Journal of Neurophysiology

Abstract

The discharge patterns of respiratory neurons of the caudal ventral respiratory group (cVRG) appear to be subject to potent GABAergic gain modulation. Local application of the GABAA receptor antagonist bicuculline methochloride amplifies the underlying discharge frequency (F n) patterns mediated by endogenous excitatory and inhibitory synaptic inputs. Gain modulation can also be produced by alterations in the amplitude of spike afterhyperpolarizations (AHPs) mediated by apamin-sensitive small-conductance Ca2+-activated K+ (SK) channels. Since methyl derivatives of bicuculline (BICm) also have been shown to reduce the amplitude of AHPs, in vitro, it is possible that the BICm-induced gain modulation is due to a block of SK channels. The purpose of these studies was to determine the mechanisms by which BICm produces gain modulation and to characterize the influence of SK channels in the control of respiratory neuron discharge. Six protocols were used in this in vivo study of cVRG inspiratory (I) and expiratory (E) neurons in decerebrate, paralyzed, ventilated dogs. The protocols included characterizations of the neuronal responses to 1) BICm and apamin on the same neuron,2) BICm during maximum apamin-induced block of AHPs,3) apamin during maximum BICm-induced gain modulatory responses, 4) the specific GABAAreceptor antagonist, (+)β-hydrastine, 5) the specific GABAA receptor agonist, muscimol, and6) the GABA uptake inhibitor, nipecotic acid. For protocols 3, 5, and 6, only E neurons were studied. Four-barrel micropipettes were used for extracellular single neuron recording and pressure ejection of drugs. Cycle-triggered histograms were used to quantify theF n patterns and to determine the drug-induced changes in the gain (slope) and offset of theF n patterns. Compared to apamin at maximum effective dose rates, BICm produced a 2.1-fold greater increase in peak F n and a 3.1-fold greater increase in average F n. BICm and apamin produced similar increases in gain, but the offsets due to apamin were more negative. The responses to hydrastine were similar to BICm. During maximum apamin block, BICm produced an additional 112 ± 22% increase in peak F n. Conversely, apamin produced an additional 176 ± 74% increase in peak F n during the maximum BICm-induced response. Muscimol and nipecotic acid both decreased the gain and offset of the discharge patterns. Taken together, these results suggest that the gain modulatory effect of BICm is due to a reduction of GABAA-ergic shunting inhibition rather than a reduction in AHPs by block of SK channels in canine cVRG neurons.

INTRODUCTION

The discharge patterns of respiratory neurons of the caudal ventral respiratory group (cVRG) appear to be subject to potent GABAA-ergic gain modulation. Local application of the methyl derivatives of the competitive GABAA receptor antagonist bicuculline (BICm) (MacDonald and Olsen 1994) amplifies the underlying discharge frequency (F n) patterns that are mediated by endogenous excitatory and inhibitory synaptic inputs (Dogas et al. 1998; McCrimmon et al. 1997). These results imply a tonic GABAA-ergic mechanism that constrains the control and reflexly induced activities of these bulbospinal premotor respiratory neurons to ∼35–50% of the discharge rate without this inhibitory input. Such gain control could provide a powerful mechanism for the central respiratory control system to optimize the breathing pattern in response to changes in conditions or state (McCrimmon et al. 1997).

Some recent in vitro studies have shown that BICm can also block Ca2+-dependent, apamin-sensitive, small-conductance K+ (SK) channels in concentrations that are used for blocking GABAAreceptors (Debarbieux et al. 1998; Johnson and Seutin 1997; Seutin et al. 1997). The SK channels are activated rapidly by increases in intracellular Ca2+ concentration following a single action potential and produce a hyperpolarization that decays with a time constant of ∼150 ms (Vergara et al. 1998), termed the medium afterhyperpolarization (AHP) (Kobayashi et al. 1997; Viana et al. 1993). Blockade of the SK channels reduces the amplitude of the AHPs and causes a marked increase in the minimal repetitive firing frequency and slope of the frequency versus injected depolarizing current plot for hypoglossal motoneurons, suggesting an alternative mechanism for gain control or modulation (Viana et al. 1993). Thus it is possible that the BICm-induced gain modulation responses of the cVRG neurons may be due to a block of SK channels rather than GABAAreceptors. Alternatively, the BICm-induced gain modulation may be due to effects on both SK channels and GABAAreceptors or to GABAA receptors alone.

The purpose of the present studies was to gain further insight into the mechanisms underlying BICm-induced gain modulation using a multifaceted approach. Evidence from six protocols was used in this in vivo study of cVRG inspiratory (I) and expiratory (E) neurons in decerebrated, paralyzed, ventilated dogs. The protocols included characterizations of the neuronal responses to the separate application of BICm and apamin on the same neuron and to the sequential co-application of one during the maximum-induced effect of the other. In addition, the responses to the specific GABAA receptor antagonist, (+)β-hydrastine, the specific GABAA receptor agonist, muscimol, and the GABA uptake inhibitor, nipecotic acid, were characterized. Evidence from these protocols suggest that gain modulation is altered by both SK channel and GABAA receptor activation, but the effect of BICm on gain modulation is most likely due to a reduction of GABAAergic shunting inhibition rather than a reduction in AHPs due to a block of SK channels.

METHODS

This research was approved by the Subcommittee on Animals Studies of the Zablocki VA Medical Center, Milwaukee, WI, in accordance with provisions of the Animal Welfare Act, the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and VA policy. Data were obtained from 46 mongrel dogs of either sex weighing from 8 to 17 kg. Mask induction with a volatile anesthetic (isoflurane or halothane) was used, and anesthesia was maintained during the surgical procedure with isoflurane (1.4–2.0% end-tidal concentration). Airway, isoflurane, CO2, and O2 concentrations were continuously monitored with an infrared analyzer (POET II, Criticare Systems, Waukesha, WI). The animals were monitored for signs of inadequate anesthesia (e.g., salivation, lacrimation, and/or increases in blood pressure and heart rate), and if required, depth of anesthesia was increased immediately.

Surgical procedure

Dogs were intubated with a cuffed endotracheal tube and mechanically ventilated with an air-O2-isoflurane mixture. The surgical procedure, monitoring, and maintenance of body homeostasis have been previously described in detail (Dogas et al. 1998). Briefly, after cannulating the femoral artery (for blood pressure recording and blood-gas sampling) and vein (for continuous infusion of maintenance fluids and drugs), a bilateral pneumothorax was performed to reduce motion artifacts. A bilateral vagotomy was performed to remove the ventilator-induced effects of pulmonary mechanoreceptors on the breathing pattern. The animal was then decerebrated (Tonkovic-Capin et al. 1998). This procedure leads to an anatomically well-defined, midcollicular decerebration with modest blood loss (≤10 ml/kg). Isoflurane was discontinued, and the dogs were then ventilated with an air-O2 mixture and maintained in hyperoxic normocapnia (PO2 > 400 mmHg, PCO2: 38–45 mmHg). The dorsal surface of the medulla oblongata was exposed by an occipital craniotomy.

Phrenic nerve activity was recorded from the central end of the desheathed right C5 rootlet. The phrenic neurogram (PNG) was obtained from the moving-time average (100 ms) of the amplified phrenic nerve activity and was used to produce timing pulses corresponding to the beginning and end of the I phase. The neuromuscular blocker pancuronium (0.1 mg/kg, followed by 0.1 mg · kg−1 · h−1 ) was then given to reduce motion artifacts during neuronal recordings. Four-barrel micropipettes (10- to 30-μm composite tip diameter), consisting of one recording barrel containing a carbon filament and three drug barrels, were used for extracellular neuronal recordings and pressure ejection of nanoliter/picomol amounts of drug solutions. Bicuculline methochloride (BICm, 200 μM), apamin (0.125–0.500 μM), (+)β-hydrastine (1 mM), muscimol (10 μM), and nipecotic acid (10 mM; all from Research Biochemicals) were dissolved in an artificial cerebrospinal fluid (ACSF) vehicle that also served as a control solution. The methods used for picoejection and its limitations have been previously described in detail (Dogas et al. 1998;Krolo et al. 1999, 2000; Stuth et al. 1999). Unit activities from cVRG I and E neurons were recorded from a region extending 2–4 mm caudal and 2–4.5 mm lateral from the obex and 2–4.5 mm below the dorsal medullary surface. More than 88% of these neurons are bulbospinal premotor neurons (Bajic et al. 1992; Stuth et al. 1994).

Protocols

Following decerebration and discontinuation of the anesthetic, a period of at least an hour was allowed for washout of isoflurane (airway concentration, <0.05%) and for stabilization of the breathing pattern (i.e., PNG). Upon establishing a stable recording of single unit activity, a 2- to 3-min period was used to obtain the control discharge pattern prior to picoejection. Picoejections were made using step increases in dose rate via increases in the volume ejection-rate, which was measured as a change in meniscus height/time with a ×100 magnification microscope equipped with a reticule. Each dose rate was maintained until a steady-state neuronal response was achieved. Step increases in dose rate were continued until no further effect was observed. Picoejections of the ACSF vehicle were routinely used to verify that the ACSF constituents and/or ejected volumes were without effect.

In protocol 1, the effects of BICm and apamin were compared on the same neuron. Sufficient time (20–40 min) was allowed between picoejections for the discharge frequency to return to preejection baseline levels.

In protocol 2, we tested the ability of BICm to produce additional increases in F n during maximum block of the SK channels with apamin. For this purpose, apamin was picoejected at increasing rates until a maximum steady-state response was obtained. BICm was then picoejected concurrently with apamin at a dose rate that is known (protocol 1) to produce a near maximum response when given alone.

In protocol 3, we tested the ability of apamin to produce additional increases in F n during the maximum BICm-induced response. For this purpose, BICm was picoejected at increasing rates until a maximum steady-state response was obtained. Then during continued picoejection of BICm, the dose rate of apamin was increased until the maximum effect was seen.

In protocol 4, the neuronal responses to the competitive GABAA receptor antagonist (+)β-hydrastine were characterized and compared statistically with the responses to BICm and apamin.

In protocol 5, the neuronal response characteristics to picoejection of the specific GABAA agonist muscimol were compared for effects consistent with those produced by BICm, but opposite in direction.

In protocol 6, the neuronal responses to the GABA uptake inhibitor nipecotic acid, which is expected to increase the level of endogenously released GABA, were characterized and compared with those of muscimol.

Data analysis

Cycle-triggered histograms (CTHs; binwidth: 50 ms), triggered from either the onset of the E or I phase and based on 5–25 respiratory cycles were used to quantify the discharge frequency patterns at each dose rate. The values ofF n for each bin were calculated as the number of spikes per bin/bin duration in seconds. For each time (bin) increment within the triggered cycle, these values were averaged over the number of cycles used to generate the CTH. Drug-induced changes in the gain and offset of the discharge frequency pattern were analyzed via plots of the F n(drug) versusF n(control) values obtained from corresponding-time bin values from each CTH. Thus each point is an average based on the number of cycles used for the CTHs. This method has the advantage of being insensitive to the geometric shape of the pattern. Regardless of pattern trajectory, it can detect the amount by which a pattern is shifted parallel up or down i.e., shift in offset (offset = y intercept, Hz) and the amount by which a pattern is amplified, i.e., change in gain (gain = slope; unitless). Values of average and peakF n, gain, and offset produced by apamin, BICm, and hydrastine at maximally effective dose rates were collected for each neuron and pooled in two pools, i.e., for I and E neurons. Average F n values were obtained as the average of the bin F nvalues during the neuron's active phase. Similar data were also collected for muscimol in a separate group of E neurons but at sub-maximal dose rates to avoid complete inhibition. A repeated measures ANOVA was used for data obtained from the same neuron (apamin, BICm) and a multivariate ANOVA (MANOVA) followed by Fisher's Protected LSD post hoc test for multiple comparisons of the effects of hydrastine with those of BIC and apamin on the four variables. Differences were considered significant for P < 0.05. Values are expressed as means ± SE.

RESULTS

The results of the six protocols are based on the data from a total of 87 neurons (in 46 dogs) of which 18 were I neurons and 69 were E neurons. All I neurons had augmenting patterns, while 46 E neurons had augmenting patterns and 23 E neurons had decrementing patterns. Due to the small number of neurons per protocol, the results from the augmenting and decrementing E neurons were pooled. A statistical analysis, which used a two-way MANOVA with main effects of neuron type (I and E) and type of drug (BIC, apamin, and hydrastine) showed that there were significant differences between I and E neurons for the peakF n (P = 0.011) and slope (P = 0.0011) variables. Thus we analyzed the two neuron types separately.

Comparison of responses to apamin and BICm on the same neuron

INSPIRATORY NEURONS.

Figure 1 shows a typical example of the responses to picoejections of apamin (top) and BICm (bottom) for an I augmenting neuron. The dose rates were gradually increased until a maximal response was obtained. The effect of apamin was usually visible in <1 min from the onset of picoejection, and full recovery usually required ∼20–40 min. CTHs show that the apamin-induced increases in discharge frequency (F n) occurred mainly in the higherF n range near the end of the I phase (Fig. 2, top left). Using corresponding data from the CTHs, the plot ofF n during apamin versusF n control was highly linear and indicated that the changes in discharge pattern resulted from an increase in the gain (slope) but with a negative offset (yintercept) (Fig. 2, top right). In this example, apamin increased the average F n by 31% and peak F n by 40%. Gain was increased by 64% with a negative offset of −18.8 Hz.

Fig. 1.

Responses to picoejection of apamin (125 nM) and bicuculline (BICm, 200 μM) on the same caudal ventral respiratory group (cVRG) I neuron. Time courses of the responses are shown in the middle rate-meter traces of discharge frequency (F n). Apamin (top middle) was ejected in increasing dose rates until maximal effect was achieved at 0.6 fmol/min. No further increase in theF n was observed at more than double dose rate (1.5 fmol/min). Bottom middle: the time course of the response to BICm following complete recovery from apamin (20–40 min). BICm maximal effective dose rate: 2.3 pmol/min. Arrows: time-expanded views of the raw neuronal activity (NA),F n, and the phrenic neurogram (PNG) are shown for preejection control period and for period during maximal effect of apamin (top) and BICm (bottom).

Fig. 2.

Effects of apamin and BICm on the discharge patterns of the I neuron of Fig. 1 are compared via cycle-triggered histograms (CTHs) of unit activity (left). Right: corresponding plots of control discharge frequency pattern [F n(control)] vs. drug-induced discharge frequency pattern [F n(drug)] were used to analyze changes in the gain (slope) and offset (y intercept) of the discharge patterns. Top: apamin increased the gain by 64% and produced a negative offset of −18.8 Hz. Bottom: BICm increased the gain by 97% and produced a positive offset of 4.2 Hz. BICm also produced a much greater increase in peakF n than apamin.F n: neuronal discharge frequency. - - -, line of identity. r, correlation coefficient.

The neuronal response to BICm was usually visible in <1 min after the start of picoejection, and full recovery usually required 30–40 min (Fig. 1, bottom). The BICm-induced increases inF n were evident throughout the I phase (Fig. 2, bottom left) and were due to an increase in the gain with a positive offset (Fig. 2, bottom right). In this example, BICm increased the average F nby 104% and peak F n by 102%; gain was increased by 97% with a positive offset of 4.2 Hz.

The pooled data from nine I neurons comparing the responses to apamin and BICm are shown in Fig. 3. BICm produced a greater increase than apamin in both averageF n (97.5 ± 26.9 vs. 36.5 ± 13.4%, P < 0.05) and peakF n (103.7 ± 24.9 vs. 50.4 ± 13.6%), but only the difference in averageF n was statistically significant (P < 0.05, Fig. 3, left). The gain increases produced by apamin and BICm were not significantly different (72.0 ± 11.7 vs. 97.6 ± 18.5%, Fig. 3, top right). However, the offset produced by apamin was significantly more negative than that due to BICm (−22.1 ± 3.5 vs. −2.7 ± 8.6 Hz, P < 0.05, Fig. 3, bottom right).

Fig. 3.

Pooled data obtained from cVRG inspiratory neurons. BICm and apamin were applied on the same 9 I neurons. Hydrastine was applied on an additional 7 I neurons. Left: BICm produced the greatest average increases in averaged and peakF n. The difference between responses to BICm and apamin was only significant for the averageF n data. Top right: all 3 drugs produced significant increases in gain; however, the observed differences among these gain increases were not significantly different from each other. Bottom right: BICm and hydrastine produced significantly different offsets than apamin (*P < 0.05). Offsets produced by BICm and hydrastine were not significantly different from each other.

Expiratory neurons

The responses of the cVRG E neurons to apamin and BICm were qualitatively similar to those of the I neurons. Only the gain and peakF n increase produced by BICm was significantly greater for I neurons than for E neurons (P = 0.002 and P = 0.032, respectively). The CTHs of the responses of an E decrementing neuron showed that the apamin-induced increases inF n were also greater for the higherF n range, whereas BICm produced marked increases in both the high and low F nranges (Fig. 4, left). The average and peak F n responses to BICm were greater than those due to apamin. Both, apamin and BICm, produced marked increases in gain (50 vs. 70%, respectively), but striking differences in offset (−12.2 vs. +37.2 Hz, respectively).

Fig. 4.

Responses of a cVRG expiratory neuron to picoejection of apamin and BICm at maximally effective dose rates. Left: CTHs of unit activity during the control period and during ejection of drugs (bold patterns). Right: corresponding plots of [F n(drug)] vs. [F n(control)] used to estimate gain (slope) and offset (y intercept) changes. Top: apamin (125 nM) increased the gain by 50% and caused a negative offset of −12.2 Hz. Bottom: BICm (200 μM) increased the gain by 70% and caused, in contrast, a positive offset of 37.2 Hz.F n: neuronal discharge frequency. - - -, line of identity.

The pooled responses of 10 cVRG E neurons to apamin and BICm are compared in Fig. 5. BICm produced greater increases than apamin in average F n(76.3 ± 15.5 vs. 21.8 ± 8.4%, P < 0.05), while the increases in peak F n were not significantly different (63.4 ± 11.0 vs. 29.3 ± 7.0%). The gain increases produced by apamin and BICm were not significantly different from one another (54.0 ± 8.9 and 41.2 ± 12.2%, Fig. 5, top right). However, the offset produced by apamin was significantly different from that due to BICm (−25.0 ± 6.1 vs. +18.9 ± 9.8 Hz, P < 0.01, Fig. 5,bottom right).

Fig. 5.

Pooled data obtained from cVRG expiratory neurons. BICm and apamin were applied on the same 10 E neurons whereas hydrastine was applied on an additional 14 E neurons and muscimol was applied on another 26 E neurons. Left: BICm produced greater increases than apamin in average F n (*P = 0.034). The average F n for hydrastine was less than that for BICm (#P = 0.042). The peakF n data were similar to the averageF n data, but significant differences were not found. Muscimol produced a decrease in average and peakF n. Top right: apamin, BICm, and hydrastine produced significant increases in gain; however these gain increases were not significantly different from each other. Muscimol produced a decrease in the gain. Bottom right: the offsets produced by BICm and hydrastine were significantly different from that of apamin (**P < 0.01) but not significantly different from each other. Muscimol produced a negative offset.

BICm responses during maximum apamin-induced block of SK channels

Protocol 2 tested the hypothesis that if the BICm-induced neuronal response was due to antagonism of GABAAreceptors, then a BICm response should be observable during full block of the apamin-sensitive SK channels. Figure6 A shows an example in which apamin alone, at maximal effective dose rates, increased peakF n by 56% (i.e., from 81 to 126 Hz,left part of trace), and the concurrent application of BICm produced an additional 164% increase relative to preejection controlF n (i.e., from 126 to 259 Hz,right part of trace). To demonstrate that the block of the apamin-sensitive SK channels was complete, additional studies were performed on nine neurons comparing the effects of different concentrations of apamin (e.g., 150 and 300 nM, and 125, 250, and 500 nM, respectively). These data indicated that no further increase in peak F n occurred beyond that produced by the lowest apamin concentration. Figure 6 B shows the response of another E neuron to picoejection of 150 nM apamin (3.9 fmol/min), which produced a maximum increase of 93% in peakF n (i.e., from 75 to 145 Hz,middle). Subsequent picoejection of 300 nM apamin (10.9 fmol/min) produced no further increase (Fig. 6 B,bottom). Concurrent application of BICm increased peakF n by an additional 154% (i.e., from 145 to 260 Hz). The pooled data of eight neurons indicate that apamin increased peak F n by 67.3 ± 7.9% and BICm increased peak F n by an additional 112.4 ± 17.9% relative to the preapamin control, both at maximum effective dose rates.

Fig. 6.

BICm-induced responses of cVRG E neurons during maximum apamin-induced block of SK channels. A: apamin (100 nM) was picoejected at increasing dose rates until the maximal effect was achieved (1.5 fmol/min). No further increase in F nwas observed at a dose rate of 3.3 fmol/min. Concurrent ejection of BICm produced a substantial additional increase in theF n. B, top Fn trace: response of another cVRG E neuron to picoejection of 150 nM apamin at increasing dose rates until maximal effect was established at 1.1 fmol/min. Subsequent picoejection of 300 nM apamin at a dose rate of 10.9 fmol/min produced no further increase in F n, assuring that a maximum degree of block had taken place. Concurrent ejection of BICm produced an additional increase in theF n.

Apamin responses during maximum BICm-induced responses

Protocol 3 tested whether apamin could produce an additional neuronal response, due to a block of SK channels, during the maximum BICm-induced effect. This response was expected to be absent if BICm already blocked the SK channels. BICm was picoejected at increasing dose rates until the maximum response was obtained. While continuing BICm picoejection, co-application of apamin was initiated and continued at increasing dose rates until the maximum neuronal response was obtained. A typical example of the responses (CTHs) of an E decrementing neuron is shown in Fig. 7,left. The responses are shown for the maximum effective dose rates of BICm and the combined apamin plus BICm picoejections. Note that the slope as well as peak F nincreased. Plots of F n(BICm) versusF n(control) indicate that BICm increased the gain by 37% (Fig. 7 A). The plot ofF n(BICm + APA) versusF n(BICm) indicates that apamin produced an additional gain increase of 77% (Fig. 7 B). The slope for the effect of both drugs [F n(BICm + APA) vs.F n(control)] suggests that the two gain mechanisms act in a cascade fashion (i.e., the overall gain is the product of the gains from each mechanism, slope = 2.42 = 1.37 × 1.77, Fig. 7 C). For five E neurons, the average slope (gain) increase produced by BICm was 61 ± 16%, and apamin produced an additional 71 ± 25% increase in slope. The two cascaded gains result in an overall gain increase of 175%. BICm increased peak F n by 114.8 ± 26.6% and apamin increased peak F n by an additional 175.6 ± 74.5%, both at maximum effective dose rates.

Fig. 7.

Apamin-induced response of cVRG E neuron during the maximum BICm-induced response. Left: CTHs of unit activity during the control period and during ejection of drugs (bold patterns). BICm increased the gain of the F n pattern by an estimated 37% [plot of F n(BICm) vs.F n(control), A]. Subsequent, concurrent picoejection of apamin produced a further 77% increase in gain [plot of F n(BICm + APA) vs.F n(BICm), APA: apamin, B]. Both drugs produced an estimated increase of 142% [plot ofF n(BICm + APA) vs.F n(control), C]. The corresponding offsets from A–C were 1.4, −24.7, and −3.0, respectively.

Neuronal responses to (+)β-hydrastine

Typical examples of the neuronal responses to the specific GABAA receptor antagonist (+)β-hydrastine are shown for cVRG I and E neurons in Fig. 8,A and B, respectively. The effects of hydrastine could be observed within 1 min after the start of picoejection and lasted 10–20 min after discontinuation. The hydrastine-induced increases in F n were due to a combined increase in the gain and a positive offset as indicated by theF n(HYD) versusF n(control) plots of Fig. 8,bottom. The pooled responses from seven I neurons are shown in Fig. 3. Hydrastine produced increases of 71.1 ± 14.5 and 68.9 ± 8.8% in average and peakF n, respectively. These values were not significantly different from those produced by either apamin or BICm. Hydrastine increased the gain of the neuronal discharge pattern by 65.9 ± 10.9%, and the corresponding average offset was +6.3 ± 11.2 Hz. Similar to BICm, the hydrastine-induced offset was significantly different from that produced by apamin (P < 0.05).

Fig. 8.

Responses of a cVRG I and E neuron to picoejection of the specific GABAA antagonist (+)β-hydrastine (HYD). A: CTHs show effect of hydrastine on the discharge pattern of an I neuron. The plot of F n(HYD) vs.F n(control) indicates a 31% increase in the gain (slope) and a positive offset (y intercept) of 11.5 Hz. B: CTHs show effect of hydrastine on the discharge pattern of an E neuron. The plot of F n(HYD) vs. F n(control) indicates a 48% increase in gain and a positive offset of 13.4 Hz. - - -, line of identity

The pooled data from 14 cVRG E neurons for hydrastine were similar to those for the cVRG I neurons (Fig. 5). Hydrastine produced increases in average and peak F n of 47.7 ± 5.8 and 46.3 ± 4.4%, respectively. The increase in gain was 35.9 ± 8.6% with an offset of +8.5 ± 7.3 Hz, where the latter was significantly greater than that produced by apamin. Only average F n for hydrastine was significantly less than that for BICm (P < 0.05), while the peak F n and slope values were not significantly different from those for BICm.

Neuronal responses to muscimol

Picoejection of the specific GABAA receptor agonist muscimol resulted in a dose-dependent reduction ofF n. A typical example of the responses of an E neuron with a decrementing discharge pattern is shown in Fig. 9. In this protocol, submaximal dose rates were used to preserve the contour of theF n pattern, which can be distorted when the lower levels of activity are suppressed below discharge threshold. Higher dose rates of muscimol will completely inhibit neuronal activity. Analysis of the CTHs of this neuron's activity indicates that the reduction in F n was due to a decrease in gain to 59% of control and a negative offset of −14.4 Hz (Fig. 10).

Fig. 9.

Response of a cVRG E neuron to picoejection of the specific GABAA receptor agonist muscimol. Top: muscimol produced a dose rate dependent decrease in peakF n. ↓, time-expanded views of the PNG, raw neuronal activity (NA), and F n are shown for preejection control period and for period during 2 different dose rates of muscimol.

Fig. 10.

Effects of muscimol on the F n pattern of the E neuron of the previous figure. Top: CTHs show the effects of 2 dose rates of muscimol on theF n pattern. Bottom: the estimated decrease of 41% in gain produced by musimol at a dose rate of 180 fmol/min was obtained from the corresponding plot of the muscimol-induced F n pattern [F n(MUS)] vs. the controlF n pattern [F n(control)]. The corresponding offset was −14.4 Hz. - - -, line of identity.

The pooled data from 26 E neurons is shown in Fig. 5. Muscimol reduced the average and peak F n by 68.2 ± 2.7 and 57.8 ± 2.1%, respectively. The gain was reduced by 38.4 ± 3.2% with an offset of −26.2 ± 4.6 Hz.

Neuronal responses to nipecotic acid

A typical response of an E augmenting neuron to picoejection of the GABA uptake inhibitor nipecotic acid is shown in Fig.11. The amount of reduction ofF n was dose dependent (Fig.11 A), and at the highest dose rate, it was possible to inhibit neuronal activity almost completely. Recovery following discontinuation of the picoejection was relatively rapid (Fig.11 A, right). The analysis of the CTH data with plots of F n(NIPA) versusF n(control) (Fig. 11 B) indicates that the reduction of F n at the intermediate dose rate of 154 pmol/min was the result of a reduced gain (to 84%) and a negative offset (−39.6 Hz). The pooled mean data for the reduction of average and peakF n were 54.8 ± 23.6 and 50.2 ± 19.4%, respectively. The gain was reduced by 36.7 ± 19.3% with an offset of −11.3 ± 6.0 Hz (E neurons;n = 3). Except for the neuron of Fig. 11 in which a submaximal dose rate was used, the pooled data were analyzed at maximum effective dose rates, which were able to attenuate but not shift theF n pattern below the discharge threshold and confound the gain analysis.

Fig. 11.

Response of a cVRG E neuron to picoejection of the GABA uptake inhibitor nipecotic acid (NIPA). A: rate-meter recordings show a dose-dependent reduction ofF n and a relatively rapid recovery following discontinuation of nipecotic acid picoejection. B, left: CTHs of unit activity during the control period and during picoejection of the 2nd dose rate of nipecotic acid (154 pmol/min, bold pattern).B, right: corresponding plot ofF n(NIPA) vs.F n(control) was used to estimate the gain (0.84) and offset (−39.6 Hz) produced by nipecotic acid. - - -, line of identity.

Neuronal responses to the ACSF vehicle

A routine procedure that was incorporated in all protocols was the test of the effects of the constituents and volume of the ACSF vehicle. Very little or no noticeable effect was observed with pressure ejection of the ACSF vehicle alone at ejection rates equal to or greater than those used during the drug protocols.

DISCUSSION

The present studies provide several lines of evidence suggesting that GABAA receptors mediate a gain modulatory response in cVRG I and E neurons. An increase in the gain of the neuronal discharge frequency pattern was produced by the specific GABAA receptor antagonists, BICm and hydrastine, while a reduction in gain was produced by both the specific GABAA receptor agonist muscimol as well as an increase in endogenous GABA by the GABA uptake inhibitor, nipecotic acid.

In those studies that compared the responses of the same neuron to apamin and BICm, the changes in gain were of similar magnitude. However, the offset or y intercept associated with the gain change produced by apamin was significantly more negative than that produced by BICm. This difference in offset was on average 19.4 and 43.9 Hz for I and E neurons, respectively, and in part accounts for the smaller increase in average and peakF n produced by apamin versus BICm.

From an analysis view point, phasic discharge patterns, which are a characteristic of respiratory neurons, make it possible to distinguish between changes in the gain versus changes in baseline offset of neuronal activity. This distinction would not be possible with tonic or plateau-like, “flat,” phasic patterns without the use of additional perturbations. Mechanistically, a change in gain is analogous to a change in neuronal excitability, whereas a change in offset is analogous to a change in tonic polarization of the resting membrane potential. For example, if the Cl equilibrium potential was more negative than the resting membrane potential, then a reduction in Cl conductance would produce a relative depolarization. If, in addition, the Cl conductance was of sufficient magnitude and/or strategically located, for example on a main dendritic trunk, then the sensitivity of the neuron to other synaptic inputs would also be affected. Neuronal responses to the latter would manifest as a gain change.

During each action potential, Ca2+ enters the neuron via voltage-activated Ca2+ channels and rapidly activates small-conductance Ca2+-activated K+ channels. Because the increase in intracellular Ca2+ is rapidly buffered, this effect is transient and produces the AHP, a form of relative refractoriness. A reduction in the amplitude of the AHP by low bath Ca2+ concentration, block of Ca2+ channels, sequestration of Ca2+ by intracellular chelators, or block of SK channels results in an increase in neuronal excitability or gain (Viana et al. 1993). The mechanism by which apamin produces a negative offset is not clear, since a reduction in K+ conductance would also be expected to produce a relative depolarization of the membrane potential.

The results with hydrastine are similar to those of BICm but significantly different from those of apamin with regard to the offset of the discharge pattern. The molecular structure of hydrastine is very similar to that of bicuculline (Huang and Johnston 1990). It has been shown that bicuculline-free base can also block AHPs but only at much higher concentrations than those used for GABAA receptor antagonism (Debarbieux et al. 1998; Khawaled et al. 1999). However, bicuculline-free base is unstable in aqueous solutions. We thus used hydrastine, which is relatively stable and stays in solution at a pH of ∼7.0. Apparently, the presence of the quaternary N-methyl group, which makes BICm water soluble, is also responsible for acting as a plug of the SK channels because their block by BICm is voltage dependent (Debarbieux et al. 1998).

Different concentrations of apamin (125–500 nM) were picoejected on the same neuron to assure that a complete block of the apamin-sensitive SK channels (i.e., SK2) was achieved. The reported IC50 values for apamin at the SK2 channels are similar: 83 pM (HEK293 cells) (Strøbæk et al. 2000) and 63 pM (oocytes) (Köhler et al. 1996), and complete block was achieved at bath concentrations of <100 nM. In addition, SK3 channels that are responsible of the medium AHP in midbrain dopaminergic neurons exhibit apamin IC50values markedly greater than those for the SK2 channels: 2.0 nM (oocytes) (Ishii et al. 1997) and 9.2 nM (murine slice dopaminergic neurons) with complete block at <300 nM (Wolfart et al. 2001). Our pipette concentrations should have been sufficient to block both SK channel subtypes. Thus the additional marked increases in gain and in peak and averageF n produced by BICm during apamin block suggest that such responses are due to effects that are not mediated by a possible BICm block of SK channels but rather by GABAA receptors.

This finding alone does not rule out the possibility that BICm, when given alone, could affect both mechanisms. In this regard, the reported IC50 values for BICm for the apamin-sensitive SK channels are 26 μM for rat dopaminergic neurons (Seutin et al. 1997) and 25 μM for cloned rat SK2 channels expressed in HEK 293 cells (Strøbæk et al. 2000), but only 1.1 μM for SK2 channels expressed in Xenopus oocytes (Khawaled et al. 1999). Complete block in these preparations was achievable with bath concentrations of 100–300 μM. Thus it is possible that our BICm pipette concentration of 200 μM could have considerably blocked SK channels. However, the additional finding that apamin was capable of producing a marked increase in gain and in peak and average F n during maximum effective dose rates of BICm strongly suggests that SK channels were not blocked by BICm on these E neurons. The data from this protocol show that the two different mechanisms for gain change act in a cascade fashion (e.g., Fig. 7), that is, within the same neuron the overall gain is determined by the product of the gains from the two mechanisms.

Recent studies show that the SK channels can be subdivided into apamin-sensitive (SK2) and apamin-insensitive (SK1) channels, which may or may not coexist in the same neuron (Khawaled et al. 1999; Vergara et al. 1998). For example, the AHPs of hippocampal interneurons are sensitive to apamin, whereas those of hippocampal pyramidal neurons are not (Khawaled et al. 1999). Both, apamin-sensitive and -insensitive AHP components are present in vagal motoneurons, neocortical neurons, and pyramidal neurons of the sensorimotor cortex. Furthermore, BICm appears to block both SK1 and SK2 currents (Khawaled et al. 1999).

Apamin-sensitive AHPs are activated rapidly following a single action potential and decay with a time constant of ∼150 ms. In contrast, apamin-insensitive AHPs exhibit slow rise times, taking 0.5–1 s to peak, and decay with a time constant of ∼1.5 s. Apamin-sensitive AHPs serve to set the discharge frequency, whereas activation of the apamin-insensitive AHPs leads to spike-frequency adaptation (Sah 1996; Stocker et al. 1999). While it is not known if apamin-insensitive channels exist in the membranes of respiratory neurons, Richter and Heyde (1975) have shown that after the second spike of a spike train, induced by intracellular constant current injection in feline bulbospinal respiratory neurons, no further adaptation in F n is observed for the duration of the current. This observation is indicative of a rapid AHP process consistent with the dynamics of the apamin-sensitive SK channels. Because the properties of respiratory neurons in dogs and cats are similar, it is likely that the canine cVRG neurons also contain only the apamin-sensitive SK channels. This observation suggests that the BICm-induced responses in the present study were not due to the possible BICm block of apamin-insensitive SK channels.

In addition, it has recently been shown that apamin can also block apamin-insensitive channels, but at a slightly greater concentration (Strøbæk et al. 2000). The bath apamin IC50 for block of the SK1 channels expressed in HEK 293 cells was 3.3 nM, even though Köhler et al. (1996) found that 100 nM apamin had no effect on SK1-mediated currents expressed in oocytes. The pipette concentration of apamin used in our studies was 125 nM, which produced the maximum observable neuronal response, and 500 nM apamin did not produce further increases in the response. Therefore it is likely that if “apamin-insensitive” channels were present, they also would have been blocked along with the apamin-sensitive channels. Thus the additional increase in F n produced by BICm during maximal apamin block is probably due to antagonism of GABAA receptors.

The hypothesis that the gain change produced by BICm is mediated by GABAA receptors is further substantiated by the neuronal responses to the specific GABAA receptor agonist muscimol studied in E neurons. Picoejection of muscimol reduced gain and offset. These results are in the opposite direction to those for BICm as would be expected from antagonism based on the same mechanism. Furthermore, the increase in endogenous levels of GABA to E neurons produced by block of GABA uptake with nipecotic acid also resulted in a reduction in gain and offset.

Taken together, the results from all six protocols suggest that the mechanism of action for gain modulation produced by BICm is antagonism of GABAA receptors. Furthermore, our data suggest that BICm does not block apamin-sensitive SK channels in canine cVRG respiratory neurons, in vivo, at the pipette concentrations used in these studies (i.e., 200 μM). These studies suggest that gain changes produced by GABAA receptors and changes in SK channel conductance act in a cascade mode and are capable of potent modulation of the excitability of these respiratory neurons.

Acknowledgments

The authors are indebted to J. Tomlinson for expert surgical assistance.

This work was supported by the Office of Research and Development, the Medical Research Service, the Department of Veterans Affairs, National Institute of General Medical Sciences Grant GM-59234-01 to E. A. Stuth, and the Department of Anesthesiology of the Medical College of Wisconsin, Milwaukee.

Footnotes

  • Address for reprint requests: E. J. Zuperku, Research Service/151, Zablocki VA Medical Center, Milwaukee, WI 53295 (E-mail:ezuperku{at}mcw.edu).

REFERENCES

View Abstract