κ-Opioid receptor-mediated enhancement of the hyperpolarization-activated current (I_h) through mobilization of intracellular calcium in rat nucleus raphe magnus

Abstract

The hyperpolarization-activated current (I_h) is important in the control of resting membrane potential, in the regulation of network firing pattern and in the modulation of presynaptic transmitter release in central neurons. Recent studies on native and cloned I_h channels have demonstrated that the I_h channel is commonly modulated by cAMP through a positive shift in its voltage dependence without a change in its maximum current. The present study demonstrates that activation of κ-opioid receptors enhances I_h by increasing its maximum current in brainstem neurons in the nucleus raphe magnus. Agents that interfere with the release of intracellular calcium from calcium stores altered the maximum I_h and significantly attenuated the κ-receptor-mediated enhancement of I_h. These results suggest that κ-opioid receptors enhance the maximum I_h by mobilizing intracellular calcium from calcium stores. This provides a physiological function for κ-receptor-stimulated calcium release and may suggest another I_h-regulating mechanism by intracellular calcium in central neurons.

The hyperpolarization-activated current (I_h) is a non-selective, inward cation current that is slowly activated at membrane potentials more negative than the resting membrane potential. Although it was first described in both cardiac cells and neurons more than 20 years ago (Brown et al. 1979; DiFrancesco, 1981; Halliwell & Adams, 1982), only in recent years has there been rapid progress in our understanding of its modulation and regulatory functions.

It has now been clearly shown that I_h is increased by intracellular cAMP through a positive shift in its voltage dependence without changing its maximum in various types of central and peripheral neurons (Bobker & Williams, 1989; McCormick & Pape, 1990; Pape, 1992; Banks et al. 1993; Erickson et al. 1993; Ingram & Williams, 1994; Pedarzani & Storm, 1995; Luthi & McCormick, 1999) and in cloned I_h channels (Gauss et al. 1998; Ludwig et al. 1998; Santoro et al. 1998). A variety of functions of I_h have been described in central neurons (McCormick & Pape, 1990; Kaupp & Seifert, 2001). A membrane depolarizing current activated near the resting membrane potential, I_h is an important contributor to the control of resting membrane potential and neuronal excitability (Rainnie et al. 1994; Saitow & Konishi, 2000; Lupica et al. 2001; Williams et al. 2002). In thalamocortical relay neurons, I_h plays a key role in the burst firing pattern and controls the synchronized rhythmic activity in the thalamocortical system during sleep (Pape & McCormick, 1989; Steriade et al. 1993; von Krosigk et al. 1993; Luthi & McCormick, 1999). Recently, activation of presynaptic I_h has been shown to enhance synaptic transmission through a cAMP-dependent mechanism (Beaumont & Zucker, 2000; Saitow & Konishi, 2000; Southan et al. 2000; Lupica et al. 2001) and mediate hippocampal mossy fibre long-term potentiation (Mellor et al. 2002).

Opioid peptides are among various neurotransmitters that modulate I_h in neurons. It is interesting that many neurotransmitters, including opioids, either enhance (Colino & Halliwell, 1993) or inhibit the maximum amplitude of I_h (Jiang et al. 1993; Cathala & Paupardin-Tritsch, 1997; Parkis & Berger, 1997). However, the mechanism underlying the modulation of the maximum I_h by these neurotransmitters is unknown. The present study investigated the modulation of I_h by κ-opioid receptors and its underlying intracellular mechanisms in brainstem neurons in the nucleus raphe magnus (NRM), which are critically involved in mediating opioid analgesia (Fields et al. 1991; Pan et al. 1997).

METHODS

All procedures involving the use of animals conformed to the guidelines set by the University of Texas-MD Anderson Cancer Center Animal Care and Use Committee.

Male, neonatal (9–14 days old) Wistar rats were used for brain slice preparations and in vitro recordings. Animals were anaesthetized with halothane and then killed by decapitation. The brain was removed and cut in a vibratome in cold physiological saline at 4 °C to obtain brainstem slices (200 μm thick) containing the nucleus raphe magnus (NRM). A single slice was transferred and submerged in a shallow recording chamber with preheated (35 °C) physiological saline flowing through. The physiological saline contained (mm): NaCl 126, KCl 2.5, NaH₂PO₄ 1.2, MgCl₂ 1.2, CaCl₂ 2.4, glucose 11 and NaHCO₃ 25; saturated with 95 % O₂−5 % CO₂. Calcium-free external solution was prepared by replacing CaCl₂ with an equal concentration of MgCl₂. Neonatal rats were used for better visualization of neurons in brain slices with an infrared Nomarski microscope. It has been demonstrated that the physiological and pharmacological properties of neurons from these young rats are indistinguishable from those of adult rats (Pan et al. 1997).

Visualized whole-cell patch-clamp recordings were made from identified neurons with a glass pipette (resistance, 3–5 MΩ) filled with a solution containing (mm): potassium gluconate 126, NaCl 10, MgCl₂ 1, EGTA 11, Hepes 10, ATP 2 and GTP 0.25; pH was adjusted to 7.3 with KOH. In some experiments, BAPTA (30 mm) or heparin (500 μm) was included in the internal solution with a reduced concentration of potassium gluconate to reach a final osmolarity of 280–290 mosmol l⁻¹. An AxoPatch 1-D amplifier and AxoGraph software (Axon Instruments, Inc.) were used for data acquisition and on-/off-line data analyses. The filter and sampling frequencies for acquisition were 2 kHz. A seal resistance of 2 GΩ or above and an access resistance of 15 MΩ or less were considered acceptable. Series resistance was optimally compensated. All drugs were applied through the perfusing solution.

In all experiments, NRM cells were first classified into either a primary cell type or secondary cell type according to the criteria described in our previous study (Pan et al. 1990). Only primary cells that were hyperpolarized by κ-receptor agonists (Pan et al. 1997) were included in the results of this study. The amplitude of I_h was determined by the difference in the current values at the beginning and the end of a 1 s hyperpolarizing voltage step. Numerical data were statistically analysed using Student's paired t test or ANOVA and are presented as means ±s.e.m. All drugs used were purchased from Research Biochemicals International and Sigma-Aldrich.

RESULTS

Properties of I_h

Visualized whole-cell patch clamp recordings were made from neurons of the nucleus raphe magnus (NRM) in brainstem slices in vitro. All results were obtained from a class of cells (termed primary cells) that express κ-opioid receptors (Pan et al. 1997). In the majority of primary cells under voltage clamp, voltage steps from a holding potential of −60 mV to more negative potentials activated a slowly activating, inward sag current (I_h) (Fig. 1A). In normal extracellular potassium concentration ([K⁺]_o) of 2.5 mm, the average amplitude of the I_h at −120 mV was 18.6 ± 1.3 pA (n = 38). That is in strong contrast to the I_h amplitude in the other cell type (termed secondary cells), which displayed a much larger I_h at the same membrane potential (58 ± 8 pA, n = 18, Fig. 1B).

Figure 1. Properties of I_h in primary cells of the nucleus raphe magnus (NRM).

A, a current trace showing the I_h during a 1 s voltage step (indicated above) under voltage clamp. The holding potential was −60 mV. B, a similar current trace but from a secondary cell for comparison. Note the larger I_h amplitude than the primary cell shown in A. C, a graph of I_h amplitudes against voltage across the range for activation (n = 38). D, the activation curve of I_h tails (n = 13). Tail current amplitudes were normalized to the current amplitude at −130 mV as the maximum and fitted with the Boltzmann equation. The estimated half-maximum voltage (V_1/2) was −86.8 ± 0.4 mV. E, a graph showing the average I_h amplitudes in control (○) and in the presence of BaCl₂ (200 μm, •) in an extracellular potassium concentration of 6.5 mm ([K⁺]_o) (n = 6). F, a graph showing the average I_h amplitudes in control (○) and in the presence of CsCl (2 mm, •) in 6.5 mm[K⁺]_o (n = 8).

The activation of I_h started at potentials more negative than the holding potential and the I_h amplitude increased at more negative potentials (Fig. 1C). The time course of activation of I_h was typically fitted by a single exponential and the activation rate accelerated as the membrane potential became more negative with a time constant of 248 ± 25 ms at −120 mV (n = 8). The voltage dependence of I_h was further characterized by the activation curve of I_h tails fitted with the Boltzmann equation (Fig. 1D). The maximum tail current was reached at around −120 mV and the estimated half-maximal voltage (V_1/2) was −86.8 ± 0.4 mV (n = 13, Fig. 1D). When the [K⁺]_o was increased to 6.5 mm, the I_h amplitude was significantly increased (40.2 ± 8.8 pA at −120 mV, n = 11), suggesting that the I_h channel is permeable to potassium. The presence of BaCl₂ (200 μm) did not affect I_h (at −120 mV: control, 41.5 ± 14.1 pA; BaCl₂, 44.3 ± 10.8 pA; P > 0.05, n = 6, Fig. 1E), nor did it change the amplitude of the I_h tail at −120 mV (control, 27.4 ± 6.5 pA; BaCl₂, 26.7 ± 6.5 pA; P > 0.05, n = 4). However, the presence of CsCl (2 mm) completely blocked I_h in all cells tested (n = 8, Fig. 1F). These properties of I_h, including slow activation, voltage dependence, insensitivity to barium and blockade by caesium, are consistent with the previously reported properties of I_h in various types of neurons from other brain sites (McCormick & Pape, 1990; Uchimura et al. 1990; Khakh & Henderson, 1998; Zhu et al. 1999).

It has been reported that ZD7288 (1–100 μm) selectively blocks I_h in a variety of neurons (Harris & Constanti, 1995; Khakh & Henderson, 1998; Takigawa et al. 1998; Larkman & Kelly, 2001). In the present study, ZD7288 at 25–50 μm completely blocked I_h at all potentials (at −120 mV: control, 19.9 ± 3.9 pA; ZD7288, 1.6 ± 0.6 pA; P < 0.01, n = 16, Fig. 2A and B). Consistent with previous reports, lower concentrations of ZD7288 (< 20 μm) were also effective at blocking the I_h, though the onset of action was significantly slower. These results further confirm the identity of the I_h described here in these brainstem neurons.

Figure 2. ZD7288 blocks I_h.

A, current traces during a voltage step (−60 to −120 mV) in the absence (upper panel) and presence (lower panel) of ZD7288 (25 μm). B, I_h amplitudes at −120 mV from a group of primary cells (n = 16) in control and in the presence of ZD7288. **P < 0.01 vs. control.

κ-Receptor agonist enhances the maximum I_h

We have previously reported that these cells are hyperpolarized by κ-receptor agonists through activation of inwardly rectifying potassium channels, or GIRK channels (Pan et al. 1997). In addition to the GIRK channel-mediated hyperpolarization, it was found in the present study that the κ-receptor agonist U69593 also enhanced the I_h, particularly its maximum amplitude. In cells that displayed measurable I_h (> 5 pA, n = 31), U69593 (300 nm) induced an outward current of 26.3 ± 3.9 pA from the holding potential (−60 mV) with an average reversal potential of −93.8 ± 1.5 mV, suggesting GIRK channel activation. In the same cells (n = 31), U69593 concomitantly increased the I_h amplitude (Fig. 3A and B). The increase in the maximum I_h estimated at −120 mV varied from cell to cell and the average increase was 223 ± 16 % (16.6 ± 1.0 pA in control and 35.2 ± 2.2 pA in U69593, P < 0.01, n = 31). The effect of U69593 was reversible by wash out or by addition of naloxone (1 μm, n = 3, Fig. 3A). The selective κ-receptor antagonist norBNI (100 nm) completely blocked both the U69593-induced hyperpolarization, as shown in our previous study (Pan et al. 1997), and the U69593-induced I_h enhancement in this study (17.6 ± 4.9 pA in norBNI and 17.2 ± 5.6 pA in norBNI + U69593 at −120 mV, n = 5), suggesting a specific κ-receptor-mediated effect. In addition to the amplitude increase, U69593 accelerated the I_h activation rate with an activation time constant of 133 ± 21 ms, a significant effect when compared to 248 ± 25 ms in control (P < 0.01, n = 8). Similar I_h enhancement by U69593 was also observed in higher [K⁺]_o (6.5 mm, Fig. 3B).

Figure 3. κ-opioid receptor agonist U69593 enhances the I_h.

A, current traces during a voltage step to −120 mV in control, in the presence of U69593 (300 nm) and in the presence of U69593 + naloxone (1 μm). Note the increase in I_h in the presence of U69593. B, a graph showing the average I_h amplitudes in control (open symbols) and in U69593 (300 nm, filled symbols) in 2.5 mm (circles, n = 31) and 6.5 mm (squares, n = 5) [K⁺]_o. C, current traces similar to those in A and aligned at the holding potential from another cell showing a U69593-dependent increase in both the I_h tail and GIRK conductance. D, a graph showing the I_h tail amplitudes in control (○) and in the presence of U69593 (•), n = 7. E, activation curves of I_h tails in control (○) and in the presence of U69593 (•), n = 7. The estimated V_1/2 was −85.7 mV and −86.9 mV in control and in the presence of U69593, respectively.

The κ-receptor agonist also enhanced the maximum amplitude of the I_h tails across the activation range with an average of 155 ± 14 % increase at −120 mV (control, 26.1 ± 7.4 pA; U69593, 38.5 ± 9.2 pA; P < 0.05, n = 7, Fig. 3C and D). The I_h augmentation by U69593 was voltage dependent being more effective at negative membrane potentials (Fig. 3B and D). Analysis of the I_h activation curves in control and in the presence of U69593 revealed that the κ-receptor agonist produced only a small change in V_1/2 that was not statistically significant (control, −85.7 ± 0.4 mV; U69593, −86.9 ± 0.6 mV; P > 0.05, n = 7, Fig. 3E). The small change in V_1/2 was probably due to an underestimate (at −120 mV) of the maximum I_h that had increased in the presence of U69593. These results suggest that in addition to membrane hyperpolarization through GIRK channel activation, κ-receptor activation enhances the maximum amplitude of the I_h without a significant change in its voltage dependence in these cells.

I_h enhancement is independent of GIRK channel activation

Since the κ-receptor agonist produced I_h enhancement that was accompanied by GIRK channel activation, it is interesting and necessary to investigate whether the I_h enhancement is a separate effect of the κ-receptor, independent of the GIRK channel activation, or whether its effect on I_h is secondary to GIRK activation. Therefore, the following experiments were performed after pharmacological blockade of either GIRK channels or I_h channels.

Barium has been shown to block GIRK channels in central neurons (Watts et al. 1996; Svoboda & Lupica, 1998; Cathala & Paupardin-Tritsch, 1999). In the presence of 200 μm BaCl₂, the outward current induced by U69593 (300 nm) was blocked in all cells tested (control, 58.9 ± 10.1 pA; BaCl₂, 2.3 ± 1.1 pA; P < 0.01, n = 10). In the majority of these cells (8/10) after blockade of GIRK channels by BaCl₂, U69593 still enhanced the I_h amplitude (at −120 mV: BaCl₂, 11.3 ± 3.0 pA; BaCl₂+ U69593, 26.3 ± 4.0 pA; P < 0.01, n = 8). U69593 did not affect the I_h in the remaining two cells. The amplitude of I_h tails was also enhanced by U69593 over its entire activation range (196 ± 38 % increase at −120 mV, P < 0.05, n = 9, Fig. 4A) after the U69593-induced GIRK current had been blocked by BaCl₂ in the same cells (before BaCl₂, 55.8 ± 7.3 pA; after BaCl₂, 2.2 ± 0.9 pA; P < 0.01, n = 9, Fig. 4B). These results indicate that κ-receptor activation produces a comparable I_h enhancement after blockade of GIRK channels.

Figure 4. The κ-receptor-mediated I_h enhancement is independent of GIRK channel activation.

A, a graph showing the percentage increase in I_h tails induced by U69593 (300 nm) in the presence of BaCl₂ (200 μm, n = 9). The inset shows current traces of I_h tails at −120 mV in the presence of either BaCl₂ or BaCl₂+ U69593. B, U69593-induced outward currents in control and in the presence of BaCl₂ from the same cells as in A (n = 9). Holding potential was −60 mV. **P < 0.01 vs. control. C, current traces during a voltage step to −120 mV in the presence of either ZD7288 (1 μm) or ZD7288 + U69593 in a brain slice pre-treated with ZD7288 (1 μm). Note the absence of I_h in both conditions. D, U69593 (300 nm)-induced outward currents in normal cells (control, n = 39) and in ZD7288-treated cells (n = 14). E, current traces during a voltage step to −120 mV before (control) and during application of met-enkephalin (10 μm) in a secondary cell. Note the increase in GIRK conductance but not in the I_h.

Next, the κ-receptor-mediated effect was further investigated after blockade of the I_h channel by ZD7288. It was found that although ZD7288 at concentrations of 25–50 μm completely blocked the I_h, it also significantly reduced the U69593-induced outward current by various amounts, making the results inconclusive at those ZD7288 concentrations. Experiments were then conducted in brain slices pre-incubated with 1 μm ZD7288. Subsequent recordings were made with 1 μm ZD7288 in the perfusing solution. In these ZD7288-treated slices, no detectable I_h was observed in any of the cells tested (n = 14). In the presence of U69593 (300 nm), there was still no detectable I_h in these cells (n = 14, Fig. 4C) although U69593 produced a comparable GIRK current in the same cells (ZD7288-treated slices, 22.6 ± 3.5 pA, n = 14; all untreated slices, 24.5 ± 3.2 pA, n = 39; P > 0.05, Fig. 4D).

We have recently shown that nociceptin/orphanin FQ (N/OFQ) hyperpolarizes these NRM primary cells by activating GIRK channels (Pan et al. 2000). In the present study, while activating GIRK channels, N/OFQ (100 nm) did not significantly change the I_h amplitude in the same cells (at −120 mV: control, 15.0 ± 1.3 pA; N/OFQ, 19.1 ± 3.4 pA; P > 0.05, n = 15). In addition, as we reported previously (Pan et al. 1990, 1997), the other NRM cell type, secondary cells, are hyperpolarized by μ-receptor agonists through activation of GIRK channels. In the current study, met-enkephalin (ME, 10 μm) acting on μ-receptors (Pan et al. 1990) caused an outward current of 25 ± 3 pA in secondary cells (n = 18), but it did not alter the basal I_h amplitude in these cells (at −120 mV: control, 58 ± 8 pA; ME, 57 ± 9 pA; P > 0.05, n = 18, Fig. 4E). These findings demonstrate that membrane hyperpolarization by GIRK channel activation does not necessarily lead to an increase in I_h in these neurons under our experimental conditions.

Overall, the above results obtained using barium, ZD7288, N/OFQ and ME suggest that κ-receptor-mediated I_h enhancement is independent of κ-receptor-mediated activation of GIRK channels and may be mediated through a separate signal transduction mechanism coupled to κ-receptors.

Actions of cAMP analogues

Since it has often been reported that cAMP modulates I_h channels, an aim of the present study was to investigate whether cAMP was also involved in modulation of the I_h in these brainstem neurons. The cAMP analogue 8-bromoadenosine 3′,5′-cyclic monophosphate (8-Br-cAMP) (1 mm) caused a small increase in the I_h at sub-maximal potentials, but did not change the maximum current amplitude (at −120 mV in 6.5 mm[K⁺]_o: control, 33.3 ± 6.1 pA; 8-Br-cAMP, 33.2 ± 5.9 pA; P > 0.05, n = 5, Fig. 5A). The V_1/2 value estimated from activation curves of the I_h tails was −94.6 mV in control and −89.7 mV in the presence of 8-Br-cAMP at this higher level of [K⁺]_o (P < 0.05, n = 5, Fig. 5B). Thus, the cAMP analogue increased the I_h by a positive shift of 4.9 mV in its voltage dependence. Similar effects were observed with the adenylyl cyclase activator forskolin (10 μm, n = 5, data not shown). These results are in agreement with those previously reported in native I_h channels in other neurons (McCormick & Pape, 1990; Tokimasa & Akasu, 1990; Ingram & Williams, 1994) and similar to those in the cloned HCN1 channel, which shows a small shift in V_1/2 by cAMP (2–13 mV)(Kaupp & Seifert, 2001). Nevertheless, with no effect on the maximum I_h amplitude, the action of cAMP is apparently distinct from that of the κ-receptor agonist, which produced a large enhancement of the maximum I_h amplitude.

Figure 5. Cyclic AMP analogue increases the I_h through a positive shift in its voltage dependence.

A, a graph showing the average I_h amplitudes in control (○) and in the presence of 8-Br-cAMP (1 mm, •) in 6.5 mm[K⁺]_o, n = 5. B, activation curves of I_h tails before (control, ○) and during application of 8-Br-cAMP (1 mm, •) in 6.5 mm[K⁺]_o. The V_1/2 was −94.6 mV in control and −89.7 mV in the presence of 8-Br-cAMP.

The κ-receptor-mediated enhancement of maximum I_h requires release of intracellular calcium

Since it appeared that cAMP was not involved in the mechanism underlying κ-receptor-mediated I_h enhancement, the focus then moved to the role of intracellular calcium (Ca_i²⁺) in the I_h modulation by κ-receptors. Caffeine induces calcium release into cytoplasm from Ca_i²⁺ stores (Henzi & MacDermott, 1992). In this study, caffeine (10 mm) caused a significant increase in the I_h amplitude including its maximum estimated at −120 mV in the majority of cells tested (18/24, 75 %, Fig. 6A). Compared to controls, the average increase in I_h amplitude by caffeine at −120 mV was 257 ± 35 % (control, 11.8 ± 1.5 pA; caffeine, 23.6 ± 1.8 pA; P < 0.01, n = 18, Fig. 6B and C). The effect of caffeine was similar to that of U69593 (223 % increase by U69593, Fig. 3B). If the κ-receptor-mediated I_h augmentation also involves the release of Ca²⁺, caffeine should occlude, at least partially, the κ-receptor-mediated effect on the I_h. This hypothesis was then tested by examining the effect of U69593 in the presence of caffeine. Under this condition, the average I_h amplitude at −120 mV was 27.2 ± 4.0 pA in the presence of caffeine and 35.2 ± 8 pA in the presence of caffeine + U69593 (n = 6). Thus, the I_h-enhancing effect of U69593 was significantly reduced by caffeine (130 ± 25 % increase at −120 mV, P < 0.05, n = 6, compared to 223 ± 16 % increase by U69593 in control, Fig. 6C).

Figure 6. Caffeine and thapsigargin reduce the I_h-enhancing effect of κ-receptors.

A, current traces during a voltage step to −120 mV in control and in the presence of caffeine (10 mm). Note the similar I_h enhancement by caffeine and by U69593 in Fig. 3A. B, a graph showing the average I_h amplitudes in control (○) and in the presence of caffeine (•), n = 18. C, percentage increases of the I_h at −120 mV by caffeine (10 mm) compared to controls (P < 0.01, n = 18) and by U69593 (300 nm) in control (P < 0.01, n = 31), and in the presence of caffeine (*P < 0.05, n = 6, compared to the effect of U69593 in control). D, percentage changes of the I_h at −120 mV by thapsigargin compared to controls (n = 7), and by U69593 in control (as in C) and in the presence of thapsigargin (3 μm, *P < 0.05, compared to the effect of U69593 in control, n = 6). Con, control; Thaps, thapsigargin.

Because caffeine increased the basal I_h amplitude before addition of U65593, it was possible that the reduction of the effect of U69593 in caffeine was simply due to I_h saturation after caffeine plus U69593. To determine whether this was the case, thapsigargin was used to further confirm the role of Ca_i²⁺ in the κ-receptor-mediated effect. Thapsigargin has been shown to deplete and prevent refilling of inositol 1,4,5-trisphosphate (IP₃)-sensitive calcium stores (Jin et al. 1994; Hauser et al. 1996). In contrast to the effect of caffeine, thapsigargin (3 μm) decreased the I_h amplitude (at −120 mV: control, 17.0 ± 2.6 pA; thapsigargin, 11.1 ± 2.7 pA, n = 7, Fig. 6D). However, similar to the results with caffeine, the effect of U69593 was also significantly attenuated in the presence of thapsigargin (control, 223 ± 6 % increase, n = 31; thapsigargin, 141 ± 23 % increase, n = 6; P < 0.05, Fig. 6D). These results suggest the involvement of Ca_i²⁺ release in the mechanism underlying κ-receptor-mediated I_h enhancement.

If the κ-receptor-mediated enhancement of the I_h in these cells requires the release of Ca_i²⁺, chelation of Ca_i²⁺ released from calcium stores would be expected to block or reduce the κ-receptor-mediated effect. Therefore, the fast-acting Ca_i²⁺ chelator BAPTA was used in the following experiments. BAPTA (30 mm) was included in the recording pipette and the cell was held under voltage clamp for at least 5 min before data collection. In this condition, I_h tended to increase (31.7 ± 7.6 pA at −120 mV, n = 13, Fig. 7A). Although the reason for this is not clear, it may indirectly indicate the involvement of Ca_i²⁺ in the modulation of I_h. With the BAPTA-filled pipette, U69593 induced only a 124 ± 13 % increase in the I_h amplitude at −120 mV (33.7 ± 8.5 pA, n = 13, Fig. 7B and C), a significant reduction when compared to the effect of U69593 in control conditions (226 ± 16 %, P < 0.01, n = 31, Fig. 7C). While the I_h-enhancing effect of U69593 was significantly reduced by BAPTA, the κ-receptor agonist produced an outward current with an average amplitude of 25.3 ± 2.8 pA (n = 13), comparable to the U69593-induced current in normal conditions (26.3 ± 3.9 pA, P > 0.05, n = 30, Fig. 7D). This suggests that the GIRK channels and their coupling to κ-receptors are not affected by intracellular BAPTA.

Figure 7. BAPTA blocks the I_h-enhancing effect of κ-receptors.

The data were obtained from experiments with BAPTA (30 mm) included in recording pipettes. A, current traces during a voltage step to −120 mV in the absence and presence of U69593 (300 nm). B, a graph showing the I_h amplitudes in the absence (○) and presence of U69593 (•), n = 13. C, percentage increases of the I_h at −120 mV induced by U69593 in normal conditions (as in Fig. 6C and D) and with BAPTA-filled recording pipettes (n = 13). **P < 0.01, compared to the effect of U69593 in normal conditions. D, U69593-induced outward currents in normal conditions (n = 30) and with BAPTA-filled recording pipettes in the same cells as in C (n = 13).

It has been shown that κ-opioid receptors couple to the IP₃ pathway (Ueda et al. 1995; Smart & Lambert, 1996). It is possible that the κ-receptor agonist enhances the I_h by inducing Ca²⁺ release from IP₃-sensitive stores through activation of the IP₃ pathway. To test that hypothesis, further experiments were carried out using the recording pipette containing heparin (500 μm), which inhibits Ca²⁺ release from IP₃-sensitive stores (Kiselyov et al. 1998; Kapur et al. 2001). I_h was unaffected by the presence of heparin (16.7 ± 2.3 pA at −120 mV, n = 17, Fig. 8A). However, the I_h-enhancing effect of U69593 was blocked by internal application of heparin (120 ± 13 % of control I_h at −120 mV, P > 0.05, n = 7, Fig. 8A and C). The amplitude of I_h at −120 mV was 17.6 ± 3.8 pA in control and 19.0 ± 2.8 pA in the presence of U69593 (300 nm, n = 7). The difference between the effects of U69593 with normal and heparin-containing pipettes was also statistically confirmed (P < 0.01, Fig. 8C). While the I_h augmentation by U6953 was blocked, the U69593-induced GIRK current was not affected by heparin in the same cells (35.7 ± 2.9 pA, n = 7, Fig. 8C). Thus, κ-receptor agonists appear to augment the I_h through Ca²⁺ release from IP₃-sensitive stores.

Figure 8. Heparin and Ca²⁺-free external solution block the I_h-enhancing effect of κ-receptors.

A, current traces during a voltage step (−60 to −120 mV) in control and in the presence of U69593 (300 nm), with heparin (500 μm) included in the recording pipette. B, similar current traces to those in A, but obtained from another cell in Ca²⁺-free external solution. C, percentage changes induced by U69593 (300 nm) in the I_h amplitude at −120 mV (□) and in the GIRK currents (▪) under normal conditions (n = 31), with heparin-filled recording pipettes (n = 7), and in Ca²⁺-free external solution (n = 7). **P < 0.01, compared to the effect of U69593 in normal conditions. Con, control.

Finally, to further examine the dependence of Ca²⁺ release from intracellular stores and the effect of U69593 on extracellular Ca²⁺ (Ca_o²⁺) in these cells, the κ-receptor-mediated effect was tested in Ca²⁺-free external solution (replaced by MgCl₂). After at least 8 min in this Ca²⁺-free solution, the I_h amplitude at −120 mV was 20.4 ± 5.2 pA in control and 22.6 ± 6.2 pA in the presence of 300 nm U69593 (n = 7, Fig. 8B). Thus, U69593 was no longer effective at enhancing the maximum I_h amplitude in the absence of Ca_o²⁺ (117 ± 16 % of control at −120 mV, P > 0.05, n = 7, Fig. 8C). When removal of Ca_o²⁺ eliminated the effect of U69593 on the I_h amplitude (P < 0.01 by comparing its effects in normal and Ca²⁺-free solutions), the U69593-dependent activation of GIRK current was not significantly changed in the same cells (38.3 ± 8.5 pA, n = 7, Fig. 8C). It is interesting that after cells were reintroduced to standard physiological saline solution with Ca²⁺, U69593 was able to enhance the maximum I_h amplitude again with an average increase of 214 ± 21 % at −120 mV (control, 15.7 ± 4.8 pA; U69593, 34.0 ± 11.5 pA; n = 3). These findings suggest that the κ-receptor-induced Ca²⁺ release and augmentation of the I_h are largely dependent on extracellular Ca²⁺.

DISCUSSION

The present study demonstrates that in addition to GIRK channel activation, κ-opioid receptors enhance the maximum amplitude of I_h in NRM neurons. The current data also show that this κ-receptor-mediated effect can be significantly attenuated or blocked by various compounds that interfere with Ca²⁺ release, suggesting that this κ-receptor-mediated enhancement of maximum I_h amplitude involves mobilization of intracellular calcium released from calcium stores, particularly IP₃-sensitive stores. Thus, the κ-receptor-mediated calcium-dependent modulation of the maximum I_h amplitude may represent an intracellular mechanism for neurotransmitter regulation of the I_h channel and its associated cellular functions in central neurons.

Calcium-dependent modulation of the I_h

The current results provide evidence for the involvement of intracellular calcium released from calcium stores in the κ-receptor-mediated enhancement of maximum I_h amplitude. Besides the large number of studies on cAMP-dependent modulation of I_h channels, other studies have suggested a calcium dependence of I_h channels. In heart cells, an increase in intracellular calcium concentration enhances the I_h amplitude at all membrane potentials and the I_h enhancement is independent of protein kinase or calmodulin activities, indicating a direct regulation of the I_h channel by calcium (Hagiwara & Irisawa, 1989). A calcium-dependent, cAMP-independent potentiation of the I_h has also been reported after activation of muscarinic receptors in hippocampal neurons (Colino & Halliwell, 1993). In thalamocortical neurons, elevation of intracellular calcium concentration initiates the cAMP-mediated I_h enhancement (Luthi & McCormick, 1999), whereas the results from a separate study suggest the lack of a direct regulation of I_h by intracellular calcium in thalamic neurons (Budde et al. 1997). Furthermore, it was recently demonstrated that an increase in calcium concentration in hippocampal mossy fibre terminals enhances the I_h through calcium/calmodulin-sensitive adenylyl cyclase and cAMP (Mellor et al. 2002).

An apparent difference in the modulation of I_h by κ-receptors described here and by cAMP is that κ-receptor agonists augment the I_h by increasing its maximum current amplitude whereas cAMP increases I_h by shifting its voltage dependence to more depolarized potentials. This suggests that cAMP is not involved in the I_h enhancement by κ-receptors. Consistent with this, previously reported modulation of the maximum I_h by various neurotransmitters is independent of the cAMP pathway (Colino & Halliwell, 1993; Jiang et al. 1993; Cathala & Paupardin-Tritsch, 1997; Parkis & Berger, 1997). The current results suggest that intracellular calcium may mediate the effects of these neurotransmitters. It is not clear from this study whether calcium modulates the I_h directly or indirectly through calcium-dependent mechanisms other than the cAMP pathway. In view of the observation that the increase in maximum conductance of I_h by intracellular calcium is independent of protein kinases and calmodulins in cardiac cells (Hagiwara & Irisawa, 1989), it is possible that calcium directly regulates the maximum conductance through a calcium modulation site in I_h channels. The results of the present study may indicate a yet unidentified mechanism in the modulation of I_h channels by intracellular calcium, either directly or indirectly.

Mobilization of intracellular calcium by κ-opioid receptors

Ample biochemical evidence shows that in addition to their negative coupling to the cAMP signal transduction pathway, opioid receptors, including κ-receptors, also couple to and stimulate the phospholipase C pathway, whose activation leads to production of IP₃ and calcium release from IP₃-sensitive calcium stores (Smart et al. 1994; Ueda et al. 1995; Smart & Lambert, 1996). Consequently, κ and other opioid receptors increase IP₃ concentration and mobilize intracellular calcium from IP₃-sensitive stores in a calcium influx-dependent or -independent manner in various cell types (Jin et al. 1994; Tang et al. 1994; Fields et al. 1995; Allouche et al. 1996; Gurwell et al. 1996). These reports and the current results obtained using caffeine, thapsigargin, BAPTA and heparin strongly suggest that κ-receptors enhance the maximum I_h through mobilization of intracellular calcium, involving calcium release primarily from IP₃-sensitive stores. Consistent with this notion, inositol phospholipids have been implicated in the calcium-dependent, cAMP-independent increase of the I_h by muscarinic receptors (Colino & Halliwell, 1993; Allouche et al. 1996).

In addition, the current data obtained in Ca²⁺-free solution indicate that the κ-receptor-mediated enhancement of I_h is largely dependent on external Ca²⁺ concentration or on Ca²⁺ influx. Similar dependence on Ca²⁺ influx has been reported in muscarinic receptor-mediated Ca²⁺ release from IP₃-sensitive stores (Allouche et al. 1996). It is intriguing to consider possible mechanisms for the dependence of the κ-receptor-mediated effect on Ca²⁺ influx. Under the present experimental conditions, it is unlikely that mobilization of intracellular Ca²⁺ by κ-receptors is due to Ca²⁺ influx through voltage-sensitive or ligand-gated calcium channels. It has recently been demonstrated that depletion of Ca²⁺ stores by agonist-stimulated IP₃ production activates store-operated channels (SOCs) through interaction between IP₃-bound IP₃ receptors and SOCs, causing Ca²⁺ influx (Kiselyov et al. 1998). In view of the current evidence for κ-receptor-stimulated IP₃ production and Ca²⁺ release from IP₃-sensitive stores to increase the I_h, it appears likely that such a mechanism mediated by IP₃ receptors and SOCs is involved in the κ-receptor-mediated augmentation of the I_h, and that Ca²⁺ influx through SOCs is necessary for proper Ca²⁺ release and refill of Ca²⁺ stores induced by κ-receptors. Despite extensive biochemical data demonstrating opioid receptor-stimulated mobilization of intracellular calcium, evidence has been lacking regarding the physiological function of this signalling pathway for opioid receptors. The current findings demonstrate an example of the roles of this opioid receptor coupling in regulating neuronal activity.

Modulation of maximum I_h and activation of GIRK channels

There have been several reports of receptor-mediated inhibition of the maximum I_h and concomitant membrane hyperpolarization via activation of GIRK channels in the same cell. Although some attributed the reduced I_h to an indirect effect secondary to the GIRK channel activation in substantia nigra neurons (Watts et al. 1996; Cathala & Paupardin-Tritsch, 1999), other studies have shown that inhibition of the maximum I_h is independent of GIRK channel activation and is mediated by a separate, unknown mechanism (Rainnie et al. 1994; Svoboda & Lupica, 1998). In the present study, several lines of evidence suggest that κ-receptor-mediated enhancement of the maximum I_h is a direct effect independent of GIRK channel activation. Firstly, after blockade of GIRK channels by barium, the κ-receptor agonist still increased the maximum I_h. Secondly, after the I_h was blocked by ZD7288, κ-receptor activation induced a similar outward current without increasing the I_h, indicating that these two actions of κ-receptors are separate and independent of each other in these cells. Finally, N/OFQ and μ-receptor agonists activated GIRK channels in primary and secondary cells, respectively, but neither of them affected the I_h, making it unlikely that GIRK channel activation leads to the increase in I_h in our recording conditions.

Physiological implications

Given the general role of I_h in control of neuronal excitability and the important modulating role of these NRM neurons in opioid analgesia (Fields et al. 1991; Pan et al. 1997), the κ-receptor-mediated enhancement of the maximum I_h in these cells could have physiological significance in κ-opioid receptor-mediated acute and chronic actions. At the single cell level, it is proposed that concomitant enhancement of the I_h by κ-receptors could normally serve as a stopping mechanism to prevent extreme or over-inhibition due to GIRK channel activation by κ-receptors. It is also possible that in chronic conditions with elevated intracellular calcium levels, the activity of these cells driven by other excitatory inputs may increase due to enhanced I_h. At the system level, it is interesting to note that brain levels of intracellular calcium are significantly higher in morphine-tolerant rats (Welch & Bass, 1995). Moreover, κ-receptor agonists attenuate morphine tolerance and potentiate morphine analgesia in opioid-tolerant but not opioid-naive animals (Bhargava, 1994; Pan, 1998). Given the analgesic action of these brainstem cells on spinal pain transmission pathways (Pan et al. 1997), the κ-receptor-mediated and intracellular calcium-promoted increase in the firing activity of these cells in chronic opioid conditions may be one of the mechanisms for the anti-tolerance action of κ-receptors.