Selective potentiation of N-type calcium channels by angiotensin II in rat subfornical organ neurones

Abstract

1. Here we have characterized Ca²⁺ currents in rat subfornical organ neurones, and their modulation by angiotensin II. Currents were of the high voltage-activated (HNA) subtype, as the threshold for activation was near −30 mV (mid-point potential (V₅₀) of activation −14 mV). Using Ba²⁺ as the charge carrier, little inactivation was observed, and it occurred only at depolarized potentials (V₅₀ of inactivation −12 mV). More inactivation was observed using Ca²⁺ as the charge carrier, indicating that Ca²⁺-dependent inactivation plays a role in regulating Ca²⁺ channel function in subfornical organ (SFO) neurones.

2. The net Ba²⁺ current could be blocked by Cd²⁺ (EC₅₀ 1.6 μm), confirming that currents are of the HVA variety. By using selective antagonists, we identified the presence of both L- and N-type channels; 20 μm nifedipine blocked 22 ± 1 % of the current, while ω-conotoxin GVIA blocked 65 ± 7 %, indicating that these currents make up the net current through Ca²⁺ channels.

3. Angiotensin II potentiated the inward current throughout the range of activation. Using depolarizing voltage ramps, 1 nm angiotensin potentiated the peak current by 14 ± 5 %. We then used selective blockade of the HVA component currents; 20 μm nifedipine failed to prevent the potentiation by angiotensin II (12 ± 5 %), while blocking N-type channels with ω-conotoxin GVIA blocked the facilitation by ANG (2.3 ± 2 %). Losartan (1 μm) prevented the actions of ANG on the inward current (1.6 ± 1 %), indicating that the selective effects of ANG on N-type channels in SFO neurones are mediated by AT₁ receptors.

The control of systemic blood pressure in mammals is a tightly regulated system consisting of both endocrine and nervous components. The peripheral endocrine factors include a group of vaso- and renally active peptides that affect vascular resistance, sodium balance, and fluid volume. Angiotensin II (ANG), a circulating octapeptide with sodium retaining and vasoconstrictor effects, is also a potent central dipsogenic stimulus, acting directly at the subfornical organ (SFO) (Simpson & Routtenberg, 1973), the most rostral circumventricular structure. The SFO represents the interface between the systemic and central components of the integrated control of fluid balance and vascular resistance through direct neural connections to central nervous system (CNS) regions regulating the release of neurohypophysial hormones (Silverman et al. 1981; Lind, 1985; Smith et al. 1997) and sympathetic outflow. Receptor localization studies indicate that SFO neurones have among the highest reported density of ANG receptors in the CNS (Mendelsohn et al. 1984). The SFO angiotensinergic system has been suggested to play a role in the development of hypertension in experimental animals, as prehypertensive changes in the density of ANG receptors (Gutkind et al. 1988) and neuronal ANG immunoreactivity have been observed.

In addition to being exquisitely sensitive to ANG, SFO neurones have been shown to use ANG as a neurotransmitter in efferent connections to the paraventricular nucleus (Bains et al. 1992; Ferguson & Washburn, 1998). Thus, circulating ANG could influence the excitability of putative ANGergic neurones, and thus facilitate synaptic release of ANG in target areas of the hypothalamus and brainstem.

Although the direct excitatory actions of ANG on SFO neurones have been described extensively, both in vivo (Felix & Akert, 1974; Felix & Schlegel, 1978; Ferguson & Renaud, 1986) and in vitro (Buranarugsa & Hubbard, 1979; Schmid & Simon, 1992; Schmid et al. 1995; Ferguson et al. 1997), the mechanisms underlying the stimulatory effects have not been explored in depth. Studies from our laboratory have shown inhibition of a transient K⁺ conductance by ANG (Ferguson & Li, 1996), although this observation is not likely to explain completely the mechanism of ANG excitation, as we have also observed a decrease in input resistance induced by ANG (Ferguson et al. 1997). These observations could be reconciled by the positive modulation of an additional channel that would lead to increased membrane conductance and serve as a depolarizing drive, such as the non-selective cation channel described by Ono et al. (2000). Analysis of intracellular Ca²⁺ transients in SFO neurones by Gebke et al. (1998) suggest that ANG may influence Ca²⁺ channels, as ANG was found to trigger a rise in intracellular Ca²⁺ concentration that was abolished by removal of extracellular Ca²⁺. Patch clamp studies in cultured hypothalamic/brainstem neurones (Sumners et al. 1996) indicate that the excitatory actions of ANG on these cells occur, at least in part, due to stimulation of voltage-gated Ca²⁺ channels. Furthermore, Wang et al. (1997) have shown that a Cd²⁺-sensitive, putative Ca²⁺ current contributes to spontaneous spiking and is the target of positive modulation by ANG. Together, these previous studies set the framework for the present investigation wherein we characterize the voltage-gated Ca²⁺ channels expressed in dissociated SFO neurones and their modulation by ANG. These findings are discussed as one of the potential mechanisms through which circulating ANG acts on SFO neurones to exert its potent neuroendocrine and behavioural effects.

METHODS

Subfornical organ neurone isolation

Subfornical organ neurones were obtained from adult (125–150 g) male Sprague-Dawley rats (Charles River, PQ) as previously described (Ferguson et al. 1997). All experimental procedures are in accordance with the Queen's University animal care guidelines, and conform to the standards outlined by the Canadian Council for Animal Care. Animals were placed in a disposable rodent restrainer (Braintree Inc.) and decapitated as quickly as possible. The brains were rapidly removed and placed in ice-cold Ca²⁺/Mg²⁺-free Hanks' buffered salt solution (HBSS; Gibco). With the aid of a dissection microscope, a tissue block containing the hippocampal commissure was removed, and the SFO dissected free of all surrounding tissue. Cells were isolated by incubation in Ca²⁺/Mg²⁺-free HBSS supplemented with 1 mg ml⁻¹ trypsin (Sigma; prepared from a frozen stock before each isolation) and incubated at 37 °C in a 95 % CO₂-5 % O₂ incubator for approximately 30 min, with periodic gentle shaking. The tissue was then gently triturated with a tuberculin syringe fitted with a 20 gauge needle. The cell suspension was then centrifuged twice at 900 g, for 5 min in HBSS containing Ca²⁺, Mg²⁺ and 1 % BSA (Sigma). Cells were plated on 35 mm culture dishes (Gibco) to which they rapidly adhered. After 1 h, neurobasal A (Gibco) medium supplemented with l-glutamine (Gibco) was added to the culture dishes.

Drugs and solutions

All drugs were dissolved in artificial cerebrospinal fluid (aCSF) of the following composition (mm): 120 NaCl, 10 BaCl₂, 10 tetraethylammonium chloride, 1 4-aminopyridine, 1 MgCl₂, 10 Hepes, 10 glucose, and 0.0005 tetrodotoxin (pH adjusted to 7.4 with CsOH), and delivered through bath perfusion at a rate of 1–3 ml min⁻¹. For experiments using Ca²⁺ as the charge carrier, 8 mm CaCl₂ was added, and BaCl₂ omitted. The patch pipette solution consisted of (mm): 130 CsCl, 10 EGTA, 10 Hepes, 2 Na-ATP, and 2 MgCl₂; pH adjusted to 7.3 with CsOH. Perforated patch experiments were carried out with amphotericin B (Sigma; protected from light whenever possible) dissolved first in dimethylsulfoxide (DMSO; Sigma) and sonicated for 1 min in pipette solution at a maximum concentration of 0.1 % DMSO. Na-ATP was omitted from the pipette solution in perforated patch experiments.

A stock solution of the dihydropyridine L-type Ca²⁺ channel blocker nifedipine (Sigma) was dissolved first in DMSO and then diluted in aCSF to a final concentration of 10 μm, with DMSO at a maximum concentration of 0.1 %. Nifedipine was protected from light at all times. The specific N-type antagonist ω-conotoxin GVIA (ω-CTX; Alomone) and the P/Q-type antagonist ω-agatoxin (ω-AgTx; Alomone) were dissolved in artificial cerebrospinal fluid (aCSF) with 0.1 % cytochrome c (Sigma). Vehicle controls consisted of aCSF with 0.1 % cytochrome c.

Electrophysiology

Conventional whole-cell and amphotericin perforated patch experiments were processed with a List EPC-7 amplifier, filtered at 3 kHz with an 8 pole Bessel filter, digitized using the CED 1401 Plus interface at 10 kHz, and stored on computer for offline analysis. Data were collected using the Signal acquisition package (CED, Cambridge, UK). Patch clamp recordings were obtained using micropipettes pulled from thin-walled glass (World Precision Instruments) using a P-87 Flaming/Brown pipette puller (Sutter), with tip resistances of 1.5–3 MΩ when filled with the CsCl-based pipette solution. Junction potentials were less than 2 mV. For perforated patch experiments, series resistance (R_s) of less than 20 MΩ was required, and compensated for (70–80 %). In whole-cell recording mode, R_s was monitored and compensated throughout the course of the experiment. Experiments were terminated in all cases where R_s was not stable for the entire protocol. In all pharmacological and ANG experiments, run-down/up of currents was controlled for by assuring that stable steady-state currents were achieved before the onset of the experiment, and therefore experiments were not initiated until at least 10 min after achieving access. Control experiments indicate that in our hands, SFO HVA currents do not run-down, and in fact often showed a small amount (up to 15 %) of run-up when tested up to 30 min after achieving whole-cell access (n = 6, data not shown). In perforated patch experiments, currents were stable for up to 30 min after achieving access. Normalized conductance (G_Ca) values were calculated from the following formula:

where E_Ca represents the experimentally derived reversal potential for I_Ca/Ba (+42 mV).

RESULTS

Properties of SFO neurone calcium channels

Activation of calcium channels was studied through whole-cell voltage clamp protocols consisting of 250 ms depolarizing voltage pulses from a holding potential of −60 mV, or fast (800 mV s⁻¹) depolarizing voltage ramps. The threshold for activation of the inward current (I_Ba) was between −40 and −30 mV, resulting in a rapidly activating, slowly inactivating current (Fig. 1A), which reached a peak between −10 and +5 mV (Fig. 1B). There were no differences between the current-voltage relationships obtained with step commands versus the voltage ramps for five cells measured directly through both methods (Fig. 1B). Holding at a hyperpolarized membrane potential (−100 mV) failed to uncover a low-threshold current (inset, Fig. 1A and B; n = 5). Exclusively HVA Ca²⁺ channels carried inward Ba²⁺ currents activated by these voltage protocols, as the currents were blocked completely by Cd²⁺ (Fig. 1C; see Fig. 3). The activation profile of the Ba²⁺ conductance (G_Ba) could be fitted to a Boltzmann equation (Fig. 1D), from which a V₅₀ of activation of −14.1 mV and slope of 5.8 mV⁻¹ were derived (r² = 0.98, n = 7).

Figure 1. Activation properties of subfornical organ Ca²⁺ channels.

A, activation of Ca²⁺ channels by depolarizing voltage steps (250 ms, −50 to +70 mV, holding potential −60 mV; 0.5 μm TTX, 10 mm BaCl₂ as charge carrier) leads to inward currents typical of neuronal HVA channels (mean I–V relation shown in B; n = 5). Inset: holding at hyperpolarized potentials (−100 mV) fails to uncover a low threshold current. Shown here is an example of currents elicited from depolarizing pulses to −60 mV (1), −50 mV (2), −40 mV (3), and −20 mV (4). I–V relation is shown in the inset to B (filled symbols). B, current-voltage relation for currents obtained through both square wave pulses (filled symbols) and depolarizing voltage ramps (open symbols). C, inward Ba²⁺ currents (holding potential 60 mV, test potential 0 mV) are completely blocked by CdCl₂, indicating that they are mediated by HVA Ca²⁺ channels. D, normalized conductance plot of steady-state inward current against step potential, fitted to a Boltzmann function (V₅₀=−14.1 mV, slope 5.8 mV⁻¹). •, calculated from step voltage commands; ○, generated from voltage ramp commands.

Figure 3. Pharmacology of SFO Ca²⁺ channels.

A, shown here is the dose-response relationship for block of SFO neurone inward Ba²⁺ currents by Cd²⁺. Peak I_Ba is blocked by Cd²⁺ with an EC₅₀ of 1.6 μm, obtained from voltage ramp recordings in a total of 15 cells (each cell tested indicated by a different symbol). B, a small proportion (22 ± 1 %, see F) of I_Ba is blocked by nifedipine (20 μm). The residual current is completely blocked by 100 μm Cd²⁺, indicating that it is carried exclusively by HVA channels. C, a larger fraction (65 ± 7 %) is blocked by the specific N-type channel antagonist ω-conotoxin GVIA, (100 nm; see F). Scale bar applies to C and E. D, P/Q-type HVA channels do not contribute substantially to the net HVA current, as indicated by the relative lack of effect of 50 nmω-AgTx. E, blocking N- and L-type channels leaves a small residual, potentially R-type current. F, summary histogram showing the relative fractions of net inward current blocked by ω-CTX, nifedipine and ω-AgTx.

During 300 ms depolarizing voltage pulses, inward Ba²⁺ currents showed weak time- and voltage-dependent inactivation. The steady-state voltage dependence of inactivation was studied through protocols consisting of 300 ms prepulses (from −110 to 0 mV) delivered prior to the test pulse to +10 mV (Fig. 2Aa). Relative peak I_Ba was fitted to a Boltzmann function (Fig. 2Ab, filled symbols) with V₅₀ and slope values of −12.6 mV and −10 mV⁻¹ respectively. The voltage-dependent inactivation of the net inward I_Ba was incomplete, showing a mean inactivation of 20 ± 3 % with the 0 mV prepulse (Fig. 2Ab). In light of the relatively modest inactivation of I_Ba, we tested the possibility of Ca²⁺-dependent inactivation by using 8 mm CaCl₂ as the charge carrier. Steady-state voltage-dependent inactivation was found to be more extensive with Ca²⁺ as the charge carrier, as shown by the greater sensitivity to voltage prepulses (Fig 2Ab, open circles). The plot of relative current against prepulse potential did not fit a single Boltzmann function, but rather appeared to follow two distinct processes. We next studied the time course of inactivation of HVA currents and the effect of charge carrier. Normalizing peak inward currents (Fig. 2Ba) indicated that I_Ca exhibits a transient component inactivating more rapidly than I_Ba, although not completely. Inactivation was more complete in Ca²⁺ conditions across the voltage range tested (−20 to +30 mV; Fig. 2Bb; analysed by ANOVA, *P < 0.01, by Neumann-Keuls post hoc test). Together, these data suggest that inactivation of SFO HVA channels is strongly dependent upon modification by Ca²⁺ of voltage-dependent processes, as using Ca²⁺ as the charge carrier appeared to facilitate voltage-dependent inactivation. The inactivating component of I_Ca activated in the same voltage range as I_Ba, suggesting that it is due to Ca²⁺-mediated inactivation, and not the emergence of a new conductance.

Figure 2. Inactivation properties of SFO Ca²⁺ channels.

Aa, inactivation protocol (300 ms prepulse, −110 to 0 mV, 300 ms test step to +10 mV) reveals relative insensitivity of SFO Ba²⁺ currents to voltage-dependent inactivation. Note the lack of change in inward current amplitude with progressively more depolarized voltage prepulses. Ab, a plot of relative current (elicited during the test pulse to +10 mV) against prepulse potential showing the small degree of voltage-dependent inactivation occurring with Ba²⁺ as the charge carrier (•, n = 10). Voltage dependence of inactivation is enhanced with 8 mm Ca²⁺ (○, n = 8) as the charge carrier. Ba, substitution of 8 mm Ca²⁺ as the charge carrier facilitates inactivation. Note the more rapid, yet still incomplete, time-dependent inactivation. Traces are currents normalized to peak, generated from voltage steps to +10 mV from a holding potential of 80 mV. Bb, mean inactivation of HVA currents generated from depolarizing voltage pulses (−20 to +30 mV) measured 200 ms into the pulse with Ba²⁺ (n = 10) or Ca²⁺ (n = 8) as the charge carrier (*P < 0.01 by ANOVA, followed by Neumann-Keuls post hoc test).

Pharmacology of SFO calcium channels

We next explored the pharmacology of Ca²⁺ channels present in SFO neurones. The net inward I_Ba was found to be completely Cd²⁺ sensitive, in a dose-dependent manner that fitted well to a single modulus function (Fig. 3A) obtained from testing Cd²⁺ on a total of 15 cells, some of which were tested with several doses of Cd²⁺. In each case, the relative block by Cd²⁺ was calculated by normalizing the remaining current to the pretreatment peak current. To control for the possibility of run-down/up, complete washout of the previous dose of Cd²⁺ was ensured before application of the succeeding dose. External Cd²⁺ was an effective blocker of I_Ba, with an EC₅₀ of 1.6 μm. To identify the specific Ca²⁺ channel subtypes present in SFO neurones, we used more specific Ca²⁺ channel antagonists. Nifedipine (20 μm) blocked 22.1 ± 1 % of I_Ba (Fig. 3B and F; n = 5) in a voltage-independent manner, indicating the presence of an L-type current. The specific N-type Ca²⁺ channel antagonist ω-conotoxin GVIA (ω-CTX) blocked the largest proportion of the inward current, inhibiting 65 ± 7 % (Fig. 3C and F; n = 7). N-type channel block by ω-CTX developed slowly and was generally irreversible. The presence of P/Q-type Ca²⁺ channels was investigated through the specific blocker ω-agatoxin IVA (ω-AgTx; 50 nm; Fig. 3D and F). P/Q-type Ca²⁺ channels contribute to only a small proportion of the net HVA current, as 50 nmω-AgTx blocked only 3.0 ± 2 % of the inward current (n = 5). Four out of five cells tested showed no significant attenuation of the current in the presence of ω-AgTx, while one cell showed a small reduction of current. Together, these data indicate that the HVA current in SFO neurones is comprised largely of L- and N-type Ca²⁺ currents. There is, however, the potential for some degree of heterogeneity, as it would not be expected that all SFO neurones would have an identical proportion of component currents. Application of 20 μm nifedipine and 100 nmω-CTX in combination (Fig. 3E) blocked 89 ± 5 % of the total HVA current in four cells tested, suggesting that the remaining current may be a resistant ‘R-type’.

Actions of ANG on SFO calcium channels

We next tested the hypothesis that Ca²⁺ channels in SFO neurones are one of the ionic-specific targets for ANGergic modulation. Currents elicited by depolarizing voltage pulses (300 ms, −50 to +40 mV, amphotericin perforated patch) were potentiated by 1 nm ANG (Fig. 4A). Using depolarizing voltage ramps, we found that the inward I_Ba was potentiated by ANG (1 nm) in a voltage-independent manner, as shown by the ANG-induced difference current (Fig. 4B) that was present throughout the range of HVA channel activation. Under these conditions, 1 nm ANG potentiated the peak inward current by 14 ± 5 % (n = 12). Peak facilitatory effects of ANG on the inward HVA current occurred between 3 and 4 min, as shown in the time course plot (Fig. 4C; n = 5) relative to time-matched controls (perforated patch; n = 7). The time course of effect on Ca²⁺ channels was found to be similar to that reported for excitation in slice preparations (Schmid & Simon, 1992). Although our goal in this study was to use physiologically relevant levels of ANG, we also tested doses of ANG up to two orders of magnitude greater; 10 nm (n = 4) and 100 nm (n = 4) ANG were found to potentiate inward HVA currents by 26 ± 4 % and 36 ± 6 % respectively. Given our previous identification of two main components of the net inward current mediated by voltage-gated Ca²⁺ channels, we next sought to identify which of these subtypes were modulated by ANG as observed through perforated patch voltage clamp experiments. Using 20 μm nifedipine (a dose effective in blocking neuronal L-type channels), we found that 1 nm ANG was able to facilitate the ramp-evoked I_Ba by 12 ± 5 % (n = 5, Fig. 5A). These findings suggest that dihydropyridine-sensitive L-type channels are not the major target of ANG-induced facilitation, as the increase in current was still evident under blocking conditions. Conversely, when N-type channels were selectively blocked with 100 nmω-CTX (Fig. 5B), the facilitation of current by ANG was essentially abolished. Under these conditions, 1 nm ANG changed the peak inward current by only 2.3 ± 2 % (n = 5, P < 0.05 vs. control, ANOVA followed by Neumann-Keuls post hoc test). These data are summarized in Fig. 5D. Pretreatment with the AT₁ receptor antagonist losartan (1 μm) prevented potentiation of I_Ba by ANG (1.6 ± 1 %; n = 5), indicating the involvement of this receptor subtype. These data suggest that ANG selectively potentiates ω-CTX-sensitive N-type channels through an AT₁ receptor-dependent mechanism.

Figure 4. ANG potentiates inward HVA currents in SFO neurones.

A, Ba²⁺ currents from 300 ms voltage pulses (−50 to +40 mV; holding potential 60 mV) under control (top traces) and 1 nm ANG (middle traces) conditions. Bottom: peak currents (−10 mV step) overlaid from control and ANG conditions. B, currents evoked from depolarizing voltage ramps under the same conditions as in A indicate that ANG potentiates the inward Ba²⁺ current through the range of activation, as shown in the ANG difference current (bottom panel). C, time course of ANG facilitation of inward ramp-evoked currents (•; n = 5). ○, relative control currents (n = 7).

Figure 5. Angiotensin II selectively targets N-type channels through an AT₁ receptor mechanism.

A, pretreatment with nifedipine (20 μm) failed to prevent potentiation of inward current by ANG, suggesting that L-type channels are not a target of ANGergic modulation. In this, and all subsequent panels, trace 1 indicates the control antagonist condition, and trace 2 indicates treatment with ANG. B, pretreatment with the specific N-type Ca²⁺ channel antagonist ω-CTX (100 nm) completely occludes the stimulatory effect of 1 nm ANG on inward HVA current. C, effects of ANG on SFO Ca²⁺ channels are mediated by an AT₁ receptor, as pretreatment with 1 μm losartan prevented the potentiation. D, summary histogram indicating the selective potentiation of N-type channels through an AT₁ receptor mechanism. *P < 0.05, ANOVA, followed by Neumann-Keuls post hoc test.

DISCUSSION

Since the initial observation that the SFO represents the central locus through which circulating ANG acts to induce drinking behaviour (Simpson & Routtenberg, 1973), considerable attention has been paid to this circumventricular structure, with particular emphasis on its role as the interface between the protected brain and the milieu interieur. Although there exists a significant body of literature describing the behavioural, autonomic, and neuroendocrine roles of the SFO, the cellular mechanisms through which the various chemical messengers initiate these effects at the SFO remain largely unexplained. Through in vitro patch clamp studies in recent years we have begun to understand the complement of ion channels (Washburn et al. 1999a; Johnson et al. 1999; Washburn et al. 2000b) that control the firing patterns of SFO neurones and their modulation by various extracellular messengers, including vasopressin (Washburn et al. 1999a), osmolarity (Anderson et al. 2000), and Ca²⁺ ions (Washburn et al. 1999b). To date, the only electrophysiological studies exploring the ionic mechanisms of ANG-induced excitation of SFO neurones show an AT₁ receptor-mediated inhibition of the transient outward K⁺ current (Ferguson & Li, 1996). Here we characterize the previously undescribed voltage-gated Ca²⁺ channels present in SFO neurones and their positive modulation by ANG. In this study we report that SFO neurones express a complement of HVA Ca²⁺ channels, with N-type channels making up the largest component, and L- and P/Q-type channels rounding out the net inward current. Furthermore, these studies show that the ω-CTX-sensitive N-type current is the preferential target for AT₁-receptor-mediated potentiation.

The HVA currents found in SFO neurones are typical of the known endogenous and cloned N-type Ca²⁺ channels in both their activation parameters and pharmacological sensitivity. The relative insensitivity of SFO Ca²⁺ channels to voltage-dependent inactivation processes are, however, more unique to specific N-type channel subtypes. N-type Ca²⁺ channels are formed as heteromultimeric protein complexes around an α_1B backbone that typically results in a rapidly inactivating inward current (Zhang et al. 1994). Although variations in the α₁ subunit confer the characteristic activation, inactivation, and pharmacological properties of HVA channels, accessory β subunits can have profound effects on the kinetics of voltage-dependent inactivation (Stotz & Zamponi, 2001). Association with β₁-subunits can, for example, result in dramatic slowing of inactivation in N-type/α_1B channels. Although not explored directly in these studies, the pharmacological and inactivation profiles of the SFO neurone HVA currents most closely resemble those that would be anticipated to arise from a channel comprised of α_1B and β₁ subunits (G. Zamponi, personal communication).

Rates of inactivation of the HVA current in SFO neurones were shown to be strongly dependent upon Ca²⁺ in the extracellular solution. Ca²⁺-dependent inactivation processes have been described for N-, L- and P/Q-type HVA channels, although they appear to be mediated through different mechanisms (Cox & Dunlap, 1994). Both L- (Zuhlke et al. 1999) and P/Q-type (Lee et al. 1999) Ca²⁺ channels have been reported to inactivate through Ca²⁺-dependent actions of calmodulin.

Physiological implications

High voltage-activated Ca²⁺ channels have been implicated in the regulation of firing patterns in many neuronal systems, including bursting cells of the striatum (Phillips & Stamford, 2000), subthalamic nucleus (Beurrier et al. 1999) and hypothalamic neurosecretory cells (Fisher & Bourque, 1996; Boehmer et al. 2000). Our present study suggests that ANG potentiation of N-type channels may be one of the ionic mechanisms through which circulating ANG is involved in the excitation SFO neurones. Although HVA Ca²⁺ currents would not be expected to be activated at rest (and therefore probably do not trigger the initial depolarization) it would be predicted that potentiation of these currents which would be activated during depolarization (such as that attributed to a non-selective cation channel regulated by ANG as well; Ono et al. 2000) could contribute to the overall spike patterning produced during excitation. In bursting SFO neurones, for example, ANGergic recruitment of an inward current could contribute to either burst maintenance (inward current would support the depolarization) or termination (Ca²⁺ influx could trigger Ca²⁺-activated K⁺ channels). These potential mechanisms would suggest important roles for these channels in shaping the bursts of action potentials which we have suggested may contribute to burst-mediated ANGergic neurotransmission by SFO neurones (Washburn et al. 2000a). This conceptual framework centres on the idea that extrinsic modulation of SFO neuronal ion channels allows precise control of action potential discharge patterns that may allow the preferential release of ANG and/or a more traditional amino acid neurotransmitter in the appropriate target nucleus.

Similarly, modulation of presynaptic N-type channels are yet another potential target for control of neuroendocrine output. Since N-type channels are known to mediate synaptic neurotransmitter release, the potential exists for ANG to have stimulatory presynaptic effects in targets of ANGergic projections. It is feasible that ANG released from PVN-projecting SFO neurones could act upon their own or neighbouring terminals in a facilitatory fashion. A positive modulatory effect on presynaptic N-type channels could potentially increase the probability of either amino acid (i.e. glutamate, GABA) or peptide (ANG) transmitter release, as a result of presynaptic [Ca²⁺] elevation. This model would be consistent with the observed facilitatory effects of ANG on evoked EPSPs in the nucleus tractus solitarius (Kasparov & Paton, 1999).

In addition to contributing to electrical activity of neurones, voltage-gated Ca²⁺ channels have been implicated in the control of various regulatory enzymes and activity-dependent gene expression. Angiotensin II has been shown to increase the expression of inducible transcription factors (ITFs) through AT₁ receptor-dependent mechanisms in the SFO and other brain regions (Moellenhoff et al. 1998). Additionally, experimental models in which circulating ANG levels are elevated (spontaneously hypertensive rats, chronic dehydration) also show strong expression of ITFs and target genes (e.g. AT₁ receptor). The possibility exists that HVA Ca²⁺ channels play a role in regulating gene expression in these situations, thus serving as a focal point for ANGergic regulation of both electrical activity and protein expression. Although most of the work to date implicates L-type channels in the synaptic control of nuclear gene expression (Murphy et al. 1991), recent work (Brosenitsch & Katz, 2001) indicates that N-type channels may in fact mediate gene expression induced by physiological neuronal firing patterns. Thus, it is possible that in SFO neurones, where the majority of the somatically localized channels appear to be N-type, these channels may in fact serve as the conduit through which Ca²⁺ transients could arise and affect nuclear transcription and regulatory processes.

The present study indicates that the complement of voltage-gated Ca²⁺ channels present in dissociated SFO neurones are exclusively of the HVA variety, as no evidence of a low-threshold T-type current could be found. Furthermore, ω-CTX-sensitive N-type channels appear to be both the predominant subtype expressed in these cells, and the exclusive Ca²⁺ channel target for potentiation by ANG.