Attenuated outward potassium currents in carotid body glomus cells of heart failure rabbit: involvement of nitric oxide

Abstract

It has been shown that peripheral chemoreceptor sensitivity is enhanced in both clinical and experimental heart failure (HF) and that impairment of nitric oxide (NO) production contributes to this enhancement. In order to understand the cellular mechanisms associated with the alterations of chemoreceptor function and the actions of NO in the carotid body (CB), we compared the outward K⁺ currents (I_K) of glomus cells in sham rabbits with that in HF rabbits and monitored the effects of NO on these currents. I_k was measured in glomus cells using conventional and perforated whole-cell configurations. I_K was attenuated in glomus cells of HF rabbits, and their resting membrane potentials (−34.7 ± 1.0 mV) were depolarized as compared with those in sham rabbits (−47.2 ± 1.9 mV). The selective Ca²⁺-dependent K⁺ channel (K_Ca) blocker iberiotoxin (IbTx, 100 nm) reduced I_K in glomus cells from sham rabbits, but had no effect on I_K from HF rabbits. In perforated whole-cell mode, the NO donor SNAP (100 μm) increased I_K in glomus cells from HF rabbits to a greater extent than that in sham rabbits (P < 0.01), and IbTx inhibited the effects of SNAP. However, in conventional whole-cell mode, SNAP had no effect. N^ω-nitro-L-arginine (L-NNA, NO synthase inhibitor) decreased I_k in sham rabbits but not in HF rabbits. The guanylate cyclase inhibitor 1H-[1,2,4]oxadiazole[4,3-a]quinoxalin-1-one (ODQ) inhibited the effect of SNAP on I_k. These results demonstrate that I_K is reduced in CB glomus cells from HF rabbits. This effect is due mainly to the suppression of K_Ca channel activity caused by decreased availability of NO. In addition, intracellular cGMP is necessary for the K_Ca channel modulation by NO.

Activation of the sympathetic nervous system is one of the most profound and reproducible changes in chronic heart failure (HF) (Floras, 1993; Zucker et al. 1995; Esler et al. 1997). Previous studies have shown that peripheral chemoreceptor activation is an excitatory input that results in increased sympathetic outflow and blood pressure (Marshall, 1994). Studies from our laboratory have shown that in a rabbit model of pacing-induced HF, peripheral chemoreceptor activity is enhanced even under normoxic conditions. This enhanced sensitivity of the peripheral chemoreflex contributes, at least in part, to the sympathetic activation in this rabbit model of HF, because inhibition of peripheral chemoreceptor activity decreased renal sympathetic nerve activity (RSNA) in HF, but not in sham rabbits (Sun et al. 1999a). However, the cellular and molecular mechanisms involved in enhancing peripheral chemoreceptor sensitivity during HF still remain unclear.

The carotid body (CB) is the primary sensor site of the peripheral chemoreflex, and is composed of neurotransmitter-enriched glomus cells and glial-like sustentacular cells among other neuronal structures. Currently it is believed that glomus cells, which lie in synaptic apposition with the afferent nerve ending, are the initial sites of sensory transduction in the CB. The glomus cell membrane contains many voltage-gated ion channels. The suppression of outward K⁺ channels is proposed to contribute sequentially to the initial depolarization, activation of voltage-gated Ca²⁺ channels, release of neurotransmitter, and activation of discharge in the carotid sinus nerve (Gonzalez et al. 1994; Prabhakar, 1994).

NO is a gas molecule that functions as a transmitter in both the peripheral and central nervous systems. It is synthesized during the catalytic conversion of arginine to citrulline by the enzyme nitric oxide synthase (NOS) (Snyder, 1992). The extensive plexus of nerve fibres and vessels surrounding glomus cells contains neuronal NOS (nNOS) and endothelial NOS (eNOS) (Prabhakar et al. 1993; Wang et al. 1993), but NOS is not found in detectable levels in glomus and type II cells. However, these other neural and vascular cells may produce paracrine effects of their NO production on neighbouring glomus cells.

Our previous study has shown that a down-regulation of NO synthesis in the CB in HF rabbits contributes to the enhancement of CB chemoreceptor sensitivity (Sun et al. 1999b). We hypothesize that NO affects I_K, which contributes to the altered peripheral chemoreceptor function in HF. To test this hypothesis, we compared I_K of CB glomus cells in sham and pacing-induced HF rabbits and examined the effects of NO on I_K. Our results show that I_K was attenuated in CB glomus cells from HF rabbits and that down-regulation of NO attenuated I_K via the inhibition of Ca²⁺-dependent K⁺ (K_Ca) channels. Furthermore, the ability of NO to activate IK_Ca in glomus cells appears to be mediated by a cGMP-dependent pathway. Preliminary results from this study have been published in an abstract form (Li & Schultz, 2003).

Methods

Pacemaker implant and production of HF

All experiments were carried out on male New Zealand White rabbits weighing 2.5–3.5 kg. Experiments were approved by the University of Nebraska Medical Center Institutional Animal Care and Use Committee and were carried out in accordance with the National Institutes of Health and the American Physiological Society's Guides for the Care and Use of Laboratory Animals.

Rabbits were anaesthetized with a cocktail consisting of 1.2 mg acepromazine, 5.9 mg xylazine and 58.8 mg ketamine, given as an i.m. injection. Using sterile technique, a left thoracotomy was performed as previously described (Sun et al. 1999a). The pericardium was opened and wire loop electrodes were attached to the left ventricle for pacing. Two sonomicrometer crystals (Sonometrics Corp., London, ON, Canada) were attached to opposing walls of the lateral left ventricle for measuring external diameter. All leads exited the chest between the 3rd and 4th ribs. The chest was closed in layers and evacuated. Rabbits were placed on an antibiotic regimen consisting of 5 mg kg⁻¹ Baytril i.m. for 5 days. After the rabbits recovered from the thoracotomy (about 2 weeks), baseline left-ventricular end-systolic and end-diastolic external diameter (D), fractional shortening, and shortening velocity (dD/dt_max) were measured by sonomicrometry (Triton Technology Inc., San Diego, CA, USA). An arterial blood gas sample (0.25 ml) was taken by needle puncture of an ear artery periodically to monitor arterial blood gases. The pacing was started at 320 beats min⁻¹, which was held for 7 days, and then the rate was gradually increased to 380 beats min⁻¹, with an increment of 20 beats min⁻¹ each week. The pacemaker was of our design, with its output usually being set at 4–5 V and 0.5 ms. Sonograms and blood gas (ABL5, Radiometer, Copenhagen) were done weekly, with the rabbits sitting quietly in a Plexiglas box and with the pacemaker turned off for at least 30 min before recordings were made. Sham-operated animals underwent a similar period of sonographic measurements with the pacemaker turned off. Rabbits with >40% reductions in dD/dt_max and shortening fraction were considered in HF (generally after 3–4 weeks). Any rabbit exhibiting abnormal arterial blood gases (P_a,O2 < 85 mmHg; 45 mmHg <P_a,CO2 < 30 mmHg) were excluded from study (Sun et al. 1999a).

Isolation of CB glomus cells

The carotid bifurcations on both sides were removed surgically from sham or HF rabbits anaesthetized with the cocktail described above. The rabbits were then killed with an intravenous injection of 150 mg kg⁻¹ sodium pentobarbital. The removed tissue was placed in ice-cold Ca²⁺/Mg²⁺-free solution (mm: NaCl, 140; KCl, 5; Hepes 10; glucose 5; pH 7.2), and then minced with microscissors. The CBs were then subjected to a two-step enzymatic digestion protocol. Each step lasted for 30 min at 37°C. The enzymatic solution for the first step of digestion contained trypsin (type II, 2 mg ml⁻¹) and collagenase (type IV, 2 mg ml⁻¹); and that for the second step contained collagenase (4 mg ml⁻¹) and bovine serum albumin (BSA, 5 mg ml⁻¹). Other solutes in the solutions used for enzymatic digestion were the same as the above Ca²⁺/Mg²⁺-free solution. The digested tissue fragments were gently triturated for 1 min every 10 min during the process of digestion. CB cells were obtained after centrifugation of the digested tissue at 150 g for 5 min. The isolated cells were then resuspended in culture medium and plated onto culture wells. The culture medium consisted of a 50/50 mixture of Delbecco's modified Eagle's medium (DMEM) and Ham's F12 medium supplemented with antibiotics and 10% fetal bovine serum. The CB cells including glomus cells, other neural cells and vascular cells were cocultured at 37°C in a humidified atmosphere of 95% air–5% CO₂ (Zhong et al. 1997) and studied within 24 h of dissociation.

Identification of glomus cells

Short-term cultured cells freshly isolated from CB in sham and HF rabbits contain several cell types, including glomus and type II cells. It has been shown that rabbit glomus cells are excitable cells and they exhibit Ca²⁺, K⁺ and Na⁺ currents, but type II cells are non-excitable and exhibit only small outward K⁺ currents. Therefore, at the beginning of each experiment, cells were superfused with an extracellular solution containing Na⁺. Those cells that exhibited Na⁺ current were considered to be glomus cells (Overholt & Prabhakar, 1997). Once the presence of Na⁺ current was confirmed, the extracellular solution was changed to solution containing 0.5 μm TTX to record I_K (see below).

Recording of whole cell K⁺ currents and resting membrane potential

K⁺ currents were measured in the whole-cell configuration of the patch-clamp technique (Hamill et al. 1981) using a Warner PC-501A patch-clamp amplifier (Warner Instrument Corp. Hamden, CT, USA), which allowed the same cell to be recorded in either the voltage-clamp or current-clamp mode. Patch pipettes had resistances of 4–6 MΩ when filled with (mm) 105 potassium aspartate, 20 KCl, 1 CaCl₂, 10 EGTA, 5 Mg-ATP, 10 Hepes and 25 glucose, pH 7.2. In the perforated whole-cell configuration, nystatin was added to the patch-pipette solution at a final concentration of 300 μg ml⁻¹ immediately before recording. The extracellular solution had the following composition (mm): 140 NaCl, 5.4 KCl, 2.5 CaCl₂, 0.5 MgCl₂, 5.5 Hepes, 11 glucose, 10 sucrose, pH 7.4. Na⁺ channels were blocked by TTX (0.5 μm).

Cell membrane capacitance (C_m) was determined by integrating the capacitive current evoked by a 5 mV voltage step and dividing the resulting charge by the voltage step. Mean series resistances, as determined from the decay of the capacitive transient, were 7.4 ± 0.5 MΩ (conventional whole-cell mode) and 13.2 ± 0.7 MΩ (perforated whole-cell mode), and were compensated electronically by 80–90%. Current traces were sampled at 10 kHz and filtered at 5 kHz. Holding potential was −80 mV. Current–voltage (I–V) relations were elicited by 400 ms test pulses from −80 mV to +80 mV applied in 10 mV increments (5 s between steps). Peak currents were measured for each test potential and were plotted against the corresponding test potential.

To study steady-state inactivation, we used two-step pulse protocols from a holding potential of −80 mV. First, a 10 s prepulse was applied at various potentials ranging from −110 to −10 mV (10 mV step) followed by a 400 ms test pulse to +20 mV. Inactivation curves were constructed from currents at the corresponding prepulse voltages and fit to the Boltzmann function:

where I is the peak current at each prepulse potential V; I_max is the maximum current recorded; V_1/2 is the prepulse potential where I is half of the maximum; K₁ is the slope factor at V_1/2; and S is the non-inactivating component. To measure the resting membrane potentials, the amplifier was switched from voltage-clamp mode to current-clamp mode with zero current being delivered. pCLAMP 8.1 programs (Axon Instruments Inc., Union City, CA, USA) were used for data acquisition and analysis. All experiments were done at 22°C and the external solution equilibrated with air (P_O2 of about 150 Torr). In order to measure the effect of each treatment, six to ten glomus cells from five to six rabbits were used in each group and the effects of one or two treatments were measured in one cell.

Test chemicals were all delivered to the recording bath through a fast-flow apparatus consisting of a linear array of borosilicate glass tubes (Overholt et al. 1995). Chemicals were added into syringe reservoirs, which contained extracellular solution and connected to the fast-flow apparatus. All chemicals were obtained from Sigma (St. Louis, MO, USA).

Data analysis

All data are presented as means ±s.e.m. Statistical significance was determined by Student's paired or unpaired t test for haemodynamic parameters, or a two-way ANOVA, with a Bonferroni procedure for post hoc comparisons of I_K. Statistical significance was accepted when P < 0.05.

Results

Induction of heart failure

Rapid left ventricular pacing induced HF by the 3rd or 4th week of pacing. In paced rabbits, LV dD/dt_max and LV shortening fraction were reduced after 3 or 4 weeks of pacing, compared with prepaced baselines (P < 0.05, Table 1). There was no significant change in the LV dD/dt_max and LV shortening fraction between the baseline and the 3rd or 4th week in sham rabbits (Table 1).

Table 1.

Cardiac diameters and contractility in sham and HF rabbits

	Sham (n= 15) Baseline	3–4 week	HF (n= 16) Baseline	3–4 week
ESD (% of control)	100 ± 0.0	100.2 ± 1.4	100 ± 0.0	113.4 ± 2.4
EDD (% of control)	100 ± 0.0	100.4 ± 0.9	100 ± 0.0	110.3 ± 1.8
dD/dtmax (mm s−1)	−11.2 ± 0.9	−11.0 ± 1.1	−10.9 ± 1.3	−4.3 ± 0.8*#
% shortening	10.3 ± 0.9	10.1 ± 0.8	10.1 ± 0.8	4.6 ± 0.5*#

Data are means ±s.e.m. ESD, left ventricular end-systolic diameter; EDD, left ventricular end-diastolic diameter; dD/dt_max, 1st derivative of change in diameter; %shortening = (EDD – ESD)/EDD × 100%.

^*P < 0.05 versus baseline;

^#P < 0.05 versus sham rabbits at the same period.

Electrophysiological properties of CB glomus cells from sham and HF rabbits

The resting membrane potential of glomus cells in HF rabbits (−34.7 ± 1.0 mV, n= 21 cells from 20 rabbits) was depolarized (P < 0.05), compared with that in sham rabbits (−47.2 ± 1.9 mV, n= 18 cells from 18 rabbits). There was no significant difference in the whole-cell capacitance between sham (2.59 ± 0.18 pF) and HF rabbits (3.23 ± 0.35 pF). Thus, absolute currents measured in our recordings reflected current density in the cells and were not normalized.

Figure 1Aa–c illustrates typical I_K recordings and I–V curves obtained from glomus cells from sham and HF rabbits, with voltage steps from −80 mV under conventional whole-cell patch conditions. As illustrated in current traces in Fig. 1Aa and b, I_K was attenuated in HF rabbits, compared with those in sham rabbits. Figure 1Ac shows the peak current–voltage relationships evoked by 400 ms depolarizing test pulses from −80 mV to +80 mV in 10 mV increments. From this curve, the attenuated I_K in HF rabbits began near −30 mV and appeared to be voltage dependent.

Figure 1. Outward K⁺ currents of carotid body glomus cells from sham and HF rabbits.

A, outward K⁺ currents under conventional patch condition evoked in glomus cells from sham (Aa) and HF (Ab) rabbits by 400 ms depolarizing pulses from a holding potential of –80 mV to test pulses over the range of −80 to +80 mV in 10 mV increments. Ac, peak I–V relationships obtained in glomus cells from sham (n= 10 cells from 6 rabbits) and HF rabbits (n= 10 cells from 6 rabbits). B, outward K⁺ currents under perforated-patch condition evoked in glomus cells from sham (Ba) and HF (Bb) rabbits using the same pulse protocol above. Bc, peak I–V relationships obtained in glomus cells from sham (n= 8 cells from 6 rabbits) and HF rabbits (n= 9 cells from 6 rabbits). Data are means ±s.e.m. in each group. *P < 0.05versus sham. •, sham; ○, HF.

A potential limitation to the conventional whole-cell patch-recording technique used above is the possibility of the loss into the recording pipette of intracellular constituents that may influence K⁺ channel function. To assess this possibility we performed experiments using the perforated-patch technique, which restricts movement of intracellular contents into the pipette. From Fig. 1Ba–c it is evident that I_K under perforated-patch conditions was larger than that under conventional patch conditions in glomus cells from sham rabbits, but I_K was similarly decreased in glomus cells from HF rabbits with either recording technique. These results suggest that the activity of outward K⁺ channels in the glomus cells was influenced by an intracellular factor and this factor was absent in glomus cells from HF rabbits. We performed additional experiments described below to assess the role of cGMP as the candidate factor. Because I_K was affected by the patch conditions, the remaining experiments were performed under perforated-patch conditions except as specifically mentioned.

Steady-state inactivation was evaluated with a two-pulse protocol (Fig. 2A). The inactivating prepulse was 10 s in duration in order to reach a steady state of inactivation. The interpulse interval was 60 s to allow full recovery of inactivation. Figure 2B illustrates the mean inactivation curves obtained in eight cells from sham and HF rabbits. The mean inactivation curve exhibited a threshold at about −100 mV and failed to reach full inactivation at values greater than 0 mV. There was no difference in the inactivation curves between sham and HF rabbits (sham, V_1/2=−45.6 mV, K₁= 13.56, S= 0.16; HF, V_1/2=−47.5 mV, K₁=−15.44, S= 0.18).

Figure 2. Steady-state inactivation of outward K⁺ currents in carotid body glomus cells from sham and HF rabbits under perforated-patch conditions.

A, records obtained with the two-step pulse protocol (inset) in glomus cells. During the interpulse interval, the membrane was held at −80 mV for 60 s. B, the steady-state inactivation curves were plotted by normalizing the peak current evoked with the test pulse (+20 mV) against the potential of the prepulse. Data are means ±s.e.m., n= 8 cells from 6 rabbits in each group. Lines represent the best fit of the data to Boltzmann functions (see text). •, sham; ○, HF.

Pharmacological characterization of I_K was carried out by the application of 1 mm 4-AP plus 1 mm TEA to the external bath. As shown in Fig. 3, 4-AP plus TEA markedly decreased I_K to the same level in glomus cells from sham and HF rabbits (P < 0.05).

Figure 3. Effects of 4-AP plus TEA on outward K⁺ currents of carotid body glomus cells from sham and HF rabbits under perforated-patch conditions.

A, outward K⁺ currents of glomus cell from sham rabbit evoked by 400 ms depolarizing pulses from holding potential −80 mV to test pulses (−80 to +80 mV, 10 mV steps) in control conditions (Aa) and with 1 mm 4-AP plus 1 mm TEA (Ab). Ac, peak I–V relationships (n= 6 cells from 5 rabbits) obtained before and after administration of 1 mm 4-AP plus 1 mm TEA. B, outward K⁺ currents of glomus cell from HF rabbit evoked by same pulse protocol in control conditions (Ba) and with 1 mm 4-AP plus 1 mm TEA (Bb, n= 7 cells from 5 rabbits). Bc, peak I–V relationships obtained before and after administration of 1 mm 4-AP plus 1 mm TEA. *P < 0.05versus control. •, control; ○, 4-AP + TEA.

Effects of iberioxin (IbTx) on I_K of CB glomus cells from sham and HF rabbits

It has been shown that an IK_Ca component exists in rabbit carotid body glomus cells (Lopez-Lopez et al. 1993). Therefore, IK_Ca in glomus cells was further characterized by using the selective IK_Ca blocker, iberiotoxin (IbTx). In glomus cells from sham rabbits, 100 nm IbTx (saturated concentration) decreased I_K (P < 0.05, Fig. 4A and C). However IbTx had no effect on I_K in glomus cells from HF rabbits (Fig. 4B and D). These results suggested that the decrease of I_K in glomus cells from HF rabbits was due to suppression of the K_Ca channel.

Figure 4. Effects of iberioxin (IbTx, Ca²⁺-dependent K⁺ channel blocker) on outward K⁺ currents of carotid body glomus cells from sham and HF rabbits under perforated-patch conditions.

A (sham rabbit) and B (HF rabbit), outward K⁺ currents elicited in the glomus cells by 400 ms test pulse from −80 to +70 mV before and after administration of 100 nm IbTx. C (sham rabbit, n= 7 cells from 6 rabbits) and D (HF rabbit, n= 8 cells from 6 rabbits), peak I–V relationships obtained from glomus cells before and after administration of IbTx. *P < 0.05versus control. •, control; ○, IbTx.

Effects of NO on I_K of CB glomus cells from sham and HF rabbits

The modulation of I_K by NO was investigated by application of the NO donor S-nitroso-N-acetylpenicillamine (SNAP) and a NOS inhibitor, L-NNA, to the recording bath. As shown in Fig. 5, 100 μm SNAP significantly increased I_K of glomus cells from HF rabbits and IbTx completely inhibited this effect (P < 0.05, Fig. 5B and D). Similarly, SNAP also enhanced the I_K of glomus cells from sham rabbits (P < 0.05, Fig. 5A and C). At the same time, following the changes of I_K induced by SNAP and IbTx, corresponding changes were induced in resting membrane potential. In glomus cells from both sham and HF rabbits, SNAP increased the resting membrane potential and IbTx inhibited this effect of SNAP (Fig. 6). However, in conventional whole-cell mode, SNAP did not change I_K (Fig. 7) or membrane potential in either group (data not shown).

Figure 5. Effects of SNAP (NO donor) on outward K⁺ currents of carotid body glomus cells from sham and HF rabbits under perforated-patch conditions.

A (sham rabbit) and B (HF rabbit), outward K⁺ currents elicited in the glomus cells by 400 ms test pulse from −80 to +70 mV in control condition, with 100 μm SNAP alone and with 100 μm SNAP plus 100 nm IbTx. C (sham rabbit, n= 8 cells from 6 rabbits) and D (HF rabbit, n= 8 cells from 6 rabbits), peak I–V relationships obtained from glomus cells in control condition, with 100 μm SNAP alone and with 100 μm SNAP plus 100 nm IbTx. *P < 0.05 versus control; #P < 0.05versus SNAP. •, control; O, SNAP; ▴, SNAP + IbTx.

Figure 6. Effects of SNAP (NO donor) on resting membrane potential in carotid body glomus cells from sham and HF rabbits.

Data are means ±s.e.m., n= 8 cells from 6 rabbits in each group. *P < 0.05 versus control; #P < 0.05 versus SNAP.

Figure 7. Effects of SNAP (NO donor) on outward K⁺ currents of carotid body glomus cells from HF rabbits under conventional patch conditions.

A, outward K⁺ currents elicited in the glomus cells by 400 ms test pulse from −80 to +70 mV before and after administration of 100 μm SNAP. B, peak I–V relationships (n= 7 cells from 5 rabbits) obtained from glomus cells before and after administration of 100 μm SNAP. •, control; ○, SNAP.

Since NOS inhibition by L-NNA can take several minutes, cells in the recording dish were preincubated for 30 min in normal extracellular solution containing 10 μm L-NNA. In sham rabbits, we examined the effect of L-NNA in 13 different glomus cells and its effect was clearly seen in 10 cells. As shown in Fig. 8A and C, L-NNA decreased I_K (P < 0.05). However, L-NNA did not affect I_K in glomus cells from HF rabbits (Fig. 8B and D). These results suggest that down-regulation of NOS activity resulted in the suppression of K_Ca channel activity in glomus cells from HF rabbits, and this effect contributed to the attenuated macroscopic I_K in these cells. In addition, NO appeared to modulate the K_Ca channel via some intracellular factor(s) since SNAP was ineffective in conventional whole-cell recordings.

Figure 8. Effects of N^ω-nitro-l-arginine (L-NNA, NOS inhibitor) on outward K⁺ currents of carotid body glomus cells from sham and HF rabbits under perforated-patch conditions.

A (sham rabbit) and B (HF rabbit), outward K⁺ currents elicited in the glomus cells by 400 ms test pulse from −80 to +70 mV, control and administration of 10 μm L-NNA. C (sham rabbit, n= 10 cells from 6 rabbits) and D (HF rabbit, n= 8 cells from 6 rabbits), peak I–V relationships obtained from glomus cells in control and administration of L-NNA. *P < 0.05versus control. •, control; ○, L-NNA.

Involvement of cGMP in the effect of NO on I_K

To test whether enhancement of I_K by NO is related to activation of guanylate cyclase, we examined the effects of SNAP on I_K in the presence of 1H-[1,2,4]oxadiazole[4,3-a]quinoxalin-1-one (ODQ), a guanylate cyclase inhibitor. As shown in the current traces in Fig. 9A and B, and I–V curves in Fig. 9C and D, ODQ (1 μm) alone reduced I_K in glomus cells from sham rabbits (P < 0.05) but not in those from HF rabbits. More importantly, ODQ inhibited the effects of SNAP on I_K in cells from both sham and HF rabbits (P < 0.05, Fig. 10). We further inferred that if the effects of NO were mediated by cGMP, then 8-bromoguanosine-3′,5′-cyclic monophosphate (8-Br-cGMP), a cGMP analogue, should mimic the effects of SNAP on I_K. Therefore we tested this possibility in seven additional cells from each group. Treatment with 8-Br-cGMP (1 mm) for 20 min increased I_K in glomus cells from sham and HF rabbits by 132 ± 11% and 176 ± 19% of control, respectively (P < 0.01), at a test potential of +70 mV. These results support the idea that NO enhanced I_K via a cGMP-dependent pathway.

Figure 9. Effects of 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ, guanylate cyclase inhibitor) on outward K⁺ currents of carotid body glomus cells from sham and HF rabbits under perforated-patch conditions.

A (sham rabbit) and B (HF rabbit), outward K⁺ currents elicited in the glomus cells by 400 ms test pulse from −80 to +70 mV, before and after administration of 1 μm ODQ. C (sham rabbit, n= 8 cells from 6 rabbits) and D (HF rabbit, n= 7 cells from 6 rabbits), peak I–V relationships obtained from glomus cells before and after administration of ODQ. *P < 0.05versus control. •, control; ○, ODQ.

Figure 10. Effects of SNAP on outward K⁺ currents through a cGMP-dependent mechanism.

A (sham rabbit) and B (HF rabbit), outward K⁺ currents elicited in the glomus cells by 400 ms test pulse from −80 to +70 mV in control condition, with 100 μm SNAP alone and with 100 μm SNAP plus 1 μm ODQ. C (sham rabbit, n= 6 cells from 6 rabbits) and D (HF rabbit, n= 6 cells from 6 rabbits), peak I–V relationships obtained from glomus cells in control condition, with 100 μm SNAP alone and with 100 μm SNAP plus 1 μm ODQ. *P < 0.05versus control; #P < 0.05versus SNAP. •, control; ○, SNAP; ▴, SNAP + ODQ.

Discussion

The results of the present study demonstrate that: (1) I_K is attenuated in CB glomus cells from HF rabbits, and the resting membrane potential of glomus cells in HF rabbits is depolarized as compared with that in sham rabbits; (2) the selective K_Ca channel blocker IbTx decreased I_K in glomus cells from sham rabbits, but had no effect on I_K from HF rabbits; (3) the NO donor SNAP increased I_K in glomus cells from sham and HF rabbits and IbTx inhibited the effect of SNAP; (4) the non-selective NOS inhibitor L-NNA reduced I_K in sham rabbits but not in HF rabbits; (5) the guanylate cyclase inhibitor ODQ inhibited the effects of SNAP on I_K. From these results, we conclude that: (1) the attenuated I_K in glomus cells from HF rabbits is due to suppression of K_Ca channel activity; (2) NO activates K_Ca channels in rabbit glomus cells; (3) the decreased activity of K_Ca channels in glomus cells in HF rabbits is due to decreased NOS activity or bioavailability of NO; and (4) NO modulates the K_Ca channels via intracellular cGMP.

Functionally, K⁺ channels are classified into five families: (1) voltage-gated K⁺ (K_V) channels, (2) K_Ca channels, which are intracellular Ca²⁺-sensitive and also voltage-sensitive, (3) ATP-sensitive K⁺ channels, (4) inward rectifier K⁺ channels, and (5) voltage-insensitive background K⁺ channels (Jan & Jan, 1997; Lesaga & Lazdunski, 2000). These channels are extraordinarily diverse and are expressed in many different cell types. In adult rabbits to the extent known, carotid body glomus cells express K_V channels, K_Ca channels and HERG-like K⁺ channels (inward rectifier K⁺ channel; Lopez-Lopez et al. 1993, 1997; Overholt et al. 2000).

Of these K⁺ channels, which type is likely to account for attenuated I_K in carotid body glomus cells from HF rabbits, and what regulates the change of resting membrane potential observed? The present study showed that IbTx inhibited I_K in sham rabbits but had no effect in HF rabbits. These results suggest that K_Ca channel activity is markedly suppressed in glomus cells from HF rabbits, and the attenuated I_K in HF rabbits is mainly due to decreased activity of these K_Ca channels. On the other hand, our results also do not imply that suppressed K_Ca channel activity exclusively caused the attenuated I_K in HF rabbits. After administration of K_V channel blockers, 4-AP and TEA, I_K was reduced to the same level in both sham and HF rabbits (Fig. 3). However, in sham rabbits IbTx did not reduce the I_K level to that seen in HF rabbits (Fig. 4). These results suggest that other K_V channels are also involved in the attenuated I_K in HF rabbits, but to a lesser extent.

Our results also suggest that the attenuated K_Ca activity contributed to the elevation in resting membrane potential observed in the HF glomus cells. SNAP lowered the resting potential to normal levels in HF cells, and this effect was reversed by IbTx. However, a HERG-like K⁺ channel has been reported in rabbit glomus cells which could also influence the resting membrane potential (Overholt et al. 2000). The contribution of HERG channels to the elevated membrane potential in the HF glomus cells remains to be determined.

Much evidence indicates that NO produced within the CB is an inhibitory modulator of chemoreception (Prabhakar et al. 1993; Chugh et al. 1994; Wang et al. 1994; Alcayaga et al. 1997; Prabhakar, 1999; Sun et al. 1999b; Iturriaga et al. 2000a). The administration of the precursor l-arginine, NO donors (Prabhakar et al. 1993; Chugh et al. 1994; Wang et al. 1994; Iturriaga et al. 2000b), and NO gas (Iturriaga et al. 2000a) to the cat carotid body perfused in vitro reduces the chemosensory response to hypoxia. Non-specific NOS inhibitors, such as -nitro-l-arginine methyl ester (L-NAME), enhance the chemosensory response to hypoxia (Wang et al. 1994) or nicotine and NaCN (Valdes et al. 2003). In a previous study we demonstrated that the NO donor (SNAP) inhibited the baseline discharge of the carotid sinus nerve (CSN) and the chemosensitivity in sham and HF rabbits (Sun et al. 1999b). Our present results confirmed that SNAP enhanced I_K and hyperpolarized the resting membrane potential in glomus cells from sham and HF rabbits, and IbTx completely abolished the effects of SNAP. We have shown that NO production and protein expression of NOS in CB from HF rabbits is lower than that from sham rabbits (Sun et al. 1999b; Zeng et al. 2002). From these data, we propose that decreased NO production contributed to the attenuated the I_K in glomus cells from HF rabbits by preventing the activation of K_Ca channels.

In contrast to our study, Hatton & Peers (1996) found that SNAP did not affect I_K recorded in CB glomus cells. One possibility for this discrepancy is that their study employed the conventional whole-cell patch-clamp technique. The present study also found that SNAP did not change the I_K of glomus cells in the conventional whole-cell mode, but SNAP was effective in the perforated-patch mode. Taken together these results suggest that some intracellular factor(s) that is dialysed in conventional whole-cell recording is necessary for K_Ca channel modulation by SNAP. This notion is supported by Silva & Lewis's study (2002), in which K_Ca channel activity in rat CB glomus cells was enhanced by SNAP in cell-attached mode.

Our results indicate that this intracellular factor is cGMP. NO activates soluble guanylate cyclase, and many of its effects are mediated by stimulation of cGMP production (Snyder, 1992). cGMP is well known to activate K⁺ channels in vascular smooth muscle, and is the major mediator of NO-induced vasodilatation in many vascular beds (Vaandrager & de Jonge, 1996). The NO donor nitroglycerine has been shown to elevate cGMP levels in glomus cells (Wang et al. 1994). In rat CB glomus cells, NO enhances K_Ca channel activity through cGMP-dependent protein kinase G (Silva & Lewis, 2002). However, recently it has been shown that NO inhibits L-type Ca²⁺ channels in rabbit type I cells via a cGMP-independent mechanism (Summers et al. 1999). In rabbit aortic smooth muscle cells (Bolotina et al. 1994) and rat brain (Shin et al. 1997), NO activates K_Ca channels directly. The heterogeneity of these results could be due to the difference of species and tissues. Our results confirmed that the effect of NO on the I_K in CB glomus cells was cGMP-dependent because ODQ inhibited the effects of SNAP on the I_K and 8-Br-cGMP could mimic the effect of SNAP on the I_K in both sham and HF rabbits.

It is known that NOS isoforms (eNOS, nNOS and iNOS) exist in the extensive plexus of nerve fibres and vessels surrounding glomus cells (Wang et al. 1993) but not in detectable levels in CB glomus cells and type II cells. Thus, a reasonable question is how NO production and NOS inhibitors can influence I_K in isolated CB glomus cells. There may be at least two possible explanations for this phenomenon. First, NOS may be present in glomus cells at low levels that are difficult to detect (Prabhakar et al. 1993), but sufficient to influence K⁺ channel activity, or sequestered in mitochondria inaccessible to immunofluorescence antibodies (Alvarez et al. 2003). Secondly, even though glomus cells are dispersed in cell culture, they continue to coexist with other neural and vascular cells of the CB that possess substantial amounts of NOS. These cells may produce sufficient NO to exert a paracrine effect on the neighbouring glomus cells from which we recorded. This notion is supported by other electrophysiological studies in which paracrine effects can be observed from neighbouring cells in the recording media (Chapleau et al. 2001). Although we cannot ascertain the cellular origin of NOS activity in the present study, we observed that L-NNA reduced the I_K in glomus cells from sham rabbits but not from HF rabbits, which demonstrated that lower activation of NOS was responsible for the attenuated I_K in glomus cells from HF rabbits.

Our previous study has shown that the baseline discharge of CB chemoreceptor afferents in the normoxic state and their response to hypoxia are enhanced in pacing-induced HF rabbits as compared with sham rabbits (Sun et al. 1999b). Our present results indicate that I_K is attenuated in CB glomus cells from HF rabbits in the normoxic state and the resting membrane potential of glomus cells from HF rabbits is depolarized as compared with that in sham rabbits. The depression in I_K from the CB glomus cells in HF rabbits is consistent with the enhanced peripheral chemoreceptor function observed for CB afferent responses in HF rabbits. NO donors inhibited the baseline discharge of CSN (Sun et al. 1999b) and enhanced the I_K in glomus cells in sham and HF rabbits. L-NNA increased the baseline discharge of CSN in sham rabbits but not HF rabbits (Sun et al. 1999b) and attenuated the I_K in sham rabbits but not in HF rabbits. Based on these results, it is reasonable to assume that the attenuated I_K in glomus cells induced by HF contributes to the enhancement of CB chemoreceptor function. Nevertheless, it is important to point out that other ion channels in glomus cells may be altered in HF due to changes in NO or other neuro-endocrine factors. These effects must be taken into consideration when fully assessing the functional significance of the results we describe here.

In conclusion, our results indicate that I_K is attenuated in CB glomus cells from HF rabbits, and the attenuated I_K is due mainly to the decreased activity of K_Ca channels. In the normal state, NO maintains this K_Ca channel activity via a cGMP-dependent signalling pathway, and during HF a decrease in NO induced by an attenuated NOS activity is responsible for the inactivation of the K_Ca channels.