The Diverse Actions of Volatile and Gaseous Anesthetics on Human-cloned 5-Hydroxytryptamine³Receptors Expressed in Xenopus  Oocytes

Human-cloned 5-HT^3Acomplementary DNA (cDNA) was kindly provided by Akira Miyake, Ph.D. (Institute for Drug Discovery Research, Yamanouchi Pharmaceutical Co., Ltd., Ibaragi, Japan). 13The cDNA encoding the human 5-HT^3Areceptor was subcloned into pBluescriptII (Stratagene, La Jolla, CA) and was linearized by EcoR1 (Wako Pure Chemical Industries, Ltd., Osaka, Japan) to create the template cDNA. Capped complementary RNA (cRNA) was synthesized in vitro  from cDNA using T³RNA polymerase (T3 mMESSAGE mMACHINE KIT; Ambion, Austin, TX) according to the manufacturer's instructions. In accordance with the study protocol approved by the Animal Research Committee of Osaka University Medical School (Osaka, Japan), female Xenopus laevis  were anesthetized on ice with 1% 3-aminobenzoic ethyl ester (Tricaine; Sigma, St. Louis, MO). Oocytes were harvested through a laparotomy incision, manually defolliculated with forceps, and treated with 1.5 mg/ml collagenase type 1A (Sigma) for 30 min at room temperature in modified Barth saline (MBS: 88 mm NaCl, 1 mm KCl, 10 mm HEPES, 1 mm MgCl², 2.4 mm NaHCO³; pH 7.4). Between 10 and 50 ng cRNA was injected into an oocyte with a glass capillary, using a Nanoject injector (Drummond Scientific, Broomall, PA). Oocytes were incubated at 20°C in MBS containing 1.8 mm Ca²⁺until the start of the electrophysiologic experiment.

Between 24 and 48 h after cRNA injection, the oocytes were placed in a 0.2-ml chamber and were continu-ously superfused with MBS containing 1.8 mm CaCl²at 5–10 ml/min. The electrophysiologic recordings were made using a two-electrode, voltage clamp technique. The oocytes were impaled with 1- to 5-MΩ electrodes filled with 3 m KCl solution and were voltage clamped at −70 mV (CEZ-1250; Nihon Kohden, Tokyo, Japan). Drugs were dissolved in MBS and applied to the perfusate. Volatile and gaseous anesthetics were added to the perfusate by bubbling anesthetic-containing gases with or without 5-hydroxytryptamine (5-HT). A 50-ml conical tube (Corning 430291; Corning, NY) was filled with solution, and anesthetic gases were continuously bubbled into the tube at a rate of 100 ml/min. For volatile anesthetics (halothane, isoflurane, and sevoflurane), the air was passed through the following calibrated vaporizers: Fluotec3 for halothane (Ohmeda, Steeton, West Yorkshire, United Kingdom), Forawick for isoflurane (Murako Medical, Tokyo, Japan) and Sevotec3 for sevoflurane (Ohmeda). For gaseous anesthetics (N²O and Xe), gas mixtures at different concentrations of each gas were prepared using precise flowmeters (PMG-1; KOFLOC Co., Ltd., Kyoto, Japan). The concentrations of N²O and Xe in the gas mixtures were measured using a Datex Capnomac Ultima (Datex Instrumentarium Corp., Helsinki, Finland) and a Xe meter (Riken, Tokyo, Japan), respectively. The concentration of oxygen or nitrogen in the control solution was adjusted to the same level as the oxygen concentration in the anesthetic solution. Anesthetic solutions were bubbled with the anesthetic-containing gas mixture for 30 min. Anesthetic solutions were preapplied to oocytes before exposure to 5-HT to allow time for equilibrium. Each drug application was separated by intervals of a few minutes and by longer intervals after application of high drug concentrations to eliminate receptor desensitization. Cumulative desensitization was excluded by confirming that the same response was induced by a low concentration of 5-HT during an experiment with one oocyte. The current was digitally recorded using AxoScope software (Axon Instruments, Burlingame, CA), running on an IBM personal computer (IBM Aptiva, Armonk, NY). All electrophysiologic experiments were performed at room temperature.

The different concentrations of volatile and gaseous anesthetics in solution were determined by gas chromatography–mass spectrometry–selected ion monitoring (GC-MS-SIM). 14,15A gas chromatograph (Trace GC2000, ThermoQuest; CE Instruments, Austin, TX) and mass spectrometer (GCQ plus, ThermoQuest; CE Instruments) equipped with a data processing system (Xcaliber, ThermoQuest; CE Instruments) were used. The capillary column used was DB-5MS (0.25 mm ID × 30 m; J&W Scientific, Folsom, CA). Using the head space sampling technique, 14the concentrations of anesthetic in the solution and gas phase were measured using an automatic gas sampler (COMBI PAL; CTC Analytics AG, Zwingen, Switzerland). The ion intensities of isoflurane, sevoflurane, N²O, and Xe were monitored using mass/charge values (m/z) of 51, 51, 30, and 129, respectively. The anesthetic peak spectra areas were measured and compared to determine the gas/MBS partition. The gas/solution partition coefficient (P) of each anesthetic at 25°C was calculated using the following equation: P = C^Ring/C^gas, where C^gasand C^Ringare the equilibrated concentrations of anesthetic in the gas phase and MBS. Then the actual concentration of each anesthetic in the experimental solution was determined by using the absolute calibration curve established for each anesthetic in solution.

Peak amplitudes of the current elicited by the drugs were measured directly from digital recordings stored in AxoScope. To obtain the concentration–response curve for 5-HT–induced currents, observed peak amplitudes were normalized and plotted and then fitted to the Hill equation below using Sigmaplot software (Jandel Scientific, San Rafael, CA):

where I is the peak current at a given concentration of 5-HT, I^maxis the maximum current, EC⁵⁰is the concentration of 5-HT eliciting a half-maximum response, and n denotes the Hill coefficient.

Statistical analyses were performed using the Student t  test, with significance levels set at P < 0.05. All data were expressed as mean ± standard error of the mean.

5-Hydroxytryptamine was purchased from Wako Pure Chemical Industries, Ltd. The selective 5-HT³antagonist ramosetron was donated by the Institute for Drug Discovery Research, Yamanouchi Pharmaceutical Co., Ltd. The anesthetics were sourced as follows: halothane (Takeda Chemical Industries, Osaka, Japan), isoflurane (Abbott Laboratories Ltd., Chicago, IL), sevoflurane (Maruichi Pharmaceutical Co., Ltd., Osaka, Japan), N²O (Teisan Ltd., Tokyo, Japan), and Xe (99.995%; AirWater Co. Ltd., Wakayama, Japan).

The gas/MBS partition coefficient of each anesthetic at 25°C was calculated based on measurements made using GC-MS-SIM, and was 1.12 ± 0.02 for halothane, 1.04 ± 0.01 for isoflurane, 0.63 ± 0.01 for sevoflurane, 0.94 ± 0.03 for N²O, and 0.12 ± 0.01 for Xe. The actual concentrations of the anesthetics in the MBS used in this study are summarized in table 1.

We confirmed that the 5-HT–induced currents in oocytes injected with human 5-HT^3Areceptor cRNA were reversibly blocked by the selective 5-HT³receptor antagonist ramosetron (data not shown). The peaks of the 5-HT–induced currents were concentration dependent, and the 5-HT concentration–response curve fitted well to the Hill equation, with an EC⁵⁰of 2.7 ± 0.1 μm and a Hill coefficient of 1.7 ± 0.1.

Applications of all volatile (up to 4%) and gaseous (up to 100%) anesthetics without 5-HT produced no detectable current (data not shown). To exclude the involvement of oxygen and nitrogen concentrations in the 5-HT–induced current, we evaluated the effects of the perfusates, which contain 5-HT under various concentrations of these gases (0–100%). The control currents induced by 5-HT were not affected by the concentration of oxygen or nitrogen in the perfusate (data not shown). Figure 1shows the traces of 5-HT–induced currents potentiated and inhibited by the anesthetics. The potentiation and inhibition by the anesthetics were fully reversible. The volatile anesthetics halothane and isoflurane potentiated the 5-HT–induced current at clinical concentrations. The currents induced by 1 μm 5-HT, which was equivalent to EC¹⁵, were enhanced by 0.5, 1, 2, and 4% halothane to 154 ± 9, 182 ± 12, 196 ± 15, and 263 ± 14%, and by 0.5, 1, 2, and 4% isoflurane to 110 ± 2, 117 ± 2, 132 ± 4, and 151 ± 8%, respectively. Halothane potentiated the 5-HT-induced current to a larger degree than did isoflurane at the same concentration. In contrast, the volatile anesthetic sevoflurane and the gaseous anesthetics N²O and Xe inhibited 5-HT–induced currents. Sevoflurane reduced 1 μm 5-HT–induced currents to 89 ± 2, 76 ± 1, 56 ± 3, and 33 ± 2% at concentrations of 0.5, 1, 2, and 4%, respectively. N²O inhibited the current to 88 ± 1, 77 ± 4, and 64 ± 3%, and Xe inhibited the current to 51 ± 3, 34 ± 4, and 21 ± 4% at concentrations of 35, 70, and 100%, respectively. Xe inhibited the current more potently than did N²O over the range of concentrations tested. These potentiating and inhibitory effects by anesthetics on 5-HT–induced currents were all concentration dependent (fig. 2).

The concentration–response curves of 5-HT were obtained in the absence and presence of 1% sevoflurane, 100% N²O, and 100% Xe, respectively (fig. 3). Sevoflurane reduced the maximal response without changing the EC⁵⁰values (2.7 ± 0.1 and 2.8 ± 0.2 μm, respectively), but N²O and Xe shifted the concentration–response curves of 5-HT to the right without changing the maximal responses. The EC⁵⁰value of 5-HT was shifted to 4.0 ± 0.4 μm for N²O and to 4.1 ± 0.2 μm for Xe. These data indicate that these anesthetics inhibit the 5-HT³receptor according to different mechanisms, with sevoflurane acting noncompetitively and N²O and Xe acting competitively. The inhibitory effects of these three anesthetics were also examined at various membrane potentials. The effects of 2% sevoflurane, 70% N²O, and 70% Xe were not significantly different for membrane potentials ranging from −90 mV to +30 mV (Student t  test, P > 0.05) (fig. 4).

The 5-HT³receptor belongs to the superfamily of ligand-gated ion channels, which includes the nicotinic acetylcholine (nACh), glycine, and GABA^Areceptors. 1Two 5-HT³receptor subunits, 5-HT^3A16and 5-HT^3B, 17have been identified. The 5-HT^3Asubunit has been cloned from various species (mouse, rat, human) and can exist as a homomeric receptor in some systems (e.g. , mouse neuroblastoma N1E-115 and dorsal root ganglion neurons). The heteromeric receptor, comprising 5-HT^3Aand 5-HT^3Bsubunits, can exist in some regions of the brain and displays distinctive pharmacologic properties, especially with an antagonist such as tubocurarine. 17Although there is no information available about the pharmacologic difference between the effect of anesthetics on homomeric and heteromeric 5-HT³receptors, research into anesthetic effects in the heteromeric 5-HT³receptor may be necessary to provide a better understanding of the clinical significance of the effect of anesthetics on the 5-HT³receptor. Miyake et al.  reported interspecies differences not only in terms of structure but also in terms of tissue distribution and pharmacologic profile. 13The affinity of the human 5-HT³receptor for the 5-HT³receptor agonist m-chlorophenylbiguanide was much lower than that seen in the rat 5-HT³receptor, and 2-methyl-5-HT, a partial agonist for the mouse 5-HT³receptor, was a full agonist for the human 5-HT³receptor. 13In native neurons, the dissociative anesthetic ketamine potentiated 5-HT³receptor function in a rabbit nodose ganglion neuron at clinical concentrations. 18Ketamine at similar concentrations failed to modulate recombinant murine 5-HT³receptor expressed in oocytes, whereas high concentrations produced inhibition of function. 19Therefore, the pharmacologic responses of the 5-HT³receptors to anesthetics seem to differ among species. Our data showed 1% (0.41 mm) halothane and 1% (0.48 mm) isoflurane potentiated 1 μm (EC¹⁵) 5-HT–induced currents of a human recombinant 5-HT³receptor to 181.9 and 116.9%, respectively. These results were consistent with those of previous reports. Machu et al.  5reported 0.75 μm 5-HT–induced currents of a NCB-20 recombinant 5-HT³receptor were potentiated by 0.15 mm halothane and isoflurane to 133 and 111%, respectively. Jenkins et al.  4reported that in N1E-115 neuroblastoma cells, 0.21 mm halothane, and 0.31 mm isoflurane potentiated 1 μm 5-HT–induced currents to almost 150 and 115%. The larger potentiations with halothane than with isoflurane at clinically relevant concentrations were found in 5-HT³receptors from various species. Anesthetic action on the 5-HT^3Areceptor seems to show no difference between species.

The 5-HT³receptor has received considerable attention recently in relation to general anesthetics. 20Most volatile anesthetics, such as halothane, isoflurane, enflurane, and methoxyflurane, at clinical concentrations potentiated 5-HT–induced currents in 5-HT³receptors in N1E-115 neuroblastoma cells and recombinant 5-HT³receptors expressed in oocytes. 4,5,13Intravenous anesthetics, such as thiopental, etomidate, alfaxalone, and propofol, inhibited 5-HT³receptor–mediated currents in N1E-115 cells, 7,21whereas propofol has been shown to have no significant effect at clinical concentrations in recombinant 5-HT^3Areceptors. 5Jenkins et al.  4reported diverse effects of n-alcohols on the 5-HT³receptor; the lower alcohols, such as butanol and hexanol, potentiated the 5-HT³receptor at low concentrations but inhibited it at high concentrations, and the higher alcohols, such as octanol and decanol, inhibited the 5-HT³receptor at any concentration. Little is known about the effects of gaseous anesthetics, such as N²O and Xe, on the 5-HT³receptor. Recently, Yamakura et al.  12reported that a recombinant 5-HT³receptor was slightly inhibited by N²O. N²O and Xe had a similar type of effect on other ligand-gated ion channels, although there are no data available about the interaction between Xe and the 5-HT³receptor. According to the results of this study, isoflurane and halothane had a potentiating effect on the 5-HT³receptor, whereas sevoflurane, N²O, and Xe produced inhibitory effects. Like other ligand-gated ion channels, the 5-HT³receptor can be directly modulated by general anesthetics, but it produces diverse effects. Interestingly, this is in contrast to the effect on the GABA^Areceptor, which is potentiated by most anesthetics. 2 

It is interesting to note that the inhibitory effects of N²O and Xe on the 5-HT³receptor are competitive with no voltage dependency, whereas that of sevoflurane is noncompetitive. The evidence that both thiopental 4and alfaxalone 21inhibit the 5-HT³receptor in a noncompetitive manner suggests that sevoflurane may share a similar inhibitory mechanism of action with these anesthetics. Yamakura et al.  12investigated the effects of N²O and Xe on several kinds of ligand-gated ion channels, and the inhibition of NMDA and nACh receptors by N²O was noncompetitive with the voltage dependencies. It has been reported that the inhibitory mechanism of Xe was also noncompetitive for the NMDA receptor. 22The mechanisms of action of N²O and Xe at the 5-HT³receptor might be different from those at the NMDA and nACh receptors. The molecular determinant of the sensitivities to N²O and Xe has been identified in the nACh receptor 23as a single amino acid near the middle of the second transmembrane segment of the β²or β⁴subunit of the nACh receptor. There are no data available for the molecular site of action of N²O and Xe in the NMDA receptor. According to our results on the 5-HT³receptor, N²O and Xe might act at the agonist recognition site for 5-HT rather than at the channel pore in the 5-HT³receptor. It is also possible that there might be multiple modulatory sites for these anesthetics in the agonist recognition domain of the 5-HT³receptor. 4,5Specific amino acid residues that seem to form part of the 5-HT binding site in the recombinant 5-HT³receptor have been identified as a glutamate residue 24at 106 and tryptophan residues at 90, 183, and 195 in the N-terminal loops. 25Further mutagenesis studies on these residues of the 5-HT³receptor will be required to reveal the detailed site of action for N²O and Xe.

Recently, much interest has focused on the role of the 5-HT³receptor in PONV because specific 5-HT³receptor antagonists have been reported to clinically decrease the incidence of PONV. 9–11These results suggest that anesthetics that inhibit the 5-HT³receptor might cause less PONV than those that substantially potentiate 5-HT³receptor function. In the current study, we showed that halothane and isoflurane potentiated the 5-HT³receptor but, surprisingly, that sevoflurane, N²O, and Xe inhibited the 5-HT³receptor. According to clinical research about PONV after sevoflurane and isoflurane anesthesia, the incidence of PONV was lower in the sevoflurane group than in the isoflurane group. 26Potentiation by halothane was much greater than that of isoflurane, consistent with a clinical study that showed the incidence of PONV after halothane anesthesia also to be greater than that of sevoflurane. 27However, N²O, which is thought to cause clinical emesis, 28and Xe did not potentiate the 5-HT³receptor but rather inhibited it at clinical concentrations. Although little is known about the effects of Xe in terms of PONV, a study examining the analgesic potency of Xe reported nausea in 30% of volunteers. 29Therefore, it is unlikely that all the effects of anesthetics on the 5-HT³receptor have an important role in producing PONV, although the potentiation of the 5-HT³receptor by halothane and isoflurane might be partly involved in the mechanism associated with their anesthesia. No clear conclusion is possible as to whether the effects of anesthetics on the 5-HT³receptor can be related to the clinical cause for PONV.

In this study, we showed volatile and gaseous anesthetics had diverse actions on the 5-HT³receptor. Both halothane and isoflurane were confirmed as enhancing homomeric 5-HT³receptor function in a concentration-dependent manner. In contrast, we revealed that sevoflurane inhibited the 5-HT³receptor noncompetitively, and N²O and Xe inhibited the 5-HT³receptor competitively.

The authors thank Kazuro Nakano (Technician, Central Laboratory for Research and Education, Osaka University Medical School, Osaka, Japan), and Shoko Nampo, B.S. (Maruishi Pharmaceutical Co., Ltd., Osaka, Japan).