The Journal of Neurophysiology Vol. 79 No. 6 June 1998, pp. 2885-2894
Copyright ©1998 by the American Physiological Society

Contribution of Ca²⁺-Activated K⁺ Channels to Central Chemosensitivity in Cultivated Neurons of Fetal Rat Medulla

M.-C. Wellner-Kienitz, H. Shams, and P. Scheid

Institut für Physiologie, Ruhr-Universität Bochum, 44780 Bochum, Germany

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Wellner-Kienitz, M.-C., H. Shams, and P. Scheid. Contribution of Ca²⁺-activated K⁺ channels to central chemosensitivity in cultivated neurons of fetal rat medulla. J. Neurophysiol. 79: 2885-2894, 1998. Neurons in fetal rat medullary slices that exhibited spontaneous electrical activity after blockade of synaptic transmission were investigated for their response to decreases in extracellular pH. Increases in [H⁺] (induced either by fixed acid or increases in PCO₂) induced a significant increase in the frequency of action potentials, associated with a membrane depolarization, and/or increases in the slope of the interspike depolarization. In addition, CO₂/H⁺ prolonged the repolarizing phase of action potentials and reduced the afterhyperpolarization, suggesting that K⁺ channels were the primary site of CO₂/H⁺ action. The type of K⁺ channel that was modulated by CO₂/H⁺ was identified by application of agents that inhibited Ca²⁺-activated K⁺ channels either directly (tetraethylammonium chloride, TEA) or indirectly (Cd²⁺ ions) by inhibiting Ca²⁺ influx. CO₂/H⁺ effects on neuronal activity were abolished after application of these blockers. The contribution of Ca²⁺-activated K⁺ channels to H⁺ sensitivity of these neurons was confirmed further in voltage-clamp experiments in which outward rectifying I-V curves were recorded that revealed a zero current potential of 70 mV. CO₂/H⁺ induced a prominent reduction in outward currents and shifted the zero current potential to more positive membrane potentials (mean 63 mV). The CO₂/H⁺-sensitive current reversed at 72 mV and was blocked by external application of TEA. It is concluded that CO₂/H⁺ exerts its stimulatory effects on fetal medullary neurons by inhibition of Ca²⁺-activated K⁺ channels, either directly or indirectly, by blocking voltage-dependent Ca²⁺ channels, which in turn results in a reduction of K⁺ efflux and in cell depolarization.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Chemosensitive neurons respond to changes in PCO₂ and/or pH with a change in membrane potential and/or in spike frequency (Kawai et al. 1996; Neubauer et al. 1991; Richerson 1995). Whereas the general neuronal response to acidic stimuli is depression of activity (Balestrino and Somjen 1988; Jodkowski and Lipski 1986), specialized neurons have been identified that respond to increases in PCO₂ with a membrane depolarization and/or increases in spike frequency.

CO₂ sensitivity appears to be a widespread property of neurons within areas involved in control or modulation of respiration. CO₂-sensitive neurons, which are proposed to mediate the central chemosensitivity of breathing, have been identified below the ventral surface of the medulla oblongata (Issa and Remmers 1992; Kawai et al. 1996). In addition, CO₂-stimulated neurons have been recorded in the medullary raphe (Richerson 1995) and in the locus coeruleus (Ballantyne et al. 1997), suggesting that CO₂ sensitivity also is present in nuclei that are not primarily involved in the control of respiration. Beside the modulation of respiration, CO₂ is shown to affect other CNS functions such as cardiovascular regulation (Millhorn and Eldridge 1986), cerebral blood flow (Madden 1993), pain sensitivity, and arousal (Gronroos and Pertovaara 1994).

Several studies have demonstrated an intrinsic chemosensitivity of neurons, i.e., a response to changes in CO₂/H⁺ even after all synaptic inputs has been blocked. Such neurons have been found in the ventrolateral medulla (Kawai et al. 1996; Onimaru et al. 1989), the nucleus tractus solitarii (Dean et al. 1990), and rostral medullary raphe (Richerson 1995). These data give rise to the hypothesis that CO₂/H⁺ may modulate ion channels expressed in chemosensitive cells. However, the underlying ionic mechanisms of these neuronal responses have not yet been identified. Voltage-clamp recordings of Richerson and Pizzonia (1995) in medullary raphe neurons demonstrated a transient outward current (I_A), calcium currents, Ca²⁺-activated K⁺ currents, and an inward rectifying K⁺ current. Any of these currents could be the site of action of CO₂ and H⁺, but a detailed analysis of CO₂-induced modulation of these ion currents has not been performed. Dean et al. (1989) reported a hypercapnia-induced decrease in membrane conductance in dorsal medullary neurons attributed to a decreased outward K⁺ conductance which was, however, not further identified. Experiments on H⁺-sensitive type I cells of the rat carotid body (Peers 1990) indicated that H⁺ effects on Ca²⁺-activated K⁺ channels result in a reduction of K⁺ efflux and membrane depolarization. In their study of chemosensitive neurons of the solitary complex and the ventrolateral medulla, Southard et al. (1995) determined the reversal potential of the CO₂-sensitive current at 90 mV and found that the application of K⁺ channel blocker 4-aminopyridine abolished the CO₂-induced membrane depolarization in these cells. These data suggest an inhibitory effect of CO₂ on I_A.

However, a comprehensive study investigating the contribution of different ion currents to the neuronal chemosensitivity is not yet available. We performed experiments on cultivated CO₂-stimulated neurons of the fetal rat medulla to study which ion currents are modulated by CO₂ and/or H⁺ and thus might be responsible for the chemosensitivity of these neurons.

	METHODS

Abstract Introduction Methods Results Discussion References

Slice preparation

Organotypic cultures of the fetal rat medulla were prepared as described earlier (Wellner-Kienitz and Shams 1998). Briefly, pregnant Sprague-Dawley rats were anesthetized with halothane and killed by cervical dislocation on the 16th day of gestation. After removing the fetuses, further preparation steps were performed under sterile conditions. The fetuses were transferred to sterile Hank's balanced salt solution (HBSS without NaHCO₃ and phenol red) and decapitated. The medulla was prepared from its rostral part (adjacent to the pons, near the outlet of the hypoglossus) to the caudal region, where both vertebral arteries merge into the basilar artery (near the outlet of N. abducens) and cut into 225-µm-thick slices using a conventional tissue chopper. The slices were transferred to the Hank solution and kept there for 1 h at 4°C. After this incubation time, the slices were attached on sterile glass slides by using 20 µl thrombin and 40 µl chicken plasma (Sigma). The attached slices were placed into 15 ml centrifuge tubes containing medium A [59% Dulbecco's modified Eagle's medium (DMEM, GIBCO), 29% HBSS (GIBCO), 9% fetal calf serum (FCS, GIBCO), 2% glucose (20 g/l), 1% antibiotic/antimycotic solution (GIBCO) and 20 µg/l nerve growth factor (Biermann, Germany) and cultivated at 37°C and 5.3% CO₂ under continuous rotation. Five days after preparation, medium A was substituted by medium B (89% DMEM, 10% FCS, 1% antibiotic/antimycotic solution, 5 µg/ml nerve growth factor) supplemented with a solution containing 10 µM 5-fluoro-2-desoxyuridine, 10 µM uridine, and 10 µM cytosine--D-arabino-furanoside hydrochloride (referred to as FUA; all substances from Sigma) to reduce growth of glial cells. Medullary slices were incubated with FUA for 24 h, then stored in medium B for further cultivation. During cultivation, medium B was refreshed all 3 days.

As described previously (Wellner-Kienitz and Shams 1998), outgrowing cells that originated from slices, cultivated for 5 days, were identified as glial cells and immature neurons that exhibited a large sodium inward current but only small potassium outward currents. The membrane potential of these cells varied between 10 and 30 mV (hyperpolarizing with increasing cultivation time). During further cultivation time (days 11-17), outgrowing cells formed a neuronal network at the dorsal and to a smaller extent at the ventral side of the medullary slice. The membrane potential of these outgrowing cells was significantly more negative (around 50 mV) and some of these cells exhibited a spontaneous, regular firing pattern.

Electrophysiological experiments presented in this study were predominantly performed on neurons that originated from rostral medullary slices, cultivated for 11 days.

Electrophysiological measurements

Electrophysiological recordings were performed in the current- and voltage-clamp configuration of the patch-clamp technique. Electrodes were made from borosilicate glass (Clark Electromedical Instruments, Reading, UK; 5-6.5 M tip resistance) and filled with a solution containing (in mM) 120 K-gluconate, 10 KCl, 10 ethylene glycol-bis(-aminoethyl ether)N,N,N',N'-tetraacetic acid (EGTA), N-(2-hydroxyethyl) piperazine-N'-(2-ethane-sulfonic acid) (HEPES), 1 CaCl₂, and 1 MgCl₂ or with a solution containing 50 KCl, 80 K-gluconate, 10 EGTA, 10 HEPES, 1 CaCl₂, and 1 MgCl₂, adjusted to pH 7.2 with KOH. Signals were amplified (Axopatch 200A, Axon Instruments), filtered (1 kHz), and stored on disk for further analysis. Spikes and integrated firing rate also were recorded with a multichannel pen-recorder (model 2800S, Gould). Membrane potential, action potential parameters, and membrane currents were analyzed with the software ISO2 (MFK, Germany). All experiments were performed at 37°C.

For recordings, the attached medullary slices were transferred into the experimental chamber (volume 300-500 µl) and superfused with an artificial cerebrospinal fluid (CSF I) at a flow rate of 3 ml/min. CSF I contained (in mM) 125 NaCl, 0.5 NaH₂PO₄, 4 KCl, 10 MgSO₄, 26 NaHCO₃, 30 glucose, and 2 CaCl₂ and was equilibrated with gas mixtures of low (70% N₂, 28% O₂, 2% CO₂, superfusate pH 7.8) or high P_CO2 (70% N₂, 21% O₂, 9% CO₂, superfusate pH 7.2). Although the Ca²⁺ concentration was not reduced to preserve Ca²⁺-dependent cell functions, increasing the bath Mg²⁺ concentration to 10 mM was sufficient to block synaptic transmission. As described previously (Wellner-Kienitz and Shams 1998), the increase in the bath Mg²⁺ concentration from 1 to 10 mM completely blocked both the electrical activity and the response to CO₂ in one type of chemosensitive medullary neuron. Although in these cells, both the spike generation and the chemosensitivity were entirely dependent on synaptic inputs, other chemosensitive neurons retained their spontaneous electrical activity and sensitivity to changes in the bath CO₂ in the presence of 10 mM Mg²⁺. In this type of neuron, we frequently observed that the regular firing pattern was superimposed by synaptically transmitted spikes and excitatory postsynaptic potentials (EPSPs) in the presence of 1 mM Mg²⁺. Increasing the bath Mg²⁺ concentration from 1 to 10 mM abolished the superimposing spikes and EPSPs without affecting the spike generation or chemosensitivity. In the present study, electrophysiological experiments were performed on neurons perfused with a bath solution containing 10 mM Mg²⁺ for 20 min. Under these experimental conditions, superimposed spikes were not observed.

Electromagnetic valves controlled the solution flow from each reservoir, allowing rapid changes of CO₂ by switching from one perfusion solution to the other. In another series of experiments, Ca²⁺-activated K⁺ channels were blocked by superfusion of the modified CSF (CSF II) containing (in mM) 10 tetraethylammonium chloride (TEA), 120 NaCl, 0.5 NaH₂PO₄, 4 KCl, 10 MgSO₄, 30 glucose, 2 CaCl₂, and 10 piperazine-N,N'-bis(2-ethanesulfonic acid) (PIPES) (Na₂-salt), adjusted to pH 6.8 or 7.8 with NaOH and equilibrated with a gas mixture of 60% N₂ and 40% O₂. Control experiments for this series were performed with CSF II in the absence of TEA.

Space clamp problems that cause distortions of membrane currents might result in misinterpretation of the I-V relationships. To ensure that the neurons investigated in this study were appropriately clamped, we determined both the spike rise time and spike amplitude. Only neurons that exhibited spike amplitudes of >60 mV and a dV/dt of >50 V/s (at 2% CO₂) were selected for analysis. In the voltage-clamp configuration, the constancy of the I_Na threshold indicated that the cells were appropriately clamped. However, despite these precautions to select cells that were adequately clamped, in a few experiments, small current deflections superimposed the membrane currents that occurred in response to depolarizing voltage steps. Because we cannot exclude the possibility that outgrowing neurons are electrically coupled, these current deflections might be due to the activation of sodium channels in neighboring neurons.

	RESULTS

Abstract Introduction Methods Results Discussion References

Electrophysiological experiments were performed on neurons that were cultivated for 11-17 days. Most of the cells studied were localized within groups of 10-15 cells. Within these cell groups, several types of neuron were distinguished in respect of their firing pattern and CO₂/H⁺ sensitivity. In addition to chemosensitive cells in which both the electrical activity and the response to CO₂ were dependent on synaptic inputs, CO₂-inhibited and -stimulated neurons with a regular firing pattern were identified that retained their electrical activity and CO₂/H⁺ sensitivity in the absence of synaptic transmission (see Wellner-Kienitz and Shams 1998). For this study, only CO₂-stimulated cells that generated action potentials after blockade of synaptic transmission and thus revealed an intrinsic chemosensitivity were investigated(n = 42). A high percentage of these regularly firing neurons (60%, n = 25) was found in the outgrowing margins of the rostroventral medulla (in contrast, 80% of the CO₂-inhibited cells were located in the dorsal part of the tissue explant).

CO₂-stimulated neurons

The spontaneously active neuron in Fig. 1A displayed a regular firing pattern at 2% CO₂ with a spike frequency averaging 141/min. As shown on an expanded time scale in Fig. 1B, each spike was preceded by a gradual interspike ramp that depolarized the neuron up to the threshold for action potential generation, and was followed by an afterhyperpolarization. This neuron responded to a reduction in the superfusate pH (by increasing the bath CO₂-level to 9%) with an increase in spike frequency, due to increases in the interspike ramp slope, and membrane depolarization (from 55 to 49 mV, see Fig. 1B, right). The comparison of two typical spikes in Fig. 1C shows that the spike amplitude decreased with high CO₂ and that the spike duration was prolonged. In addition, the afterhyperpolarization level was reduced at 9% CO₂.

View larger version (45K): [in this window] [in a new window]   FIG. 1. Responses of a medullary pacemaker neuron to different levels of CO2 in the absence and presence of Ca2+ channel blockers. A: pen-recording of the electrical activity of a CO2-stimulated cell at 2 and 9% CO2. B: computer playbacks of action potentials in an expanded time scale. Spikes were shown at 2% CO2 (left) and 9% CO2 (right). C: comparison of single spikes at different CO2 levels. D: pen-recording of the same neuron in the presence of 50 µM Cd2+ and Ni2+. Membrane potential was set to the same level as in A by constant injection of 10 pA. E: single spikes in the absence and presence of Cd2+ and Ni2+ were superimposed (CO2 level: 2%). F: membrane currents recorded during a depolarizing test pulse from a holding potential of 80 mV to a test potential of 0 mV (duration, 100 ms) at 2 and 9% CO2 (different cell as in A). Note the superimposition of INa with K+ outward currents. CO2 increase from 2 to 9% has no effect on INa amplitude but decreases the amplitude of outward currents.

The reduction in spike amplitude is suggestive of a hypercapnia-induced decrease of a sodium inward current, but as shown in voltage-clamp experiments (Fig. 1F), the amplitude of the Na⁺ current was not affected by hypercapnia. Therefore, the decrease in spike amplitude is most likely due to partial inactivation of sodium channels as a result of depolarized membrane potential at the high level of CO₂. Both the prolongation of repolarization and the reduction of afterhyperpolarization may be caused by the inhibition of K⁺ outward currents. Because it is known that the activation of Ca²⁺-activated K⁺ channels contribute to the repolarization and afterhyperpolarization in neurons (e.g., in rat CA1 hippocampal interneurons) (Zhang and McBain 1995), we investigated whether hypercapnia inhibits Ca²⁺-activated K⁺ channels and thus blocks the observed CO₂ effects on spike frequency and membrane depolarization.

Indirect block of Ca²⁺-activated K⁺ channels

APPLICATION OF CA²⁺ CHANNEL BLOCKERS CD²⁺ AND NI²⁺. To inhibit Ca²⁺-activated K⁺ channels indirectly by reducing the Ca²⁺ influx, we blocked voltage-dependent Ca²⁺ channels by Cd²⁺ and Ni²⁺ and recorded the neuronal activity in the presence of both blockers. Application of both CdCl₂ (50 µM) and NiCl₂ (50 µM) resulted in a membrane depolarization accompanied by an increase in spike frequency and a reduction in afterhyperpolarization (not shown here). For an appropriate comparison of action potential parameters recorded during control with those recorded in the presence of Cd²⁺ and Ni²⁺, we corrected the Cd²⁺- and Ni²⁺-mediated change in membrane potential by injection of a constant negative current and then tested the neuronal response to CO₂ (Fig. 1D). In comparison with control, the spike frequency was significantly reduced at both levels of CO₂ (compare A with D), and the CO₂ effects on spike frequency and membrane potential were abolished completely after application of Cd²⁺ and Ni²⁺ (Fig. 1D, n = 5). In addition, the level of afterhyperpolarization was diminished and the repolarization of action potentials was prolonged by Cd²⁺ and Ni²⁺ (Fig. 1E).

To determine whether these effects were produced alone by one of these ions, Ni²⁺ or Cd²⁺, or by both, we applied these agents separately. The chemosensitive neuron of Fig. 2A responded to decreases in bath CO₂ with hyperpolarization (from 50 to 54 mV) and a decrease in spike frequency. Application of Ni²⁺ alone (Fig. 2B) resulted in a decrease in spike frequency due to a slope reduction of interspike depolarization, but neither the action potential duration nor the level of afterhyperpolarization were significantly affected (Fig. 2C). The effects of CO₂ on spike frequency as well as the CO₂-mediated membrane depolarization were retained in the presence of Ni²⁺ (Fig. 2B). In all four cells tested, application of Ni²⁺ (50-100 µM) was ineffective in blocking neuronal chemosensitivity.

View larger version (38K): [in this window] [in a new window]   FIG. 2. CO2-induced effects on membrane potential and spike frequency maintained in the presence of Ni2+. A: electrical activity of a CO2-stimulated neuron under control conditions. B: pen-recording of the same neuron in the presence of 50 µM Ni2+. C: superimposition of spikes in the absence and presence of Ni2+. Note the reduction in slope of the interspike depolarization in the presence of Ni2+ (CO2 level: 9%).

The effect of Cd²⁺ alone is shown in Fig. 3. Under control conditions, this neuron responded to hypercapnia with a slight membrane depolarization (from 64 to 60 mV) and an increase in spike frequency. These effects were reversible when switching back to 2% CO₂. Superfusion with Cd²⁺ resulted in a membrane depolarization that was corrected by injection of a negative DC current. Cd²⁺ induced a reduction in the afterhyperpolarization level (compare Fig. 3, A with B) and completely blocked the CO₂ sensitivity of the neuron (Fig. 3B). The CO₂-dependent modulation of spike frequency and membrane potential as well as a significant afterhyperpolarization reappeared when Cd²⁺ was removed by washout (Fig. 3C). The Cd²⁺-mediated block of the CO₂ effects was repeatable and independent of the level of neuronal activity at different membrane potentials (compare Fig. 3, B and D). These effects of Cd²⁺ were observed in all four spontaneously active neurons tested.

View larger version (67K): [in this window] [in a new window]   FIG. 3. Block of chemosensitivity by Cd2+. A: pen-recording of neuronal activity under control conditions. B: electrical activity in the presence of 50 µM Cd2+ during constant current injection of 15 pA. C: washout. D: electrical activity (at a different membrane potential as in B during 2nd application of Cd2+.

APPLICATION OF BA²⁺. Ba²⁺ is known to permeate Ca²⁺ channels instead of Ca²⁺, but in contrast to Ca²⁺, it does not activate the Ca²⁺-dependent K⁺ channels (Yellen 1987). The neuron in Fig. 4A responded to an increase in the bath CO₂ level with an increase in spike frequency without a significant membrane depolarization. Substituting external Ca²⁺ with equimolar concentration of Ba²⁺ depolarized the cell and increased its spike frequency, but the CO₂ sensitivity of the cell was abolished (Fig. 4B). Even the correction of membrane potential back to its control level (induced by current injection of 25 pA) did not restore the CO₂ sensitivity of the neuron (Fig. 4C). Again, as Ba²⁺ was replaced by Ca²⁺ (washout, Fig. 4D) the CO₂ effects were restored. The comparison of single spikes in the absence and presence of Ba²⁺ (Fig. 4E) demonstrates that the repolarization of action potentials was significantly prolonged, and the afterhyperpolarization was diminished markedly by Ba²⁺, whereas the Ba²⁺-induced changes in spike amplitude were small and the spike rise time was not affected. Qualitatively identical results were obtained in all cells studied (n = 5).

View larger version (48K): [in this window] [in a new window]   FIG. 4. CO2 sensitivity is reversibly blocked by Ba2+. A: control conditions. B: pen-recording of neuronal activity in the presence of 2 mM Ba2+. C: same neuron recorded during constant current injection of 25 pA. D: washout. E: comparison of single spikes in the absence (control) and presence of Ba2+ at a CO2 level of 9%.

Direct block of Ca²⁺-activated K⁺ channels by TEA

In these experiments, the spontaneous activity and H⁺ sensitivity of the neurons were recorded in a CO₂-free bath solution (CSF II, n = 28) while the H⁺ concentration was modified by fixed acids. Jarolimek et al. (1990) proposed that hypercapnia potentiated the pH-induced responses of medullary neurons and demonstrated that the same pH decrease was much more effective when raising P_CO2 than when decreasing [HCO₃]. Because we performed our experiments in a CO₂-free bath solution, we decided to enhance the pH reduction (by using bath solutions adjusted to pH 7.8 and pH 6.8) to maintain adequate H⁺-induced effects on neuronal activity in comparison with those experiments in which the bath P_CO2 was increased.

In the experiment shown in Fig. 5A, increasing bath pH from 6.8 to 7.8 resulted in membrane hyperpolarization and an almost complete cessation of firing. Subsequent reduction of bath pH to 6.8 depolarized the membrane from 54 to 50 mV and increased the spike frequency of the neuron back to its control level (35 spikes/min).

View larger version (63K): [in this window] [in a new window]   FIG. 5. Tetraethylammonium (TEA)-induced block of H+ effects on membrane potential and spike frequency. A: neuronal responses to pH reduction under control conditions and during application of 10 mM TEA. Note the strong membrane depolarization and increase in spike frequency after application of TEA. B: pen-recording of electrical activity during current injection of 5 pA in the presence of TEA. C: washout of TEA. D: superimposition of spikes in the absence and presence of TEA at pH 6.8.

Application of TEA (10 mM) further depolarized the cell membrane to 44 mV accompanied with increases in the action potential frequency and a reduction in the afterhyperpolarization level (see Fig. 5A). In addition, the repolarization of the action potentials was significantly prolonged (Fig. 5D). TEA induced also a slight increase in the spike amplitude (Fig. 5D), which could be due to the delayed spike repolarization, which in turn allows the full appearance of Na⁺ current amplitude.

In the presence of TEA, changes in bath pH had no effect on either the membrane potential or spike frequency of the neuron (Fig. 5A). Even at more negative membrane potentials (induced by current injection, B), at which pH-induced effects on neuronal activity have been shown to be augmented (Wellner-Kienitz and Shams 1998), changes in pH did not modulate the regular firing pattern or membrane potential of the neuron (Fig. 5B). The H⁺ effects on neuronal activity were restored after TEA was removed during washout (Fig. 5C). TEA either completely blocked the cellular response to pH (9 cells) or attenuated the H⁺-induced effects (1 cell).

Voltage-clamp experiments

Voltage-clamp experiments were performed to investigate whether acidification is accompanied with a decrease in K⁺ currents. H⁺-stimulated neurons were depolarized by voltage ramps from 120 to + 80 mV within 3 s (Fig. 6IA) or from 120 to 0 mV within 5 s (Fig. 6IC). At 2% CO₂, the resulting I-V curve reversed at 73 mV [see Fig. 6IA,bottom; mean value of 0 current potential: 70 ± 10(SD) mV, n = 5] and exhibited a strong outward rectification (Fig. 6IA, top). In the same cell, increasing CO₂ induced a prominent decrease in outward currents and a shift in the zero current potential to more positive membrane potentials (63 mV, see Fig. 6IA, bottom; mean value: 63 ± 9 mV). The I-V curves recorded at 2 and 9% CO₂ crossed each other at approximately 75 mV (mean 72 ± 9 mV, n = 5). Although the reversal potential of the H⁺-sensitive current (see difference trace in Fig. 6IB) is more positive than the calculated K⁺ equilibrium potential, E_K = 88 mV, a H⁺-induced modulation of ion conductances other than K⁺ (e.g., a chloride channel) is extremely unlikely because increasing the chloride concentration in the pipette from 14 to 54 mM (thereby shifting E_Cl from 56 to 22 mV) did not significantly shift the reversal potential of the H⁺-sensitive current (n = 2).

View larger version (19K): [in this window] [in a new window]   FIG. 6. Voltage-clamp recordings of CO2/H+-stimulated cells. IA, top: I-V curves of a CO2-stimulated neuron at 2 and 9% CO2. Voltage ramps polarized the cell from 120 to +80 mV within 3 s (holding potential 80 mV). I-V curves revealed a crossover at 75 mV (see arrow). IA, bottom: I-V curves at increased vertical resolution to show the 0 current potential at different levels of CO2. IB: difference trace of the I-V curves in IA (2  9% CO2). H+-sensitive current reversed at 75 mV. IC: I-V curves obtained during a voltage ramp from 120 to 0 mV (duration 5 s, holding potential 80 mV) at pH 7.8 and 6.8. IC, bottom: I-V curves at increased vertical resolution. II: same neuron as in IC. IIA: I-V curves recorded during application of 10 mM TEA at pH 7.8 and 6.8. IIB: difference trace.

In another series of experiments, the bath pH was reduced from 7.8 to 6.8 (Fig. 6IC). The pH reduction shifted the zero current potential from 54 to 49 mV (Fig. 6IC, bottom) and reduced the outward currents. The reversal potential of the H⁺-sensitive current was 60 mV (Fig. 6ID). Applying TEA in the same cell results in a strong decrease in outward membrane currents (compare Fig. 6, IC and IIA). In the presence of TEA, membrane currents were identical at both levels of pH (Fig. 6, IIA and IIB), e.g., H⁺-sensitive currents were blocked completely by TEA. In addition, TEA induced a positive shift of the zero current potential, consistent with the membrane depolarization recorded under current-clamp conditions. While the zero current potential was 54 mV at pH 7.8 during control (Fig. 6IC), it was shifted to 46 mV in the presence of TEA.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

We investigated the effects of CO₂/H⁺ on spontaneously active neurons with intrinsic chemosensitivity in cultures of the fetal rat medulla. A stimulatory effect on the neuronal activity was observed irrespective of whether changes in bath pH were induced by fixed acids or CO₂. Increasing H⁺ induced increases in the slope of the interspike depolarization and prolonged the spike repolarization phase together with a reduction in the afterhyperpolarization. Furthermore, hypercapnia induced a significant reduction in outward current (voltage-clamp experiments) with a reversal potential of the H⁺-sensitive current of 72 mV (Fig. 6). Similar results were described in a study of Buckler and Vaughan-Jones (1994) in type I cells of rat carotid body in which a reversal potential for the H⁺-sensitive current of 75 mV was found. Our data suggest a CO₂/H⁺-induced inhibition of K⁺ currents that contribute to the resting potential, which, in turn, results in a positive shift in the zero current potential (see Fig. 6), membrane depolarization and increasing firing rate.

The modulation of a chloride influx as a cause of the cell response to H⁺ is unlikely because shifting the E_Cl from 56 to 22 mV did not affect the reversal potential of the H⁺-sensitive current.

In several studies, the CO₂/H⁺-mediated inhibition of K⁺ channels is proposed to account for H⁺ chemosensitivity of these cells. The H⁺-inhibitory effect on a K⁺ channel is suggested to be the primary step in transduction of hypercapnic and/or acidic stimuli in type I cells of the carotid body (for review, see Peers and Buckler 1995) as well as in neurons of the nucleus tractus solitarii (Dean et al. 1989) and of the ventrolateral medulla (Southard et al. 1995). Whereas the H⁺-induced inhibition of the large Ca²⁺-activated K⁺ channel appears to be responsible for the H⁺ chemoreception in rat type I cells (Peers 1990), the H⁺-induced inhibition of the A current contributes to the H⁺ sensitivity in neurons of the solitary complex and the ventrolateral medulla (Southard et al. 1995).

In our study, however, Ca²⁺-sensitive K⁺ channels were involved in the process of neuronal H⁺ sensitivity in as much as the neuronal response to pH was abolished by the application of Cd²⁺. In the neurons of our study, the Cd²⁺ block of Ca²⁺ channels induced prolongation of spike duration without affecting the spike rise time and amplitude. The lack of Cd²⁺ effects on spike rise time and amplitude suggests that Ca²⁺ influx did not contribute to the generation of action potentials in CO₂/H⁺-stimulated neurons. The application of Cd²⁺ was accompanied by a membrane depolarization and a reduction of the afterhyperpolarization, indicating that K⁺ channels were blocked in this condition. Because it is known that Cd²⁺ has no direct blocking effect on K⁺ channels, the effect of Cd²⁺ is proposed to be mediated indirectly by blocking the Ca²⁺ influx, which would result in an inhibition of one or more types of Ca²⁺-activated K⁺ channel.

In contrast to Cd²⁺, application of Ni²⁺ (50-100 µM) alone affected neither the spike duration nor the amplitude of afterhyperpolarization and failed to block neuronal chemosensitivity. Therefore the contribution of Ni²⁺-sensitive Ca²⁺ channels to the chemosensitivity of CO₂/H⁺-stimulated neurons can be ruled out, but we hypothesize that Ni²⁺-sensitive Ca²⁺ currents contribute to the spontaneous interspike depolarization because the application of Ni²⁺ reduced the spike frequency by a reduction in the slope of the interspike depolarization (Fig. 2).

It is known that Cd²⁺ in a concentration of 50 µM preferentially blocks the Ca²⁺ current through L-type channels, whereas Ni²⁺ in a concentration of 100 µM effectively blocks T-type Ca²⁺ channels without affecting L-type Ca²⁺ channels (Fox et al. 1987, experiments on chick sensory neurons). However, both Cd²⁺ and Ni²⁺ are not absolute selective blockers for the different types of Ca²⁺ channel. Therefore, additional voltage-clamp experiments and the application of more specific blockers (e.g., dihydropyridines, -conotoxin) are required to identify the types of Ca²⁺ channel that contribute to the K⁺ conductance in our study. Although our data suggest that one type of these channels may be a L-type Ca²⁺ channel, we have to assume that at least one more type of Ca²⁺ channel contributes to the activation of Ca²⁺-dependent K⁺ channels, one that is active near the resting potential. A persistent Ca²⁺ current that can account for the slow ramp-like depolarization has been described in dopaminergic neurons of the substantia nigra (Kang and Kitai 1993). It remains to be investigated whether a similar Cd²⁺-sensitive Ca²⁺ current is also present in our preparation.

We have, however, experimental evidence that Ca²⁺-activated K⁺ currents contribute to the membrane currents at negative membrane potentials because the application of TEA induced a decrease of membrane currents at potentials negative to 30 mV [Fig. 6: reduction from +42 pA (without) to +24 pA (with TEA) at 40 mV, pH 7.8]. Although we propose that Ca²⁺-activated K⁺ channels are sufficiently active near the resting potential such that its H⁺-induced inhibition can account for a membrane depolarization, the activity of Ca²⁺-activated K⁺ channels at negative membrane potentials in vivo may be greater than assumed here. Because our recording solution strongly buffered the intracellular free Ca²⁺ concentration, Ca²⁺-dependent K⁺ channels only may be activated partially. Although the slow Ca²⁺ buffer EGTA does not modify the I_K(Ca) in other preparations (Protti and Uchtel 1997, experiments on mouse motor nerve terminals), we cannot exclude that, in our preparation, the neuronal response to increases in the bath CO₂ and pH reduction may be enhanced in vivo or under experimental conditions that reduce Ca²⁺ buffering.

The idea that CO₂/H⁺-induced inhibition of Ca²⁺-activated K⁺ channels primarily accounts for central chemosensitivity is further supported by our observation that Ba²⁺ and TEA completely block the neuronal response to CO₂/H⁺. Both TEA and Ba²⁺ are known to block Ca²⁺-activated K⁺ channels in various tissues either directly (TEA) or indirectly (Ba²⁺) (Yellen 1987). From our data, we cannot determine whether the neuronal chemosensitivity is attributed to an indirect inhibition of Ca²⁺-activated K⁺ channels after H⁺-induced block of voltage-dependent Ca²⁺ channels or if H⁺ ions are acting directly on the Ca²⁺-activated K⁺ channel. Although an H⁺-induced modulations of L-type Ca²⁺-currents has been described in rat CA1 neurons, in which extracellular acidosis (pH 6.9-6.0) reversibly depressed the Ca²⁺ current amplitude and caused a positive shift in the voltage dependence of current activation (Tombaugh and Somjen 1996), further voltage-clamp experiments have to investigate whether a similar CO₂/H⁺-induced inhibition of one or more types of Ca²⁺ channel occurs in our preparation.

In the case of a direct H⁺ inhibition of Ca²⁺-activated K⁺ channels, it remains to be elucidated whether the site of H⁺ action is extracellular or intracellular. In rat type I cells of carotid body, the K⁺ inhibition appears to be mediated by a fall in intracellular pH (see Peers and Buckler 1995 for review). Because Erlichman et al. (1995) observed a hypercapnia-induced decrease in pH_i in medullary neurons, a similar mechanism might be proposed for the inhibition of Ca²⁺-activated K⁺ channels in our preparation, but this hypothesis remains to be tested.

	ACKNOWLEDGEMENTS

  This work was supported by the Deutsche Forschungsgemeinschaft (We 2073/1-1).

	FOOTNOTES

  Address reprint requests to M.-C. Wellner-Kienitz.

  Received 18 September 1997; accepted in final form 11 February 1998.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• BALESTRINO, M., SOMJEN, G. G. Concentration of carbon dioxide, interstitial pH and synaptic transmission in hippocampal formation of the rat. J. Physiol. (Lond.) 396: 257-266, 1988.
• BALLANTYNE, D., OYAMADA, Y., MÜCKENHOFF, K., AND SCHEID, P. Chemosensitive pontine (loecus coeruleus, LC) neurones in the neonatal rat brainstem in vitro. Proc. XXXIII Intl. Congr. Physiol. Sci. St. Petersburg, Russia L030.08, 1997.
• BUCKLER, K. J. AND VAUGHAN-JONES, R. D. Effects of hypercapnia on membrane potential and intracellular calcium in rat carotid body type I cells. J. Physiol. (Lond.) 478: 1: 157-171, 1994.
• DEAN, J. B., BAYLISS, D. A., ERICKSON, J. T., LAWING, W. L., MILLHORN, D. E. Depolarization and stimulation of neurons in nucleus tractus solitarii by carbon dioxide does not require chemical synaptic input. Neuroscience 36: 207-216, 1990.[Medline]
• DEAN, J. B., LAWING, W. L., MILLHORN, D. E. CO₂ decreases membrane conductance and depolarizes neurons in the nucleus tractus solitarii. Exp. Brain Res. 76: 656-661, 1989.[Medline]
• ERLICHMAN, J. S., RITUCCI, N. A., PUTNAM, R. W., HUANG, R.-Q., DEAN, J. B. In vitro measurements of pH_i in neurons in medullary chemosensitive areas. Proc. 25th Annu. Meet. Soc. Neurosci., San Diego, CA 21: 738.10, 1995.
• FOX, A. P., NOWYCKY, M. C., TSIEN, R. W. Kinetic and pharmacological properties distinguishing three types of calcium currents in chick sensory neurons. J. Physiol. (Lond.) 394: 149-172, 1987.[Abstract/Free Full Text]
• GRONROOS, M., PERTOVAARA, A. A selective suppression of humain pain sensitivity by carbon dioxide: central mechanisms implicated. Eur. J. Appl. Physiol. Occup. Physiol. 68: 74-79, 1994. [Medline]
• ISSA, F. Q., REMMERS, J. E. Identification of a subsurface area in the ventral medulla sensitive to local changes in PCO₂. J. Appl. Physiol. 72: 439-446, 1992.[Abstract/Free Full Text]
• JAROLIMEK, W., MISGELD, U., LUX, H. D. Neurons sensitive to pH in slices of the rat ventral medulla oblongata. Pflügers Arch. 416: 247-253, 1990.[Medline]
• JODKOWSKI, J. S., LIPSKI, J. Decreased excitability of respiratory motoneurons during hypercapnia in the acute spinal cat. Brain Res. 386: 296-304, 1986.[Medline]
• KANG, Y., KITAI, S. T. A whole cell patch-clamp study on the pacemaker potential in dopaminergic neurons of rat substantia nigra compacta. Neurosci. Res. 18: 209-221, 1993.[Medline]
• KAWAI, A., BALLANTYNE, D., MÜCKENHOFF, K., AND SCHEID, P. Chemosensitive medullary neurons in the brain-stem-spinal cord preparation of the neonatal rat. J. Physiol. (Lond.) 492,1: 277-292, 1996.
• MADDEN, J. A. The effect of carbon dioxide on cerebral arteries. Pharmacol. Ther. 59: 229-250, 1993.[Medline]
• MILLHORN, D. E., ELDRIDGE, F. L. Role of ventrolateral medulla in regulation of repiratory and cardiovascular systems. J. Appl. Physiol. 61: 1249-1263, 1986.[Abstract/Free Full Text]
• NEUBAUER, J. A., GONSALVES, S. F., CHOU, W., GELLER, H. M., AND EDELMAN, N. H. Chemosensitivity of medullary neurons in explant tissue cultures. Neuroscience 45: 1: 701-708, 1991.
• ONIMARU, H., ARATA, A., HOMMA, I. Firing properties of respiratory rhythm generating neurons in the absence of synaptic transmission in rat medulla in vitro. Exp. Brain Res. 76: 530-536, 1989.[Medline]
• PEERS, C. Effect of lowered pH on Ca²⁺-dependent K⁺ currents in type I cells from the neonatal rat carotid body. J. Physiol. (Lond.) 422: 381-395, 1990.[Abstract/Free Full Text]
• PEERS, C., BUCKLER, K. J. Transduction of chemostimuli by the type I carotid body cell. J. Membr. Biol. 144: 1-9, 1995.[Medline]
• PROTTI, D. A., UCHITEL, O. D. P/Q-type calcium channels activate neighboring calcium-dependent potassium channels in mouse motor nerve terminals. Pflügers Arch. 434: 406-412, 1997.[Medline]
• RICHERSON, G. B. Response to Co₂ of neurons in the rostral ventral medulla in vitro. J. Neurophysiol. 73: 933-944, 1995.[Abstract/Free Full Text]
• RICHERSON, G. B., PIZZONIA, J. H. Electrophysiology of medullary raphe neurons in slices and tissue culture. Proc. 25th Annu. Meet. Soc. Neurosci., San Diego, CA 21: 737.11, 1995.
• SOUTHARD, T. L., HUANG, R. Q., DEAN, J. B. Electrophysiological properties of neurons in chemosensitive areas of the dorsal and ventral brain-stem. Proc. 25th Annu. Meet. Soci. Neurosci., San Diego, CA 21: 738.12, 1995.
• TOMBAUGH, G. C., SOMJEN, G. G. Effects of extracellular pH on voltage-gated Na⁺, K⁺ and Ca²⁺ currents in isolated CA1 neurons. J. Physiol. (Lond.) 493: 719-732, 1996.[Medline]
• WELLNER-KIENITZ, M. C., SHAMS, H. CO₂-sensitive neurons in organotypic cultures of the fetal rat medulla. Respir. Physiol. 111: 137-151, 1998.[Medline]
• YELLEN, G. Permeation in potassium channels: implications for channel structure. Annu. Rev. Biophys. Chem. 16: 227-246, 1987.[Medline]
• ZHANG, L., MCBAIN, C. J. Potassium conductances underlying repolarization and afterhyperpolarization in rat CA1 hippocampal interneurones. J. Physiol. (Lond.) 488: 661-672, 1995.[Medline]

0022-3077/98 $5.00 Copyright ©1998 The American Physiological Society