Isoflurane Differentially Modulates Inhibitory and Excitatory Synaptic Transmission to the Solitary Tract Nucleus

All animal procedures were performed with the approval of the Institutional Animal Care and Use Committee at Oregon Health and Science University, Portland, Oregon, and conform to the guidelines of the National Institutes of Health publication Guide for the Care and Use of Laboratory Animals .

In all experiments, adult male Sprague-Dawley rats (>160 g; Charles River Laboratories, Inc., Wilmington, MA) were used. Animals were housed under 12-h light/12-h dark conditions and were fed standard pellet chow ad libitum .

Brain stem slices were prepared from isoflurane anesthetized rats as previously described.15Briefly, the medulla was removed and placed in cooled artificial cerebrospinal fluid (aCSF) containing 125 mm NaCl, 3 mm KCl, 1.2 mm KH²PO⁴, 1.2 mm MgSO⁴, 25 mm NaHCO³, 10 mm dextrose, and 2 mm CaCl², bubbled with 95% O²–5% CO². The tissue was trimmed rostrally and caudally to yield a tissue block centered on obex. To orient the ST axons with the NTS in a common plane for cutting, the ventral surface of the brain stem block was cut. A single 250-μm-thick horizontal slice resulted that contained the ST together with the neuronal cell bodies of the medial NTS region. Slices were cut with a sapphire knife (Delaware Diamond Knives, Wilmington, DE) mounted in a vibrating microtome (VT1000S; Leica Microsystems Inc., Bannockburn, IL). Slices were secured with a fine polyethylene mesh in a custom perfusion chamber (Siskyou Design Instruments, Grants Pass, OR) and perfused with aCSF constantly bubbled with 95% O²–5% CO²at 32°C and 300 mOsm.

The anatomical landmarks preserved in the horizontal slices allowed targeting of neurons within the medial subnucleus of the caudal NTS. Recorded neurons were located medial to the ST and within 200 μm rostral of obex. Patch electrodes were visually guided to neurons using infrared illumination and differential interference contrast optics (40× water immersion lens) on an Axioskop 2 microscope (Zeiss, Thornwood, NJ) with digital camera (Hamamatsu Photonic Systems, Bridgewater, NJ). Recording electrodes (3–4 MΩ) were filled with one of two intracellular solutions depending on the experimental protocol. For miniature inhibitory postsynaptic current (mIPSC) measurements, a high-Cl⁻(54 mm, E^Cl=−25 mV) intracellular solution was used containing 10 mm NaCl, 40 mm KCl, 70 mm K-gluconate, 11 mm EGTA, 1 mm CaCl², 1 mm MgCl², 10 mm HEPES, 2 mm Na²ATP, and 0.2 mm Na²GTP. For stimulated excitatory postsynaptic current (EPSC) and for miniature excitatory postsynaptic current (mEPSC) measurements, a low-Cl⁻(10 mm, E^Cl=−69 mV), intracellular solution contained 6 mm NaCl, 4 mm NaOH, 130 mm K-gluconate, 11 mm EGTA, 1 mm CaCl², 1 mm MgCl², 10 mm HEPES, 2 mm Na²ATP, and 0.2 mm Na²GTP. Intracellular solutions were pH 7.3 and 296 mOsm. All neurons were studied using voltage clamp conditions with an Axopatch 200A or MultiClamp 700B amplifier (Axon Instruments, Foster City, CA) and held at V^H=−60 mV using pipettes in open, whole cell patch configuration. Signals were filtered at 10 kHz and sampled at 30 kHz using p-Clamp software (version 8.2; Axon Instruments).

A concentric bipolar stimulating electrode (200-μm outer tip diameter; Frederick Haer Co., Bowdoinham, ME) was placed on distal portions of the visible ST rostral to the recording region. This remote placement of the stimulating electrode (1–3 mm) from recorded neurons minimized the likelihood of activation of non-ST axons or local neurons.15Consistent current shocks were delivered to ST every 3 s (shock duration 0.1 ms) using a Master-8 isolated programmable stimulator (A.M.P.I., Jerusalem, Israel). In some assays, responses to bursts of five shocks (20-ms intershock interval) assayed frequency-dependent amplitude depression and synaptic failures for ST-evoked EPSCs. Latency was measured as the time between the ST shock and the onset of the resulting EPSC.15Synaptic variability was calculated as the SEM of 30–40 ST-evoked EPSC latencies within each neuron. An SEM of less than 200 μs reliably identified direct, monosynaptic afferent contacts.15 

All drugs were administered through the aCSF bath perfusion. GABAergic neurotransmission was isolated by antagonizing ionotropic glutamate receptors (both N -methyl-d-aspartate and α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid) with D-2-amino-5-phosphonovalerate (AP-5, 100 μm) and 2,3-dihydroxy-6-nitro-7-sulfonyl-benzo[f ]quinoxaline (NBQX, 20 μm), respectively. Glutamatergic neurotransmission was isolated by administration of gabazine (SR95531, 3 μm) to block synaptic GABA^Areceptors. Nonspecific GABA and glycine receptor antagonists picrotoxin (100 μm; Sigma, St. Louis MO) and bicuculline (100 μm; Sigma) were administered to block all GABAergic receptor activity. Miniature events were defined by the presence of tetrodotoxin (3 μm). All drugs were purchased from Tocris (Ellisville, MO) unless otherwise noted. Isoflurane was purchased from Abbott Labs (N. Chicago, IL).

A stock of saturated isoflurane solution was produced by mixing equal volumes of isoflurane and aCSF in a sealed volumetric flask at room temperature. This saturated isoflurane stock measured approximately 12–14 mm and for experiments was then diluted to nominal final concentrations between 10 μm and 1 mm. For each exposure, isoflurane was perfused for 5 min at a flow rate of 1.5–2 ml/min. Bath samples (100 μl) were collected at four time points for each experiment; one baseline sample was taken 2 min before isoflurane, two samples were taken during isoflurane exposure (at 2 and 5 min), and one recovery sample was taken 2 min after return to bath solution. Slices were washed for 20 min after isoflurane exposure. All samples were extracted using glass Hamilton syringes and immediately added into glass vials, with heptane (500 μl) containing halothane as an internal standard, and capped. Through the use of sealed glass syringes and Teflon tubing (Small Parts Inc., Miami Lakes, FL), the perfusion system was optimized to prevent loss of isoflurane from the apparatus. Consistent amounts of isoflurane were delivered over time using this closed system such that only the bath was open to the atmosphere. Even so, the actual bath concentration of isoflurane varied from exposure to exposure, with the greatest variability occurring at the highest concentrations tested. As such, bath samples were taken for every isoflurane exposure to accurately attribute phenomena observed to the actual isoflurane concentration.

Gas chromatography–mass spectrometry was used to accurately measure isoflurane concentrations of samples taken from the bath. Analyses were based on Kharasch et al.  16and were performed using a Finnigan TRACE DSQ GC/MS system (Thermo Electron Corporation, Austin, TX) configured with an AS 2000 autosampler and a split/splitless injector. Isoflurane and halothane were separated on an AT-5ms fused silica capillary column (30 m × 0.25 mm ID, 0.25 μm film thickness, Alltech Associates, Deerfield, IL) using helium as the carrier gas at a constant flow rate of 1 ml/min. Samples (1 μl) were injected in the split mode with a split rate of either 1:100 or 1:200 depending on the concentration. The injector was maintained at 150°C, and the ion transfer line was maintained at 250°C. The Finnigan TRACE GC Ultra oven was held constant at 40°C for 3 min, and then the temperature increased at 40°C/min to 120°C. The mass spectrometer was operated in the electron impact mode with an ionization energy of 70 eV and a source temperature of 250°C. Data were acquired in the selected ion monitoring mode. The ions at m/z = 117 and 148 for isoflurane and m/z = 117 and 198 for halothane were used for quantitation. Typical retention times were 1.61 min for isoflurane and 1.93 min for halothane. Instrument control and data acquisition and analysis were accomplished using Xcalibur Version 1.4 software (Thermo Scientific, Waltham, MA). Samples were analyzed and area ratios were compared with a standard calibration curve prepared from known amounts of isoflurane. For isoflurane concentrations in the range of 100–1,000 μm, 500 μm halothane was used as the internal standard, and for concentrations in the range of 10–100 μm, 50 μm halothane was used.

Experimentally one brain slice from each animal was obtained, and from each brain slice a single neuron was recorded. For statistical testing, each observation was considered independent and grouped into a fixed concentration. Concentration groups were compared using unpaired, non–repeated-measure assumptions. Digitized waveforms were analyzed using an event detection and analysis program (MiniAnalysis; Synaptosoft, Decatur, GA) for all miniature synaptic currents and Clampfit 10 (Axon Instruments) for all ST-stimulated currents. To obtain accurate decay-time constants, amplitude, and baseline values for miniature synaptic currents, those events smaller then 15 pA and those with multiple peaks were excluded. All events greater than 15 pA were counted for frequency values. Decay-time constants were acquired by fitting a single exponential between the 10% and 90% peak amplitude portion of the current decay. Isoflurane concentrations for each 5-min exposure were averaged across the two samples taken (2 and 5 min) and were fixed into concentration ranges accordingly. Statistical comparisons of waveform parameters (decay-time constant, amplitude, area, and baseline) were averaged across each concentration group and compared by a one-way analysis of variance each with Bonferroni post hoc  testing (SigmaStat, San Jose, CA). Comparisons of interevent interval within individual neurons was nonparametric and used a Kolmogorov-Smirnov test. All group data are represented as the normalized mean ± SEM with P < 0.05 considered statistically significant.

Our studies focused on NTS neurons from the medial subnucleus (fig. 1) because this region is a common site of cardiorespiratory afferent terminations.11The horizontal slice preparation retained key anatomical landmarks including the fourth ventricle and ST that guided recording electrodes to this medial region (fig. 1A) while infrared illumination and differential interference contrast optics outlined individual neurons for recording (fig. 1B). After stable intracellular recordings were established, we identified neurons that were directly contacted by primary visceral afferents by their characteristic responses to orthodromic ST activation. Electrical shocks remotely delivered to the ST recruited EPSCs that were nearly invariant in their latency, shape, and amplitude, and rarely failed (<0.1%) (fig. 1C). Latency values from repeated trials within such neurons produced narrow distributions whose synaptic variability (SEM of latency) was less than 200 μs and thus were considered to have direct ST afferent connections (second order). Peak amplitude averaged 247 ± 29 pA, and latency averaged 5.05 ± 0.45 ms (range, 2.67–7.29 ms) with an average SEM of 110 ± 20 μs in the 78 neurons included in these studies. As previously reported for second-order NTS neurons, stimulus trains of five ST shocks (20-ms interval) produced substantial frequency-dependent depression of EPSC amplitude (fig. 1D), and EPSCs were blocked by ionotropic glutamate receptor antagonists NBQX (20 μm) and AP-5 (100 μm) (data not shown). All NTS neurons included in the current study were initially characterized in this manner as second order by ST testing.

To investigate the effects of isoflurane on phasic inhibitory events, tetrodotoxin and ionotropic glutamate receptor blockers were used to eliminate EPSCs, leaving mIPSCs isolated for study (n = 30 neurons, fig. 2A). Bath application of isoflurane prolonged the decay phase of mIPSCs without changes in amplitude or frequency (fig. 2A). Aggregate waveforms indicated prolongation of the decay phase beginning at 118 ± 8 μm isoflurane, and at very high concentrations (836 ± 80 μm), isoflurane depressed the amplitude as well as prolonged the decay phase (fig. 2B). On average, the mIPSC decay-time constant and total integrated area of the mIPSCs increased in a concentration-dependent manner (fig. 2C, top and bottom panels, n = 3–6 neurons/concentration, P < 0.001 and P < 0.05, respectively). In the most sensitive neurons, 50 μm isoflurane shifted the mIPSC time constant. The mIPSC amplitude (fig. 2C, middle panel) decreased significantly at nominally 1,000 μm isoflurane (control: 36 ± 8 pA vs.  isoflurane: 28 ± 7 pA, n = 6 neurons, P < 0.05). Isoflurane did not to alter the frequency of mIPSCs (figs. 3A and B) at any concentration tested (n = 3–6 neurons/concentration, P > 0.05) (fig. 3C). In these adult animals, all mIPSCs were blocked by gabazine (3 μm), suggesting that glycinergic events are absent in adult medial NTS neurons (not shown). Holding currents were uniformly small and unaltered by isoflurane despite substantial Cl⁻driving force (E^Cl=−25 mV; 6 ± 4, 4 ± 3, 5 ± 4, and 6 ± 1 pA observed in 50 ± 5, 118 ± 8, 310 ± 39, and 782 ± 83 μm isoflurane, respectively). These findings suggest that isoflurane acts postsynaptically to enhance inhibitory transmission chiefly by slowing the kinetics of phasic, GABA^A-mediated events with no discernible presynaptic actions on GABA release or postsynaptic actions on “extrasynaptic” tonic GABA^Areceptors.

In second-order NTS neurons (n = 21 neurons), ST shocks activate large EPSCs with consistent amplitudes over time (fig. 4A). Isoflurane rapidly and reversibly reduced ST-evoked EPSC amplitudes and, at high concentrations, produced synaptic failures (zero amplitude) (fig. 4A). Note that full GABAergic and glycinergic blockade had no effect on ST-EPSC amplitude. On average, ST-evoked EPSC amplitude and integrated current were significantly depressed by 244 ± 35 μm isoflurane (fig. 4B and C, middle panel, n = 3–8 neurons/concentration, P < 0.05). Isoflurane had no effect on ST-evoked EPSC kinetics, and the decay-time constant remained unchanged at all isoflurane concentrations tested (fig. 4C, top panel, n = 3–8 neurons/concentration, P > 0.05). Isoflurane reversibly increased ST-evoked EPSC latency in 6 of 10 neurons tested with greater than 300 μm isoflurane (fig. 5A), indicating decreased ST axonal conduction velocity. The average change in latency was 550 ± 60 μs (n = 6 neurons), represented by a rightward shift in the latency distribution as well as a flattening of the distribution (fig. 5B). In half (5 of 10) of the neurons tested, the higher isoflurane concentrations (>300 μm) induced complete ST-evoked EPSC failures (fig. 5A). Therefore, isoflurane decreases ST-evoked EPSC amplitude and alters conduction properties of ST afferent axons.

To test for presynaptic and postsynaptic actions on glutamatergic transmission, mEPSCs were isolated in second-order NTS neurons (n = 27 neurons) using tetrodotoxin and combined GABA/glycine receptor antagonism. Isoflurane rapidly decreased the frequency of mEPSCs (figs. 6A and B) and shifted the frequency distribution to higher interevent intervals (fig. 6C). On average (n = 5–6 neurons/concentration, P < 0.001), isoflurane decreased the frequency of mEPSC events significantly from approximately 150 μm isoflurane and higher, with maximal suppression occurring at 687 μm (n = 5–6 neurons/concentration, P < 0.05 vs.  control, fig. 6D). The average frequency of mEPSCs under control conditions was 11.01 ± 3.01 Hz, where 700 μm isoflurane reduced the rate by 35% to 7.46 ± 2.51 Hz. This effect of isoflurane was reversed during wash with aCSF (fig. 6B). In contrast to mIPSCs, the kinetic properties of mEPSCs were unchanged in isoflurane (fig. 7). Neither the decay-time constant nor the amplitude of mEPSCs changed at any concentration of isoflurane tested (n = 5–6 neurons/concentration, P > 0.05, figs. 7B and C). Miniature EPSC events were blocked by ionotropic glutamate receptor antagonism (data not shown). Together, the mEPSC and ST-evoked EPSC observations suggest selective presynaptic actions of isoflurane on glutamate release, and indicate differences between the presynaptic release process for glutamate compared with GABA.

The current studies focused on neurons within the medial subregion of the NTS, which is known to receive direct cardiorespiratory afferent innervation,12,17with the goal of assessing the integrated actions of isoflurane on GABA- and glutamatergic synaptic transmission. This in vitro  slice preparation and electrophysiologic approach allow for the determination of both presynaptic and postsynaptic mechanisms of isoflurane to modulate the ST–NTS synapse in these autonomic reflex circuits of adult animals.18,19Isoflurane dose-dependently enhanced GABA-mediated inhibitory synaptic events while depressing glutamate-mediated excitatory neurotransmission to these NTS neurons. Our simultaneous sampling of the perfusion bath matched isoflurane concentrations to the neuronal effects. Significant IPSC enhancement occurred by 118 μm, whereas EPSC transmission depressed beginning at 166 μm or greater. The threshold concentrations for these effects were less than 1 minimum alveolar concentration for the rat (approximately 310 μm).20The changes in decay kinetics of mIPSCs and lack of an effect on event frequency support postsynaptic sites of isoflurane action at GABA^Areceptors. In contrast, decreased mEPSC frequency combined with no change in event kinetics indicates that glutamatergic transmission was depressed by presynaptic processes that included decreased synaptic glutamate release, independent of action potentials. Further, increased latency of ST-evoked EPSCs and event failures suggest disrupted axonal conduction likely via  voltage-dependent ion channels along ST afferents.5,21Taken together, our results indicate that isoflurane attenuates synaptic neurotransmission within the NTS via  multiple distinct mechanisms that are within the clinically relevant concentration range.

As commonly reported for general anesthetic actions at GABA^Areceptors, we found that isoflurane significantly increased the total integrated current of phasic IPSC events through a prolongation of the event decay-time constant. This effect on the IPSC decay rate was relatively selective, considering that isoflurane did not significantly alter event amplitudes or frequency at moderate concentrations. Although somewhat variable across individual neurons, 118 μm isoflurane increased the decay-time constant of these phasic events, on average, by more than 50% and in a concentration-dependent manner. The sensitivity of mIPSCs in NTS neurons was greater than in hippocampal neurons, which show significant changes in mIPSC kinetics at 300 μm.22At higher isoflurane concentrations (>300 μm), the decay-time constant increased dose dependently. In contrast, the relative increase in integrated current was maximal at 118 μm isoflurane and did not increment with further increases in concentration. This difference may result from the concomitant decrease in event amplitude at high concentrations. Banks et al.  and others report similar findings in other neurons and implicate a direct GABA^Achannel block.22–24 

In contrast to the changes in kinetic parameters, isoflurane did not change the mIPSC frequency; findings consistent with selective postsynaptic actions at GABA^Areceptors on NTS neurons. Direct measurements of transmitter release from isolated cortical synaptosomes indicate that high concentrations of isoflurane (IC⁵⁰= 1,000 μm) can decrease GABA release but to a lesser extent than glutamate release.25–27Interestingly, our data in NTS neurons found that isoflurane selectively decreased glutamate release (see discussion below) but not GABA release. In contrast to results from neurons in the hippocampus,28isoflurane did not alter the holding current or activate a tonic, gabazine-sensitive current even at high concentrations. We have observed the activation of tonic GABA^Acurrents in similar second-order NTS neurons by the intravenous anesthetic propofol.29This differential action of isoflurane on phasic and tonic GABA^Amechanisms may reflect the particular heterogeneous expression and assembly of native GABA receptor complexes across this region. Although speculative at this time, an interesting possibility in NTS might be ρ¹GABA^Asubunits, which are expressed in NTS30and may be resistant to isoflurane.31 

Electrical stimulation of the ST produces large-amplitude, low-variability EPSCs which reflect visceral afferent to NTS neuronal communication.18Non–N -methyl-d-aspartate receptors mediate visceral afferent synaptic transmission in NTS.32,33We found that isoflurane dose-dependently decreased the amplitude and integrated current of ST-evoked EPSCs. These decreases began at 280 μm, and by 837 μm isoflurane, EPSC amplitude and integrated current were reduced by 33%—similar to findings in other neurons.34–36Isoflurane acted rapidly and reversed after wash. Postsynaptic modulation of glutamate receptors and/or presynaptic inhibition of vesicular glutamate release may contribute to attenuation of EPSC signaling. In NTS neurons, postsynaptic glutamate receptor kinetics were remarkably isoflurane resistant, consistent with work on expressed glutamate receptors in which isoflurane even at 1,200 μm did not alter α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor responses.37Considering that isoflurane decreased mEPSC frequency beginning at 166 μm and had no affect on the amplitude or kinetic parameters at any concentration tested, we conclude that inhibition of glutamatergic transmission by isoflurane resulted from presynaptic decreases in glutamate release.

Attenuation of glutamate release from presynaptic terminals may result from modulation of the synaptic bouton or altering the propagating action potential. Even when action potentials were blocked by tetrodotoxin, isoflurane still decreased the frequency of mEPSCs. This is consistent with isoflurane acting at the synaptic bouton to disrupt constitutive vesicular docking and release independent of action potentials. At calyx of Held synapses, isoflurane inhibited both the synaptic vesicular release and the propagating action potential.36We observed both increases in the latency to onset and failures of ST-evoked EPSCs in a majority of neurons at high concentrations of isoflurane (>300 μm). Such observations are consistent with conduction delays and block, respectively, in the ST axons leading to synaptic terminals. Such delayed arrival and/or blockade of transmission along ST afferents can occur independently of changes in the probability of glutamate release, a property of the synaptic terminals themselves.38Afferent axonal conduction is a voltage-dependent process that relies strongly on fast sodium and potassium channels, both of which have unique expression patterns that differ between peripheral cranial afferent neurons and central NTS neurons.39–43The effects of isoflurane on action potential propagation are likely a result of isoflurane interactions with these voltage-gated ion channels.5,21Taking these findings together, we conclude that isoflurane inhibits glutamate release both at the synaptic bouton and by attenuating action potential propagation.

In conclusion, isoflurane, at clinically relevant concentrations, enhanced GABA-mediated inhibitory currents and suppressed glutamate-mediated excitatory currents on NTS neurons receiving direct visceral afferent innervation. These isoflurane actions occurred via  distinct presynaptic and postsynaptic mechanisms. These differential effects of isoflurane within the NTS may underlie depressed cardiorespiratory function.