Selective modulation of membrane currents by hypoxia in intact airway chemoreceptors from neonatal rabbit

Abstract

1. We previously described voltage-dependent ionic currents and hypoxia chemosensitivity in cultured pulmonary neuroepithelial body (NEB) cells isolated from fetal rabbit. Here we use fresh neonatal rabbit lung slices (200–400 μm thick) to characterize the electrophysiological properties of ‘intact’ NEBs with patch-clamp, whole-cell recording.

2. Under voltage clamp, outward currents were partially inhibited by TEA (20 mm), 4-amino pyridine (4-AP; 2 mm) and cadmium (Cd²⁺; 100 μm), suggesting the presence of both Ca²⁺-dependent (I_K(Ca)) and Ca²⁺-independent (I_K(V)) components.

3. Inward currents, carried by voltage-dependent Ca²⁺ channels and also, in occasional cells (∼11%), by TTX-sensitive Na⁺ channels, were also detected in intact NEB cells.

4. Hypoxia (P_O2= 15–20 mmHg) reduced the outward K⁺ current by ∼34% during voltage steps from −60 to +30 mV, while inward Ca²⁺ or Na⁺ currents were not affected by hypoxia. Hypoxia suppressed roughly equally both I_K(Ca) and I_K(V) components of outward current, and no further inhibition of K⁺ currents was seen with either TEA and 4-AP + hypoxia.

5. Diphenylene iodonium (DPI; 1 μm) suppressed outward K⁺ current by ∼42%, and DPI + hypoxia had no additional effect on the K⁺ current.

6. Direct application of H₂O₂ augmented outward K⁺ current; for a voltage step from −60 mV to +30 mV, 0.25 mm H₂O₂ increased K⁺ current by ∼37%.

7. These results indicate that intact neonatal NEB cells express hypoxic chemosensitivity and introduce the rabbit lung slice preparation as an new model for investigating the role of airway O₂ chemoreceptors.

Previous studies in our laboratory have shown that cultures of pulmonary neuroepithelial bodies (NEBs) isolated from fetal rabbit lung exhibit membrane properties of excitable cells since they possess voltage-activated K⁺, Na⁺ and Ca²⁺ currents (Youngson et al. 1994). Upon exposure to hypoxia (P_O2, 25–30 mmHg) there was a reversible reduction in K⁺ current but no effect on Na⁺ or Ca²⁺ currents. The prominence of NEBs during the perinatal period indicates that their proposed function as airway O₂ sensors may be important during the transition from fetal to neonatal life. There is morphological evidence for vagal innervation of NEB cells, thus providing a pathway by which responses of NEB cells could be integrated in the central nervous system (Lauweryns & Van Lommel, 1987; Adriaensen & Scheuermann, 1993; Van Lommel et al. 1995).

In mammalian carotid bodies (CBs), the main peripheral arterial chemoreceptors involved in the control of breathing, inhibition of voltage-gated K⁺ channels by hypoxia is well documented for parenchymal type I or glomus cells (Delpiano & Hescheler, 1989; Gonzalez et al. 1994). Two main types of oxygen-sensitive K⁺ channel have been described: a high-conductance calcium-activated (BK_Ca) channel and a voltage-insensitive K⁺ leak channel (Buckler, 1997) in rat glomus cells (Peers, 1990; Lopez-Lopez et al. 1997) and a lower conductance (40 pS) calcium-insensitive channel (termed the K_O2 channel) in rabbit glomus cells (Lopez-Lopez et al. 1989; Ganfornina & Lopez-Barneo, 1991). However, the precise mechanisms of inhibition of K⁺ current by hypoxia in CB or NEB cells is at present unknown.

Here we report characterization of voltage-activated ion channels, and their sensitivity to hypoxia, in NEB cells in neonatal rabbit using a fresh lung slice preparation. This novel technique allows direct patch-clamp recording from NEB cells localized in the airway epithelium in situ, without potential artifacts induced by enzymatic treatment, cell dissociation and/or long-term culture. Our data demonstrate that hypoxia-sensitive K⁺ currents are functional in intact neonatal rabbit NEB cells. We also report the presence in these cells of Ca²⁺-dependent and Ca²⁺-independent components of K⁺ current, both of which were found to be inhibited by hypoxia. In addition we observed modulation of NEB outward K⁺ current by H₂O₂ in the neonatal lung slice, as previously demonstrated in dispersed cultures of NEB cells from fetal lung (Youngson et al. 1993; Wang et al. 1996).

METHODS

Lung slice preparation

Neonatal New Zealand White rabbits of both sexes were used between 1 and 10 days of age. All experiments were carried out with the approval of the local ethics committee and in accordance with the Institutional Guidelines for Animal Care. The rabbits were killed by an intraperitoneal Euthanyl injection (100 mg kg⁻¹). The lungs were perfused with Krebs solution (mM: 140 NaCl, 3 KCl, 1.8 CaCl₂, 1 MgCl₃, 10 Hepes, 5 glucose; pH 7.3 adjusted with HCl) and then embedded in 2 % agarose (FMC Bioproducts, Rockland, ME, USA). Transverse lung slices (200–400 μm) were cut with a Vibratome (Ted Pella, Inc. Redding, CA, USA). Sectioning was performed with the tissue immersed in ice-cold Krebs solution.

Electrophysiological techniques and solutions

For electrophysiological recordings the lung slices were transferred to a recording chamber mounted on the stage of a Nikon microscope (Optiphot-2UD, Nikon, Tokyo, Japan). The ‘perfusing’ Krebs solution had the following composition (mM): 130 NaCl, 3 KCl, 2.5 CaCl₂, 1 MgCl₂, 10 NaHCO₃, 10 Hepes, 10 glucose (pH 7.35–7.4). To identify NEB cells in fresh lung tissue, the slices were incubated with the vital dye Neutral Red (0.02 mg ml⁻¹) for 15 min at 37°C as previously described (Youngson et al. 1993). Whole-cell K⁺ currents were recorded using the above ‘perfusing’ Krebs solution as the extracellular medium. For studies of Ca²⁺ channel currents, Ba²⁺ was used as the charge carrier and the external solution was switched to solution containing (mM): 140 NaCl, 2.7 KCl, 1.2 MgCl₂, 10 BaCl₂, 10 glucose, 10 Hepes (pH 7.4). The chamber, which had a volume of 0.2 ml, was perfused continuously with ‘perfusing’ Krebs solution at a rate of 6–7 ml min⁻¹. All recordings were made from submerged lung slices at room temperature (∼22°C).

Drugs were applied to the perfusate and their delivery to the cells was controlled by separate valves. Tetraethylammonium (TEA), 4-aminopyridine (4-AP), diphenylene iodonium (DPI), tetrodotoxin (TTX) and hydrogen peroxide (H₂O₂) were obtained from Sigma. Outward K⁺ currents, Ca²⁺ currents and Na⁺ currents in NEB cells were recorded in voltage-clamp experiments with patch electrodes, made from 1.1 mm o.d. and 0.8 mm i.d. thin-walled glass tubing (Kimax-51, Kimble, Pittsburgh, PA, USA) drawn with a vertical puller (PP-83, Narishige, Tokyo, Japan). Two different internal solutions were used in whole-cell recordings. When K⁺ currents were recorded, the pipette solution contained (mM): 30 KCl, 100 potassium gluconate, 1 MgCl₂, 4 Mg-ATP, 5 EGTA, 10 Hepes (9). The pH of the solution was adjusted to 7.25 with KOH. To isolate inward currents, an internal solution with following composition was used (mM): 130 CsCl, 5 TEACl, 2 MgCl₂, 10 Hepes (pH adjusted to 7.2 with CsOH). Whole-cell patch recording was performed as described by Hamill et al. (1981). The recording electrodes were advanced with a hydraulic manipulator (ONM-1, Narishige) to the NEB cells at a 45 deg angle under visual guidance (× 400 magnification). The seal resistance was typically 1–2 GΩ. Hypoxic solution was prepared by bubbling 95 % N₂ into the reservoir which fed the perfusion chamber via low-gas permeability tubing. The samples of perfusion medium from the recording chamber were taken for measurements of P_O2 levels using a P_O2 electrode (1610 pH/Blood Gas Analyser, Instrumentation Laboratory, Lexington, MA, USA). The P_O2 of the control normoxic perfusion medium in the recording chamber was 145–155 mmHg. The level of P_O2 of the hypoxic solution in the recording chamber varied between 15 and 20 mmHg.

An Axopatch 200B (Axon Instruments) amplifier was used for recording in whole-cell voltage-clamp mode. The data were filtered at 5 kHz. The level of fluid over the slices was kept low to minimize stray capacitance. Voltage commands and data acquisition were done using pCLAMP6 software and DigiData 1200 interface (Axon Instruments). The leak current was subtracted from all current records using pCLAMP software (P/4 subtraction protocol).

The resting membrane potential was measured soon after rupture of the cell membrane. The resting input resistance (R_m) was estimated from linear current-voltage relationships over the range -80 to -50 mV in voltage-clamp mode. The cell capacitance (C_m) was determined from the transient current response during a small hyperpolarizing voltage step. The time constant (τ_m) was calculated from the input resistance and capacitance (τ_m=R_m×C_m; Fu et al. 1996). All data are given as means ±s.e.m. Statistical analysis was performed using Student's paired and unpaired t tests. Differences were considered to be statistically significant when P < 0.05.

RESULTS

In lung slices incubated with Neutral Red dye (0.02 ml⁻¹), NEB cells appeared as reddish-pink cell clusters within the airway epithelium which remained unstained (indicated by arrowheads in Fig. 1).

Figure 1. Staining of NEB cells.

Arrowheads show Neutral Red staining of NEB cells in a fresh slice of neonatal rabbit lung (2-day-old). Scale bar represents 5 μm.

Passive membrane properties of NEB cells

Patch-clamp recordings were performed on 63 NEB cells sampled from 55 lung slice preparations. The mean resting potential was -51.2 ± 1.6 mV (n= 21). The mean resting input resistance and capacitance were 1.08 ± 0.06 GΩ (n= 22) and 2.42 ± 0.1 pF (n= 46), respectively. The mean membrane time constant was 2.5 ± 0.17 s (n= 22). Assuming NEB cells are spherical and specific membrane capacitance is 1 μF cm⁻², our results correspond to a mean NEB cell diameter of ∼8.7 μm.

Pharmacology of outward K⁺ currents

Pharmacological characterization of K⁺ currents in NEB cells of rabbit neonatal lungs was carried out by testing the effects of the K⁺ channel blockers TEA and 4-AP on the amplitude of currents evoked during steps from -60 to +30 mV. The effects of various concentrations of TEA (5, 20 and 30 mM) and 4-AP (1, 2 and 4 mM) were tested. The maximum effects, using 20 mM TEA and 2 mM 4-AP, on the same NEB cell are shown in Fig. 2A-C and E-G, respectively. Both blockers significantly reduced the current amplitude at voltages between -30 and +30 mV (P < 0.05, n= 6, Fig. 2D and H). Current amplitudes at a test potential of +30 mV before and during 20 mM TEA application were 422.5 ± 9.6 pA and 223.7 ± 50 pA, respectively, corresponding to a reduction by ∼47 %. Current amplitudes at a test potential of +30 mV before and during 2 mM 4-AP application were 422.5 ± 9.6 and 200.7 ± 50 pA, respectively, corresponding to a reduction by 52.6 %. There was no significant change in current amplitude when the concentration of TEA was increased from 20 to 30 mM (223 ± 35.8 pA in 30 mM TEA, n= 4). or when the concentration of 4-AP was increased from 2 to 4 mM (243.4 ± 28.7 pA in 4 mM 4-AP, n= 4).

Figure 2. Effects of 4-AP and TEA on K⁺ current in neonatal rabbit NEB cells in lung slices.

A, outward K⁺ current evoked by depolarizing steps from -60 mV to +30 mV in control Krebs solution. B, K⁺ current was reduced by perfusing 20 mM TEA. C, washout of the TEA caused a recovery of the outward K⁺ current. D, I-V relationship was plotted under control conditions (•), after perfusing 20 mM TEA (▵); data are shown as means ±s.e.m. for a sample of 6 cells. E, outward K⁺ current evoked by depolarizing steps from -60 mV to +30 mV in control Krebs solution. F, K⁺ current was reduced by perfusing 2 mM 4-AP. G, washout of the 4-AP caused a recovery of the outward K⁺ current. H, I-V relationship was plotted under control conditions (•), after perfusing 2 mM 4-AP (⋄); data are shown as means ±s.e.m. for a sample of 5 cells.

Hypoxia suppresses K⁺ current in NEB cells from neonatal rabbit lung slices

Previous studies on NEB cells from rabbit fetal lung culture preparations have shown that hypoxic transduction involves reversible blockade of K⁺ channels (Youngson et al. 1993). Voltage-gated K⁺ currents were prominent in NEB cells in neonatal rabbit lung slices (Fig. 2). Depolarizing steps from a holding potential of -60 to +30 mV evoked outward K⁺ currents in the majority of cells tested (95 %). The onset of the current was relatively slow and inactivation was not apparent during 600 ms pulses. The current-voltage relationship is shown in Fig. 3D, using current values measured near the end of the 600 ms depolarizing step. These delayed outward currents were apparent during steps from a holding potential of -60 mV to potentials between -40 and +30 mV. Exposure to hypoxia resulted in a rapid and reversible reduction in amplitude of the outward K⁺ currents in NEB cells (Fig. 3B and C). This reduction of the current amplitude was observed over a voltage range between -40 and +30 mV, and was significant between -10 and +30 mV (Fig. 3D). Current amplitudes at a test potential of +30 mV before and during hypoxia were 349.4 ± 42.8 and 230.8 ± 38.3 pA, respectively (P < 0.01, n= 10), corresponding to a reduction by ∼34 %. A similar response to hypoxia was observed in four NEB cells from fetal lung slices; an ∼34 % reduction of K⁺ outward current was observed when the cell was exposed to hypoxia at a test potential of +30 mV. The outward current was fitted by a two-component exponential function with time constants of 257.6 ± 46.3 and 27.8 ± 3.5 ms in control conditions and 221.5 ± 45.3 and 21.3 ± 3.1 ms during hypoxia (P < 0.05, n= 6).

Figure 3. Effect of hypoxia on K⁺ current in neonatal rabbit NEB cells.

A, outward K⁺ current evoked by depolarizing steps from -60 to +30 mV in control normoxic Krebs solution. B, outward current evoked by same voltage steps as in A was reduced by hypoxia. C, washout of the hypoxic solution caused a recovery of the outward K⁺ current. D, I-V relationships for the current in control solution (•), and in hypoxic solution (^) are plotted together with recovery K⁺ current (▵). Holding potential was -60 mV; * significant difference from control (P < 0.05). Data represent means ±s.e.m. for a sample of 10 cells.

In order to assess the relationship between oxygen-sensitive K⁺ current and TEA- or 4-AP-sensitive K⁺ current in NEB cells, the outward K⁺ current was evoked by depolarization of the holding potential from -60 to +30 mV. The current amplitude was reduced by applying either 20 mM TEA (Fig. 4A and B) or 2 mM 4-AP (Fig. 4F and G). In the presence of TEA or 4-AP, hypoxic solution failed to reduce the residual outward K⁺ current (Fig. 4C and H). The I-V curves for the TEA and TEA + hypoxia conditions are shown in Fig. 4E. The mean control K⁺ current, TEA-insensitive K⁺ current, and TEA + hypoxia-insensitive K⁺ current were 403.7 ± 10, 219 ± 9.8 and 207 ± 6.8 pA (n= 4), respectively. The I-V curves for the 4-AP and 4-AP + hypoxia conditions are shown in Fig. 3J. The mean control K⁺ current, 4-AP-insensitive K⁺ current and 4-AP + hypoxia-insensitive current were 406.5 ± 22.8, 250 ± 22.7 and 238.6 ± 28.9 pA (n= 4), respectively. The effects were reversible (Fig. 4D and I). These data indicate that the TEA/4-AP-sensitive K⁺ current may correspond to the O₂-sensitive K⁺ current in NEB cells.

Figure 4. Hypoxic solution with TEA and 4-AP failed to reduce K⁺ current.

A, outward K⁺ current evoked by depolarizing steps from -60 to +30 mV in control Krebs solution. B, K⁺ current was reduced by perfusing 20 mM TEA. C, K⁺ current in B was not altered by further exposure to both TEA and hypoxia. D, washout of the TEA + hypoxic solution caused a recovery of the outward K⁺ current E, I-V relationships obtained from four cells under control conditions (•), after perfusing the cells with 20 mM TEA (▵) and after perfusing with 20 mM TEA and hypoxic solution (^). F, outward K⁺ current evoked by depolarizing steps from -60 to +30 mV in control Krebs solution. G, K⁺ current was reduced by perfusing 2 mM 4-AP. H, K⁺ current in B was not altered by further exposure to both 4-AP and hypoxia. I, washout of the 4-AP + hypoxic solution caused a recovery of the outward K⁺ current. J, I-V relationships obtained from four cells under control conditions (•), after perfusing the cells with 2 mM 4-AP (⋄) and after perfusing with 2 mM 4-AP + hypoxic solution (□). Data are shown as means ±s.e.m.

Ca²⁺-dependent and Ca²⁺-independent hypoxia-sensitive K⁺ currents

Outward current in NEB cells was carried mainly by K⁺ ions, and two different components could be distinguished: a Ca²⁺-dependent K⁺ current (I_K(Ca)) sensitive to Cd²⁺, and a Ca²⁺-insensitive voltage-dependent K⁺ current (I_K(V)). To test the calcium dependence of the hypoxia-sensitive K⁺ current, NEB cells were exposed to hypoxia before and after blockade of voltage-gated Ca²⁺ channels with 100 μm Cd²⁺ (n= 6). An example is shown in Fig. 5. Voltage-dependent K⁺ currents were evoked by depolarization from a holding potential of -60 mV to +30 mV. Current amplitude at a test potential of +30 mV during bath application of 100 μm Cd²⁺ was reduced by 11.9 ± 0.4 % (n= 6). In the presence of Cd²⁺, hypoxic solution further reduced the residual Ca²⁺-independent K⁺ current by ∼20 % in all cells tested (n= 6, e.g. Fig. 5E). This indicated that at least a Ca²⁺-independent component of K⁺ current (I_K(V)) was inhibited by hypoxia. To isolate the hypoxia-sensitive components of K⁺ current, a procedure based on difference current measurements was used. For example, in Fig. 5G, the total hypoxia-sensitive current was isolated by subtraction of the remaining currents in hypoxia (Fig. 5B) from the control currents recorded in normoxia (Fig. 5A). This hypoxia-sensitive component is shown as a difference current for each voltage step in Fig. 5G. The Ca²⁺-independent hypoxia-sensitive K⁺ current component (Fig. 5H) was isolated by subtraction of the K⁺ currents recorded in the presence of Cd²⁺+ hypoxia (Fig. 5E) from control currents recorded with Cd²⁺ alone (Fig. 5D). Since the difference currents in the absence of Cd²⁺ (Fig. 5G) are larger than those in its presence (Fig. 5H), it appears that an additional Ca²⁺-dependent component of K⁺ current is sensitive to hypoxia. This component was isolated in Fig. 5I by subtraction of the Ca²⁺-independent hypoxia-sensitive K⁺ current component (Fig. 5H) from the total hypoxia-sensitive K⁺ current (Fig. 5G). The I-V curves for the different hypoxia-sensitive K⁺ currents are shown in Fig. 5J and K, where the Ca²⁺-dependent and Ca²⁺-independent components contribute roughly equally to the overall hypoxia-sensitive K⁺ current. The mean hypoxia-sensitive K⁺ current was 121.2 ± 13.2 pA (n= 6), of which the Ca²⁺-independent component was 60.2 ± 8.3 pA (n= 6), and the Ca²⁺-dependent component 63.41 ± 5.8 pA (n= 6) for a voltage step to +30 mV. The I-V curves of the means ±s.e.m. of the normalized currents elicited in six cells with or without hypoxia and Cd²⁺ is shown in Fig. 5K. These findings suggest that both Ca²⁺-independent and Ca²⁺-dependent components of K⁺ current contribute to hypoxia chemosensitivity in rabbit neonatal NEB cells.

Figure 5. Effects of hypoxia on K⁺ currents recorded in Cd²⁺-free and Cd²⁺-containing medium.

A, outward K⁺ current evoked by depolarizing steps from -60 to +30 mV in control Krebs solution. B, outward current evoked by same voltage steps as in A was reduced by hypoxia. C, washout of the hypoxic solution caused a recovery of the outward K⁺ current. D, after perfusion with 100 μm Cd²⁺ outward currents evoked by same voltage steps as in A decreased. E, the remaining I_K(V) current, after blockade of I_K(Ca) by Cd²⁺, was also reduced by hypoxic solution. F, washout of Cd²⁺+ hypoxic solution caused a recovery of the major outward K⁺ current. G, total hypoxia-sensitive K⁺ current was obtained by subtracting currents in B from those in A. H, the hypoxia-sensitive I_K(V) current was obtained by subtracting currents in E from those in D. I, the hypoxia-sensitive I_K(Ca) current was obtained by subtracting currents in H from those in G. J, I-V relationships obtained from the total hypoxia-sensitive K⁺ current (▵), hypoxia-sensitive I_K(V) current (□), and hypoxia-sensitive I_K(Ca) current (^). All recordings were obtained from the same NEB cell. K, I-V relationships of the means ±s.e.m. of the normalized currents (I_step/I_max) elicited in 6 NEB cells in total hypoxia-sensitive K⁺ current (▵), hypoxia-sensitive I_K(V) current (□), and hypoxia-sensitive I_K(Ca) current (^).

Effects of DPI and H₂O₂ on outward K⁺ current

Previous studies on cultured NEB cells from rabbit fetal lungs have shown that K⁺ currents were suppressed by DPI (0.4–4 μm), an inhibitor of NADPH oxidase (Youngson et al. 1993), and increased by H₂O₂ (Wang et al. 1996). We tested the effects of these agents on NEB cells in neonatal rabbit lung slices. Outward K⁺ currents and the I-V relationship obtained from NEB cells before and during bath application of 1 μm DPI are shown in Fig. 6A, B and E. Current amplitudes at a test potential of +30 mV before and during DPI application were 421.4 ± 48.0 and 243.0 ± 67.3 pA, respectively (P < 0.01, n= 6); this corresponds to a significant K⁺ current suppression (∼42 %) by DPI (Fig. 6A and B). Furthermore, after treatment with 1 μm DPI, K⁺ currents in NEB cells no longer responded to hypoxia stimulus (Fig. 6C and E). The effect of 1 μm DPI was reversible (Fig. 6D). Application of 0.25 mM H₂O₂ increased K⁺ currents over the voltage range -3 to +30 mV, and enhanced K⁺ current by 37.5 % at a test potential of +30 mV (Fig. 7). The currents were significantly increased between test potentials of -10 and +30 mV (P < 0.05, n= 4, Fig. 7D). The effect of H₂O₂ was also reversible (Fig. 7C). Thus, as for cultured NEB cells (Wang et al. 1996), H₂O₂-sensitive K⁺ channels are functional in NEB cells in the neonatal rabbit lung slice preparation.

Figure 6. Effect of DPI on K⁺ current.

A, outward K⁺ current evoked by depolarizing steps from -60 to +30 mV in control Krebs solution. B, outward current was reduced after bath application of 1 μm DPI. C, K⁺ current in B was not altered by further exposure to both DPI and hypoxia. D, washout of the DPI + hypoxic solution caused a recovery of the outward K⁺ current. E, I-V relationships obtained from six cells under control conditions (•), after perfusing the cells with 1 μm DPI (^) and after perfusing with 1 μm DPI + hypoxic solution (□). Recorded currents in DPI were significantly (*P < 0.05, n= 6) reduced relative to control; there was no significant difference between currents recorded in DPI and DPI + hypoxia. Data are shown as means ±s.e.m.

Figure 7. Effects of H₂O₂ on K⁺ current.

A, outward K⁺ current evoked by depolarizing steps from -60 to +30 mV in control Krebs solution. B, the K⁺ outward currents were increased significantly (*P < 0.05, n= 4) after perfusing with 0.25 mM H₂O₂. C, washout of the H₂O₂ caused a recovery of the outward K⁺ current. D, I-V relationships for a sample of 4 cells were plotted under control conditions (•) and after perfusing with 0.25 mM H₂O₂ (^). Data are shown as means ±s.e.m.

Effect of hypoxia on inward currents

The inhibition of the Ca²⁺-dependent K⁺ currents by hypoxia could be due to a direct effect of this stimulus on Ca²⁺-dependent K⁺ channels or a direct block of Ca²⁺ channels. As previously reported in fetal NEB cell cultures (Youngson et al. 1993), hypoxia had no effect on the Ca²⁺ channel currents studied under voltage clamp during blockage of K⁺ currents. Currents were evoked during steps from a holding potential of -60 mV to test potentials between -70 and +40 mV for 500 ms. The amplitudes of the calcium channel current at 0 mV ranged between 65 and 117 pA when 10 mM Ba²⁺ was used as the charge carrier, with a mean of 85.02 ± 9.5 pA (n= 5). Perfusion of the cells with hypoxic solution had no significant effects on the amplitude of these currents in all cells tested (n= 5). Exposure to 100 μm Cd²⁺ reduced the Ba²⁺ (Ca²⁺) currents in our preparation by 90 % (not shown). An example of the current traces at three different test potentials, before and during perfusion of hypoxic solution are shown in Fig. 8A and B. The corresponding I-V curve is shown in Fig. 8C.

Figure 8. Lack of effect of hypoxia on Ca²⁺ channel currents in NEB cells.

A, example of Ca²⁺ channel currents obtained using 10 mM Ba²⁺ as charge carrier in a cell at the indicated potentials in control Krebs solution. B, currents obtained after hypoxic solution was applied to the same cell as in A. C, I-V relationships for Ba²⁺ current obtained from this cell in control conditions (•) and in hypoxic solution (^). Holding potential was -70 mV.

Among all the cells tested, 11 % (7 in 63) exhibited TTX- sensitive currents indicating the presence of voltage-gated Na⁺ channels (Fig. 8). Inward Na⁺ currents were found using test potential protocols similar to those used for the study of outward K⁺ currents (Fig. 9A). The mean peak current measured at -10 mV in seven cells exhibiting I_Na was 42.1 ± 5.4 pA. Application of 0.1 μm TTX completely blocked these inward currents during recordings from the same cell (Fig. 9B). Hypoxic solution failed to modify the amplitude of Na⁺ current (Fig. 9C and D). The I-V curve shows that I_Na had an activation threshold around -50 mV and reached maximal amplitude at -10 mV (Fig. 9E). The mean densities of the K⁺, Na⁺ and Ca²⁺ channel currents in NEB cells of the neonatal rabbit lung slice preparation were 91.3 ± 10.3 pA pF⁻¹ (n= 21), 18.0 ± 3.9 pA pF⁻¹ (n= 7), and 36.7 ± 6.6 pA pF⁻¹ (n= 5), respectively, for a voltage step to +30 mV. Taken together, these findings suggest that hypoxia selectively modulates I_K(V) and I_K(Ca).

Figure 9. Lack of effect of hypoxia on Na⁺ current in NEB cells.

A, Na⁺ inward current (downward deflection) was evoked by voltage steps from -60 to +30 mV in control Krebs solution. B, 0.1 μm TTX completely blocked Na⁺ currents in the same cell as in A. C, Na⁺ current recorded in control conditions at a test potential of 0 mV. D, hypoxic solution failed to modify peak Na⁺ current in the same cell as shown in C. E, I-V relationship of peak Na⁺ current (•) elicited during a step from -60 to +30 mV in 6 cells. F, comparison of K⁺, Na⁺ and Ca²⁺ (Ba²⁺) current density in NEB cells, obtained by dividing peak currents by whole-cell capacitance. For K⁺ currents, a voltage step to +30 mV was used.

DISCUSSION

The electrophysiological properties of pulmonary NEB, presumed airway chemoreceptors, have been partially characterized using a fetal rabbit lung NEB cell culture model (Youngson et al. 1993, 1994; Wang et al. 1996). The present study is the first report of whole-cell patch-clamp recordings from intact NEB cells in situ utilizing a neonatal lung slice preparation. Using this model, we show that depolarization of NEB cells activated voltage-dependent K⁺ channels, at least some of which are O₂ sensitive. An O₂-sensitive K⁺ current was previously demonstrated in cultures of NEB cells isolated from fetal rabbit lung (Youngson et al. 1993). Inward currents, carried by Na⁺ and Ca²⁺ ions, did not respond to hypoxia in either NEB cell cultures (Youngson et al. 1993) or lung slice preparations (present study). With regard to these voltage-activated currents, some differences in magnitude were observed when comparing data obtained from NEB cells in culture and in the lung slices. One explanation could be the differences in cell size. For example, the capacitance of fetal NEB cells maintained in culture for a few days was 6.1 pF (Youngson et al. 1993; cell diameter ∼10 μm) compared with 2.42 pF (cell diameter ∼8.7 μm) in the neonatal lung slices. This difference in cell size could explain the more prominent inward and outward currents recorded from NEB cultures (Youngson et al. 1993, 1994) compared with the lung slices. The data presented in Fig. 4 indicate the presence of at least two types of voltage-dependent K⁺ channel. One of them, I_K(Ca), exhibits properties of a Ca²⁺-dependent K⁺ channel that is maximally activated at a voltage where Ca²⁺ influx through voltage-gated Ca²⁺ channels occurs and is sensitive to Ca²⁺ channel blockers such as Cd²⁺ (Lopez-Lopez et al. 1997). The K⁺ current remaining in the presence of 100 μm Cd²⁺, I_K(V), is similar in its kinetic and pharmacological properties to other delayed rectifier K⁺ channels found in neurons from different preparations (Segal & Lewis, 1984; Cobbert et al. 1989; Fu et al. 1996) and also has kinetics resembling the kinetics of the KV3.3a K⁺ channel expressed in Xenopus oocytes (Vega-Saenz de Miera & Rudy, 1992). Low P_O2 inhibits both Ca²⁺-independent K⁺ current and Ca²⁺-dependent K⁺ current in NEB cells. Both of these hypoxia-sensitive currents were activated at a test potential around -40 mV. The hypoxia-sensitive I_K(Ca) current represented ∼55 % of the total hypoxia-sensitive K⁺ current, whereas I_K(V) accounted for the remaining 45 %. Previous data from cultured fetal NEB cells argued against the view that the effect of hypoxia is mediated solely by the Ca²⁺-activated K⁺ current (Youngson et al. 1994). In other studies, I_K(Ca) was selectively inhibited by hypoxia in glomus cells from the carotid body of neonatal (Peers, 1990) and adult (Lopez-Lopez et al. 1997) rats. On the other hand, I_K(V) appears to be selectively inhibited by hypoxia in rat and canine pulmonary artery myocytes (Yuan et al. 1993; Post et al. 1995) and in PC12 cells (Zhu et al. 1996).

We have reported oxygen-sensitive K⁺ currents in cultures of rabbit fetal NEB cells (Youngson et al. 1993, 1994; Wang et al. 1996) and several groups have examined the O₂-sensing properties of carotid body glomus cells (Delpiano & Hescheler, 1989; Lopez-Lopez et al. 1989; Cross et al. 1990; Peers, 1990; Stea & Nurse, 1991; Donnelly, 1995; Montoro et al. 1996; Buckler, 1997; Lopez-Lopez et al. 1997). These studies led to the proposal for a mechanism of O₂ sensing in which modulation of ion channels plays a central role (i.e. the so-called ‘membrane model’). According to this model, hypoxia-induced closure of K⁺ channels of glomus or NEB cells leads to depolarization, opening of voltage-gated Ca²⁺ channels and Ca²⁺ influx triggering neurotransmitter release (Gonzalez et al. 1994). It has been shown previously that these O₂-sensitive K⁺ channels can be directly regulated by P_O2 in excised membrane patches (Ganfornina & Lopez-Barneo, 1991), and there is some evidence that O₂ sensing may occur at the level of b-type cytochrome associated with NADPH oxidase, which generates oxygen radicals (Cross et al. 1990; Wang et al. 1996). A by-product of this reaction, H₂O₂, has been proposed to affect the redox status of glutathione, which in turn modulates K⁺ channel activity (Acker et al. 1992; Weir et al. 1994). It has been suggested that hypoxia inhibits H₂O₂ production, resulting in increased levels of reduced glutathione which in turn leads to channel closure (Acker et al. 1992). This proposal relies in part on the observation that DPI, an inhibitor of NADPH oxidase, blocks the hypoxia-induced increase in afferent chemosensory nerve discharge (Cross et al. 1990). We have previously reported that mRNAs for both the H₂O₂-sensitive voltage-gated K⁺ channel (K_H2O2 channel) subunit KV3.3a and membrane components of NADPH oxidase (gp^91phox and p^22phox) are co-expressed in NEB cells of fetal rabbit and neonatal human lung (Wang et al. 1996). In the present study, we observed augmentation of K⁺ current recorded from neonatal rabbit NEB cells in situ after direct exposure to H₂O₂, supporting the view that membrane K_H2O2 channels may be regulated by reactive oxygen intermediates produced by the putative O₂ sensor, i.e. NADPH oxidase (Youngson et al. 1993; Wang et al. 1996). Other reports indicate that at high concentrations, DPI (10 μm) is a non-selective ionic channel blocker, and the ability of DPI to block calcium currents can explain its inhibition of hypoxic pulmonary vasoconstriction (Wyatt et al. 1994). Further studies are required to determine the precise effects of DPI on the O₂-sensing mechanism in NEB cells and other O₂-sensing cells.

The inward Ca²⁺ currents observed in the present study showed voltage-dependent and slowly inactivating properties characteristic of L-type Ca²⁺ channels (Tsien et al. 1988) and were blocked by 100 μm Cd²⁺. Although O₂-sensitive Ca²⁺ channels have been reported in pulmonary artery myocytes (Franco-Obergon et al. 1995) and in CB glomus cells from adult rabbits (Urena et al. 1994; Montoro et al. 1996; Lopez-Lopez et al. 1997), hypoxia had no effect on calcium channel currents of NEB cells in our lung slice preparation. In addition, only about 11 % of neonatal NEB cells in lung slices possessed TTX-sensitive Na⁺ currents that did not respond to hypoxia. It remains to be determined whether adult rabbit NEB cells express a higher density of Na⁺ channels, as appears to be the case for their chemoreceptor counterparts (i.e. glomus cells) in the carotid body (Gonzalez et al. 1994).

The lung slice preparation reported in this study offers several advantages for electrophysiological studies on pulmonary NEBs. Although NEBs are widely scattered within a sponge-like lung parenchyma and constitute a minority population of airway epithelial cells, they can be identified in this preparation by means of supravital staining with Neutral Red. This allows specific targeting of intact NEBs for patch-clamp studies in their natural environment. In addition, since the lung slice preparation is much easier to prepare than dissociated cell cultures, this technique should facilitate comparative studies of NEB responses during different stages of development as well as among different species. It should be also feasible to study neural discharges elicited by stimulation of NEBs by hypoxia and other stimuli as has been reported for CBs using intact organ preparations (Donnelly, 1995). Such studies should help to define more fully the postulated function of NEBs as O₂-sensitive airway chemoreceptors.