Effects of nociceptin/orphanin FQ (N/OFQ), the endogenous ligand of the opioid-like orphan receptor (ORL), were investigated in the rat lateral (AL) and central (ACe) amygdala in vitro. Approximately 98% of presumed projection neurons in the AL responded to N/OFQ with an increase in inwardly rectifying potassium conductance, resulting in an impairment in cell excitability. Half-maximal effects were obtained at 30.6 nm; the Hill coefficient was 0.63. In the ACe, 31% of the cells displayed responses similar to that in the AL, 44% were nonresponsive, and 25% responded with a small potassium current with a linear current–voltage relationship. Responses to N/OFQ were reduced by 100 μmBa2+, were insensitive to 10 μmnaloxone, and were blocked by a selective ORL antagonist, [Phe1ψ(CH2-NH)Gly2]NC(1–13)NH2(IC50 = 760 nm). Involvement of G-proteins was indicated by irreversible effects and blockade of action of N/OFQ during intracellular presence of GTP-γ-S (100 μm) and GDP-β-S (2 mm), respectively, and prevention of responses after incubation in pertussis toxin (500 ng/ml). These mechanisms may contribute to the role of N/OFQ in the reduction of fear responsiveness and stress that have recently been suggested on the basis of histochemical and behavioral studies.
A heptadecapeptide, termed nociceptin (Meunier et al., 1995) or orphanin FQ (Reinscheid et al., 1995), has recently been identified as the endogenous agonist for an opioid receptor-like protein (ORL) (Bunzow et al., 1994; Chen et al., 1994; Fukuda et al., 1994; Mollereau et al., 1994; Nishi et al., 1994;Wang et al., 1994; Wick et al., 1994; Lachowicz et al., 1995). Despite its structural resemblance to opioid peptides, nociceptin/orphanin FQ (N/OFQ) selectively binds to the ORL receptor but not to μ-, δ-, and κ-opioid receptor subtypes, whereas opioid peptides were found to exert no action on the ORL receptor (Mollereau et al., 1994; Lachowicz et al., 1995; Reinscheid et al., 1995). At the cellular level, the actions of N/OFQ appear similar to those of opioid peptides (Standifer and Pasternak, 1997). The ORL receptor is coupled to G-proteins (Meunier et al., 1995; Reinscheid et al., 1995), whose activation resulted in inhibition of adenylyl cyclase activity (Meunier et al., 1995; Reinscheid et al., 1995), modulation of calcium (Knoflach et al., 1996) and potassium conductances (Connor et al., 1996a; Vaughan and Christie, 1996; Ikeda et al., 1997; Vaughan et al., 1997), and regulation of transmitter release (Liebel et al., 1997; Neal et al., 1997; Vaughan et al., 1997). At the behavioral level, systemic application of N/OFQ elicited a unique range of responses, including pronociceptive (Meunier et al., 1995; Reinscheid et al., 1995) or antinociceptive (Xu et al., 1996; Erb et al., 1997; King et al., 1997;Yamamoto et al., 1997) effects, and impairment of or increases in locomotion (Reinscheid et al., 1995; Florin et al., 1996), depending on the concentration that was administered. A more general role of N/OFQ seems to be related to blockade of stress and reversal of opioid-mediated antinociception (Mogil et al., 1996). A recent study demonstrated N/OFQ anxiolytic-like effects in mice and rats that were consistent across several behavioral paradigms generating different types of fear-like responses by exposure to various stressful environmental conditions (Jenck et al., 1997).
One important central structure involved in the integration of fear and anxiety is the amygdala (Davis, 1992; LeDoux, 1995; Maren and Fanselow, 1996). In addition, amygdaloid mechanisms have been shown to be involved in analgesic processes (Helmstetter et al., 1995; Manning and Mayer, 1995; Pavlovic et al., 1996). ORL receptors, N/OFQ and its precursor protein, are expressed in relatively high densities in the various subnuclei of the amygdala, as has been demonstrated throughin situ hybridization and immunohistochemical and autoradiographic procedures (Bunzow et al., 1994; Mollereau et al., 1994; Wick et al., 1994; Lachowicz et al., 1995; Anton et al., 1996;Nothacker et al., 1996; Schulz et al., 1996; Florin et al., 1997). In addition, autoradiographic studies showed a particularly strong effect of N/OFQ-mediated receptor-activation of G-proteins in the amygdaloid complex (Shimohira et al., 1997; Sim and Childers, 1997).
Although results from behavioral and histochemical studies thus suggest a role of N/OFQ within the integrative function of the amygdala, more direct evidence supporting this hypothesis is missing. The present study was therefore undertaken to investigate in detail the effects of N/OFQ in the amygdala on the cellular level, using electrophysiological and pharmacological techniques in the slice preparation of the rat amygdala in vitro. The experiments were focused on the lateral (AL) and central (ACe) nuclei, which represent the main input and output structures of the amygdala, respectively (Pitkänen et al., 1997).
MATERIALS AND METHODS
Preparation. Long–Evans rats of either sex [postnatal days (P) 11–16] were anesthetized with halothane and killed by decapitation. A block of tissue containing the amygdala was rapidly removed and placed in ice-cold oxygenated physiological saline containing (in mm): KCl 2.4, MgSO4 10, CaCl2 0.5, piperazine-N,N′-bis(ethanesulfonic acid) (PIPES) 20, glucose 10, sucrose 195, pH 7.35. Coronal slices (300 μm thick) were prepared on a Vibratome (Model 1000, Ted Pella, Redding, CA) and incubated in standard artificial CSF (ACSF) of the following composition (in mm): NaCl 120, KCl 2.5, NaH2PO4 1.25, NaHCO3 22, MgSO4 2, CaCl2 2, glucose 20, bubbled with 95%O2/5% CO2 to pH 7.3. Slices were kept at 34°C for 20 min and for up to 8 hr at 24–25°C. A single slice was transferred to the recording chamber and submerged in ACSF, at a perfusion rate of ∼2 ml/min. The tissue was fixed by a silk mesh around a horseshoe-shaped platinum wire.
Recording techniques. Neurons located superficially at a depth of ∼50 μm on the dorsal surface of the slice were approached under visual control by differential interference contrast (Axioskop FS, Achroplan 40/w; Zeiss, Oberkochen, Germany) infrared videomicroscopy (S/W-camera CF8/1, Kappa, Gleichen, Germany) (Dodt and Zieglgänsberger, 1994). Electrophysiological recordings were performed using the patch-clamp technique in whole-cell mode (Hamill et al., 1981) with a patch-clamp amplifier (EPC-7, List Medical Systems, Darmstadt, Germany). Patch pipettes were pulled from borosilicate glass (GC150TF-10, Clark Electromedical Instruments, Pangbourne, UK). After the whole-cell configuration was obtained, cells were held at −70 mV unless indicated otherwise. In each slice, only one neuron was recorded. Drugs were added to the external ACSF. In some experiments, the extracellular KCl concentration was elevated to 5, 7.5, or 10 mm by substitution of an equimolar amount of NaCl. In 0 Ca2+ solution, CaCl2 was replaced by MgCl2. Pipettes were filled with (in mm): potassium gluconate 95, K3 citrate 20, NaCl 10, HEPES 10, MgCl2 1, CaCl2 0.1, BAPTA 1, MgATP 3, NaGTP 0.5, pH 7.2 with KOH. In some experiments, 2 mm GDP-β-S or 100 μm GTP-γ-S were included in the internal solution in exchange for NaGTP. Typical electrode resistances were 2.5–3 MΩ in the bath, with access resistances in the range of 5 to 7 MΩ. Errors attributable to series resistance were <2.5 mV. A liquid junction potential of 10 mV was corrected (Neher, 1992). Records were low-pass-filtered at 2.5 kHz (eight-pole Bessel filter). Voltage-clamp experiments were performed using pClamp software operating via a Labmaster DMA interface (Axon Instruments, Foster City, CA) on a PC. Experiments were conducted at 24–25°C. All substances were obtained from Sigma (Diesenhofen, Germany), except for nociceptin/orphanin FQ and [Phe1ψ(CH2-NH)Gly2]NC(1–13)NH2, which were purchased from Gramsch Laboratories (München, Germany) and Tocris (Langford, UK), respectively. Data are presented as mean ± SEM.
Histological procedures. In some experiments, 0.1% biocytin was added to the intracellular solution. After recordings, slices containing a biocytin-filled cell were immersed in 4% paraformaldehyde in PBS (in mm: NaCl 120, NaH2PO4 10, K2HPO4 30, thimerosal 0.05, pH 7.4 with NaOH) for 48 hr and then cryoprotected in a solution composed of 30% sucrose in PBS for 24 hr. Slices were then sectioned at a thickness of 100 μm using a freezing microtome (Leica, Benzheim, Germany) and rinsed three times (5 min each) with PBS. Endogenous peroxidase reaction was by-passed by incubation in 0.3% H2O2 in PBS for 30 min. Sections were then treated with the avidin–biotin complex horseradish peroxidase (PK 4000, 1:100; Vector Laboratories, Burlingame, CA) in PBS supplemented with 0.03% Triton X-100 and 2% bovine serum albumin overnight. Between steps, sections were rinsed three times for 15 min each with PBS. The immunohistochemical reaction was completed by adding 3,3′diaminobenzidine (0.5 mg/ml) and H2O2(0.01%) to the incubation medium, including 0.02% (NH4)2Ni(SO4)2× 6 H2O and 0.025% CoCl2, and stopped after 6 min by rinsing with PBS (two times for 5 min). Sections were mounted on chromalaun gelatin-coated coverslips, dehydrated, and coverslipped in DePeX (Serva Feinbiochemica, Heidelberg, Germany) mounting media.
Pertussis toxin treatment. Slices were rinsed three times with sterile medium of the following composition: 200 ml Eagle’s basal medium, 100 ml HBSS, 100 ml inactivated horse serum (all from Life Technologies, Eggenstein, Germany), 2 mm glutamine, and 0.65% glucose, pH 7.3. The tissue was then affixed to coverslips with a chicken plasma (Cocalico, Reamstown, PA) clot coaggulated by thrombin (25 U/ml; Calbiochem, Bad Soden, Germany) and cultured in a roller incubator (Heraeus, Hanau, Germany) for ∼18 hr at 34°C. Slices were kept in medium supplemented with 500 ng/ml pertussis toxin, and control slices were incubated in medium with no added pertussis toxin. Control and treated slices were used for recording alternately to confirm the viability of the neurons.
Recordings were derived from a total of 94 neurons in the AL and 55 cells in the ACe. Lateral amygdaloid neurons typically possessed a pyramidal-like cell body with spiny dendrites, as revealed by intracellular staining with biocytin (Fig.1 A) (n= 14). Central amygdaloid neurons had a heterogeneous morphological appearance with piriform to ovoid cell bodies and mostly spiny dendrites (Fig. 1 B) (n = 11).
Application of N/OFQ (1 μm) through bath perfusion resulted in a transient outward current associated with an increase in membrane conductance (Fig. 1 C) in 92 of 94 recorded AL neurons. The average maximal amplitude of the outward current deflection induced by 1 μm N/OFQ at a holding potential of −70 mV was 53.3 ± 7.3 pA (n = 8). Similar responses to N/OFQ were observed during blockade of synaptic transmission in the presence of 1.5 μm TTX (61.1 ± 7.0 pA; n = 4) or in Ca2+-free solution (55.9 ± 20.3 pA; n = 3), indicating a direct postsynaptic response of the recorded neurons (data not shown).
Increase in K+ conductance through N/OFQ in AL neurons
Hyperpolarizing voltage steps of 350 msec duration applied between −80 and −140 mV in 10 mV increments elicited slowly developing currents in AL neurons (Fig.2 A). Application of N/OFQ (1 μm) increased these membrane currents. The effect of N/OFQ was isolated by subtracting currents obtained before and during action of the peptide. The difference currents showed rapid activation and no inactivation during the 350 msec voltage pulse. Ramp–voltage commands (0.2 mV/msec) applied from −70 to −140 mV revealed a membrane current with inwardly rectifying properties that was increased by 1 μm N/OFQ (Fig. 2 B). Current–voltage (I–V) relationships (Fig.2 C,D) obtained by plotting the N/OFQ-induced current amplitude at the end of the hyperpolarizing step against voltage (Fig.2 C) or by the ramp protocol (Fig. 2 D) were similar in that they demonstrated moderate inward rectification. The N/OFQ-induced current reversed at approximately −100 mV, i.e., very close to the K+ equilibrium potential as calculated by the Nernst equation (E K = −104.1 mV). To confirm the contribution of K+ as the main charge carrier of the N/OFQ-induced membrane current, the K+ concentration of the perfusing saline was varied (Fig. 3). Ramp–voltage commands were applied at 0.07 Hz to monitor the I–V relationship during the experiment. The reversal potential of the N/OFQ-evoked response induced by a 10-fold change in the external K+concentration had a slope of 53.7 mV, i.e., similar to that predicted by the Nernst equation for a pure K+ electrode (Fig.3 B). Furthermore, increases in the extracellular K+ concentration were associated with an increase in the slope of the I–V relationship. Such behavior is typical of inward rectifiers (Fig. 3 A) (Isomoto et al., 1997).
Pharmacological properties of the N/OFQ response
To obtain full recovery of N/OFQ responses during pharmacological experiments, the concentration of the peptide was reduced to 50 nm. Because the response on the first application of N/OFQ was often larger in amplitude compared with those to further applications, presumably because of desensitization phenomena (Ikeda et al., 1997), the order of application of N/OFQ alone and N/OFQ in the presence of antagonists was varied in different cells. In a first series of experiments, possible effects of N/OFQ on opioid receptors were tested through the use of the nonselective opioid receptor antagonist naloxone (Fig.4 A) (Raynor et al., 1994). N/OFQ elicited an outward current of 43.0 ± 3.7 pA (n = 5), which was not altered by the addition of 10 μm naloxone (44.8 ± 6.6 pA; n = 5). By comparison, the selective antagonist to the ORL receptor [Phe1ψ(CH2-NH)Gly2]NC(1–13)NH2(Guerrini et al., 1998) reduced the response to 50 nm N/OFQ in a concentration-dependent manner (Fig. 4 B). The antagonistic effect was quantified as reduction in the maximal N/OFQ-induced current amplitude during the presence of [Phe1ψ(CH2-NH)Gly2]NC(1–13)NH2compared with the control N/OFQ response in an individual cell. Half-maximal inhibition was obtained at 760 nm[Phe1ψ(CH2-NH)Gly2]NC(1–13)NH2with a Hill coefficient of 0.89 (n = 31) (Fig.4 B).
The effect of N/OFQ was substantially and reversibly reduced through application of 100 μm Ba2+ (Fig.4 C). The effect of Ba2+ on N/OFQ responses could not be analyzed in an individual cell, because applications of N/OFQ for periods longer than 3 min led to a current that did not return to baseline during the typical recording time of 60–80 min. This may be attributable to slow dissociation of the ligand (Ardati et al., 1997) and/or slow wash-out of the drug during bath perfusion (Muller et al., 1988). On average, Ba2+blocked 39.3 ± 6.6 pA (n = 6) of the N/OFQ-sensitive current, corresponding to a reduction of the maximal amplitude by 87.4 ±. 6.6% (n = 6).
Concentration–response relationship of the N/OFQ response
Concentration–response relationships of N/OFQ were investigated in an elevated extracellular K+ concentration (10 mm) at a holding potential of −80 mV, to increase the driving force and hence the magnitude of the N/OFQ-induced current. In an individual neuron, only one N/OFQ concentration was tested, and concentration–response relationships were constructed from maximal responses to a given concentration. The N/OFQ-induced current showed clear concentration-dependency (Fig. 5). The half-maximal effect (EC50) was elicited by 30.6 nm with a Hill coefficient of 0.63 (n = 36).
Involvement of G-proteins
Next, the possible mediation by G-proteins of the N/OFQ-induced K+ current was tested (Fig.6). The tip of the electrode was filled with standard pipette solution, and the remainder was backfilled with a solution supplemented with 100 μm GTP-γ-S, a nonhydrolyzable GTP analog (Gilman, 1987). Under these conditions, N/OFQ (50 nm) evoked an outward current shortly after establishing the whole-cell configuration, with no indication of recovery during the time course of the experiment (Fig.6 A, compare with Fig. 4 A). A second application of N/OFQ had no further effect, whereas the standing current was significantly reduced by Ba2+ (Fig.6 A) (n = 3). Inclusion in the pipette solution of 2 mm GDP-β-S, a nonhydrolyzable GDP analog (Gilman, 1987), resulted in only a small outward current (9.4 ± 1.3 pA) in response to a first application of N/OFQ, whereas further applications failed to evoke an effect (Fig. 6 B) (n = 3). Finally, after incubation of slices in a medium containing pertussis toxin (500 ng/ml) for ∼18 hr at 34°C, N/OFQ effects could not be obtained in any neuron tested (Fig.6 C) (n = 3). In slices that had been similarly treated, but with no addition of the toxin, all cells tested responded to N/OFQ (1 μm) with the typical outward current (maximal amplitude at a holding potential of −70 mV; 30.3 ± 4.2 pA; n = 4).
N/OFQ responses in central amygdaloid neurons
In 17 of 55 neurons recorded in the ACe, N/OFQ (1 μm) elicited a current with properties similar to that in AL cells, in that it possessed inwardly rectifying properties (Fig.7 A), it was reduced to 74.2 ± 5.4% by 100 μm Ba2+(n = 7) (Fig. 7 B), and its reversal potential shifted in a Nernstian manner with the external K+ concentration (Fig. 7 E). The maximal current amplitude at a holding potential of −70 mV was significantly (p < 0.01) smaller in neurons of the ACe (21.6 ± 4.1 pA; n = 9) compared with that in the AL (53.3 ± 7.3 pA; n = 8). In 24 central amygdaloid neurons, no responses to N/OFQ were detected (data not shown). The remainder of the cells recorded in the ACe responded to N/OFQ with a small current (8.6 ± 1.1 pA at a holding potential of −70 mV; n = 14) (Fig. 7 D), which displayed no inward rectification as measured by the ramp protocol (Fig. 7 C) but was also blocked by 100 μmBa2+ (Fig. 7 D).
Effect of N/OFQ on spike firing
Under current-clamp conditions, 50 nm N/OFQ induced a membrane hyperpolarization from the resting membrane potential of −70 mV, with an average maximal amplitude of 10.3 ± 1.2 mV in AL neurons (n = 7) (Fig.8 A). The input membrane resistance decreased from 535.7 ± 58.3 to 326.4 ± 30.3 MΩ during the maximal response (n = 7). This hyperpolarization and decrease in input resistance had a dampening effect on cell excitability, as was indicated by a substantial reduction in the number of spikes evoked by depolarizing current responses before and during action of N/OFQ (Fig.8 B).
The results of the present study indicate that N/OFQ acts on ORL receptors in the postsynaptic membrane of neurons in the rat lateral and central amygdaloid nuclei, thereby inducing an increase in an inwardly rectifying K+ conductance mediated via pertussis toxin-sensitive G-proteins. This promotes significant hyperpolarization of the resting membrane potential, an increase in input conductance, and a decrease in cellular excitability. Responses to N/OFQ were quantitatively different in various subnuclei of the amygdala, in that almost every presumed projection class I neuron encountered in the AL was responsive, whereas a typical response was obtained in roughly one-third of neurons in the ACe.
Mediation of N/OFQ responses through ORL receptors
The involvement of ORL receptors is indicated by three lines of evidence. First, responses to N/OFQ were reversibly blocked by the selective ORL receptor antagonist [Phe1ψ(CH2-NH)Gly2]NC(1–13)NH2(Guerrini et al., 1998), whereas naloxone, a prototypical antagonist to the μ-, δ-, and κ-subtypes of opioid receptors (Raynor et al., 1994), was ineffective even at high concentrations. In the guinea-pig ileum and in mouse vas deferens, 1 μm[Phe1ψ(CH2-NH)Gly2]NC(1–13)NH2reportedly reduced the effect of 50 nm N/OFQ by ∼80%, compared with a 61% reduction of N/OFQ-induced response in amygdaloid neurons. These results support the idea of a selective effect of N/OFQ on ORL receptors in the present study. Second, the concentration–response relationships of N/OFQ found in amygdaloid neurons (EC50 = 30.6 nm) are comparable to those reported earlier for various types of cellular responses to N/OFQ, including increases in K+ conductance (EC50 between 12 and 90 nm) (Connor et al., 1996a; Vaughan and Christie, 1996; Ikeda et al., 1997; Vaughan et al., 1997; Wagner et al., 1998), reduction of calcium currents (EC50 at 42–80 nm) (Connor et al., 1996b;Knoflach et al., 1996), and inhibition of K+-evoked glutamate release (EC50 = 51 nm) (Nicol et al., 1996). Third, mRNA-expressing ORL receptors and ORL receptor binding sites were demonstrated to be highly enriched in the amygdala (Bunzow et al., 1994; Mollereau et al., 1994; Wick et al., 1994; Lachowicz et al., 1995; Anton et al., 1996; Florin et al., 1997). In addition, the mRNA of the N/OFQ precursor protein (Houtani et al., 1996; Nothacker et al., 1996) and immunoreactivity against the peptide (Schulz et al., 1996) were detected in this limbic area. Therefore, it seems reasonable to assume a functional role of the ORL receptor/N/OFQ peptide system in the amygdala.
Involvement of G-protein-coupled K+ channels
The primary structure of ORL displays the seven putative membrane-spanning domains of a G-protein-coupled membrane receptor (Bunzow et al., 1994; Fukuda et al., 1994; Mollereau et al., 1994; Wick et al., 1994; Lachowicz et al., 1995). Indeed, N/OFQ potently inhibited forskolin-stimulated cAMP accumulation in transfected cells (Meunier et al., 1995; Reinscheid et al., 1995) and modulated Ca2+ conductances via pertussis toxin-sensitive G-proteins in acutely isolated hippocampal neurons (Knoflach et al., 1996) and SH-SY5Y neuroblastoma cells (Connor et al., 1996b). By comparison, the inhibition of the T-type Ca2+channel in rat sensory neurons induced by N/OFQ occurred through a G-protein-independent mechanism (Abdulla and Smith, 1997).
Three lines of evidence suggest that in amygdaloid neurons, the activation of ORL through N/OFQ resulting in an increase in an inward rectifier K+ conductance appears to be mediated by a G-protein-linked pathway. First, during the intracellular presence of GTP-γ-S, a nonhydrolyzable GTP analog (Gilman, 1987), N/OFQ evoked sustained responses, presumably attributable to irreversible activation of the G-proteins. Second, intracellular GDP-β-S, the nonhydrolyzable GDP analog (Gilman, 1987), significantly reduced the response amplitude to a first application of N/OFQ and prevented further responses in individual cells, presumably caused by functional inactivation of the G-proteins. Third, incubation of slices overnight with pertussis toxin, a blocker of the Gi/G0 class of G-proteins (Hille, 1994), prevented responses to N/OFQ, whereas slices that were similarly treated, but with no addition of the toxin, showed qualitatively unaltered responsiveness to N/OFQ. Taken together, these data indicated a coupling of ORL to the Gi/G0 class of G-proteins (Hille, 1994).
In amygdaloid neurons, the targets of ORL activation appear to be the inwardly rectifying types of K+ channels (Fakler and Ruppersberg, 1996; Isomoto et al., 1997; Karschin et al., 1997), as was indicated by the I–V relationship of responses to N/OFQ and their sensitivity to low concentrations of extracellular Ba2+. This hypothesis is supported by results fromin situ hybridization studies showing that the mRNA of subunits of G-protein-coupled K+ channels, namely Kir3.1, Kir3.2, Kir3.3, are expressed in high density in amygdaloid nuclei (Karschin et al., 1996). By comparison, only low levels of mRNA of “classical” inwardly rectifying K+ channels belonging to the Kir2 family are present in the amygdala (Karschin et al., 1996). The inwardly rectifying properties of membrane currents observed under control conditions (Fig. 2 A,B) may thus represent the background conductance of G-protein-gated channels (Dascal, 1997) or an H-current component rather than classical inward rectifier activity (Womble and Moises, 1993). The exact subunit composition of the G-protein-gated inward rectifier channels mediating responses to N/OFQ remains to be elucidated, although the fast but not instantaneous activation of the N/OFQ-induced current (Fig.2 A) suggests an involvement of Kir3.1 subunits.
N/OFQ and “classical” opioid effects in the amygdala
At the cellular level, N/OFQ and opioid peptides seem to share a number of common mechanisms, such as the increase in a G-protein-coupled inwardly rectifying K+conductance, whereas on the systemic level, N/OFQ can potently reverse opioid analgesia (Darland et al., 1998). In the AL, an opioid-induced hyperpolarization presumably mediated through an increase in K+ conductance reportedly occurs in the nonpyramidal type of cells (type 2), whereas cells similar in appearance to those termed class I by McDonald (1992) were nonresponsive (Sugita and North, 1993). AL neurons that were labeled with biocytin in the present study possessed morphologies indicative of class I neurons, suggesting that the effects of N/OFQ and opioids are mediated by different subsets of neurons. The observation that only a comparatively small percentage of the morphologically heterogeneous group of neurons encountered in the ACe was responsive to N/OFQ seems to be supportive of this hypothesis. A more direct correlation between morphologically defined types of cells and N/OFQ responsiveness will require immunocytochemical analyses of the cellular distribution of receptors and K+channels in identified types of cells in the various parts of the amygdala.
Possible source and functional significance of N/OFQ
The expression of N/OFQ mRNA, ORL mRNA, and the ORL protein, as determined through in situ hybridization and immunocytochemical techniques, is very dense to dense in the amygdala (Bunzow et al., 1994; Mollereau et al., 1994; Anton et al., 1996;Houtani et al., 1996; Nothacker et al., 1996; for review, see Henderson and McKnight, 1997; Meunier, 1997; Darland et al., 1998). Although these data suggest that N/OFQ may act as a transmitter of the abundant interneurons inherent to the amygdaloid complex (Nitecka and Frotscher, 1989; Smith and Paré, 1994), electrical stimulation at different sites within the amygdala or the external capsule using different stimulus protocols failed to evoke a postsynaptic potential sensitive to the selective ORL antagonist in the present study (data not shown). On the other hand, a dense expression of N/OFQ or its receptor, or both, has been found to exist in various regions with axonal projection to the amygdala, including the ventral tegmental area, the locus coeruleus, and the dorsal raphe nuclei of the brain stem (LeDoux, 1995;LeDoux and Muller, 1997), which may as well give rise to an N/OFQ-releasing system. Further analyses, such as an electron microscopic study, will be needed for the detection of N/OFQ at the synaptic level and thus the identification of the substrates of N/OFQ-mediated neurotransmission in the amygdala.
In any case, ∼98 and 55% of the cells encountered in the AL and ACe, respectively, responded to local application of N/OFQ. Morphological identification of the N/OFQ-sensitive neurons with intracellular injection of biocytin revealed characteristics indicative of projection cells (McDonald, 1992, 1996), such as pyramidal-shaped spiny cells in the AL and morphologically more heterogeneous types of cells in the ACe, all of which possessed spine-rich dendrites. Following from this is the conclusion that synaptic transmission mediated through N/OFQ is not negligible in the amygdala. The hyperpolarizing effect and associated decrease in excitability produced by N/OFQ, in turn, indicates that N/OFQ promotes a net dampening effect on amygdaloid output activity. In view of the known role of the amygdala for the integration of sensory and aversive signals (Davis, 1992; LeDoux, 1995;Maren and Fanselow, 1996), with the lateral and central nucleus serving as the main input and output station, respectively (Pitkänen et al., 1997), these conclusions support the speculation that the N/OFQ-induced dampening of amygdaloid output activity contributes to the reduction in fear responsiveness that has been observed recently during systemic application of this neuropeptide in mice and rats (Jenck et al., 1997). Studies should now be performed that combine local application of the selective ORL antagonist [Phe1ψ(CH2-NH)Gly2]NC(1–13)NH2in the amygdala, with behavioral paradigms generating various types of fear responses, to understand the exact conditions under which N/OFQ is acting as an anxiolytic.
This work was supported by the Kultusministerium des Landes Sachsen-Anhalt (FKZ 2279A/0085H) and the Deutsche Forschungsgemeinschaft (SFB 426, TP B3). We thank Professor V. Höllt for stimulating the work on N/OFQ, and A. Ritter, R. Ziegler, and A. Reupsch for expert technical assistance.
Correspondence should be addressed to Hans-Christian Pape, Institut für Physiologie, Medizinische Fakultät, Otto-von-Guericke-Universität, Leipziger Strasse 44, D-39120 Magdeburg, Germany.