transmitter release probability at the presynaptic terminal, but does not change the postsynaptic sensitivity. Furthermore, low and high doses of DA application results in an increase and decrease in action potential firing of orexin neurons, respectively. These seemingly opposite effects of DA can be observed in the same postsynaptic neuron. Thus, our results suggest that D1- and D2-like receptors can exert opposite effects on synapses converging onto the same neuron.
The hyperpolarizing effect of DA on orexin neurons is in agreement with the previous reports that used high doses (30–300 μm) (Li and van Den Pol, 2005; Yamanaka et al., 2006), resulting in cessation of firing activity (Li and van Den Pol, 2005) and reduction in intracellular calcium (Tsujino et al., 2005). Inhibition of TTX-insensitive spontaneous glutamatergic transmission shown in the present study may at least partially account for the reduced activity of orexin neurons. A direct postsynaptic effect causing a sustained outward current (Yamanaka et al., 2006) is also a potential mechanism, although we did not observe any sustained currents in the presence of DA in our preparation. This discrepancy is probably because of a lower dose used in our study [100 μm vs 300 μm (Yamanaka et al., 2006)] and different holding potential (−80 vs −60 mV). Our results suggest that the excitatory effect of 1 μm DA on spontaneous excitatory synaptic transmission also translates into the altered firing activity of orexin neurons. Indeed, it has been shown that enhancing the spontaneous excitatory transmission increases the firing rate of orexin neurons (Li et al., 2002).
The dose-dependent bidirectional effect of DA is somewhat similar to that seen in the inhibitory transmission and NMDA receptor modulation in the prefrontal cortex (PFC). In the PFC slice, DA has been shown to induce a biphasic effect on the amplitude of evoked IPSCs: D2-like mediated inhibition was followed by a long-lasting D1-like receptor-mediated increase, with the D1-mediated effect apparent at lower a dose (Seamans et al., 2001; Trantham-Davidson et al., 2004). Zheng et al. (1999) demonstrated that NMDA receptor currents were similarly modulated, with low concentration of DA inducing enhancement and high concentration inducing inhibition. The mechanism for differential dose dependency of D1 and D2 receptor-mediated modulation is unknown. It may involve different affinity state of the receptors (Seeman and Van Tol, 1993) and/or differential expression of D2 receptor isoforms that have distinct impact on excitatory transmission (Centonze et al., 2004).
The mechanisms underlying the dopaminergic modulation in the PFC and orexin neuron are likely to be different. In the PFC, presynaptic D1 receptors cause a long-lasting increase in GABA release, and postsynaptic D2 receptors modulate the phosphorylation state of postsynaptic GABAA receptors (Trantham-Davidson et al., 2004). Also, DA can modulate postsynaptic glutamatergic receptors such as AMPA receptor synaptic expression (Sun et al., 2005; Zou et al., 2005) or NMDA receptor function (Zheng et al., 1999). In orexin neurons, we did not observe any change in the amplitude of mEPSCs indicating no effect on synaptic glutamatergic receptors. In addition, the effect of the D1 receptor was readily reversible after washout of the ligand. At the presynaptic terminals impinging onto orexin neurons, D1- and D2-like receptors may be exerting opposing effects on the same signaling pathway. D1 receptor activation is known to positively affect adenylyl cyclase, whereas D2 receptor activation negatively affects it (Missale et al., 1998). Activation of adenylyl cyclase and subsequent activation of cAMP-protein kinase A signaling, in turn, may facilitate spontaneous neurotransmitter release, as shown in number of other synapses including those in the hypothalamus (Chavez-Noriega and Stevens, 1994; Chen and Regehr, 1997; Hirasawa and Pittman, 2003).
There is a functional and anatomical reciprocal communication between the LH/PFA and the mesolimbic DA system. A bulk of the DA input to the LH/PFA originate from the midbrain A8, A9, and A10 cell groups (Leibowitz and Brown, 1980; Yoshida et al., 2006) and mediate DA modulation of feeding behavior (Leibowitz and Brown, 1980). In return, orexin neurons send direct projections to the ventral tegmental area (VTA) (Fadel and Deutch, 2002) and exhibit excitatory effect there (Korotkova et al., 2003; Harris et al., 2005; Borgland et al., 2006; Vittoz and Berridge, 2006) leading to DA release in NAcc (Narita et al., 2006) and PFC (Vittoz and Berridge, 2006). Also, orexins may exert their influence through the hypothalamo-midline thalamic-striatal pathway (Kelley et al., 2005). Furthermore, NAcc inhibition results in activation of orexin neurons (Zheng et al., 2003; Baldo et al., 2004). NAcc inhibition also elicits intense feeding in satiated rats (Maldonado-Irizarry et al., 1995; Stratford and Kelley, 1997; Stratford et al., 1998) which is largely abolished by NMDA antagonist in the LH (Stratford and Kelley, 1999).
Based on the current study, in addition to previous reports demonstrating orexin neurons modulating the mesoaccumbens DA projection, we propose a dose-dependent dual feedback mechanism between the two systems. A low/moderate level of DA in the LH/PFA may excite orexin neurons through D1 receptor-mediated facilitation of excitatory input, which in turn provides excitatory influence on VTA neurons, thus creating a positive feedback mechanism. When dopaminergic activity elevates and higher concentration of DA is achieved in the hypothalamus, this will work to activate D2 receptors and inhibit orexin neurons. This will lead to decreased excitatory input to the VTA, acting as a negative feedback mechanism. This idea is supported by neurochemical and behavioral studies demonstrating that injection of sulpiride, the D2-like antagonist, into the LH/PFA induces DA efflux in the NAcc leading to locomotion and feeding (Parada et al., 1995; Morutto and Phillips, 1998).
In conclusion, the DA effect on excitatory transmission onto orexin neurons shown in this study may be one of the mechanisms by which DA intrinsic to the LH/PFA influences feeding, reward, and locomotion. It has been shown that D1 and D2 receptor agonists normalize hyperphagia as well as rectify metabolic and endocrine abnormalities independent of food intake, resulting in a improvement of obese-diabetic syndrome in leptin deficient ob/ob mice (Scislowski et al., 1999). Perhaps disruption of the physiological DA function at the level of orexin neurons partially accounts for excessive food intake and disruption of metabolism leading to obesity.
This work was supported by the Canadian Institutes of Health Research (CIHR) and Natural Sciences and Engineering Research Council of Canada. M.H. is a CIHR New Investigator.
- Correspondence should be addressed to Dr. Michiru Hirasawa, Division of Basic Medical Sciences, Faculty of Medicine, Memorial University of Newfoundland, 300 Prince Philip Drive, St. John's Campus, Newfoundland, Canada A1B 3V6.