Effects of Halothane on Excitatory Neurotransmission to Medullary Expiratory Neurons in a Decerebrate Dog Model

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

Dogs were induced by mask with halothane and intubated with a cuffed endotracheal tube. The preparation has been described elsewhere in detail. 5Briefly, after bilateral vagotomy and bilateral pneumothorax to achieve peripheral deafferentation, an occipital craniotomy was performed to expose the medulla oblongata, and the phrenic nerve was prepared for recording. Phenylephrine (0.5–5 μg · kg⁻¹· min⁻¹) was usually required during the anesthetic runs to keep the mean arterial pressure above 75 mmHg. Protocols were performed only during steady-state conditions for blood pressure. Esophageal temperature was maintained at 37.5–38.5°C.

After completion of the occipital craniotomy, the animals were decerebrated as described by Tonkovic-Capin et al.  8Briefly, the procedure leads to an anatomically well-defined mid-collicular decerebration with modest blood loss (10 ml/kg). After completion of the decerebration, the anesthetic was discontinued. Pancuronium, 0.1 mg/kg followed by 0.1 mg · kg⁻¹· h⁻¹, was then given to avoid motion artifacts during neuronal recordings.

Expiratory neurons of the cVRG were located as described elsewhere. 4,9Multibarrel 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 potential activity before and during pressure ejection of respective neurotransmitters onto the neuron of interest. 2,3,5We used the NMDA receptor agonist N -methyl-d-aspartate (200 μm; Research Biochemicals, Natick, MA) and the GABA^Areceptor antagonist bicuculline methyl chloride (BIC, 200 μM; Research Biochemicals), which were dissolved in an artificial cerebrospinal fluid.The artificial cerebrospinal fluid consisted 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, and the pH was adjusted to 7.2–7.4 by aeration with 5% CO²/95% O². The drug-free artificial cerebrospinal fluid vehicle was routinely used as a control to demonstrate lack of vehicle effect (an example of such a drug-free artificial cerebrospinal fluid vehicle run is illustrated in our previous publication 5).

Extracellular activity from individual E neurons, phrenic nerve activity, a picoejection marker, airway carbon dioxide and volatile anesthetic concentration, systemic blood pressure, and airway pressure were recorded on a digital tape system. These variables or their time averages were also continuously displayed on a polygraph during the experiments. 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 offline.

A minimum of 1 h was allowed for preparation stabilization before data collection. All experiments were performed under hyperoxic (inspired oxygen fraction > 0.8) and steady-state hypercapnic conditions (target partial pressure of alveolar carbon dioxide [Paco²], 50–60 mm Hg). This level of hypercapnia was chosen to ensure adequate phrenic nerve activity during halothane application and to allow comparison with our previous studies. Target Paco²was held constant for all experimental runs within an animal, but was allowed to differ within the set target range among animals.

After a stable signal from an E neuron was obtained, the neuronal activity was continuously recorded throughout the entire protocol. The peak neuronal discharge frequency (Fⁿ) was measured for 10–20 respiratory cycles during a preejection control period (F^con). Then the GABA^Aantagonist BIC was picoejected until complete block of GABAergic inhibition occurred, i.e.,  until an increase in dose rate did not achieve any further increase in discharge frequency. Typically, picoejection durations of 5–10 min with increasing dose rates were required. After this run, complete postejection recovery was awaited (30–45 min). Thereafter, the anesthetic was introduced to a depth of 1 MAC halothane (0.9 vol%), and after an equilibration time of 20 min, the same procedure was repeated. Peak dose rates during anesthesia always had to match or exceed those during the 0-MAC run.

The complete protocol, including the 0-MAC level, the anesthetic wash-in, the 1-MAC level, and the return to 0-MAC level (anesthesia-free end control), required about 4 h.

The method of analysis was the same as in our previous publication. 5All data were normalized to the peak activity of the neuron during maximal BIC block at the 0-MAC level. This represented the maximal discharge frequency for each neuron and was assigned a value of 100%. A two-way, repeated measures analysis of variance with main factors of level of anesthesia (0 or 1 MAC) and neurotransmitter concentration (preejection control vs.  maximal BIC block) was used.

To determine the prevailing GABA^Aergic inhibition, we defined the peak discharge frequency of the neuron (Fⁿ) for the preejection control period, F^con, and for the maximal frequency under BIC block (F^e). F^eis thus a measure of the total excitatory drive to the neuron, whereas F^conrepresents this drive reduced by the prevailing basal GABAergic inhibition.

We described the level of prevailing GABAergic inhibition by the inhibitory constant α and defined it as:α=[F^e− F^con]/F^e. The values for F^eand F^conwere obtained for the 0-MAC level (F^con0, F^e0) and the 1-MAC level (F^con1, F^e1) in the experimental protocols. These were then used in the calculation of anesthetic-induced effects on overall excitatory (ΔF^e) and overall inhibitory (Δα) neurotransmission, as described in equations 1 and 2 of table 1.

The peak neuronal discharge frequency (Fⁿ) was measured for 10–20 respiratory cycles during the preejection control period (F^con). Then the glutamate receptor agonist NMDA was applied in increasing dose rates until an increase of neuronal frequency of at least 40 Hz was achieved. Typically, picoejection durations of 6–8 min with two to three dose rates were needed. Postejection recovery was achieved within 5 min. The run was repeated at the 1-MAC level and again at the 0-MAC end-control level.

The effect of the glutamatergic agonist NMDA on E neuron frequency was quantified in the following manner: The picoejection dose rate that caused a 40-Hz net increase in neuron activity in the absence of halothane was determined from the NMDA dose–response curve and designated D^40Hz. To confirm the linearity of the NMDA dose–response curves in this range, the net increase at one half the D^40Hzdose rate was determined and designated 1/2 D^40Hz. The corresponding net increases of neuronal activity then were determined for 1 MAC halothane. All neuron response data were then normalized to the 40-Hz net increase at 0 MAC, which was assigned a value of 100%. A two-way, repeated measures analysis of variance with main factors level of anesthesia (0 or 1 MAC halothane) and neurotransmitter status (preejection control vs.  NMDA response) was used.

All data values are given as mean ± SD, and P < 0.05 was used to indicate significant differences unless stated otherwise.

Typically, one complete protocol, consisting of 0-MAC, 1-MAC, and 0-MAC end-control levels, could be obtained per animal. Data from animals with incomplete protocols were not included.

Seventeen animals were studied, and complete protocols were obtained in 14 dogs. Figure 1shows a representative example of a BIC protocol. During both anesthetic states (0 and 1 MAC), complete BIC block caused approximately a doubling of the neuron baseline activity: At 0 MAC halothane, maximal BIC increased peak Fⁿfrom 133 to 227 Hz, yielding an inhibitory factor of α⁰=[F^e0− F^con0]/F^e0, i.e., α⁰= (227 − 133)/227 = 0.42 (fig. 2). This means that the GABAergic inhibition attenuated the neuronal output by 42%. The corresponding peak Fⁿvalues for 1 MAC halothane were 100 and 183 Hz, yielding an α¹= 0.46. Thus, for this neuron, 1 MAC halothane enhanced the tonic inhibition by 9.7%[i.e., Δα= (α¹−α⁰)/α⁰= (0.46 − 0.42)/0.42 = 0.097].

At the same time, the excitatory activity, which was measured in the absence of inhibition (i.e.,  during maximal BIC block) was reduced by 19%. Thus, excitatory synaptic drive was reduced by the anesthetic, whereas the inhibitory drive was slightly enhanced. In accordance, control peak discharge frequency, F^con, was reduced by 25%.

The pooled data from 14 neuron protocols with end-control runs are shown in figure 3. Analysis of the pooled data, as exemplified in figure 2, shows that 1 MAC halothane caused a small but significant increase in the α values (from α= 0.54 ± 0.11 at 0 MAC to 0.59 ± 0.12 at 1 MAC). This represents a 11.7 ± 18.3% enhancement in overall GABAergic inhibition (Δα;fig. 3, bottom middle bar).

At the same time, 1 MAC halothane caused a significant 31.5 ± 15.5% decrease in overall excitation (ΔF^e;fig. 3, bottom left bar). Peak F^conwas significantly depressed from 46.3 ± 10.8 to 27.5 ± 9.2 (F^con;fig. 3, upper bar), or by 38.6 ± 20.6% (ΔF^con;fig. 3, bottom right bar) by 1-MAC halothane.

Sixteen animals were studied, and 13 complete protocols were obtained. Figure 4shows a representative example of an E neuronal response to picoejection of NMDA before and during 1 MAC halothane. At 1 MAC halothane, F^conwas reduced from 188 to 145 Hz (fig. 4, preejection period, and fig. 5, left). In this example, D^40Hzat 0 MAC was 3.9 pm NMDA/min. The net frequency increase at D^40Hzat 1 MAC halothane was 39.4 Hz (fig. 5, right). The 1/2 D^40Hzvalues were 23.2 Hz at 0 MAC halothane and 21.2 Hz at 1 MAC halothane (fig. 5, left and right).

The pooled normalized baseline data for 13 complete neuron protocols (0 MAC − 1 MAC − 0 MAC) of study 2 again show that halothane significantly depresses F^con27.9 ± 10.6% (data not shown). This level of depression is not statistically different from the 38.6 ± 20.6% depression of F^confound in study 1 (unpaired t  test). End-control F^conwas slightly depressed to 91.8 ± 5.0%. This may suggest incomplete recovery or a small deterioration of the preparation over time.

The pooled NMDA net increase data are summarized in figure 6, top, and the normalized data are shown in figure 6, bottom. halothane reduced the net increase from 100% to 96.7 ± 38.4% at D^40Hz, which is not statistically significant (fig. 6, lower right hatched bar). The 0-MAC end-control value for D^40Hzwas 108 ± 48.8%, which was not significantly different from the 0-MAC control. The corresponding normalized 1/2 D^40Hzdata were not statistically different (fig. 6, lower left). Slope analysis for 1/2 D^40Hzand D^40Hzvalues confirmed linearity over the full dose range.

The peak phrenic nerve activity was significantly depressed by 1 MAC halothane. Peak phrenic nerve activity was reduced by 47.8 ± 19.8% (mean ± SD) by 1 MAC halothane, whereas neural respiratory rate increased from 15.4 ± 8.4 to 23.2 ± 13.2 breaths/min.

Results of the current studies in a decerebrate canine preparation lead to the conclusion that the depressive effect of halothane on cVRG E neurons in vivo  is predominantly the result of a presynaptic reduction of glutamatergic excitatory neurotransmission. Our methodology, however, does not allow us to pinpoint specific presynaptic sites.

The lack of effect of halothane on postsynaptic glutamatergic neurotransmission (e.g.,  fig. 6) is in agreement with the in vitro  studies by Perouansky et al.  10Using whole-cell patch clamp recordings in CA1 pyramidal cells in mouse hippocampal slices, they showed that halothane depressed glutamate receptor–mediated excitatory postsynaptic potentials evoked by stimulation of afferent fibers. Halothane did not influence either the NMDA receptor–mediated or the non–NMDA receptor–mediated excitatory postsynaptic potentials evoked by bath application of the respective agonists and did not alter the receptor kinetics. The effect of halothane thus occurred most likely by inhibition of presynaptic glutamate release. 10In support of this possible mechanism, halothane has been shown to reduce glutamate release from cerebral synaptosomes. 11 

We found a small but significant enhancement of overall GABAergic inhibition at 1 MAC halothane. This is in general agreement with recent in vitro  studies on the effects of inhalational anesthetics on the miniature inhibitory postsynaptic currents in CA1 pyramidal neurons. 12Halothane reduced the amplitude but increased the duration of the miniature inhibitory postsynaptic currents, which suggests a reduction in the presynaptic release of GABA but an enhanced postsynaptic action of GABA. The overall effect was a greater than 100% increase in the total charge transferred during the miniature inhibitory postsynaptic currents, or net increase in GABAergic inhibition. With regard to our E neuronal study, it is possible that presynaptic GABAergic mechanisms were depressed in a manner similar to that of glutamate. The combination of reduced presynaptic transmitter release with an enhanced postsynaptic action could have resulted in the small overall GABAergic enhancement we observed.

Our previous data in a neuraxis-intact canine preparation, however, showed that a 1-MAC increase in halothane, from 1 MAC to 2 MAC, produced a 33.5 ± 17.2% reduction in overall GABAergic inhibitory neurotransmission. 5Although we sought to resolve this issue in our decerebrate preparation, we were unable to do so, because most animals became apneic at levels higher than 1.5 MAC halothane, even in the presence of significant hypercapnia.

It thus seems to be necessary to view modification of overall excitation and inhibition beyond synaptic actions. The observed apnea at a lower halothane level than in the intact preparation suggests that in the decerebrate preparation a suprapontine descending excitatory input to the rhythm-generation network has been eliminated.

Consideration of the functional network also helps to explain the seemingly contradictory effects of 1 and 2 MAC halothane on overall GABAergic inhibition. Perouansky et al.  13have shown that in a mouse hippocampal slice preparation halothane dose dependently depressed the glutamatergic excitation of inhibitory interneurons. An equivalent of 2 MAC halothane depressed the excitation of the interneuron by 50% and consequently would have reduced the inhibition of the secondary pyramidal cell.

Banks and Pearce 12showed for CA1 pyramidal neurons that the increase in postsynaptic GABA receptor function by volatile anesthetics peaked at the equivalent of 1 to 2 MAC. We speculate that in our in vivo  preparation at 1 MAC halothane, a near-maximal receptor-mediated increase in inhibition was present and accounted for the increase in overall inhibition to the E neurons. At 2 MAC halothane, however, the anesthetic may have decreased the presynaptic component of inhibitory input to an extent that the net effect was a reduction of overall inhibition.

Clearly, additional studies are warranted to determine the relative magnitude of presynaptic and postsynaptic GABAergic effects of halothane.

We have pointed out some of the limitations of our methodology in our previous publication 5and a detailed discussion is provided elsewhere. 3,14The decerebrate canine preparation was developed to study the effect of volatile anesthetics per se  on respiratory neurotransmission in vivo .

We found that respiratory patterns were similar to those in the intact animal. Decerebrate, vagotomized dogs, in the absence of anesthesia, had a respiratory rate of 15.4 ± 8.4 breaths/min compared with awake, resting animals (23.0 ± 6.3 breaths/min, our observation from 16 dogs at rest). Decerebrate dogs also show the tachypnea that has been typically associated with halothane at clinical levels. Respiratory rates during 1 MAC halothane were not different from our 1-MAC data obtained in intact animals (23.3 ± 13.2 vs.  21.3 ± 7.1 breaths/min). 4This data is different from that of Gautier et al.  15who described a marked reduction of halothane-induced tachypnea in decerebrate compared with intact cats and related this to a suprapontine origin of the halothane tachypnea.

Decerebration removes forebrain and midbrain influences that modify ventilatory responses. 16It is possible that the elimination of suprapontine excitatory input to the rhythm-generation network allowed apnea to occur at a lower halothane concentration than in neuraxis-intact dogs. This might explain why most of our decerebrate dogs became apneic at 1.5 MAC halothane even under hypercapnia.

However, for these cVRG E neurons, similar anesthetic-induced reductions in net excitatory drive were observed in both preparations. 5Furthermore, we found that the basic reflex inputs that we previously described in the intact preparation, including chemodrive 2and pulmonary vagal inputs, are preserved (unpublished data). Nielsen et al.  17and Mitchell 18found that respiratory rate, inspiratory duration, and expiratory duration responded to various stimuli, such as bilateral vagotomy, hypercapnia, or hypoxia, but such responses were diminished in decerebrate dogs compared with chloralose-anesthetized dogs.

In the current study, 1 MAC halothane reduced peak phrenic activity (index of tidal volume) to 52.2 ± 19.9% of control (0 MAC). This magnitude is very similar to the estimated reduction of tidal volume in awake, spontaneously breathing dogs produced by 1 MAC halothane. 19 

The current studies were performed in a newly developed decerebrate canine preparation that allows the study of volatile anesthetics on brain stem respiratory neuronal activity without the confounding effect of a background anesthetic. The studies confirmed that the depressive effect of halothane in cVRG E neurons was mainly the result of a reduction of synaptic excitation with an additional small enhancement of synaptic inhibitory mechanisms. Furthermore, because glutamatergic excitation of postsynaptic neuronal receptors was not depressed by halothane, the main depressant effects of the anesthetic halothane appear to occur at presynaptic sites.

Thus, this new decerebrate model provides additional evidence for the importance of alterations of glutamatergic neuronal transmission by volatile anesthetics in functional neural networks in vivo .

The authors thank Jack Tomlinson, Medical College of Wisconsin and Veterans Administration Medical Center, Milwaukee, Wisconsin, for excellent technical assistance.