Isoflurane Inhibits Transmitter Release and the Presynaptic Action Potential

The experiments were performed with permission from the Animal Studies Committee of Washington University (St. Louis, Missouri). All methods were described previously. 20,21Briefly, Wistar rats (7–10 days old) were decapitated. Transverse or parasagittal slices, 200 μm thick, were cut from the auditory brainstem with a vibratome. Recordings were made at room temperature (23–25°C). The holding potential was −80 mV, and the potential was corrected for a liquid junction potential of −11 mV between the extracellular and the pipette solution. Currents were low-pass filtered at 5 KHz and digitized at 20 KHz with a 16-bit analog-to-digital converter (Instrutech, Greatneck, NY). Data were expressed as mean ± SEM. The statistical test for calculation of P  values was a t  test.

For postsynaptic whole cell voltage clamp recordings of EPSCs and miniature EPSCs (mEPSCs;figs. 1–3) in transverse slices, an Axopatch 200B amplifier (Axon Instruments Inc., Foster City, CA) was used. The postsynaptic pipette (2–3 MΩ) solution contained 125 mm K-gluconate, 20 mm KCl, 4 mm MgATP, 10 mm Na²-phosphocreatine, 0.3 mm GTP, 10 mm HEPES, and 0.5 mm EGTA, at a pH of 7.2, adjusted with KOH. The series resistance (< 15 MΩ) was compensated by 95% (lag, 10 μs). The mEPSCs were analyzed with use of a software program (Jaejin Software, Leonia, NJ). The bath solution contained 125 mm NaCl, 2.5 mm KCl, 1 mm MgCl², 2 mm CaCl², 25 mm NaHCO³, 1.25 mm NaH²PO⁴, 25 mm dextrose, 0.4 mm ascorbic acid, 3 mm myo - inositol, 2 mm sodium pyruvate, 0.05 mm D-APV, 0.01 mm bicuculline, and 0.01 strychnine, at a pH of 7.4 when bubbled with 95% O²and 5% CO².

WaveMetrics, Inc., Lake Oswego, OR) using the principle of the least squares optimization, the data were fit with a Hill function (curve ) with an EC⁵⁰of 0.49 mm, a Hill coefficient of 1.7, and a maximal inhibition of 80%.

For capacitance recordings at calyces (fig. 4), an EPC-9 amplifier, together with the software lock-in amplifier (HEKA, Lambrecht, Germany), was used at parasagittal slices (see Sun and Wu 21for selection of calyces). The bath solution was the same as described in the previous paragraph (starting with “The bath solution contained …”). The series resistance (< 20 MΩ) was compensated by 60%. A sinusoidal stimulus (1,000 Hz) was applied in addition to the DC holding potential (−80 mV). The peak-to-peak voltage of the sine wave was less than 60 mV. The presynaptic pipette (3.5–5 MΩ) solution contained 125 mm K-gluconate, 20 mm KCl, 4 mm MgATP, 10 mm Na²-phosphocreatine, 0.3 mm GTP, 10 mm HEPES, and 0.05 mm BAPTA, at a pH of 7.2, adjusted with KOH.

For voltage clamp recordings of presynaptic Ca²⁺currents (figs. 5 and 6), the presynaptic pipette (3.5–5 MΩ) solution contained 125 mm Cs-gluconate, 20 mm CsCl, 4 mm MgATP, 10 mm Na²-phosphocreatine, 0.3 mm GTP, 10 mm HEPES, and 0.05 mm BAPTA, at a pH of 7.2, adjusted with CsOH. The bath solution contained 105 mm NaCl, 20 mm TEA-Cl, 2.5 mm KCl, 1 mm MgCl², 2 mm CaCl², 25 mm NaHCO³, 1.25 mm NaH²PO⁴, 25 mm dextrose, 0.4 mm ascorbic acid, 3 mm myo -inositol, 2 mm sodium pyruvate, 0.001 mm tetrodotoxin, 0.1 mm 3,4-diaminopyridine, and 0.05 mm D-APV, at a pH of 7.4 when bubbled with 95% O²and 5% CO².

Delivery of isoflurane and measurements of isoflurane concentration at the room temperature were similar to those described previously. 22,23Briefly, the 95% O²–5% CO²mixture was directed via  a flowmeter through a calibrated commercial vaporizer (Fortec; Fraser Harlake, Orchard Park, NY) containing isoflurane. The gas mixture of O²–CO²–isoflurane was then used to bubble the experimental saline in a reservoir bottle for at least 20 min before it was applied to the slice chamber with a pump (approximately 4 ml/min, Ismatec pump; Cole-Parmer, Vernon Hills, IL). The aqueous isoflurane concentration, measured with the standard gas chromatography methods, 24dropped from the saline reservoir to the slice chamber by 28 ± 2% (n = 3). The aqueous isoflurane concentration was linearly proportional to the gaseous partial pressure of isoflurane, detected by a Raman light-scattering gas analyzer (Rascal II; Ohmeda, Salt Lake City, UT; data not shown). Based on these data, the continuously measured gaseous partial pressure was used to estimate the aqueous isoflurane concentration in the slice chamber.

The postsynaptic neurons of the MNTB were whole cell voltage clamped at a holding potential of −80 mV, a potential at which the EPSC is mediated by AMPA receptors. 25A bipolar stimulating electrode was positioned at the midline of the trapezoid body, through which the presynaptic axons of MNTB synapses pass. A stimulus (5–20 V, 0.1 ms) was delivered every 20 s to induce a presynaptic action potential, 20and the resulting EPSC was recorded at the postsynaptic cell. The EPSC amplitude in control was 5.6 ± 0.7 nA (n = 8). Application of 0.35 mm isoflurane (see Materials and Methods) reduced the EPSC amplitude by 34 ± 3% (n = 5 synapses) without significantly affecting the 20–80% rise and decay time (fig. 1). The blocking effect reached steady state in less than 10 min after applying isoflurane and reversed in 5 min after washout (fig. 1A). This effect was dose dependent (e.g. , fig. 1, A and B). The percent inhibition as a function of the isoflurane concentration (eight synapses) was fit with a Hill function in which the EC⁵⁰, the Hill coefficient, and the maximal inhibition were free parameters. The fitting results were an EC⁵⁰of 0.49 mm, a Hill coefficient of 1.7, and a maximal inhibition of 80% (fig. 1C). Here, the EC⁵⁰referred to the concentration at which isoflurane achieved 50% of its maximal inhibition, i.e. , 40% (= 50%× 80%) inhibition of the EPSC. These results are quantitatively similar to those observed at hippocampal synapses. 5,23The minimum alveolar concentration (MAC) of isoflurane, at which general anesthesia occurs in 50% of animals, is 0.35 mm in rats. 26Therefore, clinically relevant concentrations (≤ 2 MAC) of isoflurane may significantly inhibit the EPSC at MNTB synapses.

When a pair of electrical stimuli with a short interval (e.g. , < 50 ms) is applied to axons, it generates a pair of action potentials, which in turn causes a pair of EPSCs. The ratio between the amplitudes of the second and the first EPSC is less than 1 at synapses, including MNTB synapses, 27which release a large number of vesicles during the first stimulus. Such a paired-pulse depression is conventionally interpreted as a result of depletion of the releasable pool by the first pulse. 28Consequently, if isoflurane reduces transmitter release, depletion is relieved, and thus, paired-pulse depression is reduced. We tested this prediction as follows. A pair of electrical stimuli (5–20 V, 0.1 ms) with an interval of 20 ms was applied to generate a pair of EPSCs (e.g. , fig. 2, A and B). The ratio between the second and the first EPSC amplitudes was 0.61 ± 0.07 (n = 16;fig. 2). This ratio was significantly increased to 0.97 ± 0.18 (n = 5; P = 0.02) and 1.42 ± 0.31 (n = 3; P = 0.0002) in the presence of 0.35 and 1.05 mm isoflurane, respectively (fig. 2). These results, similar to those observed at hippocampal synapses, 5–7are consistent with the hypothesis that isoflurane inhibits the EPSC by a presynaptic mechanism.

If isoflurane inhibits the EPSC by decreasing the sensitivity of postsynaptic glutamate receptors, it should decrease the mEPSC, the postsynaptic response to release of a single vesicle. We tested this prediction by collecting mEPSCs for 3–5 min before and after a 10-min application of 0.7 mm isoflurane (fig. 3A). Before application of isoflurane, the peak amplitude of mEPSCs was 24.1 ± 3.1 pA (n = 5 synapses). Application of isoflurane did not significantly reduce the amplitude of the mean mEPSC, 20–80% rise time and half decay time (e.g. , fig. 3, B and C; n = 5 synapses; P > 0.4). The frequency of mEPSCs ranged from approximately 1 to 5 Hz in control. Isoflurane reduced the frequency to 91 ± 4% of control, which was not statistically significant (P = 0.18; n = 5). Because AMPA receptors were not saturated by release of a single quantum at MNTB synapses, 29these results suggest that isoflurane does not inhibit the ESPC by decreasing the sensitivity of postsynaptic AMPA receptors.

To determine whether isoflurane inhibits the EPSC by reducing the number of released vesicles, we measured the membrane capacitance increase at the nerve terminal (the calyx of Held), which is proportional to the number of released vesicles. 21,30Capacitance measurements were made in the voltage clamp mode (fig. 4A). The voltage clamp mode was switched to a current clamp mode for 500 ms, during which 10 pulses of 3-ms current of 300–600 pA at 200 Hz were injected into the calyx to induce 10 action potentials (fig. 4B). We induced 10 action potentials instead of 1 to increase the signal-to-noise ratio. The capacitance jump, calculated as the difference between the capacitance values before and 500 ms after stimulation, was 293 ± 31 fF (n = 6;fig. 4A). The jump returned to the baseline in 10s of seconds (not shown, but see Sun and Wu, 21which reflects the time course of endocytosis). Therefore, slow endocytosis does not significantly affect measurements of capacitance jumps after stimulation. Consistent with our previous study, 21the capacitance change did not parallel the change of the membrane resistance or the series resistance (fig. 4A). Application of isoflurane (0.7 mm) significantly reduced the capacitance jump by 43 ± 6% (n = 6) of control. This decrease was similar to the decrease in the EPSC (approximately 50% at 0.7 mm as calculated from the fit curve in fig. 1C), suggesting that isoflurane inhibits the EPSC largely by reducing vesicle release.

The effects of isoflurane on the capacitance jump were accompanied by a decrease in the amplitude of the presynaptic action potential. For example, isoflurane at 0.7 mm reduced the presynaptic action potential amplitude from 106.3 ± 1.3 mV (n = 6) to 100.5 ± 1.7 mV (n = 6; P < 0.01;fig. 4B). That is, isoflurane reduced the action potential amplitude by 5.8 ± 1.3 mV or 5.5 ± 1.3% (n = 6). Isoflurane (0.7 mm) also slightly hyperpolarized the resting membrane potential (ranged from −56 to −68 mV) by 1.2 ± 0.3 mV (n = 6; P < 0.01). These effects were reversible on drug washout (fig. 4B). Isoflurane (0.7 mm) did not significantly affect the half-width of the action potential (0.36 ± 0.02 ms in control; 0.39 ± 0.03 ms in isoflurane; n = 6; P > 0.2). The effect of isoflurane on the action potential amplitude was dose dependent. At 0.35 mm, isoflurane reduced the amplitude by 3.3 ± 1.5% (n = 4). At 1.05 mm, isoflurane reduced the amplitude by 7.6 ± 0.8% (n = 4).

The above analysis of the action potential waveform (APW) was made at the first action potential during a train of 10 action potentials (fig. 4B). Similar results were obtained when the last action potential during the 10-action-potential train was analyzed. For example, isoflurane at 0.7 mm reduced the amplitude of the last action potential by 4.1 ± 1.1% (n = 6), without significantly affecting its half-width (P > 0.4). Therefore, the inhibition of the action potential during the 10-action-potential train was not activity dependent.

To determine the extent to which a decrease in the action potential accounts for inhibition of release, we examined the relation among the presynaptic action potential, the action potential–induced presynaptic Ca²⁺current (ICa), and the EPSC. All of these measurements were made in the absence of isoflurane. Both the nerve terminal and the postsynaptic neuron were whole cell voltage clamped at the same synapse. ICa was recorded by pharmacologically blocking Na⁺and K⁺channels (see Materials and Methods). An experimentally recorded action potential of 110 mV evoked by nerve stimulation (stimulation electrode positioned at the midline of the trapezoid body) 20was used as the voltage command to mimic an action potential. Application of this command induced a presynaptic ICa and a postsynaptic EPSC at the same synapse (fig. 5A). By scaling the typical APW to 95, 90, 85, 80, or 75%, the APW amplitude was reduced by 5.5, 11, 16.5, 22, or 27.5 mV, respectively. Consequently, the presynaptic ICa and the postsynaptic EPSC were reduced (e.g. , fig. 5A). The amplitude of the resulting EPSC was highly nonlinearly related to the presynaptic APW amplitude (fig. 5B). For example, the EPSC amplitude was decreased to 72 ± 5% (n = 7) or 43 ± 3% (n = 7) of control when the APW amplitude was decreased from 110 mV to 104.5 or 99 mV, respectively (fig. 5B). The normalized EPSC amplitude and the normalized APW amplitude obtained from seven synapses was fit with the following equation:

in which c¹= 1.15 and n = 10.2 (fig. 5B, right).

To study why there was such an extremely nonlinear relation, we plotted the ICa charge versus  the APW amplitude (fig. 5C) and the EPSC amplitude versus  the ICa charge (fig. 5D). The relation between the normalized ICa charge (Q^ICa) and the normalized APW amplitude was fit with the following equation:

in which c²= 1.04 and k = 2.8 (fig. 5C, right). Therefore, a small decrease in the APW amplitude may cause a larger decrease in the Q^ICa. The relation between the normalized EPSC amplitude and the normalized Q^ICawas fit with the following equation:

in which c³= 1.01 and m = 3.7 (fig. 5D, right). This nonlinear relation was consistent with previous studies at MNTB synapses and other synapses. 19Therefore, equations 2 and 3 explain why the EPSC amplitude was extremely nonlinearly related to the APW amplitude.

The EPSC induced by a typical APW command was 1.3 ± 0.3 nA (n = 7), which was smaller than the EPSC induced by a real action potential. 20,31This is likely caused by imperfect voltage clamp. To determine whether the results shown in figure 5still hold when the EPSC is similar to that induced by a real action potential, we induced EPSC by a 1-ms step voltage command from −80 to +7 mV (fig. 6, solid traces). We have shown such a step depolarization to induce release similar to that induced by a real action potential. 32Here, this depolarization induced an EPSC of 5.2 ± 1.4 nA (n = 4; e.g. , fig. 6). By decreasing the 1-ms step voltage command by 5 mV (the step was from −80 to +2 mV;fig. 6, dash trace), i.e. , by approximately 6% (5/87 ≈ 6%), the charge of the presynaptic calcium current and the postsynaptic EPSC decreased to 88 ± 2 and 67 ± 8% (n = 4, e.g. , fig. 6) of control, respectively. These results were similar to those shown in figure 5, confirming that a small decrease in a brief presynaptic depolarization results in a large decrease in the EPSC.

Based on the relation between the EPSC and the isoflurane concentration (fit curve in fig. 1C), 0.35, 0.7, and 1.05 mm isoflurane inhibited the EPSC by 29, 52, and 63%, respectively. Isoflurane at these concentrations reduced the action potential amplitude by means of 3.3, 5.5, and 7.6%, respectively. These results (fig. 5B, right, triangles) were plotted together with those obtained with voltage clamp experiments in control (fig. 5B, right, circles). It was clear from the plot that the decrease in the action potential amplitude may largely, but not completely account for the decrease in the EPSC induced by isoflurane. According to equation 1, 3.3, 5.5, and 7.6% decreases in the action potential amplitude would decrease the EPSC by 18, 35, and 49%, respectively. Therefore, on average, the decrease in the ESPC induced by the decrease in the action potential amplitude accounted for approximately 62% (= 18%/29%), 70% (= 35%/50%), or 78% (= 49%/63%) of inhibition of the EPSC by isoflurane at 0.35, 0.7, and 1.05 mm, respectively.

The amplitude of the APW we used was 110 mV rather than 106 mV as measured here. This was to be consistent with several previous studies that used a 110-mV APW as the voltage command. 31,33,34However, such a small difference should not significantly affect our conclusion because a similar extent of decrease in the EPSC was observed when the APW amplitude was decreased by 5.5 mV from 110 or 104.5 mV (fig. 5B).

We found that isoflurane dose-dependently inhibited the EPSC at MNTB synapses. By the use of capacitance measurements to monitor vesicle release at nerve terminals, we showed that isoflurane inhibits vesicle release induced by a train of 10 action potentials at 200 Hz. The stimulation train lasted for only 50 ms, during which mobilization of synaptic vesicles from the reserve pool to the releasable pool is negligible. 34,35Therefore, the capacitance jump induced by the stimulation train was proportional to the averaged release evoked by an action potential during the train. Because isoflurane inhibited the capacitance jump and the EPSC to a similar degree and reduced paired-pulse depression but did not affect the mEPSC amplitude, we concluded that isoflurane inhibited the EPSC predominantly by a presynaptic mechanism. These results provide strong evidence confirming many previous studies suggesting a presynaptic contribution to volatile anesthetic–induced inhibition of synaptic transmission. 3–12 

Isoflurane inhibited the presynaptic capacitance jump (43 ± 6%) and the EPSC (50%) to slightly different extents. Therefore, the possibility that isoflurane reduced the EPSC to a minor extent by a mechanism other than presynaptic mechanisms cannot be ruled out. These minor mechanisms might include effects on the number of glutamate receptors, 12glutamate receptor desensitization, and clearance of glutamate from the synaptic cleft. However, the slight difference in the inhibition of the capacitance jump and the EPSC might also be caused by the difference in the stimuli used to evoke capacitance jump and the EPSC. The EPSC was evoked by a single action potential. Such an EPSC was linearly proportional to the capacitance jump, suggesting that postsynaptic receptor saturation does not occur at such a minimal stimulus. 21In contrast, the capacitance jump was evoked by 10 action potentials at 200 Hz, which may cause a larger extent of depletion of a limited pool of vesicles immediately available for release 21,34,35and thus might reduce the percent inhibition of the capacitance jump by isoflurane.

We found that 0.35–1.05 mm isoflurane reduced the presynaptic action potential amplitude by approximately 3.3–7.6%. This is consistent with many previous studies showing little effect or minor effects of general anesthetics on the action potential. 1,2However, the EPSC was proportional to the presynaptic APW amplitude raised to a power of 10.2 (equation 1). Such an extremely nonlinear relation was caused by two highly nonlinear relations, one between the charge of the ICa induced by the APW and the APW amplitude (equation 2), and the other between the EPSC amplitude and the ICa charge (equation 3). Based on these quantitative relations, we concluded that inhibition of the presynaptic action potential accounts for approximately 62–78% of the inhibitory effects of isoflurane on the EPSC, whereas the remaining fraction might be mediated by mechanisms downstream of the action potential.

The decrease in the action potential amplitude could be due to inhibition of Na⁺channels 13,36or activation of K⁺channels. 14,37–39The mechanism downstream of the action potential may be mediated by inhibition of Ca²⁺channels 15,17,18or the release machinery downstream of the calcium influx. 16It would be of great interest to dissect each of these potential mechanisms. We attempted to perform voltage clamp recordings of Na⁺currents at the calyx and found that the current was too large (> 5–10 nA) and too fast to reliably voltage clamp. We also attempted to study whether isoflurane (0.7 mm) inhibits the presynaptic Ca²⁺currents. We did not observed consistent inhibition of Ca²⁺current by more than 15% (unpublished results, Xin-Sheng Wu, M.D., Ph.D., and Ling-Gang Wu, M.D., Ph.D., National Institute of Neurologic Disorders and Stroke, Bethesda, Maryland, May–September 2002). However, the Ca²⁺current evoked by an APW was extremely sensitive to a very small change (e.g. , 1- to 2-MΩ change) in the series resistance. As a result, stable recordings of the Ca²⁺current for as long as 10 min, which is needed for the effects of isoflurane to reach a steady state, was extremely difficult (note that experiments shown in figures 5A and 6may take < 2 min). We could not consistently resolve an inhibition of less than 10–15% in a 10 min, which may well be in the range of inhibition induced by isoflurane. In summary, resolving small changes (perhaps < 5–10%) in presynaptic Na⁺and Ca²⁺currents is a difficult and substantial undertaking that is beyond the scope of the current study.

A single calyx in the MNTB contains approximately 600 conventional active zones 40and thus is viewed as equivalent to approximately 600 boutons in parallel. 19,41Except for this difference, the calyx-type MNTB synapse shares many similar properties with central synapses, such as paired-pulse depression or facilitation, a nonlinear relation between release and Ca²⁺currents, Ca²⁺channel types and their modulation by transmitters, AMPA and N -methyl-d-aspartate receptors, and the ultrastructure of active zones. 19,41More importantly, isoflurane dose-dependently inhibited the EPSC and relieved paired-pulse depression while producing a minor reduction of the presynaptic action potential at MNTB synapses. These effects are qualitatively and quantitatively similar to those observed with volatile anesthetics at other synapses. 5–8,10,11,23,42Isoflurane inhibited the capacitance jump evoked by action potentials at MNTB synapses, which is consistent with the result that volatile anesthetics inhibit glutamate release induced by prolonged depolarization at synaptosomes. 3,4Therefore, the principles learned in the current study are likely to apply to other glutamatergic synapses.

We performed experiments in brain slices of 7- to 10-day-old rats at room temperature (23–25°C). Whether our findings can be applied to adult rats at body temperature remains unclear. We think this possibility is likely for the following reasons. The electrical properties of MNTB synapses of 9-day-old rats, such as the action potential width, the EPSC size, the quantal size, the calcium channel types, and paired-pulse depression and facilitation, are qualitatively similar but quantitatively slightly different from those of adult rats. 19Therefore, the results observed here might be quantitatively slightly different in adult rats. However, the major conclusions are likely to remain the same. Volatile anesthetics at similar aqueous concentrations inhibit excitatory synaptic transmission to a similar degree at other synapses at both room temperature and a temperature close to body temperature. 9,22,23,43–45Therefore, it is likely that the principles learned from brain slices at room temperature apply to those at body temperature. In conclusion, isoflurane-induced inhibition of the presynaptic action potential amplitude may contribute to inhibition of glutamate release, which may be relevant to certain end points of anesthesia. 1,2 

The authors thank Steve Mennerick, Ph.D. (Assistant Professor, Department of Psychiatry, Washington University School of Medicine, St. Louis, Missouri), and Joe Henry Steinbach, Ph.D. (Professor, Department of Anesthesiology, Washington University School of Medicine), for suggestions during the course of this work and comments on the manuscript.