INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

TWO DISTINCT ACTION POTENTIAL PATTERNS have been described by intracellular studies of neurons in a number of different autonomic ganglia (10, 15, 28, 32). When stimulated with a long suprathreshold constant-current depolarizing step, one population of neurons responds with a brief burst of action potentials at the onset of the depolarization, accommodating to the stimulus, and the other population responds with repetitive action potentials persisting throughout the duration of the stimulus, not accommodating to the stimulus. Neurons displaying these different discharge characteristics have been designated "phasic" and "tonic," respectively (14, 15). In most autonomic ganglia, the two types of neurons appear to have different active and passive membrane properties (10, 15) and may have different anatomic characteristics as well (6). Tonic neurons generally exhibit a prominent potassium current referred to as A-type current, which increases the interspike interval between action potentials, whereas accommodation to the stimulus in most phasic neurons is due to activation of calcium-activated potassium current(s) and/or the M-type current (1, 10, 16, 31).

Parasympathetic neurons in guinea pig bronchial ganglia display tonic or phasic action potential patterns but have similar neurophysiological and anatomic characteristics (28). Accommodative properties have also been determined for intrinsic ganglion neurons located in guinea pig (19) and rat (4) tracheae, and a preliminary study (23) identified tonic and phasic firing patterns by neurons in human bronchial parasympathetic ganglia; other studies on ferret (5, 8) or cat (22) tracheal neurons did not address these properties. In guinea pig bronchial ganglia, most phasic neurons respond to a suprathreshold current stimulus, with action potentials decreasing in amplitude during the burst similar to those of guinea pig parasympathetic phasic neurons in urinary bladder ganglia (16) or sympathetic neurons in celiac ganglia (31). By contrast, phasic neurons in sympathetic (32), vagal sensory (11), airway parasympathetic (7), and vesical pelvic parasympathetic (15) ganglia generally respond with a burst of spikes of an amplitude similar to that observed at the onset of the stimulus. With very prolonged (5-s) stimuli, most neurons in rat tracheal ganglia respond with rhythmic, high-frequency (50- to 90-Hz) bursts of action potentials of equal amplitude throughout the stimulus and display a long (3-s) calcium-dependent afterhyperpolarization (4). Neurons in guinea pig bronchial ganglia do not display a prolonged (>500-ms) calcium-dependent spike afterhyperpolarization, and only subtle differences in active and passive membrane properties between guinea pig bronchial tonic and phasic parasympathetic neurons were observed (28); consequently, the membrane properties responsible for tonic and phasic action potential patterns and firing frequency in this preparation are unknown.

Accommodative properties of neurons in bronchial parasympathetic ganglia may greatly affect their ability to relay preganglionic stimuli (24, 28) and, consequently, parasympathetic tone in the airway (24). Accommodation in guinea pig bronchial neurons is altered by immediate hypersensitivity (allergic) reactions; specific antigen challenge in vitro causes neurons that normally respond with a phasic action potential pattern to respond with a tonic pattern (29). Our understanding of the mechanism by which inflammatory mediators affect accommodation is limited by the lack of information regarding the ionic currents that govern the repetitive action potential activity. Although there have been several voltage-clamp studies (2, 4) characterizing the different currents in airway parasympathetic neurons, none have determined the role these currents play in regulating action potential accommodation or firing frequency. The present study further characterizes differences in active membrane properties of tonic and phasic neurons in guinea pig bronchial parasympathetic ganglia. The role voltage- and ionic-dependent currents have in regulating the accommodation properties and frequency of action potentials by these neurons is also examined.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Male albino guinea pigs (180-300 g) were asphyxiated with carbon dioxide and exsanguinated; the bronchus was isolated, cut longitudinally along the ventral surface, opened, and then tightly pinned as a sheet to the floor of a Sylgard-coated recording chamber (0.2-ml volume) with Z-shaped pins. Ganglia were located in the extrachondral plexus near the peribronchial nerves and were visualized, without staining, after removal of the overlying connective tissue by fine dissection. The tissue was equilibrated with flowing (5-8 ml/min) Krebs bicarbonate buffer at 36°C for at least 1 h in the recording chamber before experimentation. The composition of the Krebs buffer was (in mM) 118 NaCl, 5.4 KCl, 1 NaH₂PO₄, 1.2 MgSO₄, 1.9 CaCl₂, 25 NaHCO₃, and 11.1 dextrose and was bubbled with 95% oxygen-5% carbon dioxide at a pH of 7.4. To avoid precipitation of cadmium ions, the tissue was isolated and equilibrated as above, but, for the cadmium experiments, the tissue was temporarily superfused with a solution containing the same sodium, potassium, magnesium, calcium, and chloride ion and dextrose concentrations and was buffered with HEPES (30 mM), pH 7.4, before the addition of cadmium to this superfusate.

Micropipettes were fabricated from thick-walled capillary stock (0.5-mm ID, 1.0-mm OD; World Precision Instruments, Sarasota, FL) by a Brown-Flaming microelectrode puller (model P-87, Sutter Instruments, San Rafael, CA). Electrodes were filled with 3.0 M potassium chloride (pH 7.4), and the electrolyte in the micropipette was connected by a chloridized silver wire in an electrode holder (Axon Instruments, Foster City, CA) by a headstage to an electrometer (Axoclamp 2A, Axon Instruments). A silver-silver chloride pellet in the bath was connected to a headstage ground. The electrode DC resistance in Krebs solution ranged between 60 and 70 M. Intracellular data (voltage and current) were displayed on-line on an oscilloscope and a chart recorder and stored on digital audiotape. Data epochs were digitized by a Macintosh computer equipped with a data-translation interface and displayed and analyzed off-line with an oscilloscope simulation-and-analysis program (AxoData and AxoGraph, Axon Instruments). The delivery of constant-current pulses through the microelectrode was also controlled by the computer. Intracellular recordings were performed with the electrometer in either discontinuous current clamp (3.0- to 4.0-kHz sampling rate) or active bridge mode. Quantitative voltage clamp was not used because of the location of the ganglia deep within the tissue and bath and because of space-clamp distortion from the nonisopotential spread of voltage over the neuronal soma, dendrites, and axon(s) of the in situ guinea pig airway neuron (19, 28).

Drug and ionic substitutions were utilized as follows. CaCl₂ was substituted for with MgCl₂ on an equimolar basis to produce a nominally "calcium-free" Krebs solution. The following compounds were added to the control Krebs solution (final concentration): apamin (1 µM), 4-aminopyridine (0.1 mM), -conotoxin GVIA (1 µM), nifedipine (10 µM), tetrodotoxin (TTX; 1 µM), and verapamil (50 µM). Cadmium chloride (0.1 mM) was added to the HEPES solution as described above. All reagents used to prepare the Krebs and HEPES solutions were purchased from J. T. Baker Chemical (Phillipsburg, NJ). -Conotoxin GVIA was purchased from Research Biochemicals International (Natick, MA). All other reagents were purchased from Sigma (St. Louis, MO).

Results are presented as means ± SE. Means were compared as unpaired samples (see RESULTS), with Student's t-test statistics for two means; means were considered to differ significantly if P was <0.05. These results, as well as the slope of the change in rate of action potential frequency, were analyzed by the Statview statistics program (Abacus Concepts, Berkeley, CA).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Repetitive firing characteristics. Intracellular recordings were made from 152 neurons in parasympathetic ganglia located on the primary (predominantly right side) bronchi from guinea pigs; of these, 145 were used in these studies. The classification of bronchial neurons as either tonic or phasic was similar to that previously described (28): in response to threshold (0.5-nA) rectangular anodal constant-current steps, 100-500 ms in duration, tonic neurons (52 of 145 neurons) responded with either one or several action potentials of equal size at the onset of the stimulus (Fig. 1A) or action potentials throughout the step. Tonic neurons responded to suprathreshold (1- to 3-nA) current steps with action potentials throughout the depolarization (Fig. 1B). Phasic neurons responded to threshold (0.5-nA) rectangular anodal constant-current steps, 100-500 ms in duration, with one to several action potentials at the onset of the stimulus (Fig. 1C); 2-10 times threshold stimuli (1-5 nA) elicited either one or a burst of action potentials in the initial 50 ms of the stimulus followed by accommodation (91 of 145 neurons). In phasic neurons, action potentials decreased in amplitude during the burst until no further regenerative spikes were generated (Fig. 1D). On occasion (7 of 152 neurons), clear differentiation of tonic and phasic action potential patterns was not evident; these neurons were not included in these studies. The mean resting membrane potentials for tonic and phasic neurons used in the study were 53 ± 3 and 50 ± 2 mV, respectively. The mean resting input resistances for tonic and phasic neurons used in these experiments were 52 ± 7 and 38 ± 4 M, respectively (P < 0.05). In recording sessions that involved characterization of >1 neuron/ganglion, both tonic and phasic neurons were recorded in 15 of 27 ganglia.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Accommodating and nonaccommodating responses to threshold and suprathreshold current steps by bronchial ganglion neurons. A: response by a tonic-type neuron to threshold (1.0-nA) rectangular anodal current step, 500 ms in duration, which elicits several action potentials at onset of stimulus. B: 2 times threshold current steps (2.0 nA) evoke action potentials throughout current step. C: phasic neuron response to threshold stimuli (0.5 nA) with either 1 (data not shown) or a burst of action potentials at onset of stimulus. D: 4 times threshold stimuli (2.0 nA) elicit either 1 (data not shown) or a burst of action potentials followed by accommodation (note action potentials decrease in amplitude during burst until no further regenerative spikes were generated). Bottom traces, duration and amplitude of current step stimuli.

The interval between the first two action potentials (measured peak to peak) was used to compare the instantaneous firing rate of tonic neurons and phasic neurons that responded with two or more action potentials at a threshold (0.5-nA), 500-ms current step; the interspike interval between the next five spikes at 0.5, 1.0, 1.5, and 2.0 nA during prolonged (500-ms) current steps was measured in six bursting phasic neurons and six tonic neurons to determine the steady-state (adapted) firing rate. The instantaneous firing rate between tonic and phasic neurons was different at threshold (0.5-nA) stimuli where the interspike intervals were 23 ± 3 and 16 ± 1 ms, respectively (P < 0.05; n = 10 for both cell types; Fig. 2A), with a similar observation at four times threshold (2.0 nA; Fig. 2A). The interspike interval between successive action potentials (this "adapted" firing rate was calculated from intervals between peaks of spikes 2-5) decreased with increasing current amplitude (0.5-2.0 nA): from 41.6 to 14.0 ms for tonic neurons and from 22.2 to 10.9 ms for phasic neurons (Fig. 2B). Adapted firing rates were different at each current intensity studied (0.5, 1.0, 1.5, and 2.0 nA), with phasic neurons firing at higher frequencies than tonic neurons (P < 0.05; Fig. 2B, Table 1); the slope of the change in adapted rate for the four stimulus intensities was similar for tonic and phasic neurons (R = 0.968 and 0.966, respectively).

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Instantaneous (A) and adapted (B) firing rate of tonic and phasic neurons were determined from interval between 1st 2 and subsequent action potentials, respectively, during threshold and suprathreshold prolonged current steps. A: instantaneous firing rates at threshold and 2-4 times threshold currents (0.5-2.0 nA) ranged between 43 and 82 spikes/s for tonic neurons (n = 10; ) and between 62 and 89 spikes/s for phasic neurons (n = 10; ). B: frequency of successive action potentials (between spikes 2 and 5) increased with increasing current from 24 to 71 spikes/s for tonic neurons () and from 45 to 91 spikes/s for phasic neurons (). * P < 0.05 for differences between tonic and phasic neurons at respective current steps.

View this table: [in this window] [in a new window]   Table 1.   Adapted interspike intervals for tonic and phasic neurons: effects of apamin, 4-aminopyridine, -conotoxin GVIA and barium chloride

Effects of blocking agents and hyperpolarization on accommodation and action potential frequency. The role of sodium and calcium conductance in action potential accommodation was determined in tonic and phasic neurons. The sodium channel-blocker TTX was bath applied after the neurons were typed as either tonic or phasic with 1- to 3-nA, 100-ms depolarizing steps (Fig. 3). TTX (1 µM) reduced the amplitude and frequency of repetitive spikes in tonic neurons (n = 4; Fig. 3A, middle trace) during the depolarization. The amplitude of the remaining action potentials could be further reduced by calcium-free buffer (n = 4 neurons; Fig. 3A, right trace) and/or cadmium chloride (0.1 mM; n = 3 neurons; data not shown). TTX (1 µM) blocked action potentials in the burst of spikes in phasic neurons, leaving only one or two nonregenerative spikes at the onset of the depolarizing step (n = 4; Fig. 3B, middle trace). This remaining spike could be reduced in amplitude by calcium-free buffer (n = 4 neurons; Fig. 3B, right trace).

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Effects of blocking sodium and calcium conductances on action potential accommodation in tonic and phasic neurons. A: left, repetitive action potentials are evoked in a control tonic neuron with a 2.0-nA, 100-ms depolarizing step; middle, tetrodotoxin (TTX) reduced amplitude of repetitive spikes in this cell, leaving broad, lower-frequency spikes during depolarization (n = 4 neurons); right, amplitude of remaining spike could be further reduced by calcium-free buffer (n = 4 neurons) and/or cadmium chloride (0.1 mM; n = 4 neurons; data not shown). B: left, burst of spikes is evoked in a control phasic neuron with a 2.0-nA, 100-ms depolarizing step; middle, TTX blocked action potentials in burst of spikes, leaving only 1 nonregenerative spike at onset of depolarizing step (n = 4 neurons); right, remaining spike could be reduced in amplitude by calcium-free buffer (n = 4 neurons). Scale bars apply to respective tonic and phasic traces to right.

Reducing calcium currents increases accommodation in tonic neurons, but reducing potassium currents affected both the firing rate and accommodation in these cells. With a 500-ms, 1.0-nA depolarizing constant-current step, accommodation increased in the presence of calcium-free buffer in all tonic neurons tested, leaving action potentials of equal amplitude at the onset of the stimulus (n = 8 neurons; Fig. 4A). A similar increase in accommodation was observed during superfusion with cadmium chloride (0.1 mM; n = 10 neurons; data not shown). -Conotoxin GVIA (1.0 µM; n = 6 neurons) blocked repetitive action potentials, leaving action potentials during the beginning of the stimulus (Fig. 4B). The adapted firing rate of the remaining action potentials during prolonged depolarization steps, as described above, was unaffected by calcium-free buffer (n = 8 neurons), cadmium chloride (0.1 mM; n = 8 neurons), or -conotoxin GVIA (n = 4 neurons; Fig. 4, Table 1). Bath-applied nifedipine (10 µM; n = 4 neurons) decreased spike frequency in one of four neurons but did not block the repetitive firing pattern; verapamil (50 µM; n = 3) had a similar effect (data not shown). The potassium-channel blocker 4-aminopyridine (0.1 mM) increased accommodation in tonic neurons (n = 6; Fig. 4C, right trace, 2-nA stimulus); in the presence of this compound, the adapted interspike interval between the remaining action potentials in the burst was 8.0 ± 1 ms (P < 0.05 compared with control neurons above; Table 1). An inhibitor of calcium-activated potassium currents, apamin (1 µM), decreased the adapted interspike interval of repetitive action potentials during a 500-ms, 2-nA current step to 11.5 ± 2 ms in four tonic neurons (P < 0.05 compared with control frequencies; Table 1). Bath application of barium chloride (1.0 mM) had little effect on the adapted interspike interval (12.2 ± 3 ms) of repetitive action potentials during a 500-ms, 2-nA current step in tonic neurons (P < 0.5; n = 3; Table 1).

View larger version (45K): [in this window] [in a new window]   Fig. 4.   Blocking calcium conductance or decreasing A-type current increases accommodation in tonic neurons. A: repetitive action potentials by a control tonic neuron (left) elicited by a 500-ms, 1.5-nA depolarizing current (current trace not shown) was changed to a phasiclike pattern (right) in presence of calcium-free buffer (n = 8 neurons) or with 0.1 mM cadmium chloride (data not shown). B: a similar change in accommodation (evoked by a 500-ms, 2.0-nA current stimulus not shown) was also induced by -conotoxin GVIA (right) in 6 tonic neurons. Bath-applied nifedipine (10 µM) decreased spike frequency in 4 tonic neurons but did not block repetitive firing pattern; verapamil (50 µM; n = 2 neurons) had a similar effect (data not shown). C: 4-aminopyridine (right) also increased accommodation (n = 6 neurons; 500-ms, 2.0-nA current stimulus not shown).

After an initial hyperpolarizing step and return to the resting potential, application of a depolarization stimulus resulted in a decrease in accommodation in phasic neurons. Current clamping the resting membrane potential (50 mV) to a range of 90 to 100 mV for 1 or more seconds changed the action potential pattern of five of seven bronchial phasic neurons; after the hyperpolarizing step and return to the resting potential, the depolarization stimuli (500 ms, 0.5 nA) now elicited repetitive action potentials (Fig. 5A) similar to those observed in tonic neurons. The induction of the tonic-type firing pattern was transient; within 10-15 s, the firing pattern returned to the normal phasic pattern (data not shown). These changes in spike accommodation could be entirely blocked by bath application of 4-aminopyridine (0.1 mM; n = 4 neurons; Fig. 5B) or with bath application of cadmium chloride (0.1 mM; n = 3 neurons; data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Effects of membrane potential hyperpolarization on accommodating action potential pattern by bronchial phasic neurons. A: after neuron was typed as phasic (500-ms, 0.5-nA current trace; bottom), resting potential was clamped to 90 mV for 5 s (B; full time not shown). C: after return to preclamp resting potential, typing the same neuron results in a tonic firing pattern. D: the same neuron as in A was superfused with 4-aminopyridine (0.1 mM) and was clamped to 90 mV for 5 s (E; full time not shown) and returned to preclamp potential. F: typing the same neuron results in no change in accommodation (n = 4 neurons). Similar results were observed in presence of cadmium chloride (0.1 mM in HEPES buffer; data not shown).

Bath-applied barium chloride (1 mM; n = 6 neurons; Fig. 6A, right trace) or apamin (1.0 µM; n = 4 neurons; Fig. 6B, right trace) also decreased spike accommodation in phasic neurons evoked by a 500-ms depolarizing current step (2.0 nA), and this effect could be reversed by adding cadmium chloride (0.1 mM; n = 4 neurons) for both treatments (data not shown). Apamin (1.0 µM) had little effect on the adapted firing rate as measured above (Table 1) but did decrease the duration of the afterhyperpolarization that followed four consecutive action potentials (elicited by 20-Hz, 3.0-nA, 2-ms steps) from a control duration of 175 ± 35 to 65 ± 30 ms (P < 0.05); apamin (1 µM; n = 4 neurons) had no effect on the amplitude of the afterhyperpolarization (data not shown). -Conotoxin GVIA (1.0 µM; n = 4 neurons), calcium-free buffer (n = 8 neurons) and cadmium chloride (0.1 mM; n = 4 neurons) reduced the number of action potentials in the burst of action potentials at the onset of the stimulus in phasic neurons (data not shown). The adapted firing rate of the remaining action potentials during prolonged depolarization steps, as described above, could not be determined in the presence of calcium-free buffer (n = 8 neurons), cadmium chloride (0.1 mM; n = 4 neurons), or -conotoxin GVIA (n = 4 neurons) because too few action potentials were generated in the presence of these compounds.

View larger version (44K): [in this window] [in a new window]   Fig. 6.   Bath-applied barium chloride (A) or apamin (B) decreases accommodation in phasic neurons. A: control phasic neuron (left) fires repetitive spikes evoked during a 500-ms current step (2.0 nA) in presence of barium chloride (right). B: apamin decreased accommodation in phasic neuron (right).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In the present study, further differences in active membrane properties were observed for guinea pig bronchial tonic and phasic parasympathetic ganglion neurons, and the ionic currents regulating accommodation and action potential frequency were determined. With the use of 100- to 500-ms depolarizing steps, the interspike interval between the first two action potentials was greater (lower instantaneous firing rate) in tonic neurons than in phasic neurons at threshold and four times threshold (2.0-nA) stimulus intensities (Fig. 2A). For subsequent action potentials, tonic neurons had a more prolonged interspike interval than phasic neurons (lower adapted firing rate) at all depolarizing currents studied (Fig. 2B). These instantaneous and adapted patterns by bronchial phasic and tonic neurons were similar to guinea pig inferior mesenteric ganglion phasic type I and tonic type III and IV neurons, respectively (32). Although phasic neurons fire a burst of action potentials at a higher frequency than tonic neurons, the greater interspike interval in tonic neurons may be necessary for the maintenance of the repetitive action potential pattern observed in these cells (28). Blocking sodium conductance with TTX reduced, but did not entirely eliminate, action potentials in tonic and phasic neurons; evidence for a role of calcium channels in the remaining action potentials came from the elimination of the remaining spikes with calcium-free buffer (Fig. 3) or cadmium ions.

If the repetitive activity observed in control tonic neurons is calcium dependent, then decreasing the calcium conductance should affect repetitive action potentials in these neurons. By reducing the extracellular calcium concentration with nominally calcium-free buffer, the number of action potentials during the prolonged depolarizing stimulus decreased in tonic neurons; this indicates that calcium channels are the major charge carriers during the repetitive action potential activity that follows the initial burst of action potentials (Fig. 4). Reducing the calcium conductance, however, had no effect on the adapted firing rate of the remaining action potentials (Table 1). Similarly, cadmium ions, which are more efficacious at blocking T- and N-type calcium channels (17), decreased repetitive spikes by tonic neurons. Thus it may be inferred that the calcium-channel subtype active during the repetitive action potentials is the N-type channel because -conotoxin GVIA (relatively specific for the N-type calcium channel) blocked repetitive spikes, whereas nifedipine, relatively specific for the L-type calcium channel (reviewed in Refs. 17, 20), was without overt effect. The presence of both high-threshold (L-type) and -conotoxin-sensitive (N-type) calcium currents have been identified in dissociated neurons from rat tracheal (2) and cardiac (18) parasympathetic ganglia; N-type calcium channels have been identified in amphibian cardiac parasympathetic ganglion neurons as well (21). The characterization of calcium currents and their effects on repetitive action potential pattern is a relatively unexplored area of parasympathetic neurophysiology (reviewed in Ref. 20).

The differences in accommodation patterns by tonic and phasic neurons is likely due to different expression and/or activation of potassium channels. This study provides evidence that a potassium current may be responsible for the increase in interspike interval in tonic neurons. The A-type current blocker 4-aminopyridine reduced the number of action potentials evoked by a prolonged depolarizing stimulus, making them fire with an action potential pattern similar to that of phasic neurons (Fig. 3). These results indicate that activation of the A-type current contributes to the increased interspike interval observed in guinea pig bronchial tonic neurons because 4-aminopyridine decreased the interspike interval in tonic neurons to a level similar to that in phasic neurons (Table 1). The A-type current has also been identified in rat tracheal parasympathetic neurons (2) and is a very prominent current in tonic neurons located in guinea pig sympathetic ganglia (10, 31). Furthermore, the accommodation to the stimulus by phasic neurons could be reversed after the resting potential was current clamped to more negative levels that could be blocked by 4-aminopyridine (Fig. 5), indicating activation of the A-type current during the hyperpolarization (17). It is unlikely the hyperpolarization-activated h-type current plays a role in action potential accommodation because it is activated only at potentials more negative than 90mV, a potential that is not reached during action potential afterhyperpolarization in bronchial phasic cells (Fig. 3) (28), and, furthermore, it is not blocked by barium (17; see below).

Further evidence for the role of potassium currents in accommodation is based on the decrease in accommodation in phasic neurons by apamin or barium ions, indicating the presence of a calcium-activated potassium channel. Although it was originally suggested that the decrease in accommodation induced by barium is due to inhibition of the M-type current (12), a similar effect by apamin on bronchial ganglion neurons indicates that perhaps barium decreases accommodation by inhibiting calcium-activated potassium currents and perhaps by "uncovering" or increasing current carried by calcium channels (reviewed in Ref. 17). Furthermore, Myers and Undem previously demonstrated that muscarinic- (26) or neurokinin-receptor (25) agonists have no effect on accommodation in phasic neurons, eliminating any role of M-type current associated with these agonists in accommodation, unlike that reported for rat tracheal (4) or guinea pig intracardiac (3) parasympathetic ganglion neurons.

That similar calcium and potassium currents are present in tonic and phasic neurons indicates that these cells may not represent distinct neuronal populations as has been suggested for tonic and phasic neurons in sympathetic ganglia (6, 10, 31) but represent a single population with different states of potassium-channel activation or inactivation. Phasic neurons may have calcium channels similar to tonic neurons; the decrease in accommodation in guinea pig bronchial phasic neurons induced by hyperpolarization, barium, or apamin in all instances could be reversed by blocking the calcium current. Such results indicate the presence of calcium current in phasic neurons, which may be shunted by an outward current with similar temporal kinetics. Thus blocking the antagonistic outward potassium conductance with apamin or barium may unmask or amplify the calcium currents necessary for repetitive action potential activity. However, calcium-activated potassium currents may also be active but have a different role in tonic neurons, i.e., regulating the frequency of repetitive action potentials, as apamin increased the firing rate in these neurons (Table 1). As mentioned above, guinea pig bronchial ganglion neurons have similar anatomic characteristics (28); furthermore, chemical coding for neurotransmitters in guinea pig bronchial ganglion neurons is similar as well: all neurons in the guinea pig bronchial ganglia synthesize acetylcholine, i.e., are choline acetyltransferase positive (9), and do not synthesize nitric oxide (13).

The results from these studies may explain the mechanism for the decrease in accommodation in bronchial neurons observed after antigen challenge (29): inflammatory mediators such as prostaglandin D₂ may directly or indirectly affect conductance through N-type calcium channels or decrease an opposing conductance, e.g., through calcium-activated potassium channels, as has been reported for guinea pig vagal sensory neurons (30). Thus activation or inhibition of these channels by inflammatory mediators (29) or neurotransmitters (27, 28) released near a ganglion neuron may ultimately affect the ability of that neuron to relay excitatory stimuli from the central nervous system and, consequently, regulate airway parasympathetic tone. Results from the present study suggest that regulation of accommodation properties may be on the level of a single neuron and not on the entire population of neurons within the ganglion or bronchus because recordings were made from both tonic and phasic neurons in single ganglia.