Effects of Halothane on Synaptic Neurotransmission to Medullary Expiratory Neurons in the Ventral Respiratory Group of Dogs 

This research was approved by the Medical College of Wisconsin Animal Care Committee and conformed with standards set forth in the National Institutes of Health Guide for Care and Use of Laboratory Animals.

Fifteen adult mongrel dogs (weight 8–16 kg) were studied under halothane as the sole anesthetic agent. Anesthesia in dogs was induced by mask with halothane and maintained with the same agent throughout the surgical preparation and the experimental protocols. No intravenous anesthetic agents were used in any of the studies. The trachea was intubated with a cuffed endotracheal tube without the help of muscle relaxants. During surgery the animals were artificially ventilated without the use of muscle relaxants with O²–air–halothane mixtures under normocapnic conditions. Airway CO²and volatile agent concentrations were continuously recorded with an infrared analyzer (POET II, Criticare Systems, Waukesha, WI). The femoral artery was cannulated for continuous blood pressure monitoring (Gould-Statham P23 ID transducer) and periodic arterial blood gas sampling. A triple-lumen catheter was placed in the femoral vein for drug administration and for continuous infusion of maintenance fluids (lactated Ringer's solution at 4 ml · kg⁻¹· h⁻¹). Additional sodium bicarbonate to correct metabolic acidosis was given as determined by blood gas analysis but was rarely required. In addition, 3 ml of the maintenance solution was infused to replace each milliliter of estimated blood loss. Phenylephrine (0.5–5.0 μg · kg⁻¹· min⁻¹) was infused if necessary to keep mean arterial pressure above 75 mmHg. 20,21Phenylephrine was only required for anesthetic doses of 2.0 MAC in a few animals. The animals were monitored for signs of inadequate anesthesia, including movement, salivation, lacrimation, or increases in blood pressure and heart rate. The anesthetic depth was increased immediately if such signs were present. In general, anesthetic depths of 1.3–1.7 MAC halothane were found adequate for surgery. Esophageal temperature was measured and maintained within a target range of 37.5–38.5°C with a servocontrolled heating pad.

The animals were positioned in a stereotaxic device (model 1530; David Kopf Instruments, Tujunga, CA) with the head ventrally flexed (30 degrees). Bilateral dorsolateral neck dissections were performed. The vagus and phrenic nerves (right C5 central rootlet) were isolated and prepared for recording (phrenic) and sectioning. Some peripheral inputs, including those from the pulmonary stretch receptors and aortic arch chemoreceptors, were then removed by bilateral cervical vagotomies. Bilateral pneumothorax was performed to minimize brainstem movement and phasic inputs from chest wall mechanoreceptors. An occipital craniotomy was performed and the dura mater was cut along the midline and retracted to expose the medulla oblongata. After completion of the occipital craniotomy the animals were maintained at 1 MAC halothane level as baseline anesthesia. Pancuronium (bolus of 0.1 mg/kg, followed by a continuous infusion of 0.1 mg · kg⁻¹· h⁻¹) was then given to avoid motion artifacts during respiratory neuronal recordings.

Expiratory neurons of the caudal ventral respiratory group were located within the dorsal medulla 2–5 mm caudal to the obex, 2.5–4.5 mm lateral to the midline, at a depth of 2–4.5 mm as described previously. 21,22Over 90% of these neurons have been demonstrated to be bulbospinal premotor neurons. 12Multibarrel compound glass electrodes consisting of a recording barrel containing a 7 μm carbon filament and three drug barrels were used to simultaneously record extracellular neuronal action potentials before and during pressure picoejection of the GABA^Aneurotransmitter antagonist bicuculline methyl chloride (50–200 μM; Research Biochemicals, Natick, MA), at a constant, steady-state dose rate on the neuron of interest. 15,17The neurotransmitter antagonist was prepared in an artificial cerebrospinal fluid consisting of 124 mM NaCl, 2 mM KCl, 2 mM MgCl², 1.3 mM KH²PO⁴, 0.9 mM CaCl², 26 mM NaHCO³, and 11 mM glucose in which the p  H was adjusted to 7.2–7.4 by aeration with 5% CO²and 95% O². The drug-free artificial cerebrospinal fluid vehicle served as a control. Compressed nitrogen gas was used for pressure picoejection of the neurotransmitter antagonist and artificial cerebrospinal fluid. Ejection was accomplished with the “picospritzer” technique that allowed controlled drug volume application by regulation of ejection pressure, duration, and frequency of the ejection pressure pulses. Drug dose delivery rates were determined in vivo  by reading the change per unit time of the meniscus level in the drug barrel of the capillary glass electrode with a 50× monocular microscope equipped with a calibrated reticule (resolution 2 nl). To minimize any transient effects of the pharmacokinetics or pharmacodynamics of the neurotransmitter antagonist on neuronal response patterns, steady-state dose (constant ejection rate) picoejection techniques were used.

Single-cell expiratory neuronal activity, phrenic nerve activity, picoejection marker pulses, airway CO²and volatile anesthetic concentrations, systemic blood pressure, and airway pressure were recorded on a digital tape system (model 3000A; A. R. Vetter Co., Rebersburg, PA). These variables or their time averages were also continuously displayed on a polygraph (model 7; Grass Instruments, Quincy, MA) during the experimental runs. Timing pulses at the beginning and end of neural inspiration were derived from the phrenic neurogram and were used to aid in data analysis. The tape-recorded data were digitized and analyzed off-line.

All experiments were performed under hyperoxic (inspired oxygen fraction > 0.8) and steady-state hypercapnic conditions (target arterial carbon dioxide tension: 50–60 mmHg) in vagotomized, paralyzed, and mechanically ventilated preparations. This level of hypercapnia was chosen to ensure adequate phrenic activity even at the higher level (2 MAC) of anesthesia. Target arterial carbon dioxide tension was held constant for all experimental runs within an animal but was allowed to differ within the set target range between animals to prevent apnea in each animal at the 2 MAC level. The experimental protocols were designed to determine the amount of overall anesthetic-induced alterations of glutamatergic excitation and GABA^Aergic inhibition of expiratory neurons.

The experimental protocol is outlined in figure 2and was as follows: A minimum of 1 h was allowed for preparation stabilization before data collection. Once a stable signal from a single expiratory neuron was obtained, the neuronal activity was continuously recorded throughout the entire protocol. The peak neuronal discharge frequency, F  ⁿ, was measured for 10–20 respiratory cycles during the preejection control period (F  ^con). The GABA^Aantagonist, bicuculline, was then used to completely block the endogenous GABA^Aergic inhibition to the recorded neuron. The dose rates for bicuculline were increased in a stepwise manner until an additional increase in dose rate did not lead to a further increase in peak F  ⁿcompared with the prior dose rate. Typically picoejection durations of 5–10 min with increasing dose rates were required to achieve complete GABA^Aergic block, as evidenced by saturation of the neuronal response. Then complete postejection recovery was awaited. Control conditions were typically reestablished 20–30 min postejection of bicuculline. Thereafter the anesthetic depth was increased from 1.0 MAC halothane to 2.0 MAC and the same procedures were repeated at the new anesthetic depth once steady-state anesthesia (30 min equilibration) had been reached. The complete protocol, including anesthetic state 1 (1-MAC level), the anesthetic washin and washout between anesthetic states, anesthetic state 2 (2-MAC level), and return to anesthetic state 1 took about 4 h to complete.

The effect of halothane on peak discharge frequency, F  ⁿ, prior to and during GABA^Aergic blockade with bicuculline was quantified in the following manner: The averaged peak F  ⁿvalues for each condition were normalized to the peak activity of the neuron during maximal bicuculline block (F  ^e) at the lowest anesthetic dose (1 MAC). This represented the maximal discharge frequency obtained for each respective neuron and was assigned a value of 100%. A two-way, repeated measures analysis of variance, with main factors of anesthetic state (i.e. , 1 or 2 MAC halothane) and neurotransmitter status (preejection control vs.  maximal bicuculline block), was used to test for significant differences.

To analyze the effects of halothane on the overall excitatory and inhibitory synaptic transmission we used the neuronal model of figure 1. The index of overall inhibition, α, is obtained from the model definition (fig. 1, right) as follows:α=[F^e-F^con]/F^e, where F  ^conis the peak F  ⁿof the neuron during the prevailing basal control condition of GABA^Aergic attenuation. The values for F  ^eand F  ^conwere obtained for both anesthetic levels and were then used in the calculation of anesthetic-induced effects on the overall excitatory (ΔF^e) and the overall inhibitory neurotransmission (Δα) as summarized by equations 1 and 2 of table 1. All data values are given as mean ± SD, and P < 0.05 was used to indicate significant differences unless stated otherwise.

Figure 3illustrates the response of an expiratory neuron to picoejection of the GABA^Aantagonist bicuculline at 1 and 2 MAC halothane. Comparison of the time-expanded views of this data for the preejection control period (upper left) and during maximal bicuculline block (upper right) show that antagonism of GABA^Areceptors produces an increase in the expiratory-phase activity. Local application of bicuculline typically results in a discharge pattern that is an amplified replica of the underlying control pattern, which is a manifestation of gain modulation. The bicuculline dose-dependent responses (bottom records) show a steeper increase and a greater maximum increase in peak neuronal discharge frequency, F  ⁿ, at the maximal level of GABA^Aergic block for 1 MAC compared with 2 MAC. Quantification of these data is shown in figure 4, in which at 1 MAC halothane, the maximum bicuculline dose increased the peak discharge frequency, F  ⁿ, from 150 ± 9.9 Hz (F  ^con1) to 280 ± 12.6 Hz (F  ^e1), yielding an average α¹value of 0.46 (i.e. , α¹=(280 − 150)/280; see also table 1). This means that the tonic inhibition (gain modulation) attenuated the neuronal output by 46%. The corresponding peak F  ⁿvalues for 2 MAC were 119 ± 7.6 Hz (F  ^con2) and 154 ± 6.4 Hz (F  ^e2), yielding an average α²value of 0.23. Thus, for this neuron, the 1-MAC increase in halothane dose reduced the tonic inhibition by 50%(i.e. , Δα=(0.23 − 0.46)/0.46 =−0.50; see table 1). At the same time the excitatory activity, which was measured in the absence of tonic inhibition during maximal GABA^Aergic block with bicuculline, was reduced by 45%(i.e. , ΔF^e=(154 − 280)/280 =−0.45; see table 1). Because both excitatory and inhibitory synaptic drives were reduced by the anesthetic dose increase, overall neuronal peak, F  ⁿ, was only reduced from 150 to 119 Hz, or by 21%. This is markedly less than the 45% reduction in excitatory component of synaptic drive.

The pooled data from eight neurons with end controls are summarized in figure 5(top). At the 1-MAC halothane level complete GABA^Aergic block with bicuculline approximately doubled the control neuron activity both during the initial and the end control (1-MAC return) runs. The data suggest that these neurons were attenuated by endogenous GABA^Aergic inhibition by 48 ± 12%(i.e. , α= 0.48 ± 0.12) at the 1-MAC dose. At the 2-MAC halothane dose the control neuron frequency was only moderately decreased compared with the 1-MAC level (from 51.7 ± 11.6% to 43.5 ± 9.2%). Complete GABA^Aergic block at the 2-MAC level increased control neuron activity from 43.5 ± 9.2% to 64.8 ± 11.7%. This amounts to an α of 0.32 ± 0.13, which suggests that these neurons were attenuated by endogenous GABA^Aergic inhibition less at the 2-MAC dose (α= 0.32 ± 0.13) than at the 1-MAC dose (α= 0.48 ± 0.12). Thus, the anesthetic-induced change in overall inhibitory GABA^Aergic transmission, Δα, to these neurons was a reduction of 33.5 ± 17.2%(fig. 5, bottom). At the same time the overall excitatory neurotransmission, ΔF  ^e, was reduced by 35.1 ± 11.8%(fig. 5, bottom). For 12 neurons the average reductions of total excitatory and inhibitory neurotransmission by the 1-MAC dose increase were 38.3 ± 12.3% and 34.5 ± 16.1%, respectively. Four additional neurons are included in these latter data but not shown in the graphs of figure 5because end controls could not be not obtained because of loss of the neuronal signal. Because of electrode dislodgment, data at both 1 MAC and 2 MAC could only be obtained for 12 neurons in 15 dogs. Thus, these results suggest that the simultaneous reduction in inhibition by the 1-MAC halothane dose increase was able to preserve neuronal activity at levels much higher than would be expected after the anesthetic dose increase. There was no net enhancement of inhibition.

The vehicle in which the neurotransmitter antagonist bicuculline was dissolved did not show any effects on neuronal activity if picoejected at similar peak volume rates as bicuculline (e.g. , fig. 6). This demonstrates that not only the ejection rate but also the constituents of the artificial cerebrospinal fluid vehicle itself did not alter F  ⁿand bias our results. Such runs are important controls and were routinely performed for each experimental preparation.

These studies confirm our hypothesis that halothane's depressive effect on expiratory neurons in vivo  is caused by a depression of excitatory neurotransmission. Because the excitation of these expiratory neurons is mainly dependent on NMDA receptor activation, 15halothane appears to depress glutamatergic neurotransmission mechanisms. No enhancement of GABA^Aergic inhibitory mechanisms could be demonstrated with increasing depth of anesthesia from 1 to 2 MAC. These conclusions are based on our previously stated model assumptions that during the expiratory phase the main contributors to the discharge pattern are F  ^eand F  ⁱ(see fig. 1), and that the effects of F  ⁱwere completely antagonized by bicuculline (also see below). These results suggest that at deep levels of halothane (2 MAC) both excitatory and inhibitory neurotransmissions are depressed to a similar extent. This contrasts with in vitro  studies of strictly “postsynaptic” preparations that show that halothane enhances GABA^Aergic inhibitory currents 23–26but has little or no effect on glutamatergic receptor function at clinically relevant doses. 3,23Taken together, these findings suggest that halothane may alter presynaptic mechanisms that result in decreased neurotransmitter release of both glutamate and GABA in vivo . The combined effects of halothane on presynaptic and postsynaptic neuronal mechanisms determine the overall depressant effect of halothane on respiratory neuron activity. Although the present studies provide new information about the effects of halothane on overall excitatory and inhibitory synaptic control of respiratory neurons, clearly additional studies are necessary to determine the relative effects on presynaptic versus  postsynaptic mechanisms in this model.

Although in vivo  intracellular recordings have the advantage of detecting excitatory and inhibitory postsynaptic potentials, even during quiescent periods, such recordings are difficult to maintain under the best of conditions. Although extracellular recordings of spike discharge patterns reflect the excitability of a neuron only above its firing threshold, 27such recordings are much easier to obtain and hold for long (3–6-h) periods. It is possible, therefore, to study the effects of many different, systematically controlled interventions such as drug picoejections, changes in anesthetic depth, or alterations of afferent inputs. Usually such manipulations produce transient changes in blood pressure, which make intracellular recordings of these events very difficult. Although the picoejection technique can result in very localized agent application, it is recognized that presynaptic neurons may also be affected, which is also true for microiontophoresis or in vitro  bath application. A detailed discussion of these limitations can be found elsewhere. 19,28 

Our assumption that bicuculline produced nearly a complete block of the gain modulating input, F  ⁱ(fig. 1), has an effect on the accuracy of the amount of halothane-induced depression of both inputs. We feel relatively confident that the block was nearly complete because the bicuculline dose rate was increased until no further increase in F  ⁿwas observed compared to the prior dose rate. At the 2-MAC level the maximal dose rate was at least as large or larger than the maximal dose rate used at 1 MAC. It is possible that changes in halothane concentration may alter the affinity of bicuculline at the GABA^Areceptor. However, it is known that halothane binds to a site (or sites) on a GABA^Areceptor subunit that is distinct from the GABA recognition site. 29,30Thus, direct competition between halothane and bicuculline or GABA is not expected to occur. It has been suggested that the enhancement of GABA at the GABA^Areceptor by volatile anesthetics is caused by an increased affinity of GABA at its recognition site. 31It may be possible that the competitive antagonist, bicuculline, may be similarly affected.

Only a few studies describe the effects of volatile anesthetics on brainstem respiratory neurons. 21,22,32–37In most cases, well-defined, steady-state conditions, in terms of volatile agent concentration, were rarely established. Doi's group 36,37compared the dose-dependent depressant effects of three volatile anesthetics on inspiratory brainstem neurons in the cat, but the exact function of these neurons was not ascertained. Tabatabai et al.  33found that halothane decreased the discharge frequency of the inspiratory neurons in decerebrate cats throughout the anesthetic exposure and the anesthetic effects were readily reversible. Takeda et al.  34,35used short, 90-s exposures of 2% halothane in decerebrate cats while performing both intra- and extracellular recordings of propriobulbar respiratory neurons. This transient halothane exposure caused only a 25% depression in peak phrenic nerve activity, which is only a third of the magnitude of depression produced during steady-state conditions. Halothane decreased synaptic noise and membrane potential fluctuations in all of the respiratory neurons that were tested. This included both expiratory and inspiratory propriobulbar neurons, but their exact functions were not determined. Furthermore, the membrane responses were not uniform. A third of neurons were depolarized by the anesthetics even though neuronal discharge frequency decreased. About 15% of neurons were hyperpolarized and exhibited a uniform decrease in neuronal discharge frequency. However, in the majority of neurons, no change in membrane potential was observed. Grelot and Bianchi 32found that the combination of background anesthesia with a volatile anesthetic exerts profound depressant effects on distinct subgroups of brainstem respiratory neurons, where many of the neurons stopped firing already at halothane concentrations below 1 MAC.

In studies that employed anesthesia with a volatile agent as the sole anesthetic under steady-state conditions, Stuth et al.  delineated the dose-dependent depressant effects of halothane 21and isoflurane 22on the activity of canine respiratory premotor neurons at different levels of central chemodrive. These studies showed that the phrenic nerve discharge was depressed to a greater extent than that of the premotor neurons. A 1-MAC increase (1 MAC to 2 MAC) reduced peak phrenic activity by about 75%; peak neuronal activity was reduced to a much lesser extent. 21 

Ion-channel studies, performed with patch-clamp techniques in neuronal membrane patches or in isolated neurons in vitro,  have conclusively shown that halothane prolongs GABA^Areceptor-mediated chloride currents. 38This can enhance inhibition of neurons that receive GABA^Aergic inputs.

In addition, studies in isolated cerebral synaptosomes, a subcellular fraction containing pinched-off nerve terminals that retain the ability to take up, store, and release various neurotransmitters, have shown that halothane and also other volatile anesthetics inhibit endogenous glutamate release. 39This points to presynaptic inhibition of excitatory neurotransmitter release as a potential mechanism of anesthetic action. Halothane appeared to affect glutamate release at a step proximal to calcium influx that controls transmitter release, 39but this issue remains controversial. 40The overall effects of halothane are more variable in preparations in which both presynaptic and postsynaptic neural elements remain intact. For example, halothane depressed rather than enhanced GABA^Aergic inhibitory synaptic potentials in spinal motoneurons 41and in CA1 pyramidal cells, 42possibly by presynaptically inhibiting evoked GABA release. Recently, Perouansky et al.  43have shown in a hippocampal slice preparation that halothane blocked glutamate receptor-mediated activation of GABA^Aergic inhibitory interneurons. Both NMDA and non-NMDA subtypes of ionotropic glutamate receptors were nearly equally affected. In a similar preparation, this group previously showed that halothane had similar effects on both subtypes of glutamate receptor–mediated excitatory postsynaptic currents in CA1 pyramidal cells. 44They concluded that halothane indiscriminately reduces glutamatergic excitation, both of inhibitory interneurons and of principal excitatory neurons in the hippocampus. The studies also showed that only glutamatergic excitatory postsynaptic currents that were evoked by a presynaptic stimulus were depressed by halothane. Halothane did not reduce exogenously applied glutamatergic agonist–induced currents and did not alter the rise or decay times of non-NMDA or NMDA excitatory postsynaptic currents. This strongly suggests that the observed effects of halothane on excitatory neurotransmission occur at a presynaptic site. Findings suggestive of a presynaptic action of halothane were also observed in spinal motoneurons of decerebrate cats, in which halothane depressed excitatory postsynaptic potentials without changing their decay rate, indicative of a decrease in neurotransmitter release. 45 

In summary, picoejection of the GABA^Areceptor antagonist bicuculline in conjunction with extracellular recordings from brainstem ventral respiratory group expiratory neurons, in vivo , demonstrates that a 1-MAC increase in halothane concentration depresses both excitatory glutamatergic and inhibitory GABA^Aergic synaptic transmission. The data do not support the notion that the halothane-induced depression of respiratory neuronal activity is mediated via  enhancement of GABA^Aergic neurotransmission in this preparation.

The authors thank Jack Tomlinson for excellent technical assistance.