| ||||||||||||||
|
|
|||||||||||||




*Department of Anesthesia and Perioperative Care, University of Cologne, Cologne, Germany;
Department of Anesthesia and Perioperative Care, University of California, San Francisco, San Francisco, California; and
Department of Anesthesia, University Hospital, Basel, Switzerland
Address correspondence and reprint requests to C. Spencer Yost, MD, Department of Anesthesia and Perioperative Care, University of California, 513 Parnassus Ave., Room S-261, Box 0542, San Francisco, CA 94143-0542. Address e-mail to spyost{at}itsa.ucsf.edu.
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
Unlike other serotonin receptors that are G-protein-coupled receptors, the 5-HT3AR is a member of the ligand-gated ion channel superfamily. This group of ion channels, which comprise one or more homologous subunits, includes nicotinic acetylcholine (nAChR), type A
-aminobutyric acid (GABAA), glycine, and excitatory amino acid (glutamate, N-methyl-d-aspartic acid) receptors. The superfamily members display a range of similarity in sequence, structure, and pharmacology (2), with the nAChR
1 subunit and the 5-HT3AR subunit being the most alike (29% sequence identity and 40% similarity) (3). Both pass cation-selective currents immediately upon binding agonist to contribute excitatory input; in contrast, GABAA and glycine receptors selectively pass anions to generate inhibitory input. The muscle nAChR is composed of four homologous subunits assembled pseudosymmetrically in a pentameric structure to create a central ion-conducting pathway. The receptor is the molecular site of action of nondepolarizing muscle blockers (NDMBs), which competitively block the agonist binding of acetylcholine (ACh) to prevent neuromuscular transmission. The 5-HT3AR also appears to have a pentameric structure but can be assembled into functional ion channels from a single subunit (4). An additional subunit (5-HT3B) may also contribute to native 5-HT3 receptors.
These structural and functional similarities may account for known pharmacological overlap of compounds that act on the 5-HT3R and the nAChR. Serotonin (5-HT) antagonizes the action of ACh at the frog neuromuscular junction (5). More recently, it was reported that 5-HT as well as some serotonergic antagonists such as methysergide and spiperone which are not specific 5-HT3AR antagonists, inhibit ACh-induced currents mediated through mouse muscle nAChR and rat neuronal nAChRs expressed in Xenopus oocytes (68). Similar pharmacological cross-reactivity was also reported for the nAChR antagonist and NDMB d-tubocurarine, which potently blocks the 5-HT3R (9). Some authors have even reported a greater affinity of d-tubocurarine for the 5-HT3AR than for its principal target receptor, the nAChR (10).
These previous studies suggest that clinically used 5-HT3AR antagonists may have significant effects on muscle nAChRs. The aim of our study, therefore, was to examine the actions of the 3 widely used 5-HT3AR antagonists ondansetron, dolasetron, and granisetron as well as the primary active metabolite of dolasetron, hydrodolasetron, on ligand-gated ion channels expressed in the Xenopus oocytes. We initially determined the potency of these antiemetics on their target receptor, the 5-HT3AR, to confirm their actions in our heterologous expression system. We then determined the potency of the 5-HT3AR antagonists on muscle nAChRs. To account for developmental differences in drug action at the neuromuscular junction, we studied both the adult (
-nAChR) and fetal (
-nAChR) forms of the muscle receptor. Previous studies have shown marked functional differences between the two receptor subtypes (11). Finally, we assessed the effects of the three NDMBs, d-tubocurarine, vecuronium, and rapacuronium, on 5-HT-evoked currents at the 5-HT3R in order to further explore the pharmacological cross-reactivity between these two ion channel families.
| Methods |
|---|
|
|
|---|
, ß,
,
) or fetal (
, ß,
,
) nAChR or the 5-HT3R (single subunit), using an automated microinjector (Nanoject; Drummond Scientific, Broomall, PA). For the nAChR, the
, ß,
, and
subunits were diluted 1:1,000 and the
subunit 1:20 in ribonuclease-free water and mixed in the ratio of 2:1:1:1; for the 5-HT3R, the cRNA was diluted 1:10.
Plasmids encoding the mouse
-, ß-,
-,
-nAChR subunits and the mouse 5-HT3R cDNAs were linearized with NotI; for the
-nAChR subunit-containing plasmid, PvuII was used.
cRNA was synthesized in vitro from cDNA using either T7 (
, ß,
,
5-HT3R) or T3 (
subunit) RNA polymerase (T7 or T3 mMESSAGE mMACHINE KIT; Ambion, Austin, TX) according to the manufacturers instruction.
ACh, atropine, and 5-HT were purchased from Sigma (St. Louis, MO). Antiemetics and NDMBs were obtained in preparations for clinical use from the hospital pharmacy: ondansetron (Glaxo Wellcome, Research Triangle Park, NC), dolasetron (Aventis Pharmaceuticals, Kansas City, MO), granisetron (Roche Pharmaceuticals, Nutley, NJ), rapacuronium (Organon Inc., W. Orange, NJ), vecuronium (Baxter Healthcare Corp., Deefield, IL), d-tubocurarine (Abbott Laboratories, Chicago, IL). Hydrodolasetron was a gift from Aventis Pharmaceuticals (Bridgewater, NJ). All drugs were diluted in MBSH (1 mM stock solutions) that contained atropine for nAChR experiments to prevent stimulation of muscarinic receptors endogenously expressed in Xenopus oocytes. Solutions and their dilutions to the experimental concentrations were prepared immediately before the experiments.
Electrophysiological current recordings were performed at room temperature (20°22°C) 25 days after oocyte injection. Oocytes were placed in a recording chamber (approximately 25-µL volume) and superfused at a flow rate of 35 mL/min. They were impaled with 2 recording electrodes which were pulled from glass capillary tubing to obtain a 0.52 M
resistance when filled with 3 M KCl. Membrane potential was clamped at 60 mV and currents were recorded using a 2-electrode voltage clamp (Axoclamp 2A; Axon Instruments, Foster City, CA). Signals were filtered using an 8-pole low-pass Bessel filter (Frequency Devices, Haverhill, MA) set at a 40-Hz cutoff before sampling at 100 Hz. Resulting signals were digitized and stored on a Power Macintosh 7100 (Apple Computer, Cupertino, CA) using data acquisition software (MacLab; AD Instruments, Milford, MA).
Agonists were applied at concentrations approximating 50% of the maximal effect (EC50) for activation of the 5-HT3R and nAChRs (2 µM 5-HT, 100 µM ACh for
-nAChR and 10 µM ACh for
-nAChR) so as to obtain robust currents with little desensitization. Test solutions containing either ACh or 5-HT with or without various concentrations of NDMBs or 5-HT3R blockers were superfused for 10 (5-HT) or 20 (ACh) seconds and the peak current was determined. The baseline control response to ACh or 5-HT alone was measured before and after the application of agonist plus antagonist. The mean value of these two agonist applications was taken as the average control current, to which the antagonist response was compared. A washout time of at least 2 min (up to 6 min for larger drug concentrations) between drug applications was allowed for oocyte recovery to minimize contribution of receptor desensitization. Only oocytes in which the wash responses returned to within 90% of baseline values were included in the analysis. Antagonists were not preapplied because preliminary experiments indicated that there was no difference in the resulting current inhibition whether antagonists were preapplied or administered simultaneously with agonists. This finding is consistent with that of others (13). Currents determined during the 10- to 20-s application period can therefore be considered as equilibrium currents from steady-state action of the antagonists. Data were obtained from 57 oocytes taken from at least 2 batches of oocytes.
For each 5-HT3R antagonist or NDMB, the inhibitor concentrations for half-maximal responses (IC50), 95% confidence intervals, and Hill coefficients (nHill) for inhibition of 5-HT3R or nAChR currents were obtained by fitting the fractional block, f (f = 1 Idrug/Icontrol) at various antagonist concentrations to the Hill equation using a nonlinear least-square fitting procedure (Prism software 3.0a for Macintosh; GraphPad Software, San Diego, CA). Blocking potencies between 5-HT3R antagonists or NDMBs at each receptor were compared for significant differences by one-way analysis of variance (ANOVA) followed by Tukeys test. Potencies of 5-HT3R antagonists on each nAChR subtype were compared for each drug using the unpaired two-tailed Students t-test with the same software package. Results are represented as mean ± sd. For all tests, P < 0.05 was considered significant.
| Results |
|---|
|
|
|---|
-nAChR- or
-nAChR-expressing oocytes or application of ACh (0.1 nM1 mM) to 5-HT3AR-expressing oocytes did not evoke currents (data not shown). For 5-HT3AR-expressing oocytes, application of 5-HT induced concentration-dependent currents with an estimated concentration achieving EC50 of 2.4 ± 0.5 µM and a Hill coefficient of 1.16 (Fig. 1). In all subsequent experiments, an agonist concentration of 2 µM 5-HT (approximately EC50) was used to activate the 5-HT3R. This concentration produced robust signals without significant desensitization after repeated exposures.
|
The 5-HT3R antagonists ondansetron, dolasetron, hydrodolasetron, and granisetron reversibly inhibited 5-HT-induced inward currents through the 5-HT3R in a concentration-dependent manner (Fig. 2A). The four compounds showed high potency for their target receptor with IC50 values in the nano- (nM) to subnanomolar range (Table 1). Granisetron and hydrodolasetron were significantly more potent than ondansetron or dolasetron in blocking 5-HT-induced currents at the 5-HT3R (ANOVA, P < 0.05) with hydrodolasetron being >40-fold more potent than dolasetron for inhibiting the 5-HT3AR.
|
|
Studies have shown that NDMB drugs also inhibit 5-HT3AR. To explore this pharmacological overlap further, we determined the inhibition of the 5-HT3R by three NDMBs: rapacuronium, vecuronium, and d-tubocurarine. These compounds reversibly inhibited 5HT-induced currents in a concentration-dependent manner (Fig. 2B). The potency of d-tubocurarine on the receptor was significantly higher compared with vecuronium and rapacuronium (ANOVA, P < 0.01) as shown by their IC50 values, approximating the values obtained with the antiemetic drugs (Table 1).
Ondansetron, dolasetron, hydrodolasetron, or granisetron did not evoke currents at the
- or
-nAChR when applied alone (data not shown). However, all produced reversible, concentration-dependent inhibition of ACh-evoked currents at both receptor subtypes (Fig. 3, A and B) in the micromomolar (µM) to tens of µM range. All compounds except ondansetron inhibited the
-nAChR more potently than the
-nAChR (P < 0.05) (Table 2). As determined by their IC50 concentrations, hydrodolasetron was more potent on the
-nAChR (ANOVA, P < 0.01) and the
-nAChR (ANOVA, P < 0.05) than granisetron which was more potent than dolasetron or ondansetron. Thus, the rank order of potency at the adult receptor was hydrodolasetron >> granisetron > dolasetron
ondansetron. Hydrodolasetron was approximately 10 times more potent than dolasetron for inhibiting the
-nAChR.
|
|
The inhibitory effects of the 5-HT3AR antagonists on nAChR seem to be noncompetitive. The concentration-response relation for inhibition of
-AChR by two different antiemetics with two different agonist concentrations are shown in Figure 4. Inhibition of whole cell currents by granisetron (Fig. 4A) or by ondansetron (Fig. 4B) of
-nAChR activated by either 10 or 100 µM ACh shows near identical effect curves with IC50s that were not statistically different (t-test, P > 0.05), indicating a noncompetitive mode of inhibition.
|
| Discussion |
|---|
|
|
|---|
-nAChR) and fetal (
-nAChR) mouse muscle nAChRs in a concentration-dependant manner. In line with their clinical antiemetic potency, granisetron inhibited 5-HT3ARs more potently than ondansetron and dolasetron. The antiemetics also inhibited the nAChR with the same rank order of potency. In addition, hydrodolasetron, the active metabolite of dolasetron, displayed a much higher blocking potency on both 5-HT3AR and nAChR than that of the parent compound with an IC50 close to that of granisetron. These data support the view that pharmacological cross-reactivity can occur with receptors of the ligand-gated ion channel superfamily. The inhibiting effects of 5-HT on ACh-dependent neuromuscular transmission in amphibians have been known for many years (5) and the blockade of heterologously expressed nAChRs by serotonergic agonists as well as some antagonists (7). However, blockade of heterologously expressed mammalian adult and fetal muscle nAChR subtypes by the 5-HT3R-specific antagonists ondansetron, dolasetron, and granisetron has not been previously demonstrated. This inhibition is consistent with a noncompetitive mode of inhibition, because a 10-fold increase in agonist concentration did not shift the concentration response curve (Fig. 4). This finding corroborates earlier studies that also found noncompetitive inhibition of both muscle nAChRs and neuronal nAChRs by antagonistic serotonergic compounds (7,13,14). It was suggested that a Hill coefficient close to one and the voltage dependence of the blockade of nAChRs by serotonergic drugs indicate that they interact with one binding site that is located within the ion channel (13). Whereas neuronal nAChRs are formed by one or two different types of subunits, muscle receptors are formed by four different types of subunits. Although the binding sites for serotonergic drugs at nAChRs have not been identified, it seems likely that they bind to the same site in muscle as neuronal receptors (14).
All but one of the tested 5-HT3R antagonists displayed statistically significant higher potency for the
-nAChR versus the
-nAChR (Table 2). This finding contrasts with the results of the agonist 5-HT, which was reported to have a more potent blocking effect on the
-nAChR compared with the
-nAChR expressed in transfected BOSC 23 cells (15). It is possible that the developmental change in nAChR subunit composition could affect the affinity for serotonergic agonists versus antagonists differently.
5-HT3R antagonists were designed to block this subtype of serotonin receptors with high specificity and they are reported to have little or no affinity to other serotonin or dopamine receptors (16,17). However, pharmacological interaction of 5-HT3AR antagonists with other ion channels, such as human cardiac Na+ channels, has been reported (18) and may represent the underlying mechanism for QRS widening or QT interval lengthening associated with the clinical administration of ondansetron, dolasetron, or granisetron (19,20). The data reported here confirm that 5-HT3R antagonist antiemetics seem to be less receptor specific than proposed earlier.
It is unclear whether these results have clinical meaningthat 5-HT3AR antagonists interfere with neuromuscular transmission at clinically used doses. Muscle weakness is not a reported side effect in patients receiving even very large doses of 5-HT3AR antagonists for prophylaxis and treatment of nausea and emesis during chemotherapy treatment for cancer (16,17). However, a different situation exists in the immediate postoperative setting, where 5-HT3AR antagonists are often administered to prevent or treat PONV. Subclinical, yet significant, residual neuromuscular blockade may still exist at emergence from general anesthesia despite administration of reversal drugs. Neuromuscular transmission has a large margin of safety in that up to 75% of postsynaptic nAChRs may be blocked before muscle function decreases (21). In such a scenario, administration of 5-HT3AR antiemetics could contribute an increased blockade of nAChRs. After IV administration of a 200-mg dose of dolasetron in healthy male volunteers, transient peak plasma concentrations reached 2 µM (1.1 µg/mL), a concentration that inhibited nAChR current by 35% in our experiments (22). Peak free plasma concentrations of 480 nM ondansetron were determined after a 32-mg dose in healthy adults (23). Such peak concentrations resulting from large IV dosing could have a transient effect on neuromuscular transmission which remains clinically unnoticed or might be attributed only to the residual effect of applied NDMBs.
Nevertheless, these drugs displayed a thousand-fold lower affinity for the nAChR compared with their target receptor. Therefore, doses for successful treatment of PONV are probably much smaller than those necessary to reach drug concentrations relevant for clinically apparent interference with neuromuscular transmission. We are aware of only one clinical study examining the interaction of one 5-HT3AR antagonist antiemetic and one NDMB, which showed no additive neuromuscular blocking effect of ondansetron (8 or 16 mg) in combination with atracurium (24). However, interactions could be different with longer-acting NDMBs or larger doses of antiemetics. Further clinical studies will be necessary to explore these possibilities.
We also explored the interaction of hydrodolasetron with the nAChRs because dolasetron is quickly metabolized (t1/2 < 10 minutes) into this potent, long-acting (t1/2 approximately 6.68.8 hours) metabolite after IV administration in humans (22). Because hydrodolasetron is a 40-times more potent inhibitor of the 5-HT3AR and a 10-times more potent inhibitor of
-nAChR, this metabolite is likely responsible for clinical therapeutic effects and potential side effects of the parent compound.
Affinity of drugs to nAChRs may vary among species, so that our results established for mouse muscle nAChR may differ from those in humans. However, our previous work has established a strong correlation between human NDMB potency and the ability to block the heterologously expressed mouse nAChR (12).
In our study, we also confirmed that d-tubocurarine very potently blocks the 5-HT3R. In fact, the IC50 we found for inhibition of the 5-HT3AR (11.4 nM) was less than the IC50 (43.3 nM) we previously determined for inhibition of its clinical target, the
-nAChR (12). These data indicate that d-tubocurarine has intrinsic antiemetic properties and support the previous suggestion that the d-tubocurarine structure could serve as a lead compound for the development of new antiemetics (25). Derivatives of d-tubocurarine can block 5-HT3AR to varying degrees depending on the performed modification of its molecular structure (26). In contrast, NDMBs from the aminosteroid structural category, such as rapacuronium and vecuronium, have potencies that are 104-times less (27).
In this study, we demonstrated that the 5-HT3AR antagonist antiemetics ondansetron, dolasetron, and granisetron potently inhibited their target receptor, but also exerted marked inhibition of ACh-evoked currents in the adult and fetal form of the muscle nAChR in a concentration-dependent manner. Large doses of these compounds may clinically interfere with neuromuscular transmission. The potential for such interference may be increased in the postoperative setting, when antiemetics are typically administered, because residual neuromuscular blockade after intraoperative administration of NDMBs could still be present. Our results indicate that cross-reactivity can occur between the 5-HT3AR and nAChR through compounds that are generally considered to be highly selective. Such cross-reactivity might contribute to unexpected side effects of antagonist drugs.
The authors thank Beth Sampson for expert technical assistance, Drs. John Forsayeth, Zach Hall (Department of Physiology, University of California, San Francisco, CA), and Paul Gardner (Department of Biochemistry, Dartmouth Medical School, New Hampshire) for cDNA expression plasmids for mouse muscle nAChR subunits
, ß,
,
, and
. The cDNA encoding the mouse 5-HT3A receptor was a gift from David Julius, Department of Cellular and Molecular Pharmacology, University of California, San Francisco. Hydrodolasetron was a gift from Aventis Pharmaceuticals (Bridgewater, NJ).
| Footnotes |
|---|
This work was presented in part at the 2003 Annual Meeting of the American Society of Anesthesiologists in San Francisco, CA, October 1115.
Accepted for publication February 2, 2005.
| References |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|