Introduction

Top Abstract Introduction Methods Results Discussion References

CRF, a 41-amino acid polypeptide, plays a major role in the coordination of endocrine, autonomic and behavioral responses to stressful stimuli. In addition to activation of the HPA axis (Aguilera et al., 1992; Axelrod and Reisine, 1984; Hauger et al., 1988), CRF regulates responses to stress through activation of extrahypothalamic brain regions. Anatomically, CRF, CRF receptors, CRF-containing neurons and terminals are widely present in brain regions other than the HPA axis (De Souza et al., 1985; Palkovits et al., 1985; Sakanaka et al., 1986; Sawchenko and Swanson, 1985; Swanson et al., 1983), including the bed nucleus of the stria terminalis, septum and the medial, basolateral and central amygdala. Behavioral and autonomic effects of CRF administered centrally are also mediated through extrahypothalamic brain areas and occur in the presence of dexamethasone, a treatment that blocks HPA activation. In addition, CRF, applied in vivo or in vitro, can directly alter neuronal behaviors in extrahypothalamic brain regions. For example, CRF excites neurons of the cortex (Eberly et al., 1983), hippocampus (Aldenhoff et al., 1983) and the locus ceruleus (Valentino et al., 1983; Valentino and Foote, 1988), whereas neurons of the thalamus and lateral septal areas are inhibited (Eberly et al., 1983).

The CeA, a prominent nuclear complex within the corpus amygdaloideum, is thought to be one of the key extrahypothalamic regions in responses to stress involving CRF. Anatomically, CeA contains a relatively high amount of CRF and high density of CRF-immunoreactive (CRF-ir) neurons and fibers, as well as a moderate concentration of CRF receptors (Cassell and Gray, 1989; Cummings et al., 1983; De Souza et al., 1985; Grigoriadis and De Souza, 1992; Imaki et al., 1991; Olschowka et al., 1982; Palkovits et al., 1985; Swanson et al., 1983; Uryu et al., 1992). The CRF-ir neurons in the CeA have efferent projections to nuclei involved in central control of autonomic activity and stress (see Gray, 1989, 1990) such as the parabrachial nucleus (Moga and Gray, 1985; Sakanaka et al., 1986), the dorsal vagal complex (Veening et al., 1984) and the midbrain central gray. Furthermore, a calcium-dependent release of CRF has been measured in fetal and adult rat amygdala (Smith et al., 1986; Takuma et al., 1994) and CRF release in the CeA is increased in cocaine-treated rats (Richter et al., 1995). Electrophysiologically, CRF inhibits the slow-afterhyperpolarizing potential, hyperpolarizes the membrane and broadens the duration of Ca⁺⁺ spikes in amygdala neurons in vitro (Rainnie et al., 1992). Behaviorally, CRF microinfused into the CeA nucleus causes tachycardia (Wiersma et al., 1993), whereas infusion of CRF receptor antagonist -helical CRF_9-41 into the CeA attenuates stress-induced freezing behavior (Swiergiel et al., 1993) and reverses anxiogenic-like effects of ethanol withdrawal (Rassnick et al., 1993). Furthermore, during restraint stress, CRF mRNA content in amygdala is increased (Kalin et al., 1994).

In extrahypothalamic brain areas, CRF stimulates the release of neurotransmitters, including, GABA, dynorphin and methionine-enkephalin (Sirinathsinghji et al., 1989; Sirinathsinghji and Heavens, 1989), as well as dynorphin A (Song and Takemori, 1992). In the HPA axis, CRF increases cytosolic calcium in rat and human pituitary corticotrophs through voltage-gated calcium channels, and this calcium entry is responsible for the CRF-dependent ACTH release from those cells (Guerineau et al., 1991). L- and P-type calcium channel antagonists (Kuryshev et al., 1996) can block this increase in corticotrophs. These data suggest that CRF may affect voltage-activated Ca⁺⁺ channels, leading to neurotransmitter release, but the mechanism underlying these actions of CRF on brain neurons is not known.

Amygdala neurons possess multiple types of I_Cas (Foehring and Scroggs, 1994; Kaneda and Akaike, 1989), and we have described four different types of I_Ca in isolated CeA neurons (Yu and Shinnick-Gallagher, 1994a, 1997). This study focuses on the membrane mechanism of action of CRF acting on receptors in neurons intimately involved in integrating the autonomic, neuroendocrine and behavioral responses to stressful stimuli. The purpose of the present study was to analyze in isolated CeA neurons the actions of CRF on the I_Ca using whole-cell patch-clamp recording technique and pharmacological agents and neurotoxins. A portion of these results has been published in abstract form (Yu and Shinnick-Gallagher, 1994a).

	Methods

Top Abstract Introduction Methods Results Discussion References

Cell preparation. The techniques used to prepare 300- to 330-µm brain slices and cell dissociation were similar to those described previously (Yu and Shinnick-Gallagher, 1994b, 1997). Pregnant female Sprague-Dawley rats at 16 to 18 days of gestation were purchased from Harlan (Houston, TX). Pups of either sex ranging from 8 to 18 days were decapitated, and the brains were rapidly removed and cooled in dissecting solution (0-5°C), bubbled continuously with 100% O₂. The dissecting solution contained 120 mM NaCl, 10 mM KCl, 2 mM KH₂PO₄, 1 mM CaCl₂·2H₂O, 6 mM MgSO₄·7H₂O, 10 mM D-glucose and 10 mM piperazine-N,N-bis-[2-ethanesulfonic acid]; the pH was adjusted to 7.40 with 10 N NaOH. The osmolarity of the dissecting solution was measured using 5100 Vapor Pressure Osmometer (Wescor) and adjusted to 300 to 310 mOsM with sucrose in every experiment. In cold solution, the brain was transversely cut posterior to the first branch and anterior to the last branch of the superior cerebral vein. The resulting block of brain tissue was hemisected. Three or four serial coronal slices per hemisphere were obtained using a Vibroslice (Campden) and incubated in a beaker containing oxygenated dissecting solution at room temperature for 20 to 30 minutes. The slices were then transferred to an oxygenated enzyme solution preheated to 34.5°C and incubated for 12 to 15 min. The enzyme solution contained a mixture of 20 to 30 mg/15 ml of pronase E (Type XXV; Sigma, St. Louis, MO) and 18 to 20 mg/15 ml of trypsin (Sigma). The slices were washed twice after enzyme treatment and stored in an oxygenated dissecting solution at room temperature for 10 min. The CeA area was visualized under a stereomicroscope and dissected from the brain slices with a scalpel. The pieces of brain tissue containing the CeA were triturated gently with a series of flame-polished Pasteur pipettes of decreasing diameter to dissociate mechanically the individual neurons. The suspended cells were pipetted into the recording chamber that was mounted on the stage of a Nikon inverted microscope (Nikon Diaphot).

Electrophysiological recordings. The whole-cell patch-clamp methodology (Hamill et al., 1981) was used for recording I_Cas. Patch electrodes were made from Corning 7052 glass (1.5-mm outer diameter, Garner Glass, Claremont, CA), pulled using a Flaming-Brown micropipet puller (model P-97, Sutter Instruments, Novato, CA) or a laser puller (model P-2000, Sutter Instruments, Novato, CA) and polished using a Narishige microforge (model MF-9). The patch electrodes were coated with 20% Sylgard before polishing and had resistances of 3 to 6 M when filled with an internal solution of the following composition (in mM): 90 Cs acetate (or 100 CsF), 18 TEA, 18 HEPES, 9 BAPTA, 9 D-glucose, 5 MgATP, 0.2 NaGTP and 0.1 leupeptin. The internal fluoride solution was used in a few preliminary experiments. Data were pooled because (1) dialysis of fluoride into dissociated amygdala neurons (Kaneda and Akaike, 1989) occurs in 10 to 20 min, a time frame longer than the protocols used in the present experiments, and (2) neither the I_Ca or the effect of CRF on I_Ca recorded with acetate or fluoride was different (see Joels and Karst, 1995). The pH of the internal solution was adjusted from 7.1 to 7.2 with 1 N CsOH at room temperature and had a final osmolarity of 270 to 280 mOsM. The external solution consisted of the following: 120 mM TEA-Cl, 3 mM CaCl₂, 10 mM HEPES, 10 mM CsCl, 5 mM 4-AP, 10 mM glucose and 2 µM TTX. The pH of the external solution was titrated to 7.4 with 1 N HCl or 1 N CsOH, and the osmolarity was adjusted to 320 ± 5 mOsM with sucrose. All the experiments were performed at room temperature (22-25°C). Both the internal and external solutions were designed to suppress pharmacologically the interfering sodium and potassium currents. The I_Cas recorded in this study are thought to be in the relative physiological range because the extracellular Ca⁺⁺ concentration used (3 mM) is close to the "normal" condition (2.5 mM; Allen et al., 1993). The reference electrode was filled with the same solution as the internal recording solution; junction potentials calculated according to the method of Neher (1992) ranged from 2 to 4 mV with electrode tip diameters of 1.0 to 1.5 µm.

Drug application. CRF (human and rat; Peninsula Laboratories, Belmont, CA), -helical CRF_9-41 (Bachem, Torrance, CA), -conotoxin GVIA (RBI, Natick, MA), Aga IVA (Alamone Labs, Jerusalem, Israel) and -conotoxin MVIIC (Bachem) were prepared as concentrated stock solution in distilled water and stored in aliquots at 20°C. The stock solutions of drugs were diluted with the external solution immediately before each experiment. Nimodipine and (±)-BAY K 8644 (RBI) were prepared as concentrated stock solutions in 90% dimethylsulfoxide and protected from light. TTX was obtained from Sigma.

	Results

Top Abstract Introduction Methods Results Discussion References

CRF increased the I_Ca in CeA neurons in a concentration-dependent manner. These data were obtained from 120 neurons acutely dissociated from rat CeA. We previously described the I_Ca in CeA neurons (Yu and Shinnick-Gallagher, 1994a, 1997). The HVA I_Ca in CeA neurons is composed of at least four pharmacologically distinct components. We found that the N-type, -conotoxin GVIA (1 µM)-sensitive current accounted for 30% of the total I_Ca; the Q-type current defined as the current sensitive to -conotoxin MVIIC (250 nM) or -agatoxin IVA (Aga IVA; 1 µM) amounted to 13% to 18% of I_ca; the L-type, NIM-sensitive current represented 22% of the total I_Ca; and a resistant current, a non-L-, N- and Q-type current, comprised 37% to 53% of the total I_Ca. (Yu and Shinnick-Gallagher, 1997). Figure 1 illustrates an example of the effects of sequential applications of calcium channel blockers on I_Ca evoked by voltage-step commands to +10 or 0 mV from a holding potential of 70 mV (fig. 1A, left & 1B) or 40 mV (fig. 1A, right), respectively. The antagonists block similar proportions of total current whether held at 70 or 40 mV. L-type currents are found in only 70% of CeA neurons, and a P-type current blocked by low nanomolar concentrations of -agatoxin (Mintz et al., 1992a, 1992b) is not recorded in CeA neurons (Yu and Shinnick-Gallagher, 1997). In our previous study (Yu and Shinnick-Gallagher, 1997), we defined the resistant current as the current remaining in -conotoxin MVIIC (250 nM), NIM (5 µM) and -conotoxin GVIA (1 µM) or Aga IVA (1 µM); furthermore, increasing the concentrations of NIM (10 µM) and neurotoxins -conotoxin GVIA (2 µM) and -conotoxin MVIIC (500 nM) resulted in a similar percentage (36%) of resistant current recorded. We (Yu and Shinnick-Gallagher, 1997) have also shown that low-threshold T-type currents are not recorded from the holding potentials of 70 and 40 mV used in the present study. Our previous findings (Yu and Shinnick-Gallagher, 1997) are in agreement with other studies on unidentified (Kaneda and Akaike, 1989) and pyramidal (Foehring and Scroggs, 1994) neurons in the amygdaloid complex.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Components of ICa in CeA neurons. A, Left, ICa elicited using 100-ms depolarizing steps from 70 to +10 mV. The combination of CgTx GVIA (1 µM) plus CgTx MVIIC (250 nM) plus NIM (5 µM) reduced ICa by 64% in this neuron; the resistant current represented 36% of the total ICa. Cd++ (5 µM) reduced the resistant current. A, Right, ICa elicited using 100-ms depolarizing steps from 40 to 0 mV in another neuron. The addition of CgTx GVIA (1 µM) plus CgTx MVIIC (250 nM) plus NIM (5 µM) sequentially reduced the ICa by 69% and revealed a resistant current comprising 31% of the total ICa. Cd++ (10 µM) completely blocked the resistant current. Currents recorded in a high concentration of Cd++ (200 µM) were leak subtracted. B, Time course of NIM and neurotoxin block. Left, ICa elicited in response to a voltage step from 70 to +10 mV. ICa recorded at times indicated (right). In this neuron, NIM blocked 32% of ICa, -conotoxin GVIA blocked 23% and Aga IVA blocked 21%. In this neuron, the resistant current amounted to 23% of the total ICa.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   CRF increased the peak ICa in rat CeA neurons. Aa, CRF (50 nM) increased the peak ICa by 28% in this neuron, which was held at 70 mV and stepped to +10 mV. Ab, CRF-sensitive current was obtained by digitally subtracting the current recorded in control from that measured in CRF in the neuron shown in Aa. Ba, CRF (50 nM) increased the peak ICa by 40% when the neuron in A was stepped from 40 to +10 mV. Bb, Current enhanced by CRF at a holding potential of 40 mV. Leak currents were subtracted.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Time course of CRF action on the ICa. The effect of CRF was measured per min after the initial application. ICa was normalized to the peak current in control. Control recordings were averaged for 1 min, and CRF subsequently was added. Cells, which were treated but did not respond to CRF, were labeled "no effect." The run down of the peak whole cell ICa occurs in these experiments because there was no ATP-regenerating system added to the solution in the electrode. Peak ICa was elicited from 80 to 0 mV or +10 mV using 100-ms voltage step commands. n = number of neurons tested.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Effect of CRF on current-voltage relationships in CeA neurons. The current-voltage plot was constructed by measuring the magnitude of the peak ICa recorded before (control) and after application of CRF (100 nM) as well as during wash as a function of command potential. Currents were evoked using 100-ms depolarizing steps from 80 to +70 mV in 10-mV increments. A, Vh = 80 mV; CRF (100 nM) increased the maximal ICa (at +10 mV) by 22% in this cell. a, Original current traces recorded in control and CRF. b, Current-voltage plot of data obtained in a. c, CRF-sensitive current obtained by subtracting current in control from that recorded in CRF. The ICa was maximally enhanced by CRF (100 nM) at +20 mV. B, Vh = 40 mV; CRF (100 nM) increased the maximal ICa by 26% in another neuron held at 40 mV. Currents were evoked using 100-ms depolarizing steps from 40 to +50 mV in 8-mV increments. a, Current traces recorded in control and CRF. b, Current-voltage relationship in control and in the presence of CRF (100 nM) in same neuron as a. c, CRF-sensitive current obtained by digitally subtracting the current in control from that recorded in CRF in a neuron held at 40 mV. Leak currents in both A and B recorded in the presence of 200 µM Cd++ were subtracted.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Effect of CRF on voltage-dependent steady-state activation. Currents were measured at various test potentials (70 to +20 mV) from a holding potential of 80 mV. Istep/Imax was calculated as described in text. Curve fitting with a Boltzmann equation yielded a V1/2 of 10.5 mV in control and 11 mV in CRF; k values were 6.8 and 7.3 mV, respectively. Each point represents the mean ± S.E.M. of 4 cells. Step commands were 100 ms.

View larger version (14K): [in this window] [in a new window]   Fig. 6.   CRF does not affect steady-state inactivation. A and B were obtained from the same neuron before (trace 1) and after (trace 2) a 1-min application of CRF (100 nM). A, Effect of CRF (trace 2) on control currents (trace 1) elicited with 100-ms step commands from either 100 mV (IMax) or 40 mV (ITest) to +20 mV. B, Ratio of ITest/IMax in the absence (trace 1 in A, 39%, B1) and presence (trace 2 in A, 43%, B2) of CRF was similar. Leak currents were subtracted.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   CRF increased ICa in a concentration-dependent manner in rat CeA neurons. Concentrations of CRF tested ranged from 1 to 400 nM; the estimated EC50 value was 14.9 nM. Currents were elicited in neurons by 100-ms voltage commands from 80 to +10 mV. Currents were leak subtracted, and the value of ICa was corrected for decay of the ICa (see Methods). Numbers in parentheses are the numbers of neurons tested. Data are expressed as mean ± S.E.M. Hill coefficient = .82.

The CRF receptor antagonist -helical CRF_9-41 blocked the CRF-induced increase in I_Ca. We analyzed the effect of CRF antagonist -helical CRF_9-41 to examine whether a direct receptor mechanism mediated the CRF effect on the I_Ca. -Helical CRF_9-41, an amino-terminal-shortened analog of CRF and a synthetic competitive CRF antagonist (Rivier et al., 1984), can block many effects of CRF, including CRF-induced ACTH secretion in rat anterior pituitary (Rivier et al., 1984), CRF stimulation of neurotransmitter release (Sirinathsinghji and Heavens, 1989; Song and Takemori, 1992) and behavioral changes induced by stress and CRF (Boadle-Biber et al., 1993; Kiang, 1994; Menzaghi et al., 1994; Swiergiel et al., 1993). We first applied -helical CRF_9-41 (3 µM) and subsequently added CRF (50 nM) to the -helical CRF_9-41 (3 µM)-containing solution. In 8 of 15 neurons tested, -helical CRF_9-41 (3 µM) induced a slight but insignificant increase in the I_Ca (7 ± 2%), a possible partial agonist action. A partial agonist action for -helical CRF_9-41 has been reported previously in the amygdala (Rainnie et al., 1992) and other brain regions (Menzaghi et al., 1994). CRF (50 nM) applied in the presence of -helical CRF_9-41 (3 µM) resulted in only a 4.5 ± 1.5% increase in I_Ca (fig. 8) in 6 of 15 neurons, whereas the remaining cells (9 of 15) showed no response. This change in I_Ca (4.5 ± 1.5%, n = 6) was significantly less than that induced by CRF alone (50 nM, 25 ± 5%, n = 9, P  .05, unpaired t test). In addition, when the CRF-responding and nonresponding neurons were considered together, as shown in fig. 8B, the percent change in I_Ca in CRF alone (50 nM) was significantly different (P  .01, two-tailed Mann-Whitney rank-sum test) from that recorded in the presence of CRF (50 nM) plus CRF antagonist. These data suggest that the CRF-induced increase in the I_Ca is mediated by direct CRF receptor activation.

View larger version (17K): [in this window] [in a new window]   Fig. 8.   -Helical CRF9-41, a competitive CRF receptor antagonist, blocks the CRF-induced increase in the ICa in CeA neurons. A, Recording of ICa in a single neuron. ICa was elicited in neurons by voltage commands from 80 to +10 mV as control (trace 1); the addition of -helical CRF9-41 (3 µM) had no effect on ICa (trace 2). The subsequent addition of CRF (50 nM) in the presence of -helical CRF9-41 caused only a 5.6% increase of the whole-cell current (trace 3). B, Graph of the percentage change in ICa in neurons treated with CRF alone (50 nM) and CRF (50 nM) in the presence of -helical CRF9-41 (3 µM). The effect of CRF (50 nM) in each control CeA neuron is indicated by the solid symbols (n = 17). The effect of CRF (50 nM) in each neuron recorded in the presence of -helical CRF9-41 (3 µM) is indicated by the open symbols (n = 15). In the presence of the antagonist, CRF (50 nM) produced a 4.5 ± 1.5% (n = 6) increase in the ICa, a change significantly (P  .05, two-tail unpaired t test) less than that CRF (50 nM) induced alone (25 ± 5%, n = 9). The difference between the two populations of neurons is significant (P  .01, two-tailed Mann-Whitney rank-sum test).

CRF increased the resistant component of the I_Ca. We next tested the effect of CRF on DHP- and neurotoxin-sensitive and -resistant components of I_Ca. We recently described the presence of the multiple types of the I_Cas in acutely dissociated CeA neurons and found four components of HVA I_Ca s (Yu and Shinnick-Gallagher, 1994a, 1997): (1) NIM-sensitive, L-type (22%), (2) -conotoxin GVIA-sensitive, N-type (30%), (3) a Q-type (13-18%) current sensitive to -conotoxin MVIIC or Aga IVA and (4) resistant current recorded in the presence of NIM (5 µM), -conotoxin GVIA (1 µM) and -conotoxin MVIIC (250 nM) or Aga IVA (1 µM), which amounted to 37% to 53% (mean, 49%) of the I_Ca in CeA neurons. We examined the effect of CRF on the resistant I_Ca in the presence of all the pharmacological blocking agents. In these experiments, the I_Ca was elicited by voltage commands to 0 mV from holding potentials of 70 mV in cells pretreated with NIM (5 µM), -conotoxin GVIA (1 µM) and -conotoxin MVIIC (250 nM). Under these conditions, CRF (50 nM) still increased the remaining current by 31% (fig. 9). The CRF-sensitive current obtained in the presence of NIM, -conotoxin GVIA and -conotoxin MVIIC was similar with respect to that recorded in the absence of the antagonists (fig. 2, Ab, and fig. 8, Ab). Plotting the resistant current vs. time (fig. 9) showed that CRF increased the resistant current maximally at 1 min after application and then followed a time course of action of CRF similar to that observed with the whole-cell I_Ca (fig. 3). In a total of 7 neurons tested, CRF (50 nM) increased the resistant I_Ca by 20 ± 3%. The CRF (50 nM)-induced increase in resistant current recorded in the presence of the three calcium channel blockers was similar to that obtained in the absence of the antagonists (25 ± 6%, n = 7; P  .05). The estimated total current in the absence of the antagonists 1 to 6 min after the addition of CRF (96, 82, 79, 77, 68 and 64 pA, respectively) was not significantly greater than the CRF current (52 pA) recorded at 1 min in the presence of the antagonists. However, because the effect of CRF is numerically larger in the absence of the antagonists, it is possible that CRF may have some effect on component(s) of the DHP- and neurotoxin-sensitive current not detected under the present experimental conditions. Altogether, these data suggest that the primary effect of CRF on the I_Ca in CeA neurons is to enhance a DHP- and neurotoxin-resistant current.

View larger version (20K): [in this window] [in a new window]   Fig. 9.   CRF increased the NIM- and neurotoxin-resistant component of ICa. Data depicted in A and B were recorded from the same neuron. Aa, CRF (50 nM) increased the ICa by 31% after pretreating the cell with CgTx GVIA (1 µM) plus CgTx MVIIC (250 nM) plus NIM (5 µM) for 2 min. Ab, CRF-sensitive current was computed by subtracting the current in control from that in the presence of CRF. B, Time course of CRF action on the resistant ICa recorded from the same neuron as A. The holding potential was 70 mV, and the command potential was +10 mV.

	Discussion

Top Abstract Introduction Methods Results Discussion References

The present results show for the first time that CRF enhances HVA I_Cas in brain neurons and that this increase is mediated through DHP- and neurotoxin-resistant components in CeA neurons. The CRF enhancement of I_Ca was concentration dependent with an estimated EC₅₀ value of 14.9 nM. The effect of CRF was maximal at the peak of the current-voltage relationship and not mediated through alteration of the voltage dependence of steady-state activation or inactivation. These data also provide evidence that the effect of CRF on the I_Ca is through a receptor-mediated mechanism.

The finding here that CRF increased the I_Ca is consistent with the following studies implicating a calcium mechanism for CRF action: (1) CRF stimulates the release of neurotransmitters in central nervous system, including ACTH and -endorphin from the anterior pituitary (Axelrod and Reisine, 1984; Vale et al., 1981); GABA, dynorphin and methionine-enkephalin from rat neostriatum and globus pallidus (Sirinathsinghji et al., 1989; Sirinathsinghji and Heavens, 1989); and dynorphin A from mouse spinal cord (Song and Takemori, 1992). (2) CRF increases the duration of a Ca⁺⁺ action potential in amygdala neurons (Rainnie et al., 1992). (3) CRF increases cytosolic calcium of rat and human pituitary corticotrophs (Guerineau et al., 1991). (4) CRF enhances Ca⁺⁺ entry into human epidermoid A-431 cells and rat astrocytes (Kiang, 1994; Takuma et al., 1994), an action blocked by L- and P-type calcium channel blockers in rat corticotropes (Kuryshev et al., 1996).

The pharmacology of CRF receptor-mediated enhancement of I_Ca is in good agreement with that observed for native CRF receptors in the brain and cloned CRF receptors expressed in COS cells. In CeA neurons, the I_Ca was quite sensitive to CRF (EC₅₀ = 14.9 nM), suggesting the CRF receptors in the CeA neurons might bind CRF with high affinity, as reported in basolateral amygdala neurons (EC₅₀ = 40 nM; Rainnie et al., 1992) and other brain regions such as frontoparietal cortex (EC₅₀ = 36 nM; Battaglia et al., 1987) and retina (EC₅₀ = 20-30 nM; Olianas et al., 1993). Taken together, the data obtained in the present study indicate that CRF-induced increase in I_Ca occurs through activation of CRF receptors.

There are two subtypes of CRF receptors, CRF₁ and CRF₂, the latter of which has two splice variants: CRF₂, found primarily in brain, and CRF₂, localized in non-neuronal brain cells and in the periphery (Lovenberg et al., 1995). Rat/human CRF has an EC₅₀ value of 20 nM on adenylate cyclase activity in CRF₂ Ltk transiently transfected cells but an EC₅₀ value of 4 nM in CRF₁ Ltk stably transfected cells (Chalmers et al., 1996). K_i values for binding of CRF to CHO cells expressing CRF₁ vs. CRF₂ were 1 and 13 nM, respectively (Donaldson et al., 1996). Although it is not possible to distinguish between CRF receptor subtypes without the use of specific antagonists, the EC₅₀ value for CRF in the present study suggests that based on previous studies in transfected cells, the CRF effect on I_Ca in CeA neurons is more likely to be consistent with mediation through a CRF₂ receptor subtype.

The CRF-induced increase in I_Ca could be due to several mechanisms. A shift of the current-voltage relationship in the negative direction would cause more channels to open at negative potentials, thereby enhancing the I_Ca, but CRF did not shift the peak of the current-voltage relationship and voltage dependence of steady-state activation, suggesting that the CRF-induced increase in I_Ca may not be due to a change in the voltage dependence of channel activation. Furthermore, the ratio of I_Test/I_Max recorded at different holding potentials in the absence and presence of CRF was not statistically different, suggesting the CRF-induced increase in I_Ca was not mediated by a shift in the voltage dependence of steady-state inactivation over this range. It is possible that small changes in individual current components could be obscured in measurements of the voltage-dependent properties of the whole-cell current. However, these changes would have to be relatively small to not be detected in the present experiments. Further studies with single-channel recordings are required to determine definitively whether CRF enhances the I_Ca by increasing the single-channel conductance and/or increasing the channel open time.

Other neuropeptides increase HVA I_Cas; these include luteinizing hormone-release hormone in pituitary cells (Anwyl, 1991; for a review; see Rosenthal et al., 1987), angiotensin II in nodose neurons (Bacal and Kunze, 1994) and CGRP in nodose neurons (Wiley et al., 1992). In these studies, the luteinizing hormone-release hormone- and CGRP-induced enhancement of I_Cas is pertussis toxin sensitive, whereas the angiotensin II effect is insensitive and neither angiotensin II nor CGRP shift the current-voltage relationship. The CGRP-induced increase in HVA I_Ca is not caused by a change in voltage dependence of channel activation or steady-state inactivation but rather by an increase in maximal calcium conductance, probably by opening more channels (Wiley et al., 1992). The CRF-induced increase in the I_Ca in this study is similar to that of previous studies of CGRP because CRF did not shift voltage dependence of activation or inactivation of I_Ca, suggesting that the CRF effect, like CGRP, may also be due to an increase in maximal conductance resulting from an opening of more calcium channels in the presence of the peptide.

We found that CRF enhanced the DHP- and neurotoxin-resistant component of the HVA I_Ca in CeA neurons. Our data showed that CRF increased an HVA rather than an LVA I_Ca because the percent increase in currents elicited from 40 mV (26%) and 80 mV (28%) were not different and because LVA currents were not recorded under the experimental conditions. CeA neurons express multiple types of HVA Ca⁺⁺ channels (Yu and Shinnick-Gallagher, 1994a, 1997), namely, L, Q and N types as well as a resistant current accounting for the largest portion of the I_Ca. CRF increased the resistant I_Ca. The percentage increase in the whole-cell I_Ca (25%) induced by CRF (50 nM) was similar to that observed for the resistant current at equal molar concentrations of the peptide (20%). There is a possibility, however, that CRF may have an effect on some components of the DHP- and neurotoxin-sensitive currents not detected under our experimental conditions because the numerical values of the CRF currents and the percentage increase of I_Ca induced by CRF are larger in the absence than in the presence of the antagonists.

The DHP- and neurotoxin-resistant current in CeA neurons has some electrophysiological and pharmacological characteristics in common with cloned doe-1 and alpha-1E channels and native R-type I_Cas but remains unclassified (Yu and Shinnick-Gallagher, 1997). Our findings are consistent with CRF receptors being linked to resistant channels, perhaps of a doe-1, alpha-1E or R type.

The source of CRF terminals in the CeA originates intrinsically from CRF-containing cell bodies located in this nucleus and extrinsically via projections from other brain regions such as lateral hypothalamus, dorsal raphe and the medial geniculate complex (Imaki et al., 1991; Uryu et al., 1992). CRF terminals from projection neurons form synapses with non-CRF-containing neurons more frequently than with CRF-containing neurons (Imaki et al., 1991; Uryu et al., 1992). Furthermore, calcium-dependent CRF release evoked by a high concentration of potassium has been demonstrated in the amygdaloid complex (Smith et al., 1986; Takuma et al., 1994). These data along with our findings here suggest that CRF released locally at synapses may enhance the DHP- and neurotoxin-resistant I_Cas in predominantly non-CRF-containing neurons.

The CeA, as the major intra-amygdala target and output structure of the amygdaloid complex (Krettek and Price, 1978; McDonald, 1991, 1992; Stefanacci et al., 1992) projects to the nuclei involved in central control of autonomic activity and stress. The anatomic and pharmacological importance of the neuronal circuitry involving the CeA suggests that alteration in interneuronal communication within the CeA nucleus by a CRF-induced enhancement of I_Cas may be functionally relevant in regulating the autonomic, behavioral and endocrine responses to stress.