Although it does not suppress movement in response to noxious stimuli, the nonimmobilizer 1,2-dichlorohexafluorocyclobutane (F6, also known as 2N) does cause amnesia and seizures. These occur at 0.48 and 1.3 times, respectively, the concentrations that are predicted from its lipid solubility to cause immobility. The molecular and cellular basis of these effects is not known. The ionotropic gamma-aminobutyric acid type A (GABAA) receptor is modulated strongly by anesthetics, and it plays an important role in many seizure models. Also, the hippocampus is a structure central to the formation of memory and is susceptible to seizure generation. The authors therefore investigated the effect of F6 on GABAA receptor- mediated inhibition in hippocampal neurons.
Transverse hippocampal slices were prepared from young (12- to 21-day-old) Sprague-Dawley rats. Inhibitory postsynaptic currents were recorded from hippocampal CA1 pyramidal cells in the presence of ionotropic glutamate receptor antagonists. F6 was applied with the bath solution. The concentration of F6 achieved during the experiment at the location of synaptic inhibition was derived using a diffusion model.
At tissue concentrations of up to 75 microm (approximately 5 x predicted minimal alveolar concentration), F6 had no discernible effect on either the amplitude or the kinetics of GABA-mediated synaptic currents. Isoflurane, by contrast, prolonged the decay time constant of these currents at 100 microm (approximately 0.3 x minimal alveolar concentration).
At concentrations that bracket the in vivo amnestic and seizure-inducing range, F6 has no discernible effect on fast synaptic GABAA receptors in hippocampal CA1 pyramidal neurons. Synaptic GABAA receptors sharply discriminate between volatile anesthetics and a prototype nonimmobilizer. Similar in vivo effects of anesthetics and nonimmobilizers may be mediated by different cellular mechanisms.
ANESTHESIA is a complex drug-induced state that comprises multiple elements, e.g. , immobility, amnesia, and hypnosis.1Correspondingly, anesthetic drugs, particularly the volatile anesthetics, affect multiple molecular and cellular processes. Separating relevant effects from irrelevant ones and linking effects on the receptor level to desirable and undesirable manifestations in vivo are fundamental aims of anesthesia-related research.2
One approach to these aims is to compare the effects of anesthetics to those of drugs that have anesthetic-like physicochemical properties but do not produce the full spectrum of anesthetic actions in vivo . These compounds were initially termed nonanesthetics 3; subsequently, the term was changed to nonimmobilizers after it was discovered that some agents cause amnesia but do not prevent movement in response to noxious stimuli.4The volatile compound 1,2-dichlorohexafluorocyclobutane (designated F6 or 2N in the literature) is an extensively studied prototype nonimmobilizer. Like anesthetics, it produces amnesia at a concentration of approximately one third the predicted minimal alveolar concentration (MAC) at which, according to its lipid solubility, it should cause immobility to painful stimulus in 50% of subjects if it behaved as a true anesthetic (predicted MAC or MACpred).3,4At concentrations above MACpred, it induces convulsions,5a property that to some degree is shared by some anesthetic ethers, e.g. , enflurane.6
A number of ligand-gated neuronal ion channels are affected to some degree by anesthetics. Of these, the ionotropic γ-aminobutyric acid type A (GABAA) receptor is thought to play a pivotal role in general anesthetic action. In particular, GABAAreceptor modulation may contribute to the amnestic as well as the proconvulsive properties of volatile anesthetics.7,8We decided therefore to investigate in detail the effects of F6 on GABAAreceptor–mediated synaptic activity. We used whole cell patch clamp recordings from rat hippocampal CA1 pyramidal neurons. Volatile anesthetics were shown previously to exert multiple effects in this preparation. Specifically, at concentrations above 150 μm (0.5 × MAC), isoflurane prolonged the decay of fast spontaneous and miniature γ-aminobutyric acid–mediated (GABAergic) inhibitory postsynaptic currents (sIPSCs and mIPSCs, respectively), and at concentrations above 600 μm (2 × MAC), it reduced IPSC amplitude. At equivalent MAC fractions, enflurane had similar effects on inhibitory postsynaptic current (IPSC) decay, but it reduced IPSC amplitude more profoundly. These effects on decay and amplitude were proposed to underlie the amnestic and proconvulsive properties, respectively.8We hypothesized that similar effects would be observed at behaviorally equivalent concentrations of F6.
Materials and Methods
All experiments were conducted according to the guidelines in the Guide for the Care and Use of Laboratory Animals 9and were approved by the University of Wisconsin Animal Care and Use Committee (Madison, Wisconsin).
Juvenile male Sprague-Dawley rats (aged 14–24 days) were decapitated during isoflurane anesthesia, and the brain was quickly removed and immersed in cold (4°C) artificial cerebrospinal fluid (ACSF) saturated with 95% O2–5% CO2(carbogen gas). A block of tissue containing the hippocampus was removed and glued to a tissue tray using cyanoacrylate glue. Tissue slices 400 μm thick were prepared using a vibrating microtome (Leica VT1000; Bannockburn, IL), incubated at 32°C for 1 h, and then kept in carbogen-saturated ACSF at room temperature until use.
Patch Clamp Electrophysiology
Cells in the stratum pyramidale of CA1 were visualized using a video camera (Hamamatsu C2400; Hamamatsu Corp., Bridgewater, NJ) connected to an upright microscope (Zeiss Axioskop; Thornwood, NJ) equipped with an infrared bandpass filter (Chroma D775/220; Brattleboro, VT), a long working-distance water-immersion objective (Zeiss Achroplan 40×, 0.75 numerical aperture), and differential interference contrast optics (DIC, or Nomarski). Whole cell recordings were obtained at room temperature (22°–24°C) using a Multiclamp 700A patch clamp amplifier and pClamp software (Axon Instruments, Foster City, CA). Data were filtered at 5 kHz, sampled at 10 kHz (Digidata 1200; Axon Instruments), and stored on a Pentium-based computer. Patch pipettes were fabricated from borosilicate glass (KG-33; Garner Glass, Claremont, CA; 1.7 mm and 1.1 mm OD and ID, respectively) using a Flaming-Brown two-stage puller (model P-87; Sutter instruments, Novato, CA), fire polished, and coated with Sylgard (Dow-Corning, Midland, MI) to reduce electrode capacitance when necessary. Tight-seal whole cell recordings were obtained using standard techniques. Patch pipettes had open-tip resistances of 2–4 MΩ when filled with the recording solution of the following composition: 140 mm CsCl, 10 mm NaCl, 10 mm HEPES, 10 mm BAPTA, 2 mm MgATP, and 5 mm lidocaine N -methyl bromide (QX-314), pH 7.3. Access resistances were typically 10–20 MΩ and were compensated by 60–80%. The membrane potential was voltage clamped at −60 mV. GABAAreceptor–mediated IPSCs were isolated by bath application of 20 μm 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and 40 μm D,L-2-amino-5-phosphonovaleric acid to block glutamate receptors, and by the inclusion of CsCl and QX-314 in the patch pipette to block regenerative Na+and GABAB-mediated currents. Tetrodotoxin (1 μm) was used to isolate non–action potential-dependent, i.e. , miniature, IPSCs. IPSCs recorded in the absence of tetrodotoxin were considered to be a combination of both action potential dependent and independent IPSCs and were termed spontaneous IPSCs . The remaining currents were exclusively GABAAreceptor mediated because they were completely blocked by bath application of 10 μm bicuculline (data not shown). D,L-2-amino-5-phosphonovaleric acid, CNQX, tetrodotoxin, and bicuculline were prepared as 50- to 100-fold stock solutions and applied using syringe pumps (model 55-1111; Harvard Apparatus, Natick, MA) set to flow at 1 or 2% of the ACSF flow rate to achieve the desired bath concentrations.
Application of Volatile Agents
Experiments were conducted using ACSF of the following composition: 127 mm NaCl, 1.21 mm KH2PO4, 1.87 mm KCl, 26 mm NaHCO3, 2.17 mm CaCl2, 1.44 mm MgSO4, and 10 mm glucose, saturated with carbogen gas, pH 7.4. Solutions of artificial cerebrospinal fluid containing isoflurane and F6 were prepared in Chemware Teflon FEP gas sampling bags (North Safety Products, Cranston, RI) fitted with three-way stopcocks. F6 solutions were prepared by filling Teflon bags partially with ACSF that had been preequilibrated with carbogen gas and adding to the headspace appropriate quantities of F6-saturated carbogen and pure carbogen to achieve the desired aqueous concentration of F6, as calculated using a saline/gas partition coefficient for F6 of 0.026.10Isoflurane solutions were prepared similarly except that 10 mm isoflurane in ACSF (which is below the saturating concentration of 15 mm at room temperature)11was added to the Teflon bag. For electrophysiologic experiments, solutions were equilibrated by shaking on a shaker table for at least half an hour and were then allowed to stand for at least 1 h. F6-saturated carbogen stock was prepared in a separate Teflon bag by first evacuating the bag then adding carbogen, water, and an excess of liquid F6 and shaking for 1 h to equilibrate. Control ACSF solutions were prepared similarly in Teflon bags. Concentrations of F6 and isoflurane in ACSF from the Teflon bags and from the brain slice chamber were measured by gas chromatography to determine the fraction of loss and the concentrations that were actually applied to slices. Volatile agents were applied for 20 to 30 min. The data used for analysis were collected within the last 3 min of drug exposure and after at least 17 min of washing.
For F6, which in contrast to isoflurane does not reach an equilibrium within this time frame, the effective concentration around the cell soma was determined by taking into account the loss of F6 from bag to chamber, the distance of the cell from the slice surface, and the duration of F6 application, using a diffusion model as detailed previously (fig. 1).10These are the concentrations of F6 referred to in the text. For example, if the measured concentration of F6 in the Teflon bag was 81 μm, the concentration in the bath would be 48.6 μm (40% loss of F6 on the way to and within the slice chamber). After 17 min of drug application, the concentration around a cell soma located 30 μm below the slice surface was estimated by the diffusion model to be 40 μm.
Aqueous Phase Drug Concentration by Gas Chromatography
Aqueous samples (2 ml) were collected from Teflon bags or the brain slice perfusion chamber using a gas-tight glass syringe (Hamilton Co., Reno, NV) fitted with a Teflon stopcock (Hamilton Co.). Samples were transferred to 3.7-ml glass vials capped with mininert valves (Alltech, Nicholasville, KY). Aqueous and gas phases within the vial were equilibrated by shaking for 1 h. Drug concentrations in the vial headspace were determined by gas chromatography using gas phase calibration standards for F6 or isoflurane. Concentrations in aqueous samples (C aq, sample ) were calculated based on saline/gas partition coefficients (λsaline/gas ) and the relative volumes of aqueous and gas phases within the vial (V gas,vial and V aq,vial ), according to the equation
Gas phase concentrations (C gas, vial ) were measured using a Varian 3700 gas chromatograph (Varian Inc., Walnut Creek, CA) with a flame ionization detector. Separation was achieved by on-column injection into a 1.83 m × 3.2 mm stainless steel column packed with 80/100 Poropak Q.
Data were analyzed on a Pentium-based personal computer using Clampfit (Axon Instruments), Origin (MicroCal, Northampton, MA), and Instat (GraphPad Software, San Diego, CA). Data were first filtered off-line at 2 kHz. Events were then analyzed using an automated event–detection algorithm as described previously.8All events detected by the program were visually inspected, and the peak amplitude and 10–90% rise time measurements were adjusted if necessary. To characterize the decay kinetics of fast IPSCs, a subset of events also was selected for exponential curve fitting. Events were selected only if no other event occurred within 250 ms of the peak. IPSC decays were typically best described by two exponential components. To quantify the overall decay time, we computed the weighted time constant τw= (A1τ1+ A2τ2)/(A1+ A2), where τ is the time constant of decay and A is the amplitude.
For statistical comparison of interevent intervals (IEIs), data were first tested for normality using the Kolmogorov-Smirnov test. The largest groups of data (mIPSCs, n = 23, and sIPSCs, n = 7, control conditions) did pass the test of normality. Therefore, it was assumed that subsets of these data would also be normally distributed, so t tests were performed to compare group means.
These data are based on recordings from 30 cells located in the stratum pyramidale of CA1. The majority of these cells are likely to be pyramidal cells.12However, because we did not conduct histologic analysis, a small number of inhibitory cells may also be included. Under control conditions (without F6) and with isoflurane application, the recordings were usually stable for prolonged periods (greater than 60 min). Exposure to F6, by contrast, seemed to affect the recording conditions, as described in detail in the section Effect of F6 on IPSC Amplitude.
We hypothesized that F6 would either increase (at lower concentrations) and/or decrease inhibition (particularly at higher concentrations) thereby providing cellular substrates compatible with its amnestic and seizure-inducing effects in vivo . We therefore examined characteristics of inhibition at the population and the individual IPSC levels that would, either together or in isolation, translate into changes in overall inhibition.
Effect of F6 on IPSC Decay
A prominent effect of anesthetic agents on synaptic GABAAresponses is a slowing of the time constant of decay. Therefore, we investigated the effect of F6 on the decay of sIPSCs and mIPSCs. Decay was usually best fitted with two exponentials, but we report here the weighted time constant τwto simplify comparisons. Under control conditions, the τwof spontaneous events ranged from 15.7 to 21.9 ms (mean, 19.6 ± 2.8 ms; n = 7) and did not differ significantly from the τwof miniature events, which ranged from 10.4 to 24.0 ms (mean, 18.9 ± 3.4 ms; n = 18; P > 0.05, two-tailed t test). These values are similar to those previously reported in this same preparation.8We examined the effect of F6 on the decay of mIPSCs and sIPSCs in 17 cells at concentrations ranging from 15 to 75 μm (approximately 1 to 5 × MACpred). An example of the (lack of) effect is illustrated in figure 2A. Averaged mIPSCs are shown from a single cell before, during, and after washout of 36 μm F6 (> 2 × MACpred). Unlike the prominent effect produced by volatile anesthetics, no apparent effect of F6 on decay was observed in this cell. Neither the τwof sIPSCS (n = 4) nor the τwof mIPSCs (n = 13) were affected by F6 at any of the concentrations tested.
To test whether our experimental conditions and analytical techniques were suitable for detecting minor alterations in IPSC kinetics, we also tested the effect of a low concentration of isoflurane (fig. 2B). Under experimental conditions that were identical to those under which F6 was tested, isoflurane 100 μm (0.3 × MAC) slowed the decay of the averaged IPSC by almost 50% (from 15 ± 4 to 22 ± 8 ms; n = 3; P < 0.05, paired t test). A comparison of effects of isoflurane and F6 over a range of concentrations is shown in figure 2C. It is apparent that isoflurane significantly slowed mIPSC decay in a dose-dependent manner, but that F6 at concentrations up to 5 × MACpredhad no effect on either mIPSCs or sIPSCs.
Effect of F6 on IPSC Amplitude
Isoflurane, enflurane, and halothane all exert dual effects on IPSCs: In addition to prolonging the decay, they also reduce the peak amplitude of the currents.8Both effects are dose dependent, and although the effect on amplitude requires higher concentrations than the effect on decay, it is apparent at sub-MAC concentrations. Because it is more pronounced for enflurane than for isoflurane,8it has been suggested that the differential effects of isoflurane and enflurane on decay kinetics versus amplitudes of mIPSCs (more block of the amplitude for the same degree of prolongation with enflurane) could underlie the proconvulsive properties of enflurane. F6 reliably induces seizures at concentrations slightly above 1 MACpredin vivo , and we found no effect on the decay kinetics of either mIPSCs or sIPSCs (fig. 2); therefore, we examined whether it alters mIPSC amplitude.
An example of one such experiment is illustrated in figure 3A. In the presence of 1 μm tetrodotoxin, synaptic currents were measured before, during, and after the application of 40 μm F6. No gross effects on either the amplitude or the frequency of mIPSCs are apparent. Figure 3Bshows the results of a quantitative analysis of mIPSCs recorded from the cell shown in figure 3A. A minor shift of the distribution of mIPSC amplitudes toward smaller values is apparent from both the amplitude histograms (inset) and the cumulative probability histogram. However, the shift was not reversed on washout of F6. In all, we tested the effect of 15 to 62 μm F6 on mIPSCs in 13 cells. In nine of these neurons, the recordings remained stable throughout the 20-min F6 application and 20-min washout, and the results from these cells are summarized in figure 3C. We divided the cells into two groups, those exposed to a concentration of approximately 1 × MACpredand those exposed to approximately twice MACpredof F6 or higher (n = 4 and n = 5, respectively; see fig. 3, legend). The mean IPSC amplitude was reduced by 20% (from 73 ± 20 pA to 58 ± 14 pA at 1 × MACpred) and 16% (from 51 ± 11 pA to 43 ± 10 pA at ≥ 2 × MACpred), and this effect was only partially reversed by washout. To test for changes in IPSC amplitude that may occur with prolonged recordings independently of any drug exposure (so-called run-down), we recorded from five “control” cells for a similar duration but without exposure to F6. In these cells, we observed a mean reduction of the IPSC amplitude of 8% (from 61 ± 13 pA to 56 ± 9 pA). Therefore, approximately half of the change in amplitude under F6 could be attributed to run-down and half could be attributed to effects of the drug. Statistical analysis yielded a significant reduction in IPSC amplitude for the cells in the group exposed to higher F6 concentrations (P < 0.001, one-tailed paired t test).
It would be tempting to speculate that this (admittedly minor) effect on IPSC amplitude might underlie the proconvulsive effects of F6. However, we do not believe this to be the case for the following reasons. First, we tried to establish a dose–effect relation between [F6] and magnitude of block. The results are shown in figure 3D. We plotted the fractional block by F6 (IF6/Icontrol, where Icontroland IF6represent the mean amplitudes of 400–500 consecutive IPSCs before and during F6 application, respectively). It is evident that the magnitude of the block by F6 did not correlate with changes in F6 concentration. Second, during perfusion with F6-containing solution, we almost invariably observed an increase in access resistance (RS) between the recording electrode and the cell interior. In a sample of cells where the change could be analyzed, we found that, under F6, the normalized RSincreased to 1.36 (± 0.38; n = 9) times the value under control conditions. This change was independent of the F6 concentration applied. We also noted that, simultaneously, there was a slight increase in the 10–90% rise time of the IPSCs. Figure 4Aillustrates these findings in a time series plot for three cells that were exposed to 30–40 μm F6. For each individual cell, the data are averaged over 5-s intervals and normalized to the mean values obtained over the time period before application of F6. The figure illustrates that, shortly after exposure to F6, there is a reduction of mean IPSC amplitude that mirrors the increase in the mean rise time. Figure 4Bshows a plot of mean normalized rise times versus amplitudes and illustrates the correlation between these two parameters. Our interpretation of these findings is that, for unknown reasons possibly related to its physicochemical properties (extremely low aqueous and moderate lipid solubility), F6 caused an increase of the access resistance to the cell interior. The increased rise time and decreased amplitudes thus are likely to reflect a change in the recording conditions rather than physiologic effects on GABAAreceptors. Similar effects were observed in recordings of sIPSCs: Exposure to F6 (25 and 75 μm; n = 4) led to an increase in RS, but in this case, the trend toward smaller IPSC amplitude did not reach statistical significance (P > 0.05, two-tailed paired t test). Such systematic changes in RSand rise time were not seen either in control recordings (i.e. , during prolonged recordings without application of any volatile agent: normalized RS0.97 [± 0.036; n = 4] times the control value) or in recordings where isoflurane was used (normalized isoflurane RS= 1.07 [± 0.2; n = 6] times control RS). This lack of F6-mediated block of native GABAAreceptors is in general agreement with the finding by Mihic et al. ,13who did not observe any effect of F6 in a similar concentration range on responses of expressed α1β1γ2sreceptors to application of 5 μm GABA.
Effect of F6 on sIPSC and mIPSC Frequency
Changes in the frequency of mIPSCs reflect effects on the release probability, whereas the rate of sIPSCs is also affected by changes in the excitability of presynaptic interneurons. We measured the rate of sIPSCs and mIPSCs as the IEI. The IEIs of sIPSCs and mIPSCs ranged from 92 to 565 ms (mean, 274 ± 168 ms; n = 7) and from 144 to 675 ms (mean, 335 ± 175 ms; n = 23), respectively. The difference was not significant (P > 0.05, two-tailed t test), indicating that there was a relatively low rate of spontaneous action potentials in the interneuron population innervating the pyramidal cells under our recording conditions. We studied the effect of F6 on the frequency of both mIPSCs (fig. 5A) and sIPSCs (fig. 5B). Exposure to F6 (25–75 μm; n = 4) did not change the IEI of sIPSCs (313 ± 149 ms; P > 0.05, two-tailed paired t test). Likewise, F6 (15–63 μm; n = 13) had no effect on the frequency of mIPSCs (334 ± 179 ms; P > 0.05, two-tailed paired t test; fig. 5). Low isoflurane concentrations (40–100 μm) also had no effect on IEI for either sIPSCs (339 ± 237 ms and 332 ± 139 ms for control and test, respectively; n = 3; two-tailed paired t test) or mIPSCs (426 ± 158 ms and 428 ± 171 ms for control and test, respectively; n = 5; two-tailed paired t test), consistent with previous results.8
We presented data demonstrating the lack of effect of the nonimmobilizer F6, at a wide range of concentrations, on GABAAreceptor–mediated synaptic transmission. These results are relevant to ongoing investigations of anesthetic mechanisms for the following reasons: They illustrate that the GABAAreceptor, which is widely accepted to play a crucial role in the mechanisms of anesthesia, discriminates between anesthetics and nonanesthetics across a wide range of concentrations; they suggest that suppression of fast synaptic inhibition is not the mechanism of seizure induction by F6. Furthermore, in combination with other recent findings,14they raise the possibility that the clinical effects that anesthetics and nonanesthetics have in common may be mediated via different mechanisms, and therefore that the strategy of using nonimmobilizers in the search for anesthetic targets has important limitations. Each of these issues is discussed in the following sections.
Synaptic GABAAReceptors Discriminate
In striking contrast to all halogenated volatile and numerous intravenous anesthetics, F6 had no effect on the decay time course of IPSCs at any of the concentrations tested (figs. 2A and C). By contrast, the widely used volatile anesthetic isoflurane slowed IPSCs significantly at concentrations as low as 100 μm (0.3 MAC; figs. 2B and C), a concentration that impairs hippocampal-dependent learning (fear conditioning to context) but not hippocampal–independent learning (fear conditioning to tone).15A slowing of IPSC decay has at least two functional consequences: On one hand, it increases the negative charge transfer across the neuronal membrane, thereby hyperpolarizing the cell and moving it away from the firing threshold for action potential generation. On the other hand, it also prolongs the time interval that conductances shunting the membrane are active, thus reducing the membrane resistance and thereby diminishing the efficiency of excitatory synaptic inputs. Both of these effects, the hyperpolarizing and the shunting effect, diminish the excitability of the neuronal membrane.
Effects of isoflurane on synaptic inhibition are likely to at least contribute to, and perhaps even be sufficient to account for, its suppression of synaptic plasticity. In a recent study, it was found that the ability of isoflurane (0.2–0.3 mm, approximately 1 MAC) to block the induction of homosynaptic long-term potentiation in hippocampal CA1 pyramidal cells was prevented by the GABAAreceptor antagonist picrotoxin,7indicating that long-term potentiation suppression was therefore probably mediated by anesthetic modulation of the GABAAreceptor. Because synaptic plasticity has been proposed as an in vitro model of memory formation,16the authors concluded that the effect of isoflurane on GABAergic inhibition is compatible with and may be sufficient to account for its amnestic effects in vivo . Recent reports indicate that isoflurane may block synaptic plasticity at even lower amnestic (i.e. , 0.3 MAC) concentrations.17Therefore, because synaptic GABAAreceptors in the hippocampus are insensitive to F6, they cannot be instrumental in the mechanism by which F6 produces amnesia.
The Amnestic Paradox
Nonimmobilizers, known as nonanesthetics before the discovery that they could produce some effects from the anesthetic spectrum, were introduced into experimental paradigms as a tool to discriminate between relevant and irrelevant targets of anesthetic, particularly volatile anesthetic, action.3The rational behind this approach is that nonimmobilizers do not produce the full spectrum of effects in vivo that anesthetics do. For example, F6 does not produce immobility at or even well above MACpred.3Therefore, a cellular process that is similarly affected by F6 and an anesthetic cannot underlie immobility—as demonstrated for the nicotinic acetylcholine receptor.18,19By contrast, isoflurane and F6 both impair learning at concentrations of approximately 0.3 × MAC and MACpred.4Therefore, following the above argument, a process similarly affected by both drugs could contribute to the impairment of memory formation.
Our findings suggest that if isoflurane and F6 share the same mechanisms for amnesia, it is not via actions on classic fast synaptic GABAAreceptors. However, it is conceivable that other types of GABAAreceptors could play a role in the suppression of hippocampal plasticity by isoflurane. A separate class of slow dendritic inhibitory synapses exists in these cells,20,21as do extrasynaptic receptors that are a likely source of tonic inhibitory current.22,23These receptors differ in their physiologic and pharmacologic properties, and they could be susceptible to modulation by F6. These possibilities must be investigated before definitive conclusions about the interaction of this prototype nonimmobilizer with GABAergic inhibition are drawn. Alternatively, it is possible that isoflurane and F6 induce amnesia via interference with separate pathways. In this case, the usefulness of nonimmobilizers in the search for mechanisms of amnesia is further limited. The latter possibility is supported by the results of a recent in vivo study in which learning impairment caused by F6 and flurothyl were antagonized by the concomitant administration of isoflurane.14
Mechanisms of Amnesia and Seizure Generation Remain Unresolved
Previous studies in vivo demonstrated that there are numerous physiologic processes that are not affected by F6 but that are affected by volatile anesthetics.24–26However, F6 does induce amnesia at concentrations below 1 MACpredand convulsions above 1 MACpred.4,5In the current study, using the hippocampal slice preparation, we failed to detect any effects at the cellular level that could underlie either the amnestic or the convulsive properties of F6. It has been recognized that, because of its extremely low aqueous solubility and its volatility, it can be difficult to achieve targeted concentrations in experiments where F6 is delivered via the aqueous superfusate,19particularly if it must diffuse significant distances into biologic tissue—the typical conditions of experiments on brain slices. We circumvented the uncertainties posed by diffusion by measuring drug concentrations in the bath as well as the brain tissue diffusion coefficient of F6,10thus allowing us to characterize its concentration profile under our experimental conditions. We were therefore able to determine with reasonable accuracy the effective concentrations in the tissue surrounding the structures of interest—in this case, inhibitory synapses that are located primarily on the somata of CA1 pyramidal neurons.27Other parameters of GABAergic synaptic inhibition, such as the frequency of spontaneous events and postsynaptic current amplitude, are modulated by anesthetics in a less consistent manner. These parameters also remained unaffected by F6, with the caveat that under our experimental conditions, because of the low frequency of spontaneous relative to action potential–independent activity, subtle effects on action potential–dependent GABAArelease may have been missed. Such effects are not necessarily likely—F6 at concentrations up to 50 μm had no measurable effect Na+channel–mediated release of the excitatory neurotransmitter glutamate from cortical synaptosomes and had only modest effects on voltage-gated Na+currents in dorsal root ganglion cells.28
If F6 does not cause amnesia or convulsions via GABAAreceptor modulation, what might be the targets that lead to these effects? Numerous possibilities have been suggested for volatile anesthetics, and each of these could also (or instead) be a target for F6. Studies of recombinant nicotinic acetylcholine receptors expressed in Xenopus oocytes showed that F6 inhibits rat and human nicotinic acetylcholine receptors at relevant concentrations.18,19Because F6 shares this property with isoflurane and neuronal-type nicotinic acetylcholine receptors are implicated in certain forms of memory,29these receptors are one plausible target. However, if as suggested above, anesthetics and nonimmobilizers cause similar behavioral effects via independent or even antagonistic actions, identification of common molecular targets may prove to be less instructive than anticipated.
The authors thank Michael Laster, D.V.M. (Assistant Researcher), and Edmond Eger II, M.D. (Professor, Department of Anesthesia, University of California, San Francisco, California), for providing gas phase calibration standards for isoflurane and 1,2-dichlorohexafluorocyclobutane, and Mark Perkins, B.S. (Research Specialist, Department of Anesthesiology, University of Wisconsin Medical School, Madison, Wisconsin), for technical support.