Preparations

Adult male and female dwarf gourami (Colisa lalia), ∼4 cm in standard length, were purchased from a local dealer and kept at 22 ∼ 27 C° until used. The whole brain in vitro preparation was made according to the procedures described in the previous papers (Oka, 1995, 1996; Abe and Oka, 1999, 2000).

Electrophysiology

The in vitro whole brain preparation was continuously super-fused with an oxygenated Ringer solution containing (in mM) 124 NaCl, 5 KCl, 1.2 KH₂PO₄, 2.4 CaCl₂, 1.3 MgSO₄, 26 NaHCO₃, and 10 glucose (pH 7.4 adjusted with NaOH) in the silicone elastomer (Shin-Etsu Silicone No.KE-106, Shin-Etsu Chemical Co., Ltd., Japan) base of a small recording chamber. Whole cell voltage- and current-clamp recordings were carried out with the use of CEZ-2300 amplifier (Nihon Kohden, Japan) and pCLAMP software (Axon instruments). Pipette resistance was ∼8 MΩ, and seal resistance was >10GΩ. Series resistance was compensated as much as possible. The patch pipette was visually guided to the cluster of TNGnRH neurons exposed on the ventral surface of the brain under a dissecting microscope (Oka and Matsushima, 1993). After gigaohm seal formation and “break in” for the whole cell recording mode, characteristic spontaneous pacemaker activity was confirmed in the current-clamp mode (see Oka and Matsushima, 1993; Abe and Oka, 2000).

For current-clamp experiments, the patch pipettes contained (in mM) 110 K-gluconate, 3 MgCl₂, 40 N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES), 0.3 ethylene glycol-bis (β-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA), 2 Na₂ATP, and 0.2 Na₂GTP (pH 7.4 adjusted with NaOH). NiCl₂ (1 mM), LaCl₃ (5 μM), Apamin (100 nM; Alomone Labs.) and salmon GnRH (sGnRH, 200 ∼ 300 nM; RBI) were dissolved directly in the Ringer solution. Ruthenium red (50 μM; Sigma) and heparin (300 μg / ml; Sigma) were dissolved in the pipette solution. Pacemaker activities were digitized (2 kHz), displayed on-line with Axotape software or pClamp8 (Axon Instruments), and stored on a computer.

For voltage-clamp recording of Ca²⁺ currents, the extracellular solution contained (in mM) 95 choline Cl, 40 tetraethylammonium chloride (TEACl), 2.4 CaCl₂, 1.3 MgCl₂, 5 Glucose, 5 4-aminopyri-dine (4AP), and 10 HEPES (pH 7.4 adjusted with NaOH). In addition, 0.75 μM tetrodotoxin (TTX) was added to the extracellular solution. Patch pipettes were filled with a solution consisting of (mM) 90 CsCl₂, 2 MgCl₂, 20 TEACl, 10 EGTA, 10 HEPES, 2 Na₂ATP, and 0.2 Na₂GTP (pH 7.4 adjusted with CsOH). For voltage-clamp recording of tentative Ca²⁺-activated K⁺ currents, the extracellular solution contained (in mM) 115 choline Cl, 20 TEACl, 5 KCl, 2.4 CaCl₂, 1.3 MgCl₂, 5 glucose, and 10 HEPES (pH 7.4 adjusted with NaOH). 0.75 μM TTX was also added to the extracellular solution. Patch pipettes were filled with a solution consisting of (mM) 110 KCl, 3 MgCl₂, 40 HEPES, 0.3 EGTA, and 2Na₂ATP (pH 7.4 adjusted with NaOH). The linear leakage currents were digitally subtracted, either automatically with the use of the P/4 protocol, or manually, after measuring ohmic resistance in response to hyper-polarizing command pulses. The data were not corrected for the liquid junction potentials.

Data analysis

Data analysis, fitting, averaging and presentation were carried out using a combination of pCLAMP6 and 8, Axograph (Axon Instruments), Microsoft Excel (Microsoft), DeltaGraph (Polaroid), and Canvas (Deneba software) softwares. All data are present as means±SE.

Involvement in the control of pacemaker activity of Apamin-sensitive Ca²⁺-activated K⁺ currents induced by Ca²⁺ released from the intracellular store

First, in order to examine the possible involvement of Ca²⁺ released from the intracellular store in the control of pacemaker activity of TN-GnRH neuron, the cells were dialyzed with ruthenium red, an inhibitor of Ca²⁺-induced Ca²⁺ release, and heparin, that of IP₃-induced Ca²⁺ release, respectively, by including them in the patch pipette solution. After that, the effects of bath application of sGnRH on the pacemaker activity were examined. To ensure the diffusion of ruthenium red and heparin into the cytoplasm of the TNGnRH neuron, the data collection was started 10 min after the whole-cell recording was established; this time period was determined on the basis of the results of intracellular application of QX-314 (see Abe and Oka, 2000). Fig. 1 shows the pacemaker activity of a cell recorded with patch pipette containing ruthenium red (50 μM) and heparin (300μg / ml) in the pipette solution. In Ringer solution, the cells showed slightly irregular beating discharge pattern (Fig. 1Ba). Many cells tended to show such a firing pattern by intracellular application of the blockers of Ca²⁺ release from the intracellular store. In these experiments, we further applied La³⁺ (5 μM) to the bath solution in order to block some Ca²⁺ channels (see below). Bath application of La³⁺ slowed down the frequency of pacemaker activity (Fig. 1Bb). Further bath application of sGnRH (1 μM) failed to evoke transient decrease of firing frequency (Fig. 1Bc) but evoked subsequent increase of firing frequency (Fig. 1Bd). Figure 1Bc and Bd (20 and 120 s after the onset of sGnRH perfusion, respectively) were recorded during time periods similar to that of transient decrease and subsequent increase of firing frequency by sGnRH. Fig. 1Bf shows the time course of these changes before and after the application of sGnRH (during the period corresponding to Fig. 1Bb∼d).

Fig. 1

Inhibition of Ca²⁺ release from the intracellular store disrupts the transient decrease of pacemaker frequency after GnRH application. (A) Continuous recording of the pacemaker activity. Traces indicated by bars a-f are shown in (B) on enlarged time scales. Similar conventions are used in Figs. 3, 5, and 6. Intracellular application of Ruthenium Red (50 μM), which inhibits the Ca 2+-induced Ca2+ release, and heparin (300 μg/ml), which inhibits IP₃-induced Ca²⁺ release, blocked transient decrease (Bc) but not subsequent increase in the frequency of pacemaker activities (Bd), both of which should be induced by sGnRH. Compare with normal changes in the frequency of pacemaker activities that are induced by sGnRH shown in Fig. 3Ba–d.

We compared the degree of transient decrease of firing frequency in the presence and absence of the intracellular Ca²⁺ mobilization blockers. Each effect of the blockers was measured during a 20 s period that was 30 ∼ 50 s after the onset of sGnRH perfusion, which corresponded to the early transient phase of pacemaker frequency decrease, and was normalized to the frequency in Ringer solution. The transient decrease that was induced by sGnRH (0.88±0.06; n=16) was nullified by heparin (1.01±0.05; n=5). Similarly, intracellular application of ruthenium red alone or both ruthenium red and heparin nullified the transient decrease or rather increased slightly the firing frequency (1.11±0.14; n=7 and 1.16±0.11; n=4, respectively). However, we could not find statistically significant nullifying effects of blockers of Ca²⁺ release from the intracellular store due to the large SE. Anyway, these results suggest that the Ca²⁺ release from the intracellular store may be involved in the transient decrease of firing frequency that was induced by sGnRH.

In our previous studies of voltage-dependent K⁺ cur-rents (Abe and Oka, 1999), we found that a transient and 4AP-sensitive, but not TEA- or charybdotoxin-sensitive Ca²⁺-depenent outward current component was present in TN-GnRH neurons. Because this current could be observed only when Ca²⁺ was present in the external solution, we defined this 4AP-sensitive transient outward current that was dependent on the presence of extracellular Ca²⁺ ions as tentative Ca²⁺-activated K⁺ currents. When the current responses were measured in Ca²⁺-containing extracellular solutions containing 0.75 μM TTX and 20 mM TEA, a mixture of large transient currents and smaller persistent outward currents could be recorded in response to depolarizing pulses from a holding potential of −100 mV. The activation and steady-state inactivation curves of the mixture current containing tentative Ca²⁺-activated K⁺ current are shifted to more depolarized potentials compared with those of “4AP-sensitive transient current” (data not shown). These tentative Ca²⁺-activated K⁺ currents are only partially inactivated at membrane potentials of −60 to −40 mV, which correspond to the base membrane potentials of the pacemaker activity of TN-GnRH neurons. Thus, it is reasonable to think that tentative Ca²⁺-activated K⁺ current can be activated in the pacemaker range of TN-GnRH neurons.

Next, we examined the effect of sGnRH on the tentative Ca²⁺-activated K⁺ current. Fig. 2A shows the superimposed traces of the tentative Ca²⁺-activated K⁺ currents before, during, and after sGnRH applications elicited during a series of +50 mV test pulses from a holding potential of −100 mV. In this case, bath application of sGnRH (1 μM) facilitated current amplitude about 18% (Fig. 2A). The current amplitude recovered to the control level after washout by Ringer solution. Fig. 2B shows the time course of the tentative Ca²⁺-activated K⁺ current before, during, and after sGnRH applications. Test pulses were applied at 15 s intervals. The current amplitude showed a tendency to gradually decrease due to the rundown. However, it was rapidly increased by bath application of sGnRH (within 15 to 90 s; Fig. 2B, solid bars). This time course corresponded to the onset of the transient decrease of firing frequency of pacemaker activity induced by sGnRH. Averaged increase of the amplitude of tentative Ca²⁺-activated K⁺ current was about 9% (2 to 19%, n=4).

Fig. 2

Tentative Ca²⁺-activated transient K+ current is facilitated by sGnRH. (A)Voltage clamp recordings of tentative Ca2+-activated transient K+ currents. Bath application of 1 mM sGnRH increased the amplitude of tentative Ca2+-activated transient K+ current. Current traces before, during, and after sGnRH application are shown. (B) shows the time course of the peak amplitude of the tentative Ca2+-activated transient K+ current during sGnRH applications (bars).

Finally, we examined the effects of bath application of apamin, a specific SK-type Ca²⁺-activated K⁺ channel blocker. Fig. 3 shows the effect of bath application of apamin. In Ringer, TN-GnRH neurons showed regular beating discharge (3.02±0.86 Hz; n=4). Bath application of sGnRH (20 nM) transiently decreased (2.74±0.83 Hz; Fig. 3Bb) and subsequently increased (4.98±0.60 Hz; Fig. 3Bc) firing frequency of pacemaker activity. After washout of sGnRH, bath application of apamin (100 nM) facilitated the frequency of pacemaker activity (3.60±0.68 Hz; Fig. 3Be). However, further bath application of sGnRH (20 nM) failed to evoke transient decrease of firing frequency (3.70±0.74 Hz; Fig. 3Bf) but evoked subsequent increase of firing frequency (6.59±1.15; Fig. 3Bg). Fig. 3Bd and Bh show the time course of these changes before and after the application of sGnRH. We also compared quantitatively the normalized transient decrease of firing frequency in the presence and absence of apamin. The transient decrease that was induced by sGnRH (0.92±0.07; n=4) was nullified by apamin (1.02±0.03; n=4).

Fig. 3

Inhibition of SK-type Ca2+-activated K+ channel disrupts the transient decrease of pacemaker frequency after GnRH application. (A) Continuous recording of the pacemaker activity. In Ringer solution, bath application of sGnRH (20 nM) transiently decreased (Bb) and subsequently increased (Bc) the frequency of pacemaker activity (the overall time course of these changes is shown in Bd). By bath application of apamin (100 nM), which inhibits the SK type Ca2+-activated K+ channel, the frequency of pacemaker activity was increased (Be). Furthermore, bath application of apamin blocked transient decrease (Bf) but not subsequent increase in the frequency of pacemaker activities by sGnRH (Bg)(the overall time course of these changes is shown in Bh). Following washout by Ringer solution, the firing frequency decreased again.

From these data, it is suggested that sGnRH triggers the Ca²⁺ release from the intracellular store and activates apamin-sensitive Ca²⁺-activated K⁺ currents, which leads to the transient decrease of the frequency of pacemaker activity.

Involvement of Ca²⁺ currents in the pacemaker activity

Because the preliminary experiments that examined the effect of sGnRH upon INa(slow) and/or IK(V) did not show any noticeable modulation, we investigated the possible involvement of Ca²⁺ currents in the modulation of pacemaker activity. Fig. 4 shows the effects of Ca²⁺ channel blockers on the pacemaker frequency of TN-GnRH neurons. Bath application of various Ca²⁺ channel blockers slowed down the pacemaker frequency of TN-GnRH neurons. Fig. 4Aa shows the effects of the blocker of high-voltage-activated Ca²⁺ cur-rents, La³⁺ (5 μM, n=4), and Fig. 4Ab shows those of the blockers of low-voltage-activated Ca²⁺ current and store-operated Ca²⁺ current, Ni²⁺ (1 mM, n=4). The frequencies of pacemaker activities before and after these treatments were 3.7±0.8 and 2.6±0.5 Hz for La³⁺ (P<0.05, Student's two-tailed paired t test) and 4.1±0.7 and 1.7±0.2 Hz for Ni²⁺ (P<0.05, Student's two-tailed paired t test). Furthermore, when the bath application of the combination of La³⁺ (5 μM) and Ni²⁺ (1 mM) was examined, it completely blocked the generation of pacemaker activity (in 7/8 cells, Fig. 6Ba,c and d). The results are presented in Fig. 4B as the relative pacemaker frequencies after each treatment normalized to those before the treatment, which were defined as the ratio (firing frequency in Ringer solution containing Ca²⁺ channel blockers) / (firing frequency in Ringer solution). These results indicate that some kinds of Ca²⁺ currents are present in TNGnRH neurons and contribute to the generation of pacemaker activity.

Fig. 4

Blockade of Ca²⁺ currents decreases the pacemaker frequency. Aa, b: Effects of blockers of Ca2+ currents, La3+ and Ni2+, respectively. Both of these treatments decreased the frequency of pacemaker activities. Following washout, the frequency of pacemaker activities recovered. B: Comparison of the effects of Ca2+ channel blockers on the pacemaker activity. The numbers in parentheses represent the numbers of cells tested for each blocker. The ordinate indicates the relative pacemaker frequencies after each treatment normalized to those before the treatment; (Frequency of pacemaker activity in Ringer solution containing Ca2+ channel blocker) / (Frequency of pacemaker activity in Ringer solution). Bath application of Ca2+ channel blockers decreased the pacemaker frequency. It should be noted that simultaneous bath application of both La3+ and Ni2+ (the right column in B) had an additive effect of blocking the frequency of pacemaker activity (compare with the left and middle columns in B).

Next, we examined the effects of Ca²⁺ channel blockers on the modulation of pacemaker activity by sGnRH. Fig. 5 shows the effects of Ni²⁺ on the modulation of pacemaker activity by sGnRH. In Ringer solution, TN-GnRH neuron showed regular beating discharge (3.5±0.5 Hz; Fig. 5Ba). Bath application of Ni²⁺ (1 mM) decreased the firing frequency of pacemaker activity (1.8±0.2 Hz; Fig. 5Bb). Subsequent bath application of sGnRH (300 nM) transiently decreased (1.5±0.2 Hz; Fig. 5Bc) and then increased (2.4±0.8 Hz; Fig. 5Bd) the firing frequency of pacemaker activity in a qualitatively similar manner with sGnRH application in Ringer. However, the late-phase increase of pacemaker frequency was less pronounced compared with sGnRH application in Ringer (Fig. 7; compare the left and right columns). Similarly, the late-phase increase of pacemaker frequency was less pronounced under La³⁺ treatment compared with sGnRH application in Ringer (Fig. 7; compare the left and middle columns). During the simultaneous bath application of La³⁺ (5 μM) and Ni²⁺ (1 mM)(Fig. 6), the pacemaker activity was stopped (Fig. 6Bb, Bc), and further bath application of sGnRH (300 nM, in 3/4 cells, Fig. 6Bd) or depolarizing DC current injection (in 2/2 cells, data not shown) did not reinstate the pacemaker activities.

Fig. 5

The late-phase increase of pacemaker frequency by sGnRH is less pronounced under Ni²⁺ treatment. Bath application of Ni2+ (1 mM) decreased the frequency of pacemaker activity (Bb) but did not qualitatively affect either transient decrease (Bc) or subsequent increase of firing frequency (Bd) by sGnRH. However, the late-phase increase was less pronounced compared with sGnRH application in Ringer (see Fig. 7). Following washout by Ni2+-containing Ringer solution, the firing frequency decreased again (Be).

Fig. 6

Simultaneous application of La3+ and Ni2+ stops the pacemaker activity. Bath application of La3+ (5 mM) and Ni2+ (1 mM) blocked the generation of pacemaker activities (Bb, c). Furthermore, bath application of sGnRH did not induce pacemaker activities of TN-GnRH neurons (Bd). Pacemaker activities recovered after washout by Ringer solution (Bf).

Fig. 7

The late-phase increase of pacemaker frequency by sGnRH is less pronounced in the presence of Ca²⁺ channel blockers. Nomalized increase of firing frequency by sGnRH in the presence or absence of Ca²⁺ channel blockers was defined as and was calculated from the data such as those shown in Fig. 5. Normalized increase of pacemaker frequencies was less pronounced under Ni²⁺ or La³⁺ treatment. The numbers in parentheses represent the numbers of cells tested for each Ca²⁺ channel blocker.

Taken together, these data indicate that some kind(s) of Ca²⁺ channels present in the TN-GnRH neurons may be the target for the modulation by sGnRH, which somehow contributes to the late phase increase of pacemaker frequency.

Voltage-clamp recordings of voltage-dependent Ca²⁺ currents

To reveal the quantitative contribution of Ca²⁺ currents to the generation and modulation of pacemaker activity, it is important to determine the Ca²⁺ channel subtypes and kinetics. Depolarizing steps from a holding potential of −100 mV elicited a mixture of transient and sustained inward current components (Fig. 8Aa). They showed the characteristics of combined low voltage-activated (LVA) and high voltage-activated (HVA) Ca²⁺ currents. The peak inward currents (measured at peak current amplitudes during depolarizing pulses) began to appear more positive than −60 mV and reached a maximum near −20 mV. The sustained currents (measured at 200 ms after the onset of the test pulse) began to appear more positive than −20 mV and reached a maximum near 0 mV. The currents that were elicited by depolarizing steps from a holding potential of −60 mV are shown in Fig. 8Ab. The transient inward current almost disappeared, and sustained current(s) mainly remained. At this holding potential, the activation of both peak and sustained currents required depolarization more positive than −20 mV, and the maximum inward current was obtained near −10 mV. In the currents recorded from a different cell, however, the sustained current began to appear more positive than −40 mV and rapidly reached its maximum at −30 mV, when a holding potential was −60 mV (see below).

Fig. 8

Voltage clamp recording reveals the presence of one transient LVA and two types of sustained HVA Ca2+ current components. A: Currents evoked during 200 ms voltage steps from −100 to +50 mV (holding potential=−100 mV)(a) and from −60 to +50 mV (holding potential=−60 mV)(b). B: Current-voltage relations of the Ca2+ currents averaged from 40 (Vh=−100 mV) and 18 (Vh=−60 mV) TN-GnRH neurons. I/V curves were constructed by plotting the transient current amplitudes (filled squares) obtained by subtracting sustained current from peak current, and sustained current amplitudes (filled circles) measured at the end of 200 ms test pulses. The transient current probably correspond to the low-voltage activated (LVA) T type Ca2+ current. The I/V curve of the sustained currents evoked from a holding potential of −60 mV (open circles) had a shoulder at −20mV in addition to the activation threshold of −40mV (arrows). These may correspond to the activation thresholds of two HVA Ca2+ currents.

Fig. 8B shows the averaged current-voltage relationship of transient and sustained Ca²⁺ currents (n=40). We defined the current component that was obtained by subtraction of sustained current from peak current, as the ‘transient’ Ca²⁺ current. On the average, the transient and sustained currents measured at −10 mV from a holding potential of −100 mV was −1.41±0.26 and −0.87±0.12 nA, respectively (Fig. 8B ▪ and •). However, the averaged current-voltage relationship of sustained current elicited from a holding potential of -60 mV showed an activation threshold of −40 mV but also had a shoulder of current at −20 mV (Fig. 8B, arrows). This shoulder of current may reflect the difference of activation threshold and voltage dependence of sustained currents among different cells. The sustained current measured at −10 mV from a holding potential of −60 mV was −0.80±0.13 nA (n=18; Fig. 8B ○).

These results suggest that a transient inward current component that begins to activate near −60 mV and is largely inactivated when the holding potential is −60 mV represents the LVA Ca²⁺ current. On the other hand, HVA Ca²⁺ current component that are not inactivated at holding potentials of −60 mV consist of two types of sustained currents. The one current begins to activate from −20 mV and reaches its maximum amplitude near −10 mV. The other current begins to activate near −40 mV and rapidly reaches maximum at −30 mV.

Actions of GnRH on Ca²⁺ currents

Previous preliminary studies suggested that neither INa(slow) nor IK(v) is significantly modulated by sGnRH applications (our unpublished observations). Therefore, we suspected that the Ca²⁺ currents may be modulated by sGnRH. Fig. 9 shows an example of the effects of sGnRH on the Ca²⁺ currents. The modulation was examined in 10 cells held at −100 mV. In this case, when 1 μM sGnRH was added to the bath solution, the current amplitude and activation threshold of transient current were not changed (Fig. 9Ba □ and ▪). However, the amplitude of sustained current increased (−0.3 to −1.3 nA) in a wide voltage range. Furthermore, the activation threshold and maximum amplitude of sustained current were shifted to more negative potentials (Fig. 9Bb ○ and •). Responses of Ca²⁺ currents to GnRH were variable among the recorded cells. In another case, the amplitudes of both transient and sustained Ca²⁺ currents did not increase. However, the activation threshold and the peak amplitude of sustained Ca²⁺ current were shifted to more negative potentials (data not shown). On the average, the amplitude of the transient Ca²⁺ current measured at its maximum was decreased from −1.4 nA±0.3 to −1.1±0.4 nA (P>0.05, Student's two-tailed paired t-test; Fig. 10A □ and ▪), and those of the sustained Ca²⁺ current was increased from −1.6±0.4 to −1.9±0.4 (P<0.05, Student's two-tailed paired t test; Fig. 10B ○ and •), before and after bath application of 1 μM sGnRH, respectively (n=10). Moreover, the activation threshold and the voltage that evokes the maximum amplitude of the sustained Ca²⁺ currents were shifted to more negative potentials (Fig. 10B ○ and •), while those of transient Ca²⁺ current were not changed. These results indicate that the sustained HVA Ca²⁺ current is facilitated by sGnRH, i.e., the relative contribution of HVA Ca²⁺ current to the pacemaker activity may be increased by sGnRH. It is then suggested that this increased availability of HVA Ca^2+-current increases the frequency of pacemaker activity.

Fig. 9

sGnRH modulates the Ca2+ current components. A: Current traces from a holding potential of −100 mV before (a) and after the addition of 200 nM sGnRH to the bath solutions (b). B: The current-voltage relations of the transient current (a), and the sustained current measured at the end of 400 ms test pulse (b), which were constructed from the traces in A. Symbols are defined in the inset. In this cell, the current amplitude of sustained current was augmented by sGnRH, and the activation threshold of sustained current was shifted to more hyperpolarized potentials.

Fig. 10

Averaged I/V relations also show the modulation of Ca2+ current by sGnRH. I/V curves were constructed by plotting the averaged current amplitudes (n=10) evoked from holding potentials of −100 mV before and after the addition of 200 nM sGnRH to the bath solution. The current-voltage relations of transient current (A) and sustained current (B) are shown.

Involvement of Ca²⁺ release from intracellular store and apamin-sensitive K⁺ current in the transient decrease of firing frequency of TN-GnRH neurons

Intracellular application of ruthenium red and heparin inhibited the transient decrease but not late-phase increase of the pacemaker frequency, both of which were induced by bath application of sGnRH. This result suggests that [Ca²⁺]imobilization is involved in the transient decrease but not the late-phase increase of firing frequency of pacemaker activity. Then, the presence of tentative Ca²⁺-activated transient K⁺ current was suggested from the result of present voltage-clamp experiments. This current was TEA-insensitive but 4AP-sensitive, and could be only evoked when the Ca²⁺ is present in the extracellular bath solutions. However, the activation and steady-state inactivation curves of this current were shifted to more depolarized potentials compared with the 4AP-sensitive A-like currents of TN-GnRH neurons (Abe and Oka, 1999), and the two currents are considered to be different. The amplitude of the tentative Ca²⁺-activated transient K⁺ current was rapidly increased by bath application of sGnRH. Furthermore, bath application of apamin inhibited the transient decrease but not the late-phase increase of the pacemaker frequency by sGnRH. These results suggest that the apamin-sensitive Ca²⁺-activated transient K⁺ current was activated by Ca²⁺ released from the intracellular store, which had been induced by sGnRH. This may explain the early-phase transient decrease of pacemaker activity by sGnRH.

It has been reported that the Ca²⁺-activated K⁺ current exist in the embryonic GnRH neurons (Kusano et al., 1995) and GT1-7 cells (Spergel et al., 1996; Van Goor et al., 1999). The transient decrease of firing frequency by GnRH, which was induced by the activation of SK type Ca²⁺-activated K⁺ current, has been reported in the GT1-7 cells (Van Goor et al., 1999). It has been generally accepted that the activation of Gq/11-coupled GnRH receptor increases the intracellular Ca²⁺ concentration via phosphoinositide signaling pathway (Naor, 1990; Naor et al., 1998; Stojilkovic et al., 1994b), and the increased [Ca²⁺]i may open Ca²⁺-activated K⁺ current (Sah, 1996). All of these reports are in favor of the above-mentioned mechanisms for the transient decrease of pacemaker activity of TN-GnRH neurons by sGnRH.

Ca²⁺ current component is involved in the pacemaker activity

We have shown that some kind(s) of Ca²⁺ current contribute(s) to the modulation of the pacemaker activity of TNGnRH neurons. Furthermore, the result that simultaneous bath application of 5 μM La³⁺ and 1 mM Ni²⁺ completely inhibited the generation of pacemaker activity and its modulation by sGnRH, suggests that Ca²⁺ current components that were blocked by simultaneous application of La³⁺ and Ni²⁺ may be somehow involved in the generation of pacemaker potentials and may be the target of modulation by sGnRH.

The present results may appear to be partly inconsistent with those of the previous study that the Ca²⁺ currents are not essential for the generation of pacemaker potentials (Oka, 1995). One possible explanation may be the difference in the recording methods, intracellular recording (Oka, 1995) vs. whole-cell patch-clamp recording (present study). It may be possible that the intracellular dialysis of EGTA, which was introduced by the patch pipette solution, may prevent Ca²⁺ channels from Ca²⁺-dependent inactivation (Hille, 2001). Thus, a relatively large proportion of Ca²⁺ channel may be available for the generation of the pacemaker activity in the whole-cell patch-clamp recording. Second possibility is the presence of diffusion barrier. While the ventral meningeal membrane of the brain was not removed in the previous study, it is always completely removed in the patch-clamp experiments. The meningeal membrane may have served as a diffusion barrier for the drug delivery. Thirdly, the bath application of Ca²⁺ channel blockers may have simultaneously blocked other currents (for example, INa(slow)). Preliminary experiments suggest that Ni²⁺ may inhibit INa(slow) (data not shown). It has also been reported that Ni²⁺ and La³⁺ also blocks Na⁺/Ca²⁺ exchanger, store-operated Ca²⁺ current, and other voltage-gated ion channels (Nowycky, 1991; Taylor and Brond 1998). In the present study, we used supramaximum concentrations of these Ca²⁺ channel blockers to block the Ca²⁺ influx completely. Especially, the concentrations of Ni²⁺(1 mM) and La³⁺(5 μM) used here are sufficient to block store-operated Ca²⁺ currents (Skryma et al., 2000). Thus, it may be possible that these Ca²⁺ channel blockers blocked not only voltage-gated Ca²⁺ channels, but also store-operated Ca²⁺ channels, other channels, and/or exchangers. In fact, it has been reported in GT 1-7 cell-lines that store-operated Ca²⁺ current contributes to the modulation of firing activity by GnRH (Van Goor et al., 1999a). Further studies using more specific Ca²⁺ channel blockers are necessary to clarify the quantitative contributions of voltage-dependent Ca²⁺ current and store-operated Ca²⁺ current.

Both LVA and HVA Ca²⁺ currents are present in the TNGnRH neurons

Voltage-clamp experiments of whole-cell patch-clamp recording suggested that TN-GnRH neuron has at least one LVA Ca²⁺ current component and two HVA Ca²⁺ current components, and the relative prevalence of each current component seems to be different among different TN-GnRH neurons. LVA Ca²⁺ current was a transient current evoked from membrane potentials more positive than −60 mV and reached its maximum at about −20 mV (Figs. 8; Transient current). This current was almost inactivated at a holing potential of −60 mV. It most probably corresponds to the T-type Ca²⁺ current (Nowycky et al., 1985). In contrast, both of the two HVA Ca²⁺ currents inactivated slowly and could be observed at holding potentials of −100 mV and −60 mV. One of the HVA Ca²⁺ current activated at potentials more positive than −40 mV. The other HVA Ca²⁺ current activated at potentials more positive than −20 mV. Thus, the averaged I/V curve of the sustained current had a shoulder and reached its maximum around 0 mV (Fig. 8B). Preliminary experiments showed that the amplitude of HVA Ca²⁺ current was reduced (= was not completely blocked) by bath application of nifedipine (data not shown). Therefore, at least one of the HVA Ca²⁺ current may correspond to the L-type Ca²⁺ current.

It has been demonstrated that embryonic GnRH neurons (Kusano et al., 1995) and GT1 cells (Bosma, 1993; Hales et al., 1994; Javors et al., 1995; Costantin and Charles, 1999) express several types of plasma membrane Ca²⁺ channels, including transient and sustained voltage-dependent Ca²⁺ channels. Kusano et al. (1995) reported that embryonic GnRH neurons possess T- and L-type Ca²⁺ channels, and GT1 cells possess T-, N-, and L-type Ca²⁺ channels. On the other hand, Costantin and Charles (1999) and Van Goor et al. (1999b) reported that GT1 cells possess T- and L-type Ca²⁺ channels. The present study used for the first time the adult authentic GnRH neurons, and the results basically agreed well with these reports. However, pharmacological isolation and detailed analysis of kinetic properties of each Ca²⁺ current component have not yet been done in the present study mainly due to the poor space-clamp of the in vitro whole-brain preparations of TN-GnRH neurons. Therefore, the use of dissociated TN-GnRH neurons may be necessary to evaluate quantitatively the contribution of each Ca²⁺ current component to the control of pacemaker and secretory activities of TN-GnRH neurons.

Modulation of HVA type Ca²⁺ channels by GnRH

Bath application of sGnRH shifted the activation threshold of sustained Ca²⁺ current component to more hyperpolarized potentials. Moreover, the increase of sustained current amplitude was observed in some recordings, although complete recovery of Ca²⁺ current amplitude by washout could not be obtained due to the rundown of Ca²⁺ currents. It is generally accepted that GnRH receptors are coupled to G_q/11 type G-proteins, and Gq/11 type G-proteins are coupled to phosphoinositide-mediated signaling pathways (Stojilkovic et al., 1994a,b). It has also been reported that Ca²⁺ channels (especially, N- and L-type Ca²⁺ channel) are positively modulated by PKC (Bourinet et al., 1992; Dolphin, 1998; Meir et al., 1999; Stea et al., 1995; Swartz, 1993; Yang and Tsien, 1993; Zamponi et al., 1997; Zhu and Ikeda, 1994). Bosma and Hille (1992) reported that immortalized gonadotrope (aT3-1 cell), which also has GnRH receptors, express Ca²⁺ channels, and these Ca²⁺ channel were augmented by application of GnRH or phorbol ester.

From these observations and present results, it is highly possible that certain type(s) of Ca²⁺ currents that contribute to the generation of pacemaker activity are modulated by sGnRH and are involved in the late-phase increase of the pacemaker frequency.

Mechanisms of the generation and modulation of pacemaker activity and possible physiological functions

From the results of the present study, we present a model about the generation and modulation mechanisms of pacemaker activity in the TN-GnRH neuron (Fig. 11). In the intact brain, TN-GnRH neurons show regular beating pacemaker activity. The INa(slow), INa(fast), and IK(V) interact in the following manner to generate the pacemaker potentials. The INa(slow), which is persistently active in the subthreshold membrane potential range, always supplies the persistent depolarizing drive and gradually depolarizes the membrane potentials. When the membrane potential reaches the activation threshold for IK(V), outward current develops, and the net flux of current reverses to outward. Then, the membrane potential becomes hyperpolarized and deactivates the K⁺ current, and the next cycle begins. This is the subthreshold pacemaker activity, and when the membrane potential reaches the activation threshold for the INa(fast), the spiking pacemaker activities ensues. In addition, some kind(s) of Ca²⁺ currents and apamin-sensitive Ca²⁺-activated K⁺ cur-rent may be also involved in the generation of pacemaker potentials. Although INa(slow) and IK(v) are essential for the generation of the basic rhythm of pacemaker activities, they do not seem to be modulated by the process described below. When the GnRH peptide binds to the GnRH receptor located on the cell surface of TN-GnRH neurons, Gq/11 protein-coupled process starts and activates phospholipase C, which leads to the production of IP₃ and diacylglycerol. IP₃ stimulates the release of Ca²⁺ from the IP₃-sensitive intracellular Ca²⁺ store, and the increased [Ca²⁺]i then opens apamin-sensitive Ca²⁺-activated K⁺ channel. Thus, the pacemaker frequency of TN-GnRH neuron is transiently decreased. On the other hand, diacylglycerol activates protein kinase C, and protein kinase C then phosphorylates and modulates the Ca²⁺ current. Thus, the pacemaker frequency of TN-GnRH neurons is increased in the late phase.

Fig. 11

Model of the biphasic modulation of pacemaker activity of TN-GnRH neuron. The GnRH peptide, which binds to its receptor located at the cell membrane of TN-GnRH neurons, induces biphasic modulation of the pacemaker activities; (1) Facilitates Ca2+ release from the intracellular Ca2+ store. The increased [Ca2+]i activates apamin-sensitive Ca2+-activated K+ current, IK(Ca). This, in turn, decreases the frequency of pacemaker activities transiently. (2) Up-regulates the Ca2+ current(s), ICa, and increases the frequency of pacemaker activities in the late phase. INa(slow), IK(V), and INa(fast) are involved in the generation of basic pattern of pacemaker activities but are not directly modulated by GnRH.

What is the physiological significance of the pacemaker activity of TN-GnRH neurons and its modulation? Unfortunately, the functional link between the electrical activities of peptidergic neurons and the peptide release has not been firmly established thus far. A recent amperometric and RIA studies showed that the membrane depolarization triggers the secretion of GnRH from the TN-GnRH neuron in the teleost brain-pituitary slice preparation (Ishizaki and Oka, 2001 and in preparation). In the intact brain, the pacemaker activity of TN-GnRH neurons is characterized by low frequency (<10 Hz) regular beating discharge. The depolarization that is produced by single action potential of pacemaker activity may not be strong enough to induce GnRH release from the dense cored vesicles. However, it has been reported that such low frequency firing activity enables vasopressin release from rat neurohypophysis (Bondy et al., 1987), substance P or thyrotropin releasing hormone release from rat ventral spinal cord (Iverfeldt et al., 1989), and GnRH release from preganglionic C-neurons of the bullfrog (Peng and Horn, 1991). Furthermore, increase of firing frequency potentiated peptide release from these cells. Thus, it may be possible to think that the frequency of beating pacemaker activity of TN-GnRH neurons may affect the release of GnRH.

What then is the physiological significance of the modulation by sGnRH of pacemaker activity of TN-GnRH neurons? TN-GnRH neurons of the dwarf gourami make tight cell clusters with no intervening glial cells (Oka and Ichikawa, 1991; Oka and Matsushima, 1993; Oka, 1997), and the possibility of active exocytotic release from the cell body and its vicinity has also been suggested (Oka and Ichikawa, 1991). Other studies have shown that GnRH receptors are widely distributed throughout the brain (Jennes et al., 1997; Stojilkovic et al., 1994b). In addition, considerable overlap of the brain areas that contain GnRH-producing cells and GnRH receptor mRNA-expressing cells has been reported (Jennes et al., 1996). Also, cultured hypothalamic GnRH neurons have GnRH receptors (Krsmanovic et al., 1999). Furthermore, it has been reported using an in vivo analysis of multiunit activities in ovariectomized rats that GnRH injected in and around the median eminence is able to cause a population of GnRH neurons to fire synchronously (Hiruma and Kimura, 1995). From these observations and the present results, it is suggested that GnRH released from GnRH neurons facilitates the activities of their own (autocrine) and/or neighboring GnRH neurons (paracrine) and may cause synchronized positive feedback facilitation of multiple GnRH neurons. It is well known that in oxytocin neurons the release of oxytocin from single neuron into the brain environment stimulates its own activity and thus further release (Freund-Mercier and Richard, 1984; Moos et al., 1984). A similar effect has been reported for insulin-stimulated insulin release in pancreatic β cells (Aspinwall et al., 1999). Therefore, this mechanism is probably common to all neurosecretory neurons or secretory cells, whose synchronized facilitation of firing leads to facilitated release.

Comparison with the autoregulation in the other putative GnRH neurons

Possible autoregulation mechanisms of TN-GnRH cells suggested in the present study are discussed here in relation to references that described studies on putative GnRH neurons. The concept of an ultrashort feedback mechanism in the control of neurosecretion was first suggested by Hyyppa et al. (1971) in studies on the control of FSH secretion. Similarly, it was revealed from in vivo (Valenca et al., 1987) and in vitro studies (DePaolo et al., 1987) that GnRH release from hypothalamic GnRH system was negatively autoregulated. It was postulated that negative feedback could be mediated via axo-dendritic/axo-somatic synapses on adjacent GnRH or other types of neurons (DePaolo et al., 1987), and histological evidence for such connections has been reported (Leranth et al., 1985; Witkin and Silverman, 1985).

However, the results using recording of multiunit activity (MUA) in the hypothalamus, which are considered to reflect the secretory activity of GnRH, have been very complicated. Hiruma and Kimura (1995) reported that intravenous injection of GnRH or microinjection of GnRH into the median eminence immediately evoked a MUA volley of the rat. However microinjection of GnRH into the medial preoptic area did not cause these effects. Moreover, intravenous or intracerebroventricular injections of GnRH did not affect MUA volleys of the rhesus monkeys (Kesner et al., 1986; Ordog et al., 1997). Unfortunately, the neuronal elements responsible for the MUA volley, i.e., whether the MUA volley represent the activity of GnRH neurons themselves or other neuronal elements, have not been determined. Therefore, it is difficult to directly compare the results of the present paper with the results of MUA studies.

Exposure of perifused GT1-7 cells to a GnRH agonist (analog) caused a transient elevation of GnRH release and subsequent suppression of the basal pulsatile secretion. During further continuous exposure to the agonist, the cells recovered from inhibition and exhibited infrequent but increasingly prominent peaks, with a net increase in GnRH release (Krsmanovic et al., 1993). Recently, similar autoregulation of GnRH secretion has been reported in hypothalamic culture (Krsmanovic et al., 1999). However, the time courses of such autoregulation are much slower than those of TN-GnRH neurons. The autoregulation of GnRH release in GT1-7 cell and hypothalamic culture takes dozens of minutes to several hours to occur. However, the biphasic auto-regulation of pacemaker activity of TN-GnRH neurons took only a few minutes. The former has been suggested to result from the internalization of receptor molecules and downregulation of the expression of GnRH receptor gene that was induced by GnRH (Park, 1998; Stojilkovic et al., 1994a). On the other hand, the biphasic autoregulation of pacemaker activity of TN-GnRH neurons is considered to be due to the modulations of ionic channels induced by the activation of cell signaling pathways downstream of GnRH receptor activation. Thus, the two phenomena are considered to be based on quite different mechanisms.

Recently, Van Goor et al. (1999a, b) reported that the membrane excitability of GT1 cells was modulated by GnRH peptide. In this modulation of membrane excitability, they suggested that the activation of GnRH receptor induces [Ca²⁺]i mobilization, and this [Ca²⁺]i mobilization activates SK-type Ca²⁺-activated K⁺ channel and activates store-operated Ca²⁺-channel, which transiently hyperpolarizes and then persistently depolarizes membrane potentials, respectively. They further suggested that this sustained membrane depolarization is explained by complex interplay of Na⁺ channel, K⁺ channel, store-operated Ca²⁺ channel, and L-type Ca²⁺ channel (Van Goor et al., 2000). Furthermore, LeBeau et al. (2000) showed the contribution of cAMP-operated inward current to this modulation in GT1 cells. The mechanism of the transient decrease of firing frequency in TN-GnRH neurons (the present study) basically agrees well with these reports. However, detailed mechanism of the sustained increase of firing frequency seems to differ from each other. To further understand this mechanism, more detailed kinetic and pharmacological investigation of the control of pacemaker activities of TN-GnRH neurons are under way.