Synaptic properties and postsynaptic opioid effects in rat central amygdala neurons

Neuroscience 127 (2004) 871– 879

SYNAPTIC PROPERTIES AND POSTSYNAPTIC OPIOID EFFECTS IN RAT CENTRAL AMYGDALA NEURONS W. ZHUa AND Z. Z. PANa,b*

Increasing evidence suggests that the amygdala plays a crucial role in the integration and control of emotional responses, autonomic behaviors, neuroendocrine activity, and other cognitive functions such as anxiety and attention (Aggleton, 1993; Gallagher and Chiba, 1996; Cahill and McGaugh, 1998; Davis, 2000; Sah et al., 2003). Recent research has clearly established that the amygdala mediates negative emotional behaviors, especially the expression of conditioned fear responses, in both humans and animals (McKernan and Shinnick-Gallagher, 1997; Rogan et al., 1997; Garcia et al., 1999). More recently, studies also have suggested a critical role of the amygdala, particularly the central nucleus of the amygdala (CeA), in the positive emotional events represented by the reward function of the brain (Baxter and Murray, 2002; Gottfried et al., 2003; See et al., 2003). Ample evidence implicates the CeA in learning stimulus-reward responses and in mediating the motivational effects of drugs of abuse, such as opioids and alcohol (Koob et al., 1998). In addition, the CeA has attracted increasing attention due to its important role in mediating the analgesia induced by environmentally stressful conditions such as fear (environmental analgesia; Helmstetter, 1992; Aggleton, 1993; Fox and Sorenson, 1994). While the basolateral complex of the amygdala (BLA) is believed to converge information on conditioned and unconditioned environmental stimuli and to modulate the process of memory consolidation (Pitkanen et al., 1997; McGaugh, 2002), the CeA is thought to receive integrated information from other amygdala subregions including the BLA. The CeA produces its actions through extensive efferent projections to the basal forebrain, hypothalamus, midbrain, and brainstem nuclei that mediate fear response, reward behavior, and environmental analgesia (Pitkanen et al., 1997; Swanson and Petrovich, 1998; Davis, 2000). To understand the mechanisms underlying these CeA functions, it is essential to know the physiological properties and synaptic connections of neurons in the CeA and particularly, their expression profile of opioid receptors for understanding of their roles in reward behaviors and environmental analgesia. Nevertheless, little data regarding the expression of opioid receptors, and limited data regarding the fundamental properties of different types of CeA neurons, are available. In rats, CeA neurons have generally been divided into two types according to the shape or the firing pattern of their action potentials (Schiess et al., 1999; Dumont et al., 2002). A direct connection between the BLA and the CeA also has been shown both anatomically and physiologically (Nose et al., 1991; Pitkanen et al., 1997; Swanson and Petrovich, 1998; Collins and Pare,

a

Department of Symptom Research, Unit 110, The University of Texas-MD Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030, USA b Department of Biochemistry and Molecular Biology, Unit 110, The University of Texas-MD Anderson Cancer Center, Houston, TX 77030, USA

Abstract—An important output of amygdaloid nuclei, the central nucleus of the amygdala (CeA) not only mediates negative emotional behaviors, but also participates in the stimulus–reward learning and expression of motivational aspects of many drugs of abuse, and links environmentally stressful conditions such as fear to endogenous paininhibiting mechanisms. The endogenous opioid system in the CeA is crucial for both reward behaviors and environmental stress-induced analgesia. In this study using whole-cell voltage-clamp recordings, we investigated synaptic inputs and the postsynaptic effects of opioid agonists in CeA neurons. We found that synaptic inputs evoked within the CeA were mediated by both glutamate and GABA, but those evoked from the basolateral amygdala were primarily glutamatergic. Based on membrane properties, three types of cells were characterized. Type A neurons had no spike accommodation while type B neurons displayed characteristic accommodating response. Type A neurons were further classified as either A1 or A2, based on differences in resting membrane potential and the amplitude of after-hyperpolarizing potential. ␮-Opioid receptor agonists hyperpolarized a subpopulation of CeA neurons, of which the vast majority was type A1. This ␮ agonist-induced hyperpolarization was mediated by the opening of inwardly rectifying potassium channels. In contrast, the ␬-opioid receptor agonist hyperpolarized only type B neurons. These results illustrate three types of CeA neurons with distinctive membrane properties and differential responses to opioid agonists. They may represent functionally distinct CeA cell groups for the integration and execution of CeA outputs in the aforementioned CeA functions. © 2004 IBRO. Published by Elsevier Ltd. All rights reserved. Key words: ␮ receptor, ␬ receptor, glutamate synapses, potassium channels. *Correspondence to: Z. Z. Pan, Department of Symptom Research, Unit 110, UT-MD Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030, USA. Tel: ⫹1-713-792-5559; fax: ⫹1-713-745-4754. E-mail address: [email protected] (Z. Z. Pan). Abbreviations: ADP, after-depolarizing potential; AHP, afterhyperpolarizing potential; AP5, 2-amino-5-phosphonovaleric acid; BLA, basolateral complex of the amygdala; CeA, central nucleus of the amygdala; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione; DAMGO, [D-Ala2,N-Me-Phe4,Gly-ol5]-enkephalin; EPSC, excitatory postsynaptic current; GIRK, G-protein-coupled inwardly rectifying potassium (channels); ICMs, intercalated cell masses; IPSC, inhibitory postsynaptic current; [K⫹]o, extracellular potassium concentration; ME, methionine-enkephalin; NMDA, N-methyl-D-aspartate.

0306-4522/04$30.00⫹0.00 © 2004 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2004.05.043

871

872

W. Zhu and Z. Z. Pan / Neuroscience 127 (2004) 871– 879

1999; Sah et al., 2003). In this study, we first examined the membrane and synaptic properties of different types of neurons in the CeA, and then investigated the postsynaptic effects of selective ␮- and ␬-opioid receptor agonists on each neuron type.

EXPERIMENTAL PROCEDURES All procedures involving the use of animals conformed to the guidelines set by NIH and by University of Texas-MD Anderson Cancer Center Animal Care and Use Committee. Appropriate anesthesia was used in all applicable procedures to minimize pain or discomfort to animals. In vitro preparations were used whenever possible and recordings of several cells in a single slice preparation were made to minimize the number of animals used. The methods used in this study were similar to those published previously (Pan et al., 1997).

Brain slice preparations Male, Wistar rats (150 –200 g) were used for recordings in brain slice preparations in vitro. A rat was anesthetized with inhalation of halothane and then killed by decapitation. The brain was removed and cut in a vibratome in cold (4 °C) physiological saline to obtain coronal slices (200 or 300 ␮m thick) containing both the CeA and the BLA. A single slice was submerged in a shallow recording chamber and perfused with preheated (35 °C) physiological saline (in mM: NaCl, 126; KCl, 2.5; NaH2PO4, 1.2; MgCl2, 1.2; CaCl2, 2.4; glucose, 11; NaHCO3, 25, saturated with 95% O2 and 5% CO2, pH 7.2–7.4). Slices were maintained at around 35 °C throughout the experiment.

Whole-cell recordings and data analyses CeA cells were visualized on a video monitor with a Normaski microscope (Olympus America, Melville, NY, USA). Visualized whole-cell voltage-clamp recordings were made from identified CeA neurons with a glass pipette (resistance 3–5 M⍀) filled with a solution containing (mM): potassium gluconate, 126; NaCl, 10; MgCl2, 1; EGTA, 11; HEPES, 10; ATP, 2; GTP, 0.25; pH adjusted to 7.3 with KOH; osmolarity 280 –290 mosmol/L. In some experiments, potassium gluconate was replaced by an equal concentration of CsCl. An Axopatch 1-D amplifier and Axograph software (Axon Instruments, Inc., Union City, CA, USA) were used for data acquisition and on-line/off-line data analyses. A seal resistance of 2 G⍀ or above and an access resistance of 15 M⍀ or less were considered acceptable. Series resistance was optimally compensated. The access resistance was monitored throughout the experiment. Junction potential was not corrected. Electrical stimuli of constant current (0.25 ms, 0.2– 0.5 mA) were used to evoke postsynaptic currents with bipolar stimulating electrodes (FHC Inc., Bodowinham, ME, USA) placed either in the ventrolateral part of the CeA, or in the center of the BLA. Numeral data were statistically analyzed with Student’s t-test and are presented as mean⫾S.E.M. All CeA cells recorded were classified into either type A1, type A2, or type B, based on their resting membrane potential, amplitude of after-depolarization potential, and occurrence of spike accommodation. All drugs were purchased from Sigma Co. (St. Louis, MO, USA) or Research Biochemicals, Inc. (Natick, MA, USA) and were applied through the bath solution unless stated otherwise.

RESULTS Visualized whole-cell voltage-clamp recordings were made from identified neurons in the CeA in brain slice preparations in vitro. Most (⬎80%) recorded neurons were located

in the medial CeA and the rest in the lateral CeA. Since no difference was observed in either the distribution of neuron types or their responses to opioid agonists, data from these two subregions of the CeA were pooled. These neurons generally displayed no spontaneous firing activity in our recording conditions. Three cell types An obvious difference among CeA neurons in our recording conditions was spike accommodation. Of 53 neurons, 46 (87%) had no spike accommodation and seven (13%) displayed an accommodating response to a prolonged depolarizing current step (Fig. 1). This is consistent with a previous study in which CeA neurons having no accommodation were classified as type A and those displaying spike accommodation as type B (Schiess et al., 1999). In the current study, we found significantly different membrane properties in the type A neurons and therefore further classified them into two types, type A1 and type A2 (Table 1). Thus, type A1 neurons (22 of 53, 42%) had a characteristic large after-depolarizing potential (ADP) whose peak surpassed the threshold of the action potential (Fig. 1A, upper panel). The ADP was followed by a relatively slow after-hyperpolarizing potential (AHP) with a time constant of 155⫾12 ms for the decay phase (n⫽22). These cells displayed an average resting membrane potential of ⫺70 mV in 2.5 mM external potassium concentration (Table 1). By contrast, type A2 neurons (24 of 53, 45%) displayed a smaller ADP that never reached the threshold of the action potential (Fig. 1B, upper panel). The ADP was also followed by an AHP, but the AHP decay in type A2 neurons was significantly faster than that in type A1 neurons (Table 1). Furthermore, type A2 neurons had significantly less-negative resting membrane potentials (⫺62 mV) than type A1 neurons (Table 1). Nevertheless, there were no apparent differences between type A1 and type A2 cells in either the shape, which were mostly ovoid or fusiform, or the size of cell bodies (A1: 14.5⫾0.8⫻ 10.6⫾0.5 ␮M; A2: 15.3⫾0.8⫻9.7⫾0.4 ␮M). Type B neurons (7 of 53, 13%) had no or minimal ADP (Fig. 1C). Morphologically, although their pyramiform or ovoid shape could not be clearly distinguished from those of type A cells, type B cells were significantly larger (23.1⫾0.8⫻ 13.1⫾0.9 ␮M, P⬍0.01). In addition, type B cells displayed a significantly higher input membrane resistance than type A cells (Table 1). Finally, the frequency of evoked action potentials from resting membrane potential was significantly different among the three types of neurons with type A1 having the highest rate (Fig. 1, lower panels and Table 1). We found no significant difference among the three types of neurons in the width of action potentials or in the AHP amplitude. These results demonstrate that three types of CeA neurons can be identified based on their distinct characteristics in ADP amplitude, AHP decay time course, resting membrane potential and spike accommodation. More importantly, these three types of neurons generally possessed differential opioid receptor profiles and pharmacological responses to opioid agonists (see below).

W. Zhu and Z. Z. Pan / Neuroscience 127 (2004) 871– 879

873

Fig. 1. Three types of neurons in the CeA. Single (upper panels) and multiple (lower panels) action potentials were evoked from the resting membrane potential by depolarizing current pulses. Dotted lines indicate the action potential threshold.

Synaptic properties We then examined whether there were significant differences in synaptic inputs activated electrically within the CeA among the three neuron types. We found no such differences and therefore pooled the results. We also studied their synaptic inputs evoked from the BLA, as BLA is one of the main external sources of synaptic inputs in CeA neurons (Pitkanen et al., 1997; Swanson and Petrovich, 1998). When the stimulating electrode was placed in the

ventrolateral part of the CeA, single electrical stimulation evoked an excitatory postsynaptic current (EPSC) followed by an inhibitory postsynaptic current (IPSC) in every neuron recorded under voltage clamp with a holding potential of ⫺50 mV (Fig. 2A). Bath application of 6-cyano-7nitroquinoxaline-2,3-dione (CNQX; 10 ␮M) and 2-amino-5phosphonovaleric acid (AP5; 10 ␮M) completely abolished the EPSC, leaving an IPSC of 44⫾5 pA in amplitude (n⫽13). Bicuculline (10 ␮M), an antagonist of type A GABA

Table 1. Properties of the three types of neurons in the CeAa Cell type

Number (%)

Action potential (AP) ADP peak

A1

22/53 (42%)

A2

24/53 (45%)

B

7/53 (13%)

Above AP threshold Below AP threshold Small or none

AHP

Half width (ms)

Rate (Hz)

Accom

Resting membrane potential (mV)

Rin (M⍀)

␶ (ms)

Amplitude (mV)

155⫾12**

24.1⫾0.8

0.6⫾0.02

29.2⫾1.2**

No

⫺70.3⫾0.7**

265⫾8

75⫾16**

23.7⫾0.8

0.6⫾0.01

19.1⫾0.8**

No

⫺61.9⫾0.6**

248⫾10

73⫾10

29.2⫾1.6

0.7⫾0.04

4.5⫾0.7**

Yes

⫺71.5⫾1.5

566⫾19**

a Accom, accommodation. Rin, input resistance. AHP time constants were obtained by fitting the decay phase with a single exponential. AHP amplitudes were measured from the action potential threshold. ** P⬍0.01.

874

W. Zhu and Z. Z. Pan / Neuroscience 127 (2004) 871– 879

Fig. 2. Synaptic currents activated by electrical stimulation within the CeA in CeA neurons under voltage clamp. (A) Synaptic response of a CeA neuron held at ⫺50 mV, showing a glutamate receptor-mediated EPSC followed by a GABAA receptor-mediated IPSC. (B) EPSC mediated by the non-NMDA glutamate receptor. Holding potential⫽⫺50 mV. (C) EPSC mediated by the NMDA glutamate receptor in Mg⫹⫹-free solution. Holding potential⫽⫺70 mV.

receptors, blocked the IPSC (n⫽13; Fig. 2A), suggesting that this IPSC is predominantly mediated by GABAA receptors under our experimental conditions. To pharmacologically isolate the EPSC for further investigation, the following experiments were carried out with bicuculline included in the bath solution unless stated otherwise. The amplitude of the EPSC was largely blocked by CNQX (10 ␮M, control, 109⫹8 pA, CNQX, 4⫾1 pA; n⫽11; Fig. 2B), suggesting that the EPSC was primarily mediated by non-N-methyl-D-aspartate (NMDA) glutamate receptors in normal conditions. To determine whether NMDA receptors were also involved in the excitatory synaptic transmission, recordings were then made in a Mg⫹⫹-free solution with a holding potential of ⫺70 mV. In such recording conditions and after blockade of the non-NMDA component of the EPSC with CNQX, an AP5-sensitive, NMDA receptormediated EPSC was recorded in all CeA neurons tested (46⫾8 pA; n⫽11; Fig. 2C). These results are in general agreement with previous studies under similar experimental conditions (Nose et al., 1991; Sah et al., 2003). However, when the stimulating electrode was placed in the center of the BLA, the synaptic response of CeA neurons held at the same potential (⫺50 mV) was different from that elicited by stimulation within the CeA. First, BLA stimulation evoked a predominant EPSC without a detectable IPSC in every CeA neuron tested in the absence of bicuculline (n⫽10; Fig. 3A). Second, not every CeA neuron displayed an NMDA component in the EPSC evoked by BLA stimulation. The non-NMDA component similarly was present in all BLA-elicited EPSCs and predominated in normal conditions, as CNQX (10 ␮M) largely blocked the

amplitude of the EPSC (control, 96⫾9 pA, CNQX, 3⫾1 pA; n⫽5; Fig. 3A). However, in Mg⫹⫹-free solution and in the presence of CNQX, and with BLA stimulation of comparable intensity, the NMDA receptor-mediated EPSC was recorded in only two of five cells (40%; an average of 28 pA, holding potential⫽⫺70 mV; Fig. 3B). In the other three cells, CNQX blocked the EPSC with no detectable NMDA component. To further confirm the incidence of fewer NMDA receptor-mediated EPSCs elicited from the BLA, CsCl-filled pipettes were used to block all potassium channels and EPSCs were evoked with a holding potential of ⫹40 mV to completely remove Mg⫹⫹ blockade of NMDA channels (Sah et al., 2003). Under these conditions, the non-NMDA EPSC was again recorded in all cells tested, but the NMDA EPSC was present only in six of nine cells (67%) in the presence of CNQX (123⫾12 pA; Fig. 3C). In the other three cells, CNQX largely blocked the BLAelicited EPSC with little NMDA component remaining (Fig. 3D). Additionally, this NMDA-lacking EPSC did not appear to be restricted to any single cell type because the NMDA EPSCs were observed only in three of six type A1 cells, three of five type A2 cells, and two of three type B cells. It has been reported that a GABA IPSC can be evoked through BLA stimulation likely by activating the GABAergic cells in the intercalated cell masses (ICMs) interposed between the BLA and the CeA, and projecting to CeA neurons (Nose et al., 1991; Royer et al., 1999). To determine whether that could occur in our preparations, we increased the stimulation intensity and thus enlarged the activated area to include possibly ICMs. In six additional cells, BLA stimulation of usual intensities (0.51⫾0.09 mA)

W. Zhu and Z. Z. Pan / Neuroscience 127 (2004) 871– 879

875

Fig. 3. Synaptic currents activated by electrical stimulation in the BLA in CeA neurons under voltage clamp. (A) Synaptic response that was primarily mediated by the non-NMDA receptor in normal conditions (holding potential⫽⫺50 mV). (B) NMDA receptor-mediated EPSC in Mg⫹⫹-free solution (holding potential⫽⫺70 mV). (C) NMDA EPSC recorded with a CsCl-filled pipette and a holding potential of ⫹40 mV. (D) Synaptic response of another cell recorded with a CsCl-filled pipette, showing a non-NMDA EPSC with little NMDA component (holding potential⫽⫹40 mV). Experiments in B, C and D were carried out in the presence of bicuculline (10 ␮M).

evoked only EPSCs and no IPSCs. When the intensity was increased by 10 times (5.75⫾0.86 mA), a bicucullinesensitive IPSC was evoked in all six cells, consistent with the previous reports. These findings suggest that there is a differential distribution in the origin of glutamate and GABA synaptic inputs in CeA neurons with a predominant glutamatergic input directly from the BLA. In addition, it appears that some of these glutamatergic synapses of BLA origin lack apparent involvement of NMDA receptors. Postsynaptic effects of opioid agonists Next, we examined the pharmacological profile of each CeA neuron type for opioid receptors and their responses to corresponding opioid agonists. In all 53 cells of the three types shown in Table 1, we tested the effects of both ␮and ␬-selective agonists in the same cells. We found that in the great majority of type A1 neurons (18 of 22, 82%) under voltage clamp, the ␮-/␦-receptor agonist methionineenkephalin (ME; 10 ␮M) produced an outward current from

a holding potential of ⫺70 mV (Fig. 4A; Table 2). This ME effect was not affected by TTX (1 ␮M, n⫽4), but was blocked in cells loaded with CsCl through the recording pipette (n⫽4). No type A1 cells responded to U69593 (300 nM, n⫽22), a selective ␬-receptor agonist (Bie and Pan, 2003). In contrast, the majority of type A2 neurons (20 of 24, 83%) were not affected by ME (10 ␮M). The remaining four cells were inhibited by ME (Table 2). As with type A1 neurons, none of the type A2 neurons was inhibited by the ␬ agonist (300 nM, n⫽24). However, of the seven type B neurons that were not affected by ME, U69593 inhibited four of them (57%) with an outward current (Table 2). In the type A1 neurons inhibited by ME, the selective ␮-receptor agonist [D-Ala2,N-Me-Phe4,Gly-ol5]-enkephalin (DAMGO; 1 ␮M) also produced an inhibition with an average outward current of 29.1⫾7.5 pA (n⫽5; Fig. 4B). The DAMGOinduced current was completely blocked by the ␮-receptor antagonist CTAP (1 ␮M), suggesting that the inhibition of type A1 neurons is mediated by the ␮-opioid receptor. ␦-Opioid receptor agonist [D-Pen2,5]-enkephalin (1 ␮M)

876

W. Zhu and Z. Z. Pan / Neuroscience 127 (2004) 871– 879

To further identify the ionic mechanism for the ␮ receptor-mediated inhibition of type A1 neurons, we analyzed the current-voltage relationship of ME-induced currents in different concentrations of extracellular potassium ([K⫹]o). In normal, 2.5 mM [K⫹]o, the ME-induced current reversed its polarity at ⫺98.0⫾1.6 mV (n⫽10) with an increased membrane conductance (increased current slope) and clear inward rectification (Fig. 5). When the [K⫹]o was increased to 6.5 mM and 10.5 mM by adding more KCl to the external solution, the reversal potential of the ME current was shifted to ⫺78.7⫾1.2 mV (n⫽8) and ⫺63.5⫾1.2 mV, respectively (Fig. 5B). The shift in the reversal potentials by increasing [K⫹]o had a slope of ⫺58.4, in accordance with the Nernst equation (Fig. 5D). These results clearly show that activation of ␮-opioid receptors inhibits the vast majority of type A1 neurons predominantly by opening the G-protein-coupled inwardly rectifying potassium (GIRK) channels. This ␮ receptormediated inhibition through activation of the GIRK channel is consistent with previous reports of various central neurons (North et al., 1987; Kovoor et al., 1995; Osborne et al., 1996; Pan et al., 1997; Svoboda and Lupica, 1998; Alvarez et al., 2002).

DISCUSSION

Fig. 4. ␮-Opioid receptor-mediated inhibition of type A1 neurons. (A) An outward membrane current induced by ME in a neuron under voltage clamp. (B) The ␮ agonist DAMGO-induced outward current (upper trace) and its blockade by the ␮ antagonist CTAP (lower trace) in another neuron. Both traces were from the same neuron.

had no significant postsynaptic effect on the ME-sensitive type A1 neurons (n⫽4), or type A2 neurons (n⫽3), or type B neurons (n⫽2). Furthermore, the ␦ antagonist naltrindole failed to affect the inhibition mediated by ME, which produced an outward current of 31.8⫾2 pA in the presence of 1 ␮M naltrindole in type A1 neuron (n⫽4). These findings suggest that the majority of type A1 neurons in the CeA express postsynaptic ␮-opioid receptors and can be inhibited by ␮ agonists, while most type A2 neurons and all type B neurons do not contain the ␮-receptor. They also indicate that only type B CeA neurons have postsynaptic ␬-opioid receptors and are sensitive to ␬ agonists.

In the present study, we have demonstrated that CeA neurons can be generally classified into three types that possess differential physiological and pharmacological properties. Most importantly, type A1 neurons, a large cell population in the CeA, have a large ADP and are hyperpolarized, in the great majority, by agonists of ␮-opioid receptors through the opening of GIRK channels. In contrast, the ␬ agonist inhibits exclusively type B neurons, which are characterized by their unique spike accommodation. Our data also indicate that synaptic inputs directly from the BLA to CeA neurons are primarily glutamatergic, with no apparent involvement of NMDA receptors in some of these synapses. Cell types and synaptic transmission in the CeA Generally, two types of neurons have been described in the rat CeA according either to their membrane properties or to their firing patterns. Based primarily on the occurrence of spike accommodation with intracellular recording, Schiess et al. (1999) classified CeA neurons as either type A (no accommodation) or type B (with accommodation).

Table 2. Effects of opioid agonists on the three types of CeA neuronsa Cell type (N)

ME effects

U69593 effects

Yes

A1 (22) A2 (24) B (7) a

N, cell number.

No (N) (%)

N (%)

Current (pA)

18 (82%) 4 (17%) 0

30⫾3 20⫾2

Yes (N) (%)

4 (18%) 20 (83%) 7 (100%)

0 0 4 (57%)

No (N) (%) Current (pA)

19⫾2

22 (100%) 24 (100%) 3 (43%)

W. Zhu and Z. Z. Pan / Neuroscience 127 (2004) 871– 879

877

Fig. 5. The inwardly rectifying potassium currents induced by activation of ␮-opioid receptors in type A1 neurons. (A) Representative current traces during voltage steps from a holding potential of ⫺70 mV to various potentials in control and in the presence of ME (10 ␮M). (B) Current-voltage plots in control and in ME in two [K⫹]o concentrations. Note the shift in the reversal potential by the higher [K⫹]o. Currents in 6.5 mM [K⫹]o are not shown for figure clarity. (C) The ME-induced current after subtracting the control current shown in B in 2.5 mM [K⫹]o, displaying strong inward rectification. (D) A Nernst plot of reversal potentials in three [K⫹]o.

Our results are consistent with this classification of CeA cells and their general morphology, showing that nonaccommodating type A cells are smaller than accommodating type B cells, but the two types have overlapping shapes. Examining a single action potential at an extended time course of 1 s with whole-cell recording, we found that the type A neurons had significant differences in fundamental membrane properties and therefore could be further classified into types A1 and A2. Type A1 neurons significantly differ from type A2 neurons in many physiological properties, including more-negative resting membrane potential, ADP above the threshold for action potential, a higher firing rate of evoked action potentials, and a faster-decaying AHP (Table 1). The differences in the resting membrane potential and the shape of action potential, particularly AHP, between our results and those of the Schiess et al. (1999) study can be largely attributed to different recording configurations. Also using whole-cell recording, Dumont et al. (2002) have described two major cell types in rat CeA: low-threshold bursting type and regular spiking type, based on their responses to a series of hyperpolarizing and depolarizing current pulses. Because different stimulation parameters and analysis methods were used to evoke and analyze action potentials, direct

comparisons between these two cell types and the type A and type B cells in the current study are not appropriate. However, CeA neurons with a large ADP, such as type A1 neurons, could fire action potentials at a higher rate or display burst firing pattern, which would be similar to the firing pattern of low-threshold bursting neurons defined in that study. Other differences in the membrane and firing properties could also result from different recording conditions and criteria used to classify CeA neurons. Both glutamate and GABA synaptic transmission in CeA neurons have been characterized (Sah et al., 2003). We found no difference in synaptic inputs among the three types of CeA neurons with local stimulation within the CeA, but those inputs activated via BLA seem to be predominantly glutamatergic, in contrast to both glutamate and GABA inputs activated via the CeA. Although the electrical stimulation would activate both neurons and passing fibers in either site, the difference was observed under comparable stimulation and recording conditions. Thus, it may reflect anatomically differential distribution of synaptic inputs to CeA neurons. This is consistent with previous anatomical, neurochemical, and physiological evidence showing a dominant glutamatergic projection from the BLA to the CeA (Nose et al., 1991; Swanson and Petrovich,

878

W. Zhu and Z. Z. Pan / Neuroscience 127 (2004) 871– 879

1998; Sah et al., 2003). A GABA IPSC has been reported through BLA stimulation likely by activating the GABAergic intercalated cells (Collins and Pare, 1999; Royer et al., 1999). In the current study, the absence of a large GABA component in the synaptic potential elicited by BLA stimulation of low intensities indicates that the stimulation of those intensities and at the BLA location did not activate the GABAergic intercalated cells. Indeed, BLA stimulation of much higher intensities evoked a GABAergic IPSC as well as an EPSC. This is consistent with the previous reports suggesting that CeA neurons receive GABA inputs from intercalated neurons, although contributions from CeA GABAergic cells directly activated by the high intensity stimulation cannot be ruled out. In addition, a bicuculline-resistant, GABAC receptor-mediated IPSC has been reported in neurons in the lateral division of the CeA (Delaney and Sah, 2001), but was not clearly observed in this study. This can be likely attributed to the stimulation site. The previous study demonstrated that the GABAC receptors were present on dendritic synapses innervated by axons originating from the intercalated neurons, but that stimulation in a more medial site (close to our stimulation site) activated somatic synapses expressing only GABAA receptors (Delaney and Sah, 2001). Finally, in this study, we found that the NMDA EPSC was absent in some of the BLA-elicited synaptic potentials, in contrast to those elicited from the CeA under comparable stimulation conditions and antagonist concentrations. This certainly does not necessarily indicate that NMDA receptors are completely absent on these synapses. Although its physiological significance is unknown at this point, this potential differential distribution of NMDA receptors in CeA neurons may be indicative of important functional differences between glutamate synaptic inputs of different origins. Significance of opioid effects The current study identifies a subpopulation of CeA neurons, or the majority of type A1 neurons, that are inhibited by ␮-opioid agonists through activation of GIRK channels. Consistent with our results, a ␮ receptor-mediated inhibition of firing activity has been described in a subpopulation of CeA neurons in an early in vivo study (Freedman and Aghajanian, 1985). Also interesting is a previous report that activation of ␮-receptors opens a 130-ps inwardly rectifying K⫹ channel in dissociated neurons from the amygdalohippocampal area (Chen et al., 2000). Our classification of CeA neuron types based on their physiological properties is further supported generally by their pharmacological profile for different opioid receptor subtypes and corresponding opioid responses. The fact that the vast majority of ␮-receptor-containing neurons are type A1 and ␬-receptor-expressing neurons are type B makes the classification a useful criterion for initially identifying differential opioid-responding CeA neurons with distinct firing characteristics. More importantly, it represents functionally relevant cell groups in the CeA, and provides a pharmacological and physiological basis for our understanding of the CeA functions involving opioids, such as reward behaviors and environmental analgesia.

Recent research increasingly indicates that the CeA is involved in the process of stimulus-reward learning and is associated with the motivational effects of drugs of abuse, especially opioids and alcohol (Koob et al., 1998; Baxter and Murray, 2002). In fact, the CeA constitutes an important part of the neuroanatomical region termed the extended amygdala, which is believed to represent a common neural substrate for the reinforcing effects of drugs of abuse (Heimer and Alheid, 1991; Koob et al., 1998). For example, the CeA is the most effective brain site for microinjected GABAA-receptor antagonists or opioid antagonists to reduce alcohol self-administration in animals (Koob et al., 1998). Given the anatomical evidence that CeA cells and their efferent projections are predominantly GABAergic (Swanson and Petrovich, 1998; Sah et al., 2003), it is possible that the type A1 neurons defined in this study are mostly GABAergic and at least some of them send GABAergic output projections. Based on the current results, it is enticing to propose possible roles for CeA neurons in the opioid-mediated reinforcing effects of abused drugs. Thus, glutamate synaptic inputs of the CeA, such as those from the BLA, would be activated during stimulusreward learning, causing release of endogenous opioid peptides from local enkephalin-containing cells within the CeA (Day et al., 1999). Then, endogenously released opioid peptides, or exogenously administered opioids, would inhibit, through ␮-opioid receptors, those GABAergic projection neurons (type A1), and reduce their inhibitory effect on the projection targets in the brain’s reward circuitry, such as the ventral tegmental area (Baxter and Murray, 2002). The reduced inhibition from the CeA (disinhibition) could then contribute to the activation of the reward circuitry and the reinforcing effect of those drugs of abuse. Indeed, acute alcohol, like ␮-opioid agonists, also inhibits those CeA neurons by increasing local GABA synaptic transmission (Roberto et al., 2003), and induces c-Fos immunoreactivity in enkephalin-expressing CeA cells (Criado and Morales, 2000). Therefore, antagonists of both GABAA receptors and ␮-opioid receptors administered in the CeA could block this pathway and reduce drug selfadministration. In addition, this proposed mechanism may be also involved in the opioid-mediated analgesia induced by environmentally stressful stimuli, such as fear, through CeA GABAergic projections to brainstem pain-modulating sites, including the periaqueductal gray (Rizvi et al., 1991; Baxter and Murray, 2002). That is, opioid-induced reduced inhibition from CeA neurons would disinhibit, or activate neurons in the periaqueductal gray and their projection targets in the rostral ventromedial medulla, producing an analgesic effect (Pan, 1998; Fields and Basbaum, 1999). This is in line with the observation that the CeA contributes to the antinociception induced by systemically applied morphine (Manning and Mayer, 1995). In summary, the different types of CeA neurons described in the present study, with their characteristic synaptic properties and differential opioid responses, provide the fundamental elements for understanding opioidmediated CeA functions in integrating environmental re-

W. Zhu and Z. Z. Pan / Neuroscience 127 (2004) 871– 879

ward and stressful stimulation, and in controlling corresponding behavioral responses. Acknowledgements—This work was supported by the National Institute on Drug Abuse grant DA14524 and by an institutional fund of MD Anderson Cancer Center. We thank Jeanie Woodruff for reading the manuscript.

REFERENCES Aggleton JP (1993) The contribution of the amygdala to normal and abnormal emotional states. Trends Neurosci 16:328 –333. Alvarez VA, Arttamangkul S, Dang V, Salem A, Whistler JL, Von Zastrow M, Grandy DK, Williams JT (2002) mu-Opioid receptors: ligand-dependent activation of potassium conductance, desensitization, and internalization. J Neurosci 22:5769 –5776. Baxter MG, Murray EA (2002) The amygdala and reward. Nat Rev Neurosci 3:563–573. Bie B, Pan ZZ (2003) Presynaptic mechanism for anti-analgesic and anti-hyperalgesic actions of kappa-opioid receptors. J Neurosci 23:7262–7268. Cahill L, McGaugh JL (1998) Mechanisms of emotional arousal and lasting declarative memory. Trends Neurosci 21:294 –299. Chen X, Marrero HG, Murphy R, Lin YJ, Freedman JE (2000) Altered gating of opiate receptor-modulated K⫹ channels on amygdala neurons of morphine-dependent rats. Proc Natl Acad Sci USA 97:14692–14696. Collins DR, Pare D (1999) Reciprocal changes in the firing probability of lateral and central medial amygdala neurons. J Neurosci 19: 836 –844. Criado JR, Morales M (2000) Acute ethanol induction of c-Fos immunoreactivity in pre-pro-enkephalin expressing neurons of the central nucleus of the amygdala. Brain Res 861:173–177. Davis M (2000) The role of the amygdala in conditioned and unconditioned fear and anxioty. In: The amygdala: a functional analysis (Aggleton JP, ed), pp 213–287. Oxford: Oxford University Press. Day HE, Curran EJ, Watson SJ Jr, Akil H (1999) Distinct neurochemical populations in the rat central nucleus of the amygdala and bed nucleus of the stria terminalis: evidence for their selective activation by interleukin-1beta. J Comp Neurol 413:113–128. Delaney AJ, Sah P (2001) Pathway-specific targeting of GABA(A) receptor subtypes to somatic and dendritic synapses in the central amygdala. J Neurophysiol 86:717–723. Dumont EC, Martina M, Samson RD, Drolet G, Pare D (2002) Physiological properties of central amygdala neurons: species differences. Eur J Neurosci 15:545–552. Fields H, Basbaum A (1999) Central nervous system mechanisms of pain modulation. In: Textbook of pain, 4th edition (Wall P, Melzack R, eds), pp 309 –329. London: Churchill Livingstone. Fox RJ, Sorenson CA (1994) Bilateral lesions of the amygdala attenuate analgesia induced by diverse environmental challenges. Brain Res 648:215–221. Freedman JE, Aghajanian GK (1985) Opiate and alpha 2-adrenoceptor responses of rat amygdaloid neurons: co-localization and interactions during withdrawal. J Neurosci 5:3016–3024. Gallagher M, Chiba AA (1996) The amygdala and emotion. Curr Opin Neurobiol 6:221–227. Garcia R, Vouimba RM, Baudry M, Thompson RF (1999) The amygdala modulates prefrontal cortex activity relative to conditioned fear. Nature 402:294 –296. Gottfried JA, O’Doherty J, Dolan RJ (2003) Encoding predictive reward

879

value in human amygdala and orbitofrontal cortex. Science 301: 1104 –1107. Heimer L, Alheid GF (1991) Piecing together the puzzle of basal forebrain anatomy. Adv Exp Med Biol 295:1–42. Helmstetter FJ (1992) The amygdala is essential for the expression of conditional hypoalgesia. Behav Neurosci 106:518 –528. Koob GF, Sanna PP, Bloom FE (1998) Neuroscience of addiction. Neuron 21:467–476. Kovoor A, Henry DJ, Chavkin C (1995) Agonist-induced desensitization of the mu opioid receptor-coupled potassium channel (GIRK1). J Biol Chem 270:589 –595. Manning BH, Mayer DJ (1995) The central nucleus of the amygdala contributes to the production of morphine antinociception in the rat tail-flick test. J Neurosci 15:8199 –8213. McGaugh JL (2002) Memory consolidation and the amygdala: a systems perspective. Trends Neurosci 25:456. McKernan MG, Shinnick-Gallagher P (1997) Fear conditioning induces a lasting potentiation of synaptic currents in vitro. Nature 390:607–611. North RA, Williams JT, Surprenant A, Christie MJ (1987) Mu and delta receptors belong to a family of receptors that are coupled to potassium channels. Proc Natl Acad Sci USA 84:5487–5491. Nose I, Higashi H, Inokuchi H, Nishi S (1991) Synaptic responses of guinea pig and rat central amygdala neurons in vitro. J Neurophysiol 65:1227–1241. Osborne PB, Vaughan CW, Wilson HI, Christie MJ (1996) Opioid inhibition of rat periaqueductal grey neurones with identified projections to rostral ventromedial medulla in vitro. J Physiol 490 (Pt 2):383–389. Pan ZZ (1998) mu-Opposing actions of the kappa-opioid receptor. Trends Pharmacol Sci 19:94 –98. Pan ZZ, Tershner SA, Fields HL (1997) Cellular mechanism for antianalgesic action of agonists of the kappa-opioid receptor. Nature 389:382–385. Pitkanen A, Savander V, LeDoux JE (1997) Organization of intraamygdaloid circuitries in the rat: an emerging framework for understanding functions of the amygdala. Trends Neurosci 20:517–523. Rizvi TA, Ennis M, Behbehani MM, Shipley MT (1991) Connections between the central nucleus of the amygdala and the midbrain periaqueductal gray: topography and reciprocity. J Comp Neurol 303:121–131. Roberto M, Madamba SG, Moore SD, Tallent MK, Siggins GR (2003) Ethanol increases GABAergic transmission at both pre- and postsynaptic sites in rat central amygdala neurons. Proc Natl Acad Sci USA 100:2053–2058. Rogan MT, Staubli UV, LeDoux JE (1997) Fear conditioning induces associative long-term potentiation in the amygdala. Nature 390: 604 –607. Royer S, Martina M, Pare D (1999) An inhibitory interface gates impulse traffic between the input and output stations of the amygdala. J Neurosci 19:10575–10583. Sah P, Faber ES, Lopez De Armentia M, Power J (2003) The amygdaloid complex: anatomy and physiology. Physiol Rev 83:803–834. Schiess MC, Callahan PM, Zheng H (1999) Characterization of the electrophysiological and morphological properties of rat central amygdala neurons in vitro. J Neurosci Res 58:663–673. See RE, Fuchs RA, Ledford CC, McLaughlin J (2003) Drug addiction, relapse, and the amygdala. Ann NY Acad Sci 985:294 –307. Svoboda KR, Lupica CR (1998) Opioid inhibition of hippocampal interneurons via modulation of potassium and hyperpolarizationactivated cation (Ih) currents. J Neurosci 18:7084 –7098. Swanson LW, Petrovich GD (1998) What is the amygdala? Trends Neurosci 21:323–331.

(Accepted 28 May 2004) (Available online 23 July 2004)